Citation
Temporal summation of the acoustic-reflex threshold

Material Information

Title:
Temporal summation of the acoustic-reflex threshold a possible indicator of cochlear abnormalities.
Creator:
Parker, William Lee, 1945-
Publication Date:
Language:
English
Physical Description:
x, 147 leaves. : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Acoustic impedance ( jstor )
Acoustic reflex ( jstor )
Audio frequencies ( jstor )
Auditory perception ( jstor )
Ears ( jstor )
Lesions ( jstor )
Loudness ( jstor )
Reflexes ( jstor )
Signals ( jstor )
Time constants ( jstor )
Audiometry ( fast )
Dissertations, Academic -- Speech -- UF ( lcsh )
Hearing ( fast )
Speech thesis Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 131-146.
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by William Lee Parker

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
022671190 ( ALEPH )
13990469 ( OCLC )

Downloads

This item has the following downloads:


Full Text












TEMPORAL SUMMATION OF THE ACOUSTIC-REFLEX THRESHOLD:
A POSSIBLE INDICATOR OF COCHLEAR ABNORMALITIES













By

William Lee Parker

















A DISSERATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA
1973






























To my wife, Christine














ACKNOWLEDGMENTS


It is nice to be able to give thanks to all of those

who helped me complete this paper. My advisory committee,

Drs. F. O. Black, P. J. Jensen, K. C. Pollock, W. A. Yost,

never failed to extend themselves professionally or per-

sonally. Dr. Pollock, my chairman, encouraged my interest

in impedance audiometry and was instrumental in my applica-

tion for, and the award of, research funds from the

University of Florida. A special thanks is extended to

Dr. Yost, who took an inordinate interest in this project

and without whose help the project would not have been

completed. To Dr. W. N. Williams goes the award for best

encouragement and prodding, the mark of a true friend.

I also want to thank Drs. E. C. Hutchinson and D. T.

Hughes for their statistical advise and help. Christine

Parker and John Parks contributed their invaluable drawing

skills, and Ginny Parks helped with the mundane chore of

proofreading. And thanks to Sue Kirkpatrick for completing

the final typing.











iii















TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS . . . . iii

LIST OF TABLES . ... .. . vi

LIST OF FIGURES . .. . . . vii

ABSTRACT . . . . . .ix

CHAPTER

I. INTRODUCTION AND REVIEW OF THE LITERATURE 1

Review of the Literature . .... .. .. 1
Audiological Tests Commonly Used in
Differential Diagnosis . . 1
Acoustic Reflex . .. .. .. 5
Threshold of the Acoustic Reflex (ART) 14
Temporal Summation . .... .. 32

II. STATEMENT OF THE PROBLEM . . .. 45

Hypothesis .. .. . . 52

III. METHODS AND PROCEDURES . . 55

Stimuli . . ..... .. .. 57
Experimental Equipment .... . .62
Procedure ............ . . 67
Temporal Integration at Threshold of
Audibility .. . . . 67
Temporal Integration at the Threshold of
the Acoustic Reflex . . .. 68
Experimental Safeguards . . .. 71

IV. RESULTS AND DISCUSSION . . .. 76

1) Temporal Integration at the Acoustic-
Reflex Threshold . .. . 76
2) Temporal Summation at Threshold of
Audibility . . . 91
3) Temporal Integration at the Acoustic-
Reflex Threshold Compared to Temporal
Integration at Threshold of Audibility 98

iv










CHAPTER Page

V. CONCLUSIONS AND SUMMARY . . .. 100

APPENDICES . . . . . 104

A. MEDICAL HISTORY FORM. . . .. 105

B. PURE TONE HEARING LEVEL (ANSI-1969) OF ALL
SUBJECTS . . . . 107

C. VOLTAGE DIVIDER. ... . . ... 109

D. CALIBRATION DATA TABLES . . . 111

E. RAW DATA OF THE ACOUSTIC-REFLEX THRESHOLD 115

F. DISCRIMINANT ANALYSIS. .. . .. 125

G. RAW DATA OF THRESHOLD OF AUDIBILITY . 127

H. CORRELATION COEFFICIENTS OF THE TEMPORAL
INTEGRATION MEASURES. .. . ... 129

REFERENCES . . . .. 131

BIOGRAPHICAL SKETCH . . . . .. 147





























v
















LIST OF TABLES


Table Page

1. Acoustic-Reflex Threshold in dB SL . ... 16

2. Acoustic-Reflex Threshold in dB HL (ANSI-1969) 17

3. Acoustic-Reflex Threshold in dB SPL (re 0.0002
dyne/cm2) .. .. . . . 18

4. Test Stimuli Used to Elicit Threshold of Adui-
bility and Threshold of the Acoustic Reflex 58

5. Conceivable Sound Pressure Level Necessary to
Elicit the Acoustic-Reflex Threshold by a
500-Hz Burst at a Signal Duration of 10 Msec 73

6. Discriminant Analysis of the Acoustic-Reflex
Threshold . . . . 79

7. Average Slope Change Per Decade of Time of the
Acoustic-Reflex Threshold . .. 79

8. Discriminant Analysis of the Threshold of Audi-
bility Data . . . ... .93

9. Signal Duration Intervals Within Which Time
Constants of Temporal Integration at Threshold
of Audibility Occur . . .. 97


















vi















LIST OF FIGURES


Figure Page

1. Possible time, strength and direction of the
stapedius and tensor tympani muscles, and
resultant displancement of the tympanic
membrane. . . . . 8

2. The acoustic-reflex arc. . .. 10

3. Pre- and post-noise exposure thresholds 30

4. Average temporal integration slopes as a func-
tion of frequency . . .. 36

5. Examples of temporal integration . .. 43

6. Temporal summation of the acoustic reflex 48

7. Temporal integration at threshold of audi-
bility and of the acoustic reflex for normal-
hearing subjects . . .. 50

8. Acoustic-reflex contraction . .. 60

9. Spectral content of tone bursts with 5-msec
rise/decay ramp . . .. 61

10. A 20-msec signal burst with a smooth 5-msec
rise and decay . .. . ... 63

11. Block diagram of equipment used to elicit and
record the intra-aural reflex and to allow
subject control over the test situation 64

12. Typical response of the acoustic reflex at
threshold . . . .. 70

13. Upper limits of acceptable exposure to impulse
noise for 95 percent of the population to
10,000 impulses. . . .. 72

14. Mean acoustic-reflex thresholds as a function
of temporal integration . .. 78



vii










Figure Page

15. Acoustic-reflex thresholds of normal and
cochlear-impaired (bad) ears as a function
of temporal summation . . ... 81

16. Acoustic-reflex thresholds of normal (good)
and impaired (bad) ears of the Meniere's
group as a function of temporal summation 82

17. Inter-subject range of acoustic-reflex thresh-
olds as a function of temporal integration 83

18. A comparison of mean acoustic-reflex thresh-
olds as a function of temporal summation in
normal-hearing ears . . .. 85

19. Signal-duration interval in which the time
constant of temporal integration for each
test subject occurred across all frequencies 87

20. The delayed time constant in some cochlear-
impaired ears contrasted with the normal
time constant of temporal integration at the
acoustic-reflex threshold . ... 89

21. Thresholds, audibility, and acoustic reflex of
the test subjects at 500-msec signal duration 92

22. Mean thresholds of audibility as a function of
temporal integration . . .. 94

23. Inter-subject range of thresholds of audibility
as a function of temporal integration .. 96




















viii









Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of
the Requirements for the Degree of Doctor of Philosophy

TEMPORAL SUMMATION OF THE ACOUSTIC-REFLEX THRESHOLD:
A POSSIBLE INDICATOR OF COCHLEAR ABNORMALITIES

By

William Lee Parker

August, 1973

Chairman: Kenneth C. Pollock, Ph.D.
Major Department: Speech

It has been suggested that temporal summation of the

ART might be a clinical tool distinguishing normal from

cochlear-impaired ears. If the ART measure could yield

differences at least as distinctive as those of temporal

summation at threshold of audibility, the ART could be

used with patients for whom threshold measures are not

appropriate.

In this study, temporal summation of the acoustic-

reflex threshold (ART), as well as of the threshold of

audibility, was measured in five normal-hearing and five

unilaterally hearing-impaired subjects by varying signal

duration of 500 Hz, 1000 Hz, and 2000 Hz tones. The signal

durations employed were 500 msec, 200 msec, 100 msec, 20

msec, and 10 msec.

There were statistically significant differences be-

tween normal ears and cochlear-impaired ears at 2 out of

12 comparisons. In addition, there were some descriptive

differences between normal ears and cochlear-impaired ears


ix










for the ART measure. Small sample size, unexpected inter-

subject variability, and flatter integration slopes for the

normal ears contributed to the lack of consistent statistical

significance.

Temporal integration at threshold of audibility proved

to be significant in most comparisons between normal ears

and cochlear-impaired ears. Even though temporal integra-

tion at threshold of audibility evidenced a statistical

difference, the differences were not obvious enough to make

it an effective clinical tool.

There were some differences between the normal-hearing

and cochlear-impaired ears concerning temporal integration

at the threshold of the acoustic reflex. The mean integra-

tion slopes of the cochlear-impaired ears were more depressed

than those of the normal ears.

These data suggest that temporal integration of the

acoustic-reflex threshold may not be used to differentiate

normal ears from cochlear-impaired ears. It may be con-

strued that the time constant might provide a statistically

significant difference.














x














CHAPTER I

INTRODUCTION AND REVIEW OF THE LITERATURE


Auditory tests, especially those designed to determine

site of lesion, are complicated by subtle test procedures,

and patient state and sophistication. For the majority of

persons the results of audiometric procedures are of diag-

nostic value, but in a few cases even a battery of audiometric

tests gives inconclusive results. The diagnostic picture may

be incomplete because of the inability orunwillingness of the

subject to respond. The literature indicates the status of

current audiological tests and further possibilities offered

by tests of acoustic-reflex and of temporal integration by

the auditory system. Furthermore, these later two auditory

measures may be effectively combined into one diagnostic test.

These conditions, therefore, motivated the present study. It

was designed to investigate temporal summation of the acoustic

reflex as a contribution to the diagnosis of cochlear lesions.


Review of the Literature

Audiological Tests Commonly Used
in Differential Diagnosis

Audiological tests are generally divided into those

most sensitive for the detection of conductive, cochlear,

eighth nerve, and central nervous system lesions. Only those


1





2




test procedures normally employed in the assessment of end-

organ status, in particular those tests considered most

valuable in differential diagnosis, will be reviewed in this

section: Alternate Binaural Loudness Balance, Monaural

Loudness Balance, Short Increment Sensitivity Index, and

Bekesy audiometry.

The Alternate Binaural Loudness Balance (ABLB) test

(Fowler, 1936) has long been the indicator of loudness recruit-

ment.2 This abnormally rapid growth in loudness is theo-

retically regarded as pathognomonic of cochlear lesion

(Dix, Hallpike and Hood, 1948; Fowler, 1936, 1950; Hood, 1969;

Jerger, 1967, among others). In spite of its high validity

in confirming recruitment (Tillman, 1969), this test cannot

be administered to all patients. The patient must have one

normal, non-recruiting ear at each test frequency and at

least a 20 dB threshold difference between ears to obtain

meaningful data. Further, it is estimated that more bilateral

hearing losses than unilateral exist (Carver, 1972). Thus,


It is generally understood that audiometric evaluations
do not measure anatomical lesions, per se. Instead, they
reflect functional alterations in the auditory response caused
by anatomical lesions or physiological changes.

Loudness recruitment (Fowler, 1937) is an auditory
phenomenon where the supraliminal loudness growth of one ear
occurs at a faster rate than that of the contralateral ear to
equal intensity units. When the threshold differences between
the two ears is more than 20 dB and the supra-threshold
balances are considered equal within 10 dB, the loudness
growth is considered abnormal. This abnormal loudness growth
is called loudness recruitment. (Fowler, 1936; Hood, 1969;
Jerger, 1967.)





3



this standard detector of recruitment cannot be administered

to a majority of hearing-impaired persons to gain diagnostic

information.

The Monaural Bi-frequency Loudness Balance (MBLB) test

(Reger, 1936) compares the loudness above threshold at one

frequency with the loudness of another frequency. Again, the

threshold at one of the test frequencies must be normal with

the other elevated. In this way, recruitment can be tested

unilaterally, thereby allowing a larger segment of the

population to be tested diagnostically for cochlear lesions.

It is perceptually much more difficult for the patient to

perform than the ABLB and is, therefore, less reliable. This

is especially true as the difference between the standard

and comparison frequency increases (Graham, 1967).

The Short Increment Sensitivity Index (SISI) test

(Jerger, Shedd,and Harford, 1959) indicates the ability of

the cochlear mechanism to respond to small changes in signal

amplitude. It is an indicator of cochlear disorders if a

high percentage of 1 dB increments is detected by the sub-

ject, but it apparently is not as reliable a detector of

cochlear impairment as the ABLB measure (Owens, 1971; Tillman,

1969). Its lower reliability may be related to the patient

sophistication necessary to complete the procedure, which

is demonstrated by the difficulty in orienting some patients

to the auditory listening task.

The four Bekesy Types, introduced by Jerger (1960),

were based upon the difference in self-tracked threshold





4




levels using both pulsed and continuous pure-tone sweep fre-

quency stimnui. A Type II Bekesy, indicated by a slight

separation between the two tracings and a narrowing of the

tracing for the continuous tone, has been the classic

designator of cochlear pathologies (Jerger, 1960). Unfor-

tunately, other patterns such as Type I, IV, or mixed are

also obtained from subjects with confirmed end-organ lesions,

therefore contributing ambiguity to the test results (Owens,

1964). The self-tracking procedure of Bekesy audiometry is,

in addition, a long, fatiguing procedure which may make it

difficult for some patients to complete.

Even though each of these tests has certain drawbacks,

the results from each one can indicate the presence of a

cochlear lesion. Recruitment, as defined operationally by

loudness balancing, is the foremost test of a cochlear

lesion, but abnormal loudness growth is not necessarily

synonymous with recruitment (Hirsh, Palva, and Goodman, 1954;

Simmons and Dixon, 1966). All tests for cochlear lesions

mentioned may represent abnormal auditory processing in a

particular cochlear pathology but may delineate differing

physiological phenomena; otherwise there would be no need

for a battery of audiometric tests.

The four auditory tests discussed above may be valid

indicators of cochlear lesions, but their reliability is

dependent upon the full cooperation of a highly motivated

test subject. In many instances cooperation may not be










obtained because: (1) the patient cannot or does not under-

stand his responsibility in the test situation, i.e., how

to recognize or when to respond to the auditory stimuli;

(2) the motivation of the patient is minimal; or (3) the

patient, because of physical state, is incapable of an appro-

priate response (Martin, 1972). Any one of these factors can

reduce the reliability of the test procedure without con-

scious intent on the part of the patient, as frequently

occurs with such patients as psychotics, mental retardates,

cerebral palsied persons, preverbal infants, or the comatose

or severely ill individual.

There have been numerous reports on the use of the

acoustic reflex an an indicator of abnormal auditory pro-

cessing due to cochlear impairment. It is an involuntary

physiological reaction reflected in the contraction of the

intra-aural muscles to relatively intense acoustic stimuli.

In addition, the auditory system's ability to integrate

acoustic power over an interval of time has been investigated

as a measure of cochlear function and integrity. A combina-

tion of these two auditory functions may serve to quantify,

objectively, auditory assessment.


Acoustic Reflex

Anatomy and physiology of the acoustic-reflex arc

The acoustic reflex is a consensual contraction of the

middle-ear muscles producing, in persons with normal auditory





6



pathways, a bilateral change in acoustic impedance3 in both

ears to acoustic stimuli. Upon simultaneous contraction,

the two intra-aural muscles, the stapedius and tensor tympani,

act in a physiologically antagonistic manner but produce a

synergistic impedance against sound energy.

The stapedius muscle is the smallest muscle in the human

body. It originates from the pyramidal eminence on the

posterior wall of the tympanic cavity and its tendon inserts

on the head, neck, or posterior crus of the stapes. The

tensor tympani muscle arises from the bony semicanal above

the Eustachian tube. Its tendon traverses the tympanum to

insert on the mallar manubrium (Jepsen, 1963; Kobrak, 1959).

The direction of pull of these two muscles is at right angles

to the axis of their corresponding ossicles, making the func-

tional action of the two muscles almost in direct opposition

to one another (Wever and Lawrence, 1954).

If the movement caused by each of the muscles is con-

sidered independently of the other, contraction of the

stapedius muscle causes the stapes to be pulled posteriorly



It is not within the scope of this paper to differ-
entiate between the types of impedance present within the
middle ear, nor between the various contributing factors of
increased impedances (see Hung and Dallos, 1972; Lilly,
1972, 1973; Zwislocki, 1961, 1962). It is important, however,
to know that the contraction of one or both of the intra-
aural muscles will produce an opposition to energy flow
through the middle-ear cavity, thereby reducing the acoustic
energy reaching the sensory end-organ.





7



and outward from the oval window, as well as causing the

tympanic membrane to be pushed in a slightly outward direc-

tion. The malleus and tympanic membrane swing medially upon

tensor tympani activation (Jepsen, 1963; Kobrak, 1959).

This is not the case, however, upon the contraction of both

muscles simultaneously. The movement of the tympanic membrane

depends upon the relative strength, latency, and contraction

time of each muscle. Figure 1 demonstrates a possible resul-

tant of a stronger and faster-acting stapedius muscle con-

traction partially opposed by the later and weaker tensor tym-

pani muscle contraction. It is important to remember that it

does not matter acoustically which muscle predominates because,

in either case, the acoustic impedance will be increased.

The acoustic-reflex arc consists of an afferent neuron,

a reflex center, and an efferent neuron (Lorente de No, 1933,

1935; Rasmussen, 1946). The afferent portion is the same

afferent pathway for audition starting from the sensory end-

organ but terminating at the level of the superior olivary

complex (SOC). The SOC consists of at least five cellular

groups, but the accessory, or medial, nucleus is felt to be

the central mediator of both reflex arcs, the stapedial and

the tensor tympani. The accessory superior olive gives off

a few fibers to the ipsilateral motor nucleus of the facial

nerve (n. VII) (Rasmussen, 1946), which constitutes the

central portion of the stapedial-reflex arc. The facial

nerve innervates the stapedial muscle to complete the efferent

portion of this reflex arc.






8
















STAPEDIUS


E LATERAL



















~- S IMULUS
z










U

STENSOR
SMEDIAL TYMPANI

'-<- STIMULUS -*


TIME




Figure 1. Possible time, strength and direction of
the stapedius and tensor tympani.muscles, and resultant
displacement of the tympanic membrane. (Redrawn from
Mendelson, 1963.)






9



It is also surmised that fibers from the accessory

nucleus pass to and from the lateral lemniscus into the

ipsilateral motor nucleus of the trigeminal nerve (n. V)

(Rasmussen, 1946), as the central portion of the tensor tym-

pani-reflex arc. The efferent portion of this reflex arc

constitutes the tensor tympani muscle innervated by the

mandibular branch of the trigeminal nerve. The general schema

for the acoustic-reflex arc as proposed by Rasmussen (1946) is

shown in Figure 2.

Simmons (1963) does not necessarily take exception to

Rasmussen, but suggests that there are several possible re-

flex loops, both ipsilateral and crossed. Simmons speculates

that the major differences in latency and response level be-

tween the stapedius and tensor tympani muscles are due to the

relative degree of synaptic connections of their respective

reflex arcs. According to Simmons, the stapedial reflex

probably has the more compact, less diffuse, interneuron

connections, thereby explaining its greater sensitivity.

Each of the proposed schema would explain the lower reflex

threshold and shorter latency of the stapedial acoustic-

reflex arc activity in comparison to the tensor tympani

acoustic-reflex responses (as seen in Figure 1). In either

case, the SOC is the most probable reflex center. Neural

activity in the SOC increases with loudness to at least 60

dB sound pressure level (Boudreau, 1965). The acoustic re-

flex probably occurs when the neural activity in this















S-Facial
I Nucleus




P Primary Lateral Temporal
Nuclei SOCLemniscus Cortex




S.:_\ Trigeminal
i/ Nucleus








Figure 2. The acoustic-reflex arc. The stapedial-reflex arc consists of the end-
organ, the primary nuclei, the SOC, the facial nucleus, and the stapedial muscle. The
tensor tympani-reflex-arc proceeds from the SOC through the lateral lemniscus to the
trigeminal nucleus, ending in the tensor tympani muscle. (Based on Rasmussen, 1946.)






11




synaptic co..;: Lx i:ceeds some critical level of neural ex-

citation (Lo-ente Io No, 1933, 1935).

The staprdius muscle is generally considered to be the

most active and dominant muscle in response to acoustic

stimulation in humans, whereas the tensor tympani reacts to

acoustic stimulation less often4 (Flottorp and Djupesland,

1970; Holst, Ingelstedt, and Ortegren, 1963; Jepsen, 1963;

Liden, Peterson, and Harford, 1970; McRobert, 1968; Mendelson,

1961, 1966; Terkildsen, 1957, 1960a,c; Weiss, Mundie, Cashin,

and Shinabarger, 1962; and others). It appears that 1

percent to 5 percent of the population with normal hearing

and with no apparent middle-ear pathology have no demonstrable

acoustic-reflex response (Shiffman, 1972; Swannie, 1966;

Terkildsen, 1960c; Weiss et al., 1962; Wright and Btholm,

1973).


Methods of detecting the intratympanic
muscle activity in man

Some investigators of the intratympanic muscle activity

have attempted to make direct observations of muscle contrac-

tions through surgical exposure or via chronic tympanic mem-

brane perforations. For the most part, however, only the

stapedial tendon is directly available for easy visualization


4There has been considerable discussion of whether the
tensor tympani muscle is active in man to acoustic or non-
acoustic stimulation. This is a clinical question of consider-
able importance. McRobert (1968), in a critical review of
the literature,and Liden, Peterson and Harford (1970) agree
that tensor activity is present in man but that the stapedius
predominates.





12




(Kobrak, 1948; Lindsay, Kobrak, and Perlman, 1936; Lorente

de No, 1933; Lorente de No and Harris, 1933; Perlman and

Case, 1939). Since visual observation limits the accuracy

of data quantification, other approaches have been developed

for studying the reflex activity. Electromyography (EMG)

measures the individual muscle fiber action potentials and

is, therefore, the most direct method. EMG activity from

both muscles in humans has been reported in response to both

acoustic and non-acoustic stimulation (Djupesland, 1965;

Fisch and Schulthess, 1963; Salomon and Starr, 1963). This

technique is not clinically feasible.

Extratympanic manometry has been the method employed to

determine direction of tympanic membrane movement. This

technique is made possible by sealing the external auditory

meatus with a probe containing a pressure-sensing device.

If the external meatus is properly sealed, the extratympanic

air pressure should decrease or increase with respective

inward or outward movement of the tympanic membrane (Flottorp

and Djupesland, 1970; Holst et al., 1963; Liden et al., 1970;

Mendelson, 1961, 1963; Terkildsen, 1957, 1960a,c; Weiss et al.,

1962). McRobert (1968), in a review of this measuring tech-

nique,noted that the theory for the tympanic membrane response

to individual muscle contraction is valid, but that there are

many inexplicable results (viz. Mendelson, 1957) suggesting

the need for better instrumentation. Moller (1964) and

Neergaard and Rasmussen (1966) also warn that small





13



contractions of the intra-aural muscles may not produce a

measurable change in extratympanic air pressure. Instead,

these muscle contractions are reflected in a measurable in-

crease in the impedance to acoustic energy within the middle

ear.

The technique used most extensively in the past decade

to indicate intra-aural muscle contraction is acoustic-

impedance measurement. Simply stated, a probe tone is directed

perpendicularly towards the plane of the tympanic membrane.

While most of this acoustic energy is transmitted through the

tympanic membrane and attached ossicular chain to the oval

window, a portion of the acoustical energy wave is reflected

back into the external canal from the tympanic membrane.

Since the probe tone input and the reflected tone pick-up

microphone are connected to the external auditory meatus

with an airtight seal, the reflected probe tone can be

monitored accurately. As the intra-aural muscles contract,

the acoustic impedance increases, causing an increased amount

of the probe tone to be reflected back into the external canal.

It is the increased amplitude and phase change of the reflected

tone which indicate that a change in intra-aural muscle

activity has occurred.5 This type of measurement has passed



Unless specifically stated otherwise, all acoustic-
reflex data concern the ear in which the reflex is elicited.
The ear at which the reflex is elicited may not necessarily
be the same ear in which the reflex contraction is recorded;
most often the reflex ear and the measurement ear are contra-
lateral to each other when using the acoustic impedance





14



from an experimental method (Metz, 1946; Zwislocki, 1957) into

the clinical armamentarium with the introduction of several

commercially available acoustic-impedance meters (Grason-

Stadler, 1973; Madsen, n.d.; Peters, n.d.; Terkildsen and

Scott Nielsen, 1960; Zwislocki, 1963). With the increasing

interest in the United States concerning this technique as

a diagnostic tool, critical evaluations of methods and the

commercially available equipment have been made. These instru-

ments have been found to be reliable (Nixon and Glorig, 1964;

Tillman, Dallos, and Kurvilla, 1963), sensitive (Moller,

1964), and also relatively easy to utilize (Brooks, 1971).


Threshold of the Acoustic Reflex (ART)

The acoustic reflex first occurs at a predictable level

above the normal threshold of audibility. The predictability

of this reflex accounts for its clinical usefulness; there-

fore, it is forthwith described in detail along with factors

which can affect its performance and measurement.

For pure tones from 250 Hz to 4000 Hz the range of the

acoustic-reflex threshold (ART) is 70 dB to 90dB above the nor-

mal-hearing individual's threshold (sensation level: SL) with

a mean of approximately 80 dB SL (Deutsch, 1968, 1972; Jepsen,

1951). Jerger, Jerger, and Mauldin (1972) report the range


measurement technique. If the reflex threshold concerns the
left ear, the reflex eliciting tone will be delivered to the
left ear. Because both sides will contract to unilateral
stimulation, the reflex action in the above example will be
recorded in the right ear. This eliminates acoustic inter-
ference of the eliciting tone with the probe tone, which may
cause misleading results.





15




of values for the ART in 382 normal-hearing persons as being

normally distributed with a mean of 85 dB HL (ANSI-1969).

Ninety-five percent of their population fell within 70 dB to

100 dB HL and 99 percent within 65 dB to 105 dB HL. This

range of thresholds has been supported by others (Anderson

and Wedenberg, 1968; Deutsch, 1968, 1972; Harford and Liden,

1967-1968; Peterson and Liden, 1972), although 95 dB to 100

dB HL is considered the upper limit for normals because

auditory lesions outside the cochlea tend to raise the ART,

e.g., conductive problems or eighth nerve lesions (Brooks,

1971; Anderson, Barr, and Wedenberg, 1970).

As shown in Table 1, the ART in SL is rather uniform

from study to study, except for the technique using manome-

try. Weiss, et al. (1962) show ART levels which are 10 dB

to 20 dB less sensitive than those measured by acoustic

impedance. Inspection of Tables 1 and 2 does not indicate any

systematic effect of frequency in sensation level or in

hearing level in normal subjects. In Table 3 more sound

pressure level (SPL re 0.0002 dyne/cm ) is necessary to

elicit the ART at lower frequencies. In this respect, the

acoustic reflex is similar to the threshold of audibility

in response to SPL.

The range of ART levels may be due to various types of

acoustic-impedance instruments (e.g., Burke, Herer, and

McPherson, 1970,as shown in Table 2), threshold criteria

(Jerger et al., 1972 vs. Moller, 1961a, 1962a), a variety










Table 1. Acoustic-Reflex Threshold in dB SL



Sample Frequency in Hz
Sample ---------- ---------
Number 250 400 500 800 1000 1600 2000 3000 4000

Deutsch, 1968 30 ears 74 -- -- 82 -- 81

Jepsen, 1951 98 ears 83 -- 81 -- 80 -- 78 -- 76

Jepsen, 1963 88 ears 85 -- 81 -- 75 74 -- 80

Jerger et al., 1972 382 Ss -- 77 -- 78 77 -- 75

Lamb et al., 1968 19 Ss 79 -- 82 -- 82 -- 82 76 78

Weiss et al., 1962 10 Ss -- 93 -- 93 -- 97 -- 97*


*3200 Hz.














cn










Table 2. Acoustic-Reflex Threshold in dB HL (ANSI-1969)



Sample Frequency in Hz
Sample---------- ---1------
Number 250 500 1000 1500 2000 3000 4000

Anderson and Wedenburg, 1968 200 ears 85 88 86 -- 88 -- 92

Burke et al., 1970a 21 Ss -- 93 96 92

Burke et al., 1970b 21 SS -- 95 93 -- -- -- 89

Harford and Liden, 1967-1968 ? 92 95 88 84 82 87

Jerger et al., 1972 382 Ss -- 89 86 88 -- 86

Lamb et al., 1968 19 Ss 83 88 86 87 86 83

Melcher and Peterson, 1972 30 Ss -- 83 84 -- 83 -- 86

Moller, 1961a 2-5 Ss -- 85 -- 87 -- 81 --

Moller, 1962a 1-3 Ss 75c 84d 88e --

Peterson and Liden, 1972 88 ears 86 85 85 85 -- 85


belectroacoustic impedance bridge.
mechanical impedance bridge.
300 Hz.
525 Hz.
e1200 Hz.
3200 Hz.









Table 3. Acoustic-Reflex Threshold in dB SPL (re 0.0002 dyne/cm2)



Sample Frequency in Hz
Sample-----------------
Number 250 500 1000 1500 2000 3000 4000

Harford and Liden, 1967-1968 ? 117 106 95 -- 93 92 96

Jerger et al., 1972 382 Ss -- 100 96 -- 97 -- 95

Lamb et al., 1968 19 Ss 109 100 93 -- 96 96 93

Moller, 1961a 2-5 Ss -- 96 -- 93 -- 91 --

Moller, 1962a 1-3 Ss 100a 95b 95 --


a300 Hz.
525 Hz.
c1200 Hz.





19



of stimulus parameters (durations, interstimulus intervals,

intensity increments,and spectral components), and age of

the sample population (Jerger et al., 1972).

The standard deviation of the acoustic-reflex thresholds

in response to pure tones has been reported to have a range

of 6.4 dB (Jerger et al., 1972) to 9 dB (Deutsch, 1968).

Harford and Liden (1967-1968) list high Spearman rank-order

correlations for test-retest reliabilities for 250 Hz, 1000 Hz,

and2000 Hz and poorer results for 500 Hz, 3000 Hz, and 4000 Hz.

Moller (1962a) kept the acoustic reflex repeatability within

1.2 dB, but used a 10 percent change of maximum reflex con-

traction. The reported ART variability of 9 dB may be due,

in part, to the unknown physiological processes causing the

reflex at 250 Hz, 4000 Hz, and 6000 Hz to be absent or in-

consistently elicited, even in the presence of normal occur-

xring. response- to pure tones in the middle frequencies

:Deutsch, 1972; Fulton and Lamb, 1972; Jerger et al., 1972).

Itehas:been-reported .that complex stimuli of narrow band

mOcwhite noise= elicit- the- ART; at- lowers intensity levels than

pure-tone sinusoids.- Along- with the- increased- sensitivity

therB- is also- improvement of- threshold: stability (Dallos,

396-4s.Deutsch -1968,: 1972; Djupesland, Flottorp, and Winther,

1a66:Liily _1964; McRobert.- Bryan and Tempest, 1968; Moller,

1962b: -Peterson- and Liden,- 1970, 1972). -

195 There is gereral agreement that- the acoustic reflex is

significantly influenced by the energy level outside the





20



critical bandwidth. The ART is more or less constant as the

bandwidth is increased to a critical value. Beyond this

critical bandwidth the acoustic-reflex threshold will decrease

as the bandwidth increases (Flottorp, Djupesland, and

Winther, 1971; McRobert et al., 1968). Moller (1962b), in

contrast, did not see a consistent change in all subjects as

bandwidth was increased, but it is possible that he did not

extend his bandwidths far enough. Moller did obtain improve-

ment in threshold for one subject for which the bandwidth sur-

passed the critical width determined by Flottorp et al.

(1971) at 525 Hz fc. The work by McRobert et al. (1968)

established that the lowering of ART with increasing band-

width is dependent upon the center frequency, and is more

pronounced at 1000 Hz fc. Apparently, the critical bandwidth

with which the reflex mechanism responds to auditory stimuli

widens with increases in sound pressure level (Hung and

Dallos, 1972).

The acoustic-reflex threshold does not seem to differ

regardless of ascending or descending stimulus presentation

approach (Beedle, 1970; Harford and Liden, 1967-1968; Peterson

and Liden, 1970, 1972), although Deutsch (1968, 1972) reports

a systematic improvement of threshold over three test trials.

Jeutsch attributes the ART improvement of approximately 2 dB

from trial to trial to "auditory sensitization" (Hughes,

1954, 1957). Simmons (1960) labels this response condition

"post-tetanic potentiation" and attributes this sensitivity





21



to hyperexcitability of the brain-stem structures. These

results might also be ascribed to conditioned "learning,"

but Bates, Loeb, Smith, and Fletcher (1970) were unable to

condition the reflex.

Threshold of the acoustic reflex is normally defined as

the lowest stimulus level at which the reflex can be elicited

(Anderson et al., 1969, 1970; Beedle, 1970; Deutsch, 1968;

Djupesland and Zwislocki, 1971; Peterson and Liden, 1972;

and others). This minimal change in muscle contraction is

unacceptable to Moller (1962a), who uses 10 percent change

of the maximum impedance change. His criterion keeps the

ART reliability within 1.2 dB while reliability deteriorates

as the minimal detectable reflex is approached. In addition,

the ART has been elicited in successive steps of 1 dB

(Djupesland and Zwislocki, 1971), 2 dB (Lamb et al., 1968)

and 5 dB (Jerger et al., 1972), which might account for some

of the reported threshold level differences.

Stimulus duration, another variable affecting the

acoustic-reflex threshold, is reported by Lorente de No

(1935) as having an effect on the tensor tympani response in

rabbits. The strength of the muscle contraction increases

as a function of increasing stimulus duration with the

stimulus level held constant. Further studies have shown

that the acoustic-reflex threshold in SPL becomes increasingly

lower (more sensitive) as the signal duration is increased

to about 200 msec (Djupesland and Zwislocki, 1971; McRobert,





22



Bryan,and Tempest, 1968; Moller, 1962b; Simmons, 1963;

Weiss, Mundie, Cashin, and Shinabarger, 1962). A 500-msec

duration tone approximates the results of longer stimuli for

eliciting a steady change in impedance at the tympanic

membrane. With regard to the auditory response to long

duration signals, Dallos (1964) and Harford and Liden (1967-

1968) have observed that adaptation takes place only over

longer periods of time, e.g., 15 seconds to 30 seconds. This

adaptation is thought to be an afferent process, because

recovery is instantaneous upon changing to a new test fre-

quency. This afferent adaptation is also frequency-dependent,

with higher frequencies demonstrating more response decay

(Anderson et al., 1969; Melcher and Peterson, 1972).

The ear in which the reflex is recorded can also influ-

ence the level of the acoustic-reflex threshold. Recordings

made in the stimulated ear, instead of the contralateral ear,

improve threshold sensitivity (Moller, 1961b). Bilateral

stimulation improves the threshold value even more (Moller,

1962b).

The frequency of the recording probe tone can in itself

influence the acoustic-reflex values. Peterson and Liden

(1972), averaging the ART for 500 Hz and 4000 Hz, state that

a 220-Hz probe tone is about 6 dB more sensitive than a 625-

Hz probe tone in eliciting the ART, while a 800-Hz tone is

between them. In an earlier study, Harford and Liden (1967-

1968) recorded an unsystematic 2-dB threshold difference





23




between a 220-Hz and a 800-Hz probe tone frequency. This

latter study does not agree with the findings of Mehmke and

Tegtmeier (1970) in which there should be a 8 dB loss in

transformer efficienty for low frequencies. Lilly and

Shepherd (1964) and Feldman (1967) have also observed that

the acoustic impedance varies with the frequency of the probe

tone.

The choice of probe tone frequency may be dependent upon

the length of the sinusoidal wave and its relation to the

dimensions of the external auditory canal. Acoustic impedance

becomes more sensitive to changes in probe position in the

external auditory meatus and to increased diameter of the

meatus as the probe frequency increases. This may be mini-

mized by using a low frequency probe in a larger volume, e.g.,

at the external meatus (Schoel and Arnesen, 1962). This con-

tention is not supported by Djupesland et al. (1966), who

see no effects of probe position. Djupesland and his asso-

ciates used 5-dB test increments, which may have obscured any

small but significant results of plug position.

It might be possible that the attributed frequency ef-

fects of the'probe tone are due to the varied, and often

unreported, intensity levels of the probe tone. It is impor-

tant that the probe tone not be intense enough to evoke the

acoustic reflex, since it is there to indicate the existence

of intra-aural muscular contraction and not to elicit it

(see footnote 4). The probe tone level must also be above





24



the background environmental and physiological noise so that

the background noise does not cause spurious results. In

order to meet these two criteria a larger dynamic range is

available with a low frequency probe tone. The fact that low

frequency tones elicit the ART at higher intensity levels can

be seen in Table 3. It is not possible to know whether the dif-

ference in ART levels is a real effect of probe frequency or

one due to the intensity level of the probe tone (Lilly and

Shepherd, 1964; Terkildsen, Osterhammel, and Scott Nielsen,

1970).

Acoustic-reflex dynamics of magnitude and latency

The intra-aural muscle response had also been observed in

terms of how the reflex changes as the stimulus intensity is

increased above threshold. The response magnitude of the

acoustic reflex grows as a function of intensity to approxi-

mately 30 dB above the acoustic-reflex threshold (30 dB ART

SL) (Dallos, 1964). Most of this growth occurs within 16 dB

ART SL (Djupesland et al., 1966) at a near linear rate (Dallos,

1964; Moller, 1962a; Peterson and Liden, 1970; Weiss et al.,

1962) with no observable increases to pure tone stimuli be-

yond 120 dB SPL (Hung and Dallos, 1972). Low frequency tones

cause the AR to grow at a faster rate than high frequency

tones (Djupesland, et al., 1966; Harford and Liden, 1967-

1968). There does not seem to be agreement whether one par-

ticular frequency or noise band causes a larger change in

the reflex (Djupesland et al., 1966; Fisch and Schulthess,

1963; Johansson, Kylin, and Langfly, 1967).





25



Latency, the period from signal onset to contraction,

and contraction time, the interval from stimulus onset to

maximum contraction, were some of the earliest acoustic re-

flex characteristics studied (Loeb, 1964). Latency decreases

as the intensity is increased above the ART (Dallos, 1964;

Fisch and Schulthess, 1963; Perlman and Case, 1939). The

shortest latency of the stapedius muscle and tendon is about

10 msec to 15 msec in man (Fisch and Schulthess, 1963; Neer-

gaard and Rasmussen, 1966; and Perlman and Case, 1939), while

the tensor tympani has a comparatively longer latency of 50

to 120 msec. There is considerable intra- and inter-subject

variability in latency (Moller, 1958). This is due, in part,

to the obscurity of the contraction close to threshold, but

the shorter the latency the more reliable the trace (Neergaard

and Rasmussen, 1966). The latencies of the ART are among the

shortest for muscle reflexes in man. Fisch and Schulthess

(1963) conclude from an EGM study that this short latency,

especially for stapedial contraction, is probably due to the

limited number of synapses in the crossed acoustic-reflex arc.

The longer latency obtained from the tensor tympani-reflex

arc may indicate that it has additional synapses (viz. Figure

1).

The maximum impedance change may be reached within 400

msec to 500 msec after signal onset. Some contraction may be

present as long as one second after signal offset. Djupes-

land and Zwislocki (1971) contend that the growth and decay





26




of the muscle reflex is symmetrical. Decay from maximum con-

traction is apparently independent of intensity and frequency.

Although the importance of the acoustic reflex can be

obscured by the many factors affecting it, consistent thres-

hold levels have been obtained in a number of studies. The

consistency of the acoustic-reflex response contributes to

its utility as a clinical tool.


Clinical application of the acoustic reflex

Metz (1946) introduced acoustic impedance measurement as

a clinical tool. He included the presence of the acoustic

reflex and the level necessary to just elicit this reflex as

one of the useful dimensions of acoustic impedance. The

absence of the reflex can support the inference that a middle-

ear problem exists in the ear under test. Terkildsen and

Scott Nielsen (1960) and Klockhoff (1961) presented clinical

cases showing that a relatively normal middle ear is necessary

to elicit an intra-aural reflex. In other words, some authors

feel that the presence of a reflex is indicative of a normal

middle-ear (Feldman, 1967; Klockhoff, 1961). This is not

necessarily true, however. Brooks (1969, 1971) states that

a minor conductive component will not abolish the reflex but

instead elevate the acoustic-reflex threshold. He concludes,

therefore, that only subjects exhibiting a reflex to 95 dB

HL (ISO-1964) or less in the contralateral ear should be

regarded as having normal middle-ear function.





27




The manometer system in an electroacoustic impedance

bridge is able to approximate the amount of pressure in

the middle-ear. For the most part, 50 mm of equivalent

water pressure (re atmospheric pressure) is considered within

normal limits. Hearing and acoustic-reflex thresholds do not

deteriorate within this pressure range in the middle-ear

(Alberti and Kristensen, 1970; Peterson and Liden, 1970).

Negative middle-ear pressure is more detrimental to thresholds

than positive pressure (Moller, 1965), but the AR may be

stronger with slightly negative canal pressure (Terkildsen,

1964). Terkildsen (1960c) even states that the stapedial

reflex is found around normal atmospheric pressure, while the

tensor reflex is enhanced by negative pressure.

In the absence of middle-ear conductive problems, the

level at which the AR is just elicited above the threshold of

audibility is useful in confirming a cochlear abnormality.

It has been observed that in the majority of mild to moderate

sensorineural hearing losses the acoustic-reflex threshold

occurs at approximately the same hearing-threshold levels (HL)

as the normal ART (Alberti and Kristensen, 1970; Ewertsen,

Filling, Terkildsen, and Thomsen, 1958; Jerger et al., 1972;

Klockhoff, 1961; Kristensen and Jepsen, 1952; Lamb, Peterson,

and Hansen, 1968; Metz, 1946, 1952; Peterson and Liden, 1972;

Terkildsen, 1960b; Thomsen, 1955a,b; and others). If the

ART occurs at 55 dB to 60 dB SL or less, the hearing loss is

considered to be due to cochlear impairment. The ART is





28




seldom seen below 25 dB SL and never less than 5 dB to 10 dB

above pure-tone threshold (Jerger, 1970; Jerger et al., 1972;

Lamb and Peterson, 1967; Lamb, Peterson, and Hansen, 1968).

If a reflex is obtained at less than 5 dB SL it can be

assumed that the pure-tone hearing loss is due to non-organic

causes (Jepsen, 1953, 1963; Lamb, Peterson, and Hansen, 1968;

Terkildsen, 1964; Thomsen, 1955b).

Liden (1970) demonstrated that the intra-aural reflex

can be used as an objective loudness-recruitment test. He

compared 52 ears with unilateral Meniere's syndrome, 30 ears

of athetoid children and 9 ears presenting acoustic tumors.

The patients were divided into three groups according to the

level of the reflex thresholds and magnitude of separation

between their threshold of audibility and their ART. If the

ART exceeded 95 dB HL (ISO-1964), or if the intensity span

between the pure-tone threshold and the ART was less than

75 dB, the response was considered abnormal. These two values

correspond, respectively, to theninetieth and tenth percentiles

on 88 normal-hearing control subjects.

Liden maintains that a reduced intensity span between

the ART and the pure tone threshold is a result of a cochlear

lesion, and an elevated ART represents a probable higher

order lesion. A reduced span superimposed on an elevated ART

represents both areas as foci of the lesion. For the most

part, his three pathologic groups support this contention,

even though his span of 75 dB for normal hearing is 10 dB to





29




15 dB more c vaervative than the other reported studies. As

further sup11 't1 he obtained pre- and post-noise exposure

thresholds, pure tone and reflex, on 11 cats. The results

can be seen in Figure 3. An average permanent threshold shift

of 44 dB resulted with an elevation of the ART by only 1 dB

from the pre-exposure levels. The only noticeable effect upon

the ART was a slight increase in the standard deviation across

the four reflex eliciting frequencies. It appears, therefore,

that the acoustic reflex is a loudness-sensing mechanism that

can be used clinically to indicate the possible presence of

loudness recruitment.

To further authenticate the meaning of the acoustic-re-

flex level in cochlear-impaired persons, comparisons have

been made with loudness-balance tests and other audiometric

indicators of cochlear lesions. In most cases a comparison

is made with a unilateral end-organ disease and results of

the ABLB, as the standard of loudness recruitment. The re-

sults indicate that the ART level above the threshold of

hearing is as good as the ABLB and slightly better than the

SISI for determining an end-organ lesion (Alberti and

Kristensen, 1970; Ewertsen et al., 1958; Kristensen and

Jepsen,. 1952; Lamb et al., 1968; Liden, 1970; Thomsen, 1955a;

and others). It is definitely a better and more reliable

predictor of sensory damage than the Monaural Bi-frequency

Loudness Balance, the Difference Limen for Intensity or

Frequency, the Uncomfortable Loudness Level, and Bekesy Types





30











7 120-
u 110 -
M 100 -
S90 PRE-NOISE
Q 80 POST-NOISE
N 70-
a 0 60 \
00
So50 --
o 0 40 *.-
S30-
020--
w 10-
0
H 10
S-20 -
o -30-
0.1 kHz 1 kHz 10 kHz

TEST FREQUENCY


Figure 3. Pre- and post-noise exposure thresholds.
The lower thresholds are audibility and the upper thresh-
olds are acoustic reflex in 11 cats. (Redrawn from Liden,
1970.)





31



(Ewertsen et al., 1958; Niemeyer, 1971). Coles (1972), how-

ever, argues that the ART, like the Uncomfortable Loudness

Level, is a vague concept and is only diagnostically useful

when absent.

Beedle (1970) doubts that the ART is indicative of end-

organ damage. If, in fact, the ART is a measure of loudness

recruitment, it does not continue to demonstrate this rapid

growth in loudness above the ART. Instead, the reflex growth

function is less steep and slower than in normal ears.

(Also see Beedle and Harford, 1971; Peterson and Liden, 1970,

1972.)

As the hearing loss increases, the level of the acoustic-

reflex threshold also increases but not proportionately (Liden,

1970). Sensorineural hearing losses beyond 70 dB to 80 dB

HL (ISO-1964) do not normally demonstrate an acoustic reflex

at any intensity (Jerger, 1970; Jerger et al., 1972; Terkild-

sen, 1960b). Jerger makes use of this fact to predict the

probability of the hearing level at the threshold of audi-

bility. For example, in the presence of a reflex, there are

5 chances in 10 that the loss does not exceed 85 dB HL (ISO-

1964) and there is only 1 chance in 10 that it is as much as

100 dB HL. This is a nebulous approach at best, because it

lacks quantification for specific cases.

In summary, the acoustic reflex is a well-defined physio-

logical phenomenon which has diagnostic value and gives ob-

jective information. Its clinical value and true objectivity





32




occurs when the involuntary acoustic-reflex threshold is

compared with the threshold of audibility. When the dynamic

range between the two thresholds is 55 dB or less, there is

cochlear impairment. By independently varying the signal

duration at threshold of audibility, more diagnostic informa-

tion is gained. This manipulation of auditory processing, as

seen in the next section, may be applied to the acoustic re-

flex threshold as well.


Temporal Summation

There is a psychophysical assumption that the normal

auditory system can summate, or integrate, acoustic power over

some critical period of time. Since the initial work of Garner

(1947b), Garner and Miller (1947), Hughes (1946), Munson (1947),

and de Vries (1948), considerable interest has been generated

over the concept of "trading" increased acoustic power with

a decreased signal duration to maintain a constant loudness

level. When the on-time of the auditory stimulus is sequen-

tially halved, starting from some critical duration, the

signal power is reduced 2 dB to 3 dB with each decrement, i.e.,

from 200 msec to 100 msec, from 100 msec to 50 msec, etc.

The critical long duration is thought to be about 200 msec

and a linear relationship holds to a critical short signal

length of approximately 10 msec.


Perfect integration

If one assumes perfect integration (Garner, 1947a,b;





33



Garner and Miller, 1947), the ear would trade 1 log unit of

intensity (1 bel, 10 dB) for 1 log unit of duration (10 msec

to 100 msec, 20 msec to 200 msec, 15 msec to 150 msec). Even

though there is an increase of power as stimulus time in-

creases, the acoustic energy required to maintain the thres-

hold of audibility is relatively constant.6 This accumulation

of power over an average tenfold decrease in time may be

inferred from threshold data using a single-slope value (Clack,

1966; Garner and Miller, 1947; Munson, 1947; Northern, 1967;

and others). When this slope value is 10 dB it is considered

perfect temporal integration of energy. Perfect integration

is most often observed for the sinusoid of 1000 Hz.

The time constant of temporal integration (To) marks the

point at which time and intensity cease to have a linear

relationship; it is thought to occur between 200 msec to 250

msec. Goldstein and Kramer (1960) observe integration occur-

ring through 200 msec, but Harris, Haines, and Myers (1958)

report individual time constants ranging from 100 msec to

300 msec with a mean of 200 msec. Plomp and Bouman (1959) and

Hempstock, Bryan, and Tempest (1964) indicate that To is

inversely related to frequency; To changes from about 375 msec

at 250 Hz to approximately 150 msec at 8000 Hz. Regardless


Acoustic energy can be stated in the simplest form as
E = PT, where E is acoustic energy, P is acoustic power, and
T is linear for a certain period of time, the time constant.
Further increases in signal duration beyond the time constant
have less and less effect upon threshold level.





34




where To may fall, integration is essentially complete between

500 msec and 1000 msec (Zwislocki, 1960).

The phenomenon of integration of acoustic power over

time is attributed to temporal summation in the auditory

system and, most probably, is neural in nature (Zwislocki,

1960). The term summation is introduced by some authors be-

cause it is felt that the response pattern represents tem-

poral summation at the synaptic junctions in the neural path-

ways. More specifically, this apparent linear function of

log time and log power is due to an exponential decay of the

persisting neural excitation (Plomp and Bouman, 1959; Zwis-

locki, 1960). At threshold, there appears to be a direct

proportionality between an increase in the intensity of

acoustic power and the increase in neural excitation, which

is modified somewhat at suprathreshold levels by neural adap-

tation and the loudness of the stimuli (Zwislocki, 1960).

The neural mechanism for temporal integration probably exists

at the neurons above the first and, possibly, the second order,

but before the level of binaural interaction at the superior

olivary complex (Zwislocki, 1960).


Stimulus parameters

The literature indicates that, in addition to duration,

other various stimulus parameters affect temporal integration.

Those most often cited are stimulus spectrum, rise and decay

time, definition of the signal duration, inter-stimulus






35




interval, attenuation rate, method of threshold calculation,

and psychophysical procedure.

Earlier studies indicate that for a narrow band of

frequencies between 1000 Hz and 4000 Hz a 10 dB/decade slope

is generated (Garner, 1947a; Garner and Miller, 1947; Hughes,

1946; Munson, 1947), while narrow-or wide-band noise bursts

demonstrate about a 8 dB/decade change in threshold inte-

gration (Garner, 1947b; Miller, 1948; Small, Brandt, and

Cox, 1962). There are no systematic differences in the slope

between threshold determination in quiet (Garner, 1947b;

Hempstock, Bryan, and Tempest, 1964; and others) and in back-

ground noise (Garner and Miller, 1947; Blodgett, Jeffress, and

Taylor, 1958; Hempstock et al., 1964; Gengel, 1972; Plomp

and Bouman, 1959; and others).

More recently, a controversy concerning frequency effects

has developed. While 1000 Hz elicits an average slope of 10

dB per decade, lower frequencies have a steeper slope and an

increase in frequency produces a flatter slope, as seen in

Figure 4 (Brahe Pedersen and Elberling, 1972; Elliott, 1963;

Gengel, 1972; Gengel and Watson, 1971; Hattler and Northern,

1970; Hempstock et al., 1964; Miskolczy-Fodor, 1953; Northern,

1967; Olsen and Carhart, 1966; Sanders, Josey, and Kemer,

1971; Sheeley and Bilger, 1964; Simon, 1963; Tempest and

Bryan, 1971; Watson and Gengel, 1969). Some investigators

feel that there is no frequency dependence (Clack, 1966;

Martin and Wofford, 1970; Wright, 1968a, 1972; Zwicker and





36











14 250o

0 500,\
0 12
4 125'
0 1000
OH 10

S0u 2000 \
8.-


a o 4000 *



0 2 "0
4






16 32 64 128 256 512 1024
SIGNAL DURATION IN MSEC


Figure 4. Average temporal integration slopes as a
function of frequency. In general, as frequency decreases
the slope steepens. (Redrawn from Watson and Gengel, 1969.)





37



Wright, 1963). Inter-subject variability may cause the

difference in frequency dependence (Clack, 1966). Another

possibility is that frequency dependence may be observed at

shorter durations but not necessarily at the longer signal

on-times (Olsen and Carhart, 1966). Price (1972) has also

suggested that the external ear may cause a transformation of

the stimulus parameters in some subjects by raising the

energy peak in frequency and by legthening short pulses,

thereby possibly nullifying the effects of frequency.

Frequency dependence is, in part, related to the type

of psychophysical procedure used. Bekesy tracking and forced-

choice tracking show little or no frequency dependence,but

the frequency effect is demonstrated by the method of adjust-

ment, method of limits, method of constant stimuli, and

confidence ratings (Bilger and Feldman, 1969; Chamberlin and

Zwislocki, 1970; Watson and Gengel, 1969). In the majority

of studies Bekesy tracking is the method used.

Even though there is large individual variability in the

value of the slope (Clack, 1966; Gengel, 1972; Green et al.,

1957; Hattler and Northern, 1970; Martin and Wofford, 1970),

there is general agreement that the test-retest variability

is good (Doyle, 1970; Hattler and Northern, 1970; Olsen and

Carhart, 1966, Plomp and Bouman, 1959). Gengel and Watson

(1971) suggest at least 12 threshold crossings for each

data point when using Bekesy tracking in order to achieve a

reliable reading.





38



Spectral characteristics of the signal as a function of

duration cannot be divorced from the integrating bandwidth,

or the critical band of the ear (Fletcher, 1940; Scharf,

1970). The theoretical bandwidth of a pulsed tone is pro-

portional to the reciprocal of the signal's duration, or

1/t. As long as the spectrum of the test tone remains within

the critical bandwidth of the ear, all of the energy will be

integrated. If not, there should be a loss in sensitivity

(Garner, 1947a; Green et al., 1957; Olsen and Carhart, 1966;

Sheeley and Bilger, 1964; Wright, 1968a). The spread of

energy by abrupt low frequency signals upward to the more

sensitive frequencies of the ear has been misinterpreted as

an apparent increase of sensitivity within the critical band.

It has resulted in the assumption that the ear does not inte-

grate low-frequency tone pips (Campbell and Counter, 1969;

Karlovich, Lane, Smith, Tarlow, Thompson, and Vivion, 1971).

The point is that one must carefully monitor the test stimuli

so that the duration of the tone burst is maintained within

the critical bandwidth (Wright, 1968a).

Some variability in temporal integration has stemmed

from lack of defined criteria of stimulus duration. Goldstein

and Kramer (1960) measured the duration between energy onset

and cessation, while Harris (1957) designated as criteria

the half-power points on the stimulus envelope. An "equiva-

lent duration" has also be used (Brahe Pederson and

Elberling, 1972; Dallos and Johnson, 1966; Dallos and Olsen,





39



1964; Olsen and Carhart, 1966). An equivalent-duration tone

pip contains the same amount of energy as a rectangular

envelope and allows for the comparison of widely varying

envelope shapes. The use of the equivalent duration fits or

allows the conversion of the overall stimulus envelope to fit

the slope of Garner's model (1947b), which states that not all

energy is used by the ear to summate temporal information

(Dallos and Olsen, 1964). The best measure of the acoustic

waveform is still a moot question because the conversion from

the half-power points to equivalent duration of Harris' data

(1957) by Dallos and Olsen (1964) demonstrate no difference.

The time in which the stimulus envelope rises to and

decays from its effective peak is a critical variable in

auditory measures using pure-tone stimuli. If the stimulus

starts or stops too abruptly, a wide-band transient may be

generated, thereby biasing the results (Wright, 1958, 1968a).

A minimum 5 msec rise/fall time measured on the linear por-

tion of the ramp (between 10 percent and 90 percent of maximum

amplitude of the acoustic waveform) is most often suggested

(Harris, 1957; Wright, 1960, 1968a). No differences occur

with rise/fall times varying from 0 msec to 50 msec if the

equivalent duration is held constant (Dallos and Johnson,

1966; Dallos and Olsen, 1964).

The inter-stimulus interval, or the off-time between

successive stimulus envelopes, must be kept long enough to

ensure neural independence between stimulus events. Since





40



the theory of temporal summation is based on neural decay,

the minimum off-time should be 200 msec (Zwislocki, 1960).

In addition, a "critical off-time" has been calculated where-

by successive stimuli having less than a 200 msec inter-

stimulus interval take on the threshold value of one contin-

uous tone (Jerger, Jerger, Ainsworth, and Caram, 1966).

Doyle (1970) in a temporal integration study varied the inter-

stimulus interval from 100 msec to greater than 500 msec.

There were no differences in temporal-integration results as

long as the dead time was a minimum of 500 msec, but the

slopes flattened as the off-time was shortened to 100 msec.

In practice, therefore, the inter-stimulus interval should

exceed the theoretical minimum, and probably be no less than

500 msec.

The repetition rate of the signal per unit of time,

usually one second, is also based on consideration of decay

of neural excitation. If one desires to keep the repetition

rate constant, then the rate is determined by the length of

the longest-duration test signal and the desired inter-stimulus

interval. Wright (1968a) recommends the use of the same

repetition rate for all test stimuli so that sampling per unit

of time will be uniform.

Some attention has been given to attenuation rate for

those test procedures employing semi-automated equipment, for

example, Bekesy tracking. Hempstock et al. (1964) observed

an increase in the standard deviation of the threshold as the





41



number of pulses presented at each discrete level decreased

from eight pulses to one pulse. There was no significant

difference in the mean threshold. Wright (1968a) suggests

using an attenuation rate of 2.5 dB/second at a repetition

rate of 1/second as a standard.

The choice of ascending, descending,or mean thresholds

can give threshold values differing as much as 3 dB at any

one signal duration (Hempstock et al., 1964). Therefore,

for the majority of the cases, where applicable, mean values

determine threshold. The vast majority of recent publications

dealing with temporal integration have used the Bekesy track-

ing procedure, where the averaged midpoint of the crossings

is considered to be the threshold value.


Effects of end-organ lesions
upon temporal integration

Considerable interest has been generated in the use of

temporal summation as a diagnostic tool for end-organ patholo-

gies; this is otherwise known as brief-tone audiometry. A

cochlear lesion seemingly increases the integration ability

of the ear. As the duration is decreased,less acoustic power

is needed to maintain constant energy for threshold (Brahe

Pederden and Elberling, 1972, 1973; Doyle, 1970; Elliot,

1963; Gengel and Watson, 1971; Gengel, 1972; Harris et al.,

1958; Hattler and Northern, 1970; Jerger, 1955; Martin and

Wofford, 1970; Miskolczy-Fodor, 1953, 1956, 1960; Nerbonne,

1970; Olsen, Rose, and Noffsinger, 1973; Sanders and Honig,





42



1967; Sanders, Josey, and Kemer, 1971; Wright, 1968b; Wright

and Cannella, 1969). The integration slope is not affected

by either conductive problems (Miskolczy-Fodor, 1956; Harris

et al., 1958; Wright, 1968b; Wright and Cannella, 1969) or by

retrocochlear problems (Olsen et al., 1973; Sanders et al.,

1971). The critical time at which the temporal integration

begins, or the time constant of integration, also seems to

be shortened to about 100 msec or less depending on the par-

ticular frequency (Miskolczy-Fodor, 1953, 1956, 1960; Harris

et al., 1958; Sanders and Honig, 1967; Wright, 1968a). The

abnormal results of cochlear damage to slope and time con-

stant can be seen in Figure 5, i.e., the slope is flatter,

the time constant is shorter, or both.

It has also been suggested that brief-tone audiometry

may even be so sensitive as to detect incipient damage to the

cochlea (Campbell and Counter, 1969; Sanders and Honig, 1967;

Wright, 1968a) to detect recruitment (Miskolczy-Fodor, 1956,

1960), and to differentiate between various cochlear

lesions (Harris et al., 1958). Invariably, these attributes

are artifacts of unaccounted frequency effects, energy spread,

and/or inter-subject variability (Gengel and Watson, 1971;

Karlovich et al., 1971). Olsen et al. (1973) cast doubt on

the efficacy of temporal integration to differentiate be-

tween end-organ and eighth nerve lesions because of the large

overlap between populations.





43











15




S- Normal hearing
-------- Cochlear-impaired hearing
S10-














16 32 64 128 256 512 1024
5-









SIGNAL DURATION IN MSEC


Figure 5. Examples of temporal integration. Cochlear-
impaired integration functions can be flatter, can have a
shorter time constant, or both. (Redrawn from Gengel and
Watson, 1971.)





44



Nerbonne (1970), who studied temporarily fatigued ears,

and Sanders et al. (1971), who studied cochlear lesions,

demonstrated that the SISI and ABLB, as well as brief-tone

audiometry are sensitive indicators of cochlear lesions.

Brief-tone audiometry can yield definitive results when the

ABLB shows partial recruitment and negative SISI scores are

present (Sanders et al., 1971). There is some correspondence

between amount of recruitment and degree of aberration of

temporal integration,but, for the most part, temporal inte-

gration is assessing some other aspect of sensory lesions

(Nerbonne, 1970). Abnormal temporal integration seems to be

part of a syndrome of a distorted time domain of enlarged

critical bands and critical ratios (Northern, 1967; Sheeley,

1963; Simon, 1963) and poorer performance on difference limens

for frequency (Sheeley, 1963). Temporal integration, by way

of its relation to critical bands, seems to have a common

basis with several phenomena related to the tuning of the

auditory system (Licklider, 1951) that are not presently

measured clinically. The method of brief-tone audiometry,

therefore, presents clinicians with another possible test

for end-organ abnormalities, if proper standards are set

forth for the various stimulus and test parameters, and nor-

mal response limits are determined.















CHAPTER II

STATEMENT OF THE PROBLEM


Innumerable studies detailing the identification and

quantification of cochlear abnormalities indicate that, with

patients who cannot or will not cooperate fully, the present

diagnostic measures are inadequate in obtaining information

on auditory processing. Further understanding of the func-

tion of the acoustic-reflex threshold and the effect of

temporal integration upon threshold measures may fill this

void in diagnostic testing of the auditory system. Such a

measure may be obtained by maintaining a constant reflex

strength as a function of increasing acoustic power while

signal duration is decreased, or, in other words, temporal

integration of the acoustic reflex. Since there are rela-

tively few data concerning temporal integration of the acous-

tic reflex, the pertinent literature will now be reviewed.

As early as 1935, Lorente de No reported the effects of

varying short signal duration upon the contraction of the

tensor tympani muscle of the cat. He used as much as 140

dB SPL and durations from 3 msec to 100 msec. He explained

his results in terms of neural summation, which occurs at the

synapse. At some synapses one pre-synaptic impulse is not


1For further information regarding synaptic summation
see the classical treatise of Sherrington (1906, 1947);
also see Brink (1951).
45





46



enough to initiate a post-synaptic impulse, it is therefore

necessary to sunmate several incoming impulses, i.e., spa-

tial and temporal summation. Spatial summation occurs as a

result of neural impulses converging simultaneously at a

synapse but from different neural fibers. Since an increase

in the number of neural fibers transmitting impulses is

thought to be a method of coding increased stimulus intensity,

spatial summation can also be in response to increased inten-

sity. Temporal summation results from successive impulses

reaching the synapse through the same fibers. With the tem-

poral form of neural summation, the post-synaptic potential

occurs as the signal duration increases. It is the amount of

information that each fiber carries and the number of fibers

invoked which makes the difference in the type of summation.

Lorente de No has suggested, therefore, that varying the

signal duration changes the strength of the acoustic reflex

by virtue of temporal summation.

Simmons (1963) has also demonstrated temporal summation

of the intra-aural muscles in cats to acoustic stimulation.

His results are explained in terms of an on-response in the

auditory system in which there is linear integration of

acoustic power from 5 msec through 50 msec at a rate of 8 dB

per doubling of duration. There is only a 2-dB acoustic-

reflex threshold improvement from 50 msec to 100 msec which

indicates a time constant of 50 msec. In comparison to

temporal integration at threshold of audibility in man, the





47



time constant is about 200 msec and the threshold changes

2 dB to 3 db per doubling of signal duration.

There are relatively few data on temporal integration

of the acoustic-reflex threshold in humans. The systematic

published works are by McRobert, Bryan, and Tempest (1968)

and Djupesland and Zwislocki (1971). In both instances data

are plotted in terms of acoustic-power growth as signal dura-

tion decreased while maintaining acoustic-reflex threshold.

Their slopes, reported in power to maintain reflex threshold

per average decade change in signal duration, are in the

range reported by Simmons (1963). The slopes range from 14

dB/decade for 300 Hz, to 21dB/decade for 2500 Hz (McRobert

et al., 1968), and to 25 dB/decade for 1000 Hz (Djupesland

and Zwislocki, 1971). Although there are no reports of a

systematic frequency effect (McRobert et al., 1968; Simmons,

1963), the middle frequencies apparently have steeper slopes.

Figure 6 represents the effects of temporal integration

upon the threshold of the acoustic reflex in six normal hear-

ing subjects by Djupesland and Zwislocki (1971). The median

of the individual threshold means demonstrates a regular

decrease of 5 dB to 7 dB per halving of duration, or a slope

of about 25 dB/decade. The time constant is seen to be, for

the most part, about 200 msec. One subject appears to begin

integration at 100 msec with no obvious change in slope, as

seen by the dashed line connecting the lowest mean ART data.

McRobert et al. (1968) also present slopes which break

linearity between 100 msec and 200 msec.






48









u
: 60 '''""| .. i i 'i.. ...i.. i 11111

0 50 "
---- Median of 6 Ss
m 40- \ --- Lowest Mean ART


O Q "
o 20 -
2 03 2.0 .-
2C0
W M 10


... ...... .......I I .

10 100 1000 10000

SIGNAL DURATION IN MSEC


Figure 6. Temporal summation of the acoustic reflex.
The solid line intersects the median of six individual mean
scores resulting in a slope of 25 dB/decade of signal dura-
tion change. The dashed line indicates a similar slope but
a shorter time constant. (Redrawn from Djupesland and
Zwislocki, 1971.)





49


No explanation appears to be offered for the steeper

slope of the acoustic reflex, but the temporal summation ef-

fect is there without question as illustrated in Figure 7. The

range of values reported by Olsen, Rose, and Noffsinger (1973)

for temporal integration at threshold of audibility are con-

trasted with the data by Djupesland and Zwislocki (1971) for

temporal integration at threshold of the acoustic reflex.

These are both at 1000 Hz in normal ears and have been normal-

ized to 200 msec so that the slope has a common reference be-

tween the two sets of data. The slope for the acoustic-reflex

threshold at 25 dB/decade is about three times as large as

the median slope of 8 dB/decade at threshold of audibility.

Other authors have reported effects of short stimulus dura-

tions upon the acoustic reflex, but their results are not

reported systematically as temporal integration, nor in a form

easily converted to the values presented in Figures 6 and 7

(Johansson, Kylin, and Langfly, 1967; Lilly, 1964; Moller,

1962b; Weiss, Mundie, Cashin, and Shinabarger, 1962).

There have been no published reports dealing with the

effects of temporal summation of the acoustic-reflex thres-

hold in hearing-impaired ears, but there are indications of

disturbed spatial summation. Beedle (1970) and Beedle and

Harford (1971) have reported finding an effect of stimulus

intensity upon the growth to the acoustic reflex. The slope

of the reflex growth is much steeper and more rapid for nor-

mal ears than for either ear of an unilateral Meniere's group.

Niemeyer and Sesterhenn (1972) have noted the occurrence of





50













Acoustic Reflex
Threshold
-- Median for Acoustic
Reflex Threshold
o Threshold of Audi-
+30 bility
Median for Audibility
\* Threshold

0.
a +20 -


0 Q0
O 0
d O o
w oe
2 Cooooo o \^
S0 +10 0 o0 0 0
+0P 00000 0. 000000

0000 0 Y
oo oooo
ooooooooo
o* o o-
S 0 -*ooooo -o






I I I
10 20 100 200

SIGNAL DURATION IN MSEC


Figure 7. Temporal integration at threshold of audi-
bility and of the acoustic reflex for normal-hearing subjects.
The median slope of the acoustic reflex is about three times
the change in the median slope of audibility. (Based on
Djupesland and Zwislocki, 1971, and Olsen et al., 1973.)





51



a difference in reflex eliciting intensity levels for pure

tones versus white noise. The intensity level difference

between the two types of stimuli is much larger in normals

than the range in cochlear-impaired ears.

Norris, Stelmachowicz, and Taylor (1972) have reported

the most significant difference in the acoustic-reflex action

between normal and pathological cochleas. They measured the

difference in levels necessary to obtain the acoustic-reflex

threshold. The stimulus is pulsed at 2.5 pulses per second,

a 50 percent duty cycle with a 200-msec signal duration.

The resulting modulation of the reflex in response to the

pulsed tone is greater and more regular in normal ears than

in sensorineural hearing losses. This may reflect an in-

creased latency of acoustic-reflex response in cochlear-

impaired ears (Johansson et al., 1967; Simmons, 1963) or a

distorted spatial summation involving rate of reflex growth

with increased stimulus level (Simmons, 1963).

If effects due to disturbed spatial summation are present

in the acoustic reflex as a result of cochlear pathologies, it

is reasonable to assume that end-organ lesions can also

disturb temporal summation. Cochlear lesions can and do

disturb temporal integration at threshold of audibility, that

is, by flattening the slope of temporal integration and/or

shortening the critical duration. Intuitively, then, tem-

poral integration of the acoustic reflex may also be affected

by cochlear lesions, for example, a flattened slope and/or

a shortened critical duration.





52



In summary, temporal summation is a neural event origi-

nating at the cochlea and is essentially complete before the

level of the superior olivary complex (SOC), where the

acoustic reflex is activated. Growth in neural activity in

the SOC parallels the growth in loudness up to 60 dB SPL.

This neural activity decays in an exponential manner, thereby

appearing to trade energy units linearly for about 200 msec.

It is suggested that during acoustic stimulation, neural activ-

ity in the reflex center exceeds a certain level and causes

a contraction of the intra-aural muscles. Furthermore, the

mechanism suggested as controlling and/or affecting temporal

summation of threshold and loudness is the filtering charac-

teristics of the ear. Lesions in the cochlea can modify the

transmission characteristics through the auditory system in

such a way that the acoustic reflex and temporal integration

are changed in a predictable manner.

The foregoing studies indicate two important clinical

premises: (1) it is not necessary to compare the acoustic-

reflex threshold with the threshold of audibility to obtain

diagnostic information about the state of the cochlea, and

(2) there are factors affecting the acoustic reflex which may

differentiate normal ears from those with end-organ lesions.


Hypothesis

The major hypothesis, formulated for testing in this

study, is based on information concerning temporal summation

of the acoustic reflex: the acoustic reflex changes in a





53




systematic fashion in the normal ear as a function of stimu-

lus duration, i.e., a tenfold change in time can cause a

median change of 25 dB of acoustic power. The null hypo-

thesis is: temporal summation of the acoustic reflex does

not function differently in those subjects who have normal

hearing and in those subjects who have a known end-organ

auditory lesion.

The questions that this study attempted to investigate

are:

1. What is the most sensitive measure of the acoustic-re-

flex threshold when using increments of 2 dB: the first

response, the first level to occur 60 percent of the

time, a measure of central tendency of either 5 or 10

trials, or the lowest response of 10 trials?

2. What is the group average and range of acoustic-reflex

thresholds for a normal-hearing population as a function

of temporal integration?

3. What is the group average and range of acoustic-reflex

thresholds for an unilaterally cochlear impaired popula-

tion as a function of temporal integration?

4. Is there a significant difference in temporal integra-

tion of the acoustic reflex between normal hearing ears

and cochlear impaired ears?

5. Is there a significant difference in temporal integration

of the acoustic reflex between the ears of normal-hearing

subjects and the normal-hearing ears of unilaterally

cochlear impaired subjects?





54



6. Is there a significant difference in temporal integra-

tion between the normal-hearing ear and the cochlear-

impaired ear of a Meniere's group?

7. If cochlear-impaired ears demonstrate reduced integra-

tion at threshold of audibility, do they also demonstrate

reduced temporal integration at the threshold of the

acoustic reflex?














CHAPTER III

METHODS AND PROCEDURES


Temporal summation at both threshold of audibility and

threshold of the acoustic reflex were measured by the follow-

ing experimental design. First, a baseline of temporal

integration at threshold of audibility was obtained in order

to compare the results in this study to previously reported

data on normal-hearing and cochlear-impaired subjects. This

information delineated normal integration from reduced inte-

gration of cochlear-impaired ears. Second, the results of

temporal summation of the acoustic reflex determined whether

the null hypothesis should be rejected; that is, temporal

summation of the acoustic reflex does not function differ-

ently in those subjects who have normal hearing and in those

subjects who have a known end-organ auditory lesion. The

results of this second auditory procedure were obtained with-

out the subject's overt cooperation.

Ten cooperative adult subjects were tested. One-half

of the group had normal hearing and were the control subjects.

The other half of the group evidenced unilateral end-organ

hearing impairments and served as the experimental subjects.

These two groups were classified homogeneously and screened

for possible problems that might interfere with normal

acoustic-reflex function.

55





56



The subjects had a negative history of middle-ear or

inner-ear surgery and of concussion, brain damage, or

cerebral vascular accidents. Those subjects using drugs or

medication, possibly influencing normal intra-aural muscle

function, were excluded. The medical history form employed is

shown in Appendix A. None of the subjects had conductive

hearing problems. Normal compliance at the tympanic membrane

and normal tympanograms confirmed the absence of a functional

middle-ear problem.

Data from the five control subjects were collected for

each ear. These five subjects demonstrated pure-tone audio-

metric measures no worse than 20 dB HL (ANSI-1969) in either

ear for frequencies 125 Hz through 4000 Hz. The normal-hear-

ing group ranged from 23 years to 30 years of age.

Only those persons having a classical, unilateral

Meniere's disease, a pure end-organ disorder, with normal

hearing in the contralateral ear,were included in the experi-

mental group. The Meniere's disease was medically diagnosed.

The patients had, in the course of their medical history, the

classical symptoms of episodes of true vertigo, fluctuating

hearing loss, "roaring" tinnitus, and feeling of fullness in

the affected ear. These symptoms were supported by the

following objective findings: a sensorineural hearing loss,

loudness recruitment and vestibular canal paresis. This

group will be referred to as the Meniere's group or the

cochlear-impaired group. The Meniere's group ranged from 45

years to 63 years of age.





57



Since one of the characteristic symptoms of Meniere's

disease is fluctuating hearing in the impaired ear, there

were no minimum requirements of hearing level to qualify the

impaired ear as abnormal. The contralateral ear of each

experimental subject yielded pure-tone thresholds of 20 dB or

less (ANSI-1969) at a minimum of two out of the three test

frequencies. This qualified the contralateral ear as having

normal hearing. To facilitate discussion, ears with normal

hearing were referred to as "good," those ears afflicted with

Meniere's disease were denoted as "bad." This classification

of ears was subsequently used throughout this report.

Thus, there was a total of 15 good ears and 5 bad ears

in 10 subjects. Subject LS of the Meniere's group did not

have normal hearing in her good ear at 500 Hz. Nevertheless,

this particular ear is classified as good because she did

have normal hearing at two of the three test frequencies. The

audiograms of all subjects are contained in Appendix B.


Stimuli

The test stimuli were shaped sinusoids ranging from

500 msec to 10 msec in overall duration with 5-msec rise and

decay times. These signals were gated once every 1200 msec.

The stimuli used in this study are listed in Table 4. The

choice of stimuli was determined by previous temporal-inte-

gration data, ease and expediency of equipment manipulation,

and maximum utilization of the good and bad ears of the

Meniere's group.





58




Table 4. Test Stimuli Used to Elicit Threshold of Audi-
bility and Threshold of the Acoustic Reflex


Test frequency: 500 Hz, 1000 Hz, 2000 Hz
Signal durations: 500 msec, 200 msec, 100 msec, 20 msec,
10 msec
Rise/fall times: 5 msec
Inter-stimulus interval: 700 msec, 1000 msec, 1100 msec,
1180 msec, 1190 msec
Repetition rate: once every 1200 msec



More specifically, the shaped signals consisted of three

sinusoids at 500 Hz, 1000 Hz, and 2000 Hz. These test fre-

quencies were chosen because most of the Meniere's group had

normal hearing in the good ear and poorer hearing in the bad

ear at these frequencies.

Five signal durations were used: 500 msec, 200 msec,

100 msec, 20 msec, and 10 msec. The duration was measured

from the gate onset of a stimulus to the cessation of the

stimulus envelope. Five hundred milliseconds represented,

for practical purposes, an infinitely long signal. The re-

maining four durations represented two tenfold, or decade,

changes in time, i.e., 10 msec to 100 msec and 20 msec to

200 msec. These values also included the usually designated

limits of linear temporal integration from the long signal

duration of 200 msec to the short tone pulse of 10 msec.

The repetition rate of the stimulus was held constant by

gating once every 1200 msec. Neural independence was main-

tained, because the shortest inter-stimulus interval of the










test series was 700 msec. Preliminary investigation also

indicated that this rate would allow for the muscle contrac-

tion to return to the pre-contraction baseline before re-

sponding to the next signal burst. Figure 8 illustrates a

tracing of an oscilloscope recording of the acoustic reflex

evoked by a 500-msec signal at 5 dB above the ART. The

muscle activity returned to the pre-contraction baseling

within the 700-msec inter-pulse interval. It should also be

noted that this recording was made towards the end of an

hour and one-half test session, and middle-ear muscle fatigue

was not observed.

The rise and decay times of the stimulus envelope, as

controlled by a pre-set electronic switch, were approximately

5 msec. The 5-msec ramp was verified on an oscilloscope by

determining the duration between the upper 90 percent and

lower 10 percent of the slope.

Figure 9 delineates the spectral characteristics of two

1000-Hz tone burst stimuli measured at the earphone output.

The bandwidth of the 10-msec duration tone pulse is about

120 Hz wide in Figure 9a. This is about 20 Hz wider than

might be expected theoretically, indicating some incidental

spread of energy. When the tone burst is lengthened to 20-

msec duration in Figure 9b, the spectral energy fits nicely

into the theoretical bandwidth of 50 Hz, The spectrum

of these tone bursts suggests that no audible spread of energy

should be detected; this is supported by the smooth envelope





60














Repetition rate
1 stim./1200 msec

700-msec
inter-stimulus
inter-stimulus 500-msec stimulus
interval
I duration



...... Channel 1:
Sstimuli

S^\ \ Channel 2:
Sacoustic reflex
Reflex
baseline


--

Acoustic Muscle
reflex relaxation



Figure 8. Acoustic-reflex contraction. The intra-
aural muscles can contract, relax, and return to the pre-
contraction baseline before contracting again. The
repetition rate is once every 1200 msec for a 500-msec tone
burst at 5 dB above the ART.





61










I I


40 -
10-MSEC 20-MSEC
35 IMPULSE IMPULSE


30 -

z
H 25 -
0
H
20 -

z
t 15


10






950 1000 1050 950 1000 1050

a b


Figure 9. Spectral content of tone bursts with 5-msec
rise/decay ramp. (a) 1000 Hz, 10-msec impulse with half-
power width of ca. 120 Hz; (b) 1000 Hz, 20-msec impulse with
half-power width of ca. 50 Hz.





62



of a 20-m:se ton; bur:st in Figure 10. In order to confirm

this, an audibi_. check was made at the earphone. No clicks

were disce;ned from a 10-msec tone burst at 500 Hz placed

just below thres:hold audibility. It was concluded, therefore,

that a 10-msec tone pulse with a rise and fall time of 5

msec would produce a relatively undistorted signal in this

test situation,using high intensity levels.


Experimental Equipment

The block diagram in Figure 11 represents the equipment

used to generate the stimulus envelope and to record the

subject's response. A General Radio 1313A oscillator gener-

ated the test frequency which was monitored by a Monsanto

Model 100-A electronic counter. The sinusoid passed through

the Grason Stadler 3262A recording attenuator to the Grason

Stadler 829E electronic switch that shaped the signal; then

the signal was step attenuated by an Hewlett Packard 350D

attenuator. By way of switches I, II, and II, the signal was

led through either a McIntosh 162K power amplifier or through

a matching transformer to the earphone, a TDH-140 dynamic

earphone in a MX41/AR cushion. The signal from switch II was

measured on a Ballatine 321 true rms vacuum tube voltmeter

and on channel one of a Tektronix 564 oscilloscope. The

same "clock" triggered both the electronic switch and the

oscilloscope.

Three response switches were under subject control.

Response switch I controlled the recording attenuator.





63




































Figure 10. A 20-msec signal burst with a smooth 5-msec
rise and decay.









FREQUENCY VOLT
COUNTER METER


I
I I
AMPLIFIER

OSCILLATOR RECORDING ELECTRONIC STEP
ATTENUATOR SWITCH ATTENUATOR
STRANS-
I | 'FORMER -


I ,--I -I II
I OSCILLO-

BUZZER L _CLOCK OSCILLO- PONS
WITCHES
--,T -

IMPEDANCE VOLTAGE
BRIDGE DIVIDER






Figure 11. Block diagram of equipment used to elicit and record the intra-aural re-
flex and to allow subject control over the test situation.

cn





65


Response swit I -'. .:.rrupted the signal presentation at the

electronic switch, L .le response switch III activated a buz-

zer audible to tce -:xerim::nter. The reflex was measured from

the contralatcral e-ternal ear canal which was sealed with an

ear olive. The reflected probe tone passed to the Peters

AP61 electroacous-ic impedance bridge. The output of the

bridge was led through a direct current voltage dividerl and,

finally, to channel two of the oscilloscope.

To measure threshold of audibility, switches I and II were

closed so that the signal passed from the step attenuator

through the matching transformer to the earphone. Switch III

allowed the signal to be monitored at the oscilloscope and volt

meter without being audible to the subject. The recording at-

tenuator was activated by subject response switch I. Thus, a

permanent record of the subject's response was made with a

self-tracking Bekesy technique.

The same basic equipment was utilized to measure the

acoustic-reflex threshold. The recording attenuator was

turned off. The test signal was diverted through the McIntosh

power amplifier to the earphone by using the same switches.

The eliciting signal was stored on the oscilloscope of

channel one, while the acoustic reflex was stored on Channel

two. A photographic record of both oscilloscope channels was

made using a Rolleiflex SL 35 camera with a 50-mm lens.

Calibration was performed at both the output of the

earphone in SPL and across the earphone in volts. Calibration


1The d.c. voltage divider was placed in the system to in-
crease the sensitivity of the signal. See Appendix C for the
schematic diagram.





66



was made in SPL a tt: '<: inning and conclusion of the experi-

ment. The steo att!nutt:r and the recording attenuator were

checked for linearity :in r APL. A NBS type 9A coupler with a

6-cc cavity attached to 1 Bruel and Kjaer 2203 sound level

meter with a one-inch microphone were used to measure SPL

(re 0.0002 dyne/cm2). Frequency response of the earphone was

also measured in SPL. Voltage checks were made before and

after each change in test frequency. A General Radio 1900-A

wave analyzer and a General Radio 1521-B graphic level re-

corder were used to measure the spectral characteristics of

the test signals. These calibration data may be found in

Appendix D.

The Peters electroacoustic bridge was calibrated accord-

ing to the manufacturer's specifications. The probe tone

zeroed the balance meter when the input to the bridge network

was 94 dB SPL and the probe frequency was 276 Hz. Using

four different cavity volumes the compliance dial was cali-

brated at 0.2 cc, 1.0 cc, 2.0 cc, and 4.0 cc so that the

balance meter read zero. Finally, the input filter was tuned

to 276 Hz, thereby being most sensitive to the probe-tone

frequency.

Two different earphones were used in this study due to

a malfunction of the TDH-140, 10-ohm phone. A TDH-30, 10-ohm

phone elicited data on four normal-hearing subjects for their

threshold of audibility and on two normal-hearing subjects

for their acoustic-reflex threshold. Since both phones were





67



calibrated on the s;ime equipment and in a similar manner,

the data reflect tin resuective calibration curves. In other

words, the data will be reported as if one earphone were used.


Procedure

The experimental testing was conducted in three sessions

for each individual. The temporal integration at threshold

of audibility was measured for both ears during the first test

session, while temporal integration at threshold of the

acoustic reflex was measured in subsequent sessions for each

of the two ears. The three stimulus frequencies were ran-

domized for each ear,and the signal durations were presented

in random order for each test frequency. All auditory measures

were made in a single-walled IAC sound treated chamber.


Temporal Integration at Threshold of Audibility

The subject responses were recorded by using a self-

tracking, Bekesy procedure. The recording attenuator was set

at a chart speed of three-fourths of an inch per minute and

1 dB of attenuation every second. The visual mean of a

minimum of 10 threshold crossings was used to establish

threshold. This mean value was located at the nearest whole

dB,and, if the threshold crossings were not stable, the last

10 stable crossings were used to establish the mean threshold

value. The total test time for both ears, including a 15- to

30-minute rest period midway, was two to two and one-half

hours. Since signal crossover to the good ear was not observed





68



nor reported by the subjects in the preliminary or experi-

mental sessions, a masking noise was not applied to the good

ear of the Meniere's group.


Temporal Integration at the Thres-
hold of the Acoustic Reflex

Testing of the acoustic reflex commenced once the exter-

nal meatus of the reflex measurement ear was sealed with an

ear olive containing the electroacoustic bridge probe tip.

The seal was considered adequate if a positive 200 mm of

equivalent water pressure in the external auditory canal did

not break the seal. The electroacoustic bridge was periodi-

cally checked to determine if the balance dial was within

25 mm of equivalent water pressure of zero. If the balance

meter was out of the target range it was rebalanced by adjust-

ing the compliance dial. In order to maintain this criterion,

the sensitivity dial was kept on position one.2 If the

pressure dial was outside the target range,the subject was

asked to perform a Valsalva or Toynbee maneuver to increase

or decrease, respectively, the middle-ear pressure.

A modified method of limits was employed to determine

the acoustic-reflex threshold. Ten ascending ART trials were

2
Although it is suggested that the higher sensitivity
dial settings be used while determining the acoustic reflex,
i.e., two or three on the Peters electroacoustic bridge and
three or four on the Madsen electroacoustic bridge, prelimi-
nary observations suggested an ART difference of no more than
2 dB between sensitivity position one and position three. The
use of sensitivity position one reduced the amount of rebalanc-
ing necessary throughout the test procedure. If more sensi-
tivity was desired, the vertical multiplier dial or the oscil-
loscope was changed.





69



made for each signal duration at each frequency in order to

obtain a stable threshold value. The experimenter started

below the acoustic-reflex threshold and increased the

stimulus level in 2-dB steps until a minimum reflex was

detected. At this point, if the response was a definite

change from the ongoing pattern, the stimulus level was re-

corded as the ART. The experimenter started the sequence

over again by lowering the intensity 10 dB. If the response

was in doubt, then another response was evoked at the same

signal level or at the next 2-dB higher intensity increment.

If the reflex response was definite at the second response

or the 2-dB higher response, the initially observed sound

pressure level of the reflex was recorded as the ART. Figure

12 illustrates a tracing of a typical acoustic-reflex thres-

old response. In this instance, a 2-dB intensity increase

caused a definite change from the previously recorded pattern.

The tenth ascending threshold response was stored on

channel two of the oscilloscope. The eliciting signal was

increased 5 dB above the reflex threshold and superimposed

upon the ART on channel two of the oscilloscope. The two

superimposed acoustic reflex images on channel two were then

photographed to retain a permanent record.

During the acoustic-reflex testing procedure, subjects

were requested to glance quietly through some pictorial

magazines in an attempt to equate subject state of alertness

and focus of attention. Subject state and attention have





i 0

















90 92 94 96 96 98 98


S Channel 1:
S* ) a -- Stimuli of
;.. ... i .... "; 5 0 0-m se c
duration


I 1' -,,"'' t, Channel 2:
Acoustic
reflex at
threshold


Below ART ART + 2 dB
ART



Figure 12. Typical response of the acoustic reflex
at threshold. Channel 1 illustrates the 500-msec duration
tone burst increasing in 2-dB steps. Channel 2 demonstrates
the response at 96 dB SPL to be significantly different
from the ongoing activity.





71



been reported to affect the acoustic reflex (Durrant and

Shallop, 1969; Gunn, 1967). As long as the subject's atten-

tion is generally diffuse and not under intense concentra-

tion, e.g., distinguishing visual word choices or auditory

intelligibility, the variability is probably slight (Durrant

and Shallop, 1969).

Experimental Safeguards

In performing temporal integration at the threshold of

the acoustic reflex, sound pressure levels normally asso-

ciated with thresholds of discomfort, tickle and pain, and

temporary and permanent threshold shifts were required. It

was necessary to safeguard the subjects against any harm or

unnecessary discomfort. Possible response artifacts result-

ing from high intensity signals were also investigated.

It was conceivable that all subjects might be subjected

to stimulus levels in excess of 130 dB SPL, as seen in Table

5. This table demonstrates a possible intensity level neces-

sary to elicit a reflex by a 10-msec tone burst at 500 Hz.

For this hypothetical person, the maximum earphone output

would be 136.5 dB SPL.

The Committee on Hearing, Bioacoustics and Biomechanics

of the National Academy of Sciences has published upper limits

of acceptable exposure to impulse noise as a function of

sould pressure level and signal duration as seen in Figure 13

(Ward, 1968). The solid line represents the upper acceptable

limit of exposure for probably 95 percent of the population





72












U
Exposure Limits
S----- Earphone Limits
> 150

0
o 145
o0


m 140 -


135



' 130


4 125 I I I I I
.025 .05 .1 .2 .5 1 2 5 10 20 50 100 200 500 1000
a SIGNAL DURATION IN MILLISECONDS


Figure 13. Upper limits of acceptable exposure to impulse
noise for 95 percent of the population to 10,000 impulses.
(Redrawn from Ward, 1968.)





73



Table 5. Conceivable Sound Pressure Level Necessary to
Elicit the Acoustic-Reflex Threshold by a 500-Hz
Burst at a Signal Duration of 10 Msec


Normal threshold at zero hearing level
(ANSI-1969) in dB SPL 11.5 dB

The ART at the upper limit of normal
sensation level 100.0 dB

Power increase necessary to maintain thres-
hold when a tone burst is shortened to
10 msec 25.0 dB

Total SPL 136.5 dB



to 10,000 pulses in some period of time, i.e., one hour to

one day. This limit applies to impulses with a known single

rise-peak decay, not a complex impulse. The Committee

stated that, although there are many unknowns, their data for

upper limits of exposure constituted a conservative estimate.

Only the "weakest ears" of the unaccounted 5 percent might

demonstrate a temporary threshold shift because the acoustic

reflex would afford some protection in some subjects. The

horizontal dashed line in Figure 12 indicates the maximum

output of the earphone within this experimental system at

139 dB SPL and the vertical hatched lines indicate the approxi-

mate durations used in this study. As may be seen in Figure

12, the maximum output of the earphone falls below the maxi-

mum acceptable exposure limits.

To minimize possible discomfort from loud acoustic

impulses, the subjects were allowed some control over the

test situation. If the subjects felt that the stimulus pulse





74




was uncomfortable, they were instructed to press a button

(see response switch III in Figure 9), which activated a

tone audible only to the experimenter. The experimenter then

reduced the stimulus 10 dB. This situation only occurred

once because of experimenter error. A second control button

(see response switch II in Figure 9) allowed the subject to

discontinue the stimulus presentation at will. This second

button was never employed by the subjects.

Dealing with these high signal levels for long periods

of time was bound to introduce some auditory adaptation or

fatigue. The stimulus presentation rate of once every 1200

msec and a duty cycle of less than 50 percent was used as an

arbitrary compromise between long, fatiguing test sessions

and slow signal presentation rates in an attempt to minimize

neural changes. In addition, at the end of each test run

(10 thresholds per signal duration), the test signal was

turned off in order to allow a two- to five-minute rest inter-

val. During this interval, the signal duration or frequency

was changed and voltage levels checked for calibration.

Opinions of the subjects indicated that sessions of one and

one-half hours approached the maximum tolerable limit. This

was the time necessary to test the acoustic reflex of one ear.

The possibility that the test signal would have crossed

over to the opposite ear, the recording ear, was considered.

If this had occurred, the data would have been contaminated

in two ways. First, the reflex eliciting tone might have





75



crossed over and combined with the probe tone,causing a

reflex on the measuring side. This was unlikely, since a

minimum of 40-dB to 50-dB difference in hearing level is

normally necessary for crossover. The effective signal level

reaching the probe ear side would have been 40 dB to 50 dB

less than necessary to evoke the reflex. Since doubling the

signal pressure increases the SPL only 6 dB, the increase in

probe-tone pressure would have been a maximum of 6 dB and

probably only 2 dB to 4 dB. This would not have raised the

probe-tone intensity level enough to cause an acoustic reflex.

Second, the reflex-elicitng signal that crosses over may

have registered on the pickup microphone of the electroacoustic

bridging network. This network has a filter centered at

276 Hz,and it is at least 20 dB down at 500 Hz. Thus, the

filter should reduce by at least 20 dB any energy crossover

at 500 Hz and possibly even at 1000 Hz and 2000 Hz.

In conclusion, all reasonable safeguards for subject

well-being and integrity of the acoustic reflex data were

taken into account without compromising the experimental

protocol. The subjects' responses, especially from those

exhibiting loudness recruitment, indicated that the subject

safeguards were not actually needed.













CHAPTER IV

RESULTS AND DISCUSSION

The results of this study indicate that temporal summa-

tion of the acoustic-reflex does not function differently in

those subjects who have normal hearing and in those subjects

who have a known end-organ auditory lesion. A statistical

analysis of the data did not support a consistent, significant

difference between normal-hearing (good) ears and cochlear-

impaired (bad) ears, using the acoustic-reflex threshold

(ART). However, there were statistically significant differ-

ences between responses of the two groups at threshold of audi-

bility. An analysis of these data follows in a comparison of

normal-hearing and hearing-impaired ears at 1) temporal inte-

gration at the acoustic-reflex threshold, 2) temporal integra-

tion at the threshold of audibility, and 3) temporal integra-

tion at the acoustic-reflex compared to temporal integration

at threshold of audibility.

1) Temporal Integration at
the Acoustic-Reflex Threshold

A discriminant analysis2 (Rao, 1952) was performed on the

ART difference between a 10-msec and a 200-msec duration tone

burst and between a 20-msec and a 200-msec duration tone


1The raw data are reported in Appendix E.

Further information concerning discriminant analysis
is contained in Appendix F.

76





77

burst at each test frequency. In Figure 14, for example,

the mean change for the cochlear-good ear at 500 Hz from

10 msec to 200 msec is 29 dB while the mean change at the

same parameters for the cochlear-bad ears is 22dB. The

20-msec to 200-msec difference was also chosen for discrimi-

nant analysis because the change was large enough to yield

mean differences among groups (see Figure 14) and because

some subjects did not yield an acoustic reflex at 10 msec

with the maximum acoustic power of 139 dB SPL. All thresh-

old differences and the data in some figures were subtracted

from, or normalized to, 200 msec. Two-hundred milliseconds

was most often stated as the time constant of temporal inte-

gration,and there was little difference between the ART

elicited by a 500-msec and by a 200-msec duration tone. The

mean change of ART between the two durations was 2 dB to

3 dB, depending on frequency and subject group. The figures

were normalized to the signal duration which delineated the

results most clearly.

The results of discriminant analysis are reported as

an "F" ratio. Table 6 gives the statistical results of the

discriminant analysis upon the ART. Two comparisons out of

12 between good and bad ears were statistically significant,

both at 500 Hz and both at the ART difference between 20

msec and 200 msec. Ten of the 12 good-bad ear comparisons

were not significantly different.

The mean ART values of the cochlear-impaired ear, as

demonstrated in Figure 14,were always less than those of the








A-- --- Normal ears
O---- Cochlear-good
ears
.-----. Cochlear-bad
ears

9 500 Hz 1000 Hz 2000 Hz
o Z 330




mo
wA

20


\ \ \

10 \ \ \
C 20 \ ,. ...
I 0\ \ \\





I I I \ i i I
10 20 100 20d 10 20 100 200 10 20 100 200

SIGNAL DURATION IN MSEC

Figure 14. Mean acoustic-reflex thresholds as a function of temporal integration.
^^V -- <'co





79




good ear of the cochlear-impaired group and the normal-

hearing group. However, neither the mean values, nor the

average slope per decade of stimulus time, were significantly

different among the three groups (Table 7).


Table 6. Discriminant Analysis of the Acoustic-Reflex
Threshold


500 Hz 1000 Hz 2000 Hz


Normal vs. cochlear-good ears

10 msec-200 msec ns ns ns

20 msec-200 msec ns ns ns

Normal vs. cochlear-bad ears

10 msec-200 msec ns ns ns

20 msec-200 msec .05 ns ns

Cochlear-good vs. cochlear-bad
ears

10 msec-200 msec ns ns ns

20 msec-200 msec .10 ns ns

ns = not significant.
.05 = significant at the 5 percent level.
.10 = significant at the 10 percent level.


Table 7. Average Slope Change Per Decade of Time of the
Acoustic-Reflex Threshold


500 Hz 1000 Hz 2000 Hz

Cochlear-good ears 23 dB 22 dB 25 dB

Normal ears 17 dB 19 dB 19 dB

Cochlear-bad ears 15 dB 14 dB 14 dB





80



Although the group means were not statistically differ-

ent, the data from Figure 14 indicate two consistent find-

ings. First, the cochlear-impaired (bad) group always had

less mean ART change than the other two. The cochlear-good,

for the most part, had the greatest amount of ART change.

The difference between the cochlear-bad and normal ears at

20 msec increased from between 4 dB and 7 dB to 12 dB for

500 Hz, 1000 Hz,and 2000 Hz, respectively. There were no

other large systematic changes in this respect. Second,

the change from 20 msec to 100 msec (or to 200 msec) became

less as frequency increased for the cochlear-bad ears. This

decrease in mean slope caused the former results of increased

ART difference between normal and cochlear-bad groups.

Figure 15 illustrates the large inter-subject variation

of temporal integration slopes for the normal ears and the

cochlear-bad ears. Similar configurations also appear in

Figure 16, contrasting cochlear-good ears and cochlear-

bad ears. In both figures, the slope of individual sub-

jects varies over a wide range within a group, so that the

normal and cochlear-impaired groups overlap. The large inter-

subject variability is probably the reason why the mean

values were not statistically different from one another.

Figure 17 contrasts the inter-subject range among

the three groups. This graph also depicts the large amount

of overlap among the three groups, which hampered

statistical significance. The least amount of overlap

occurred at the test frequency of 500 Hz and










Nrmal ears 500 Hz Normal ears 1000 Hz Normal ears 2000 Hz







0i





o 0 Cochlear-bad ears Cochlear-bad ears .Cochlear-bad ears
I 500 Hz 1000 Hz 2000 Hz

E-iE-4


0







10 20 100 200 500 10 20 100 200 500 10 20 100 200 500
SIGNAL DURATION IN MSEC

Figure 15. Acoustic-reflex thresholds of normal and cochlear-impaired (bad) ears as
4 function of temporal summation.
ca 31











40 Cochlear-good earsa Cochlear-goodears Cochlear-good
500 Hz 1000 Hz ears 2000 Hz

30 ,


Ww 20

mo
mo 10



N 0

U 40 Cochlear-bad ears Cochlear-bad ears Cochlear-bad ears
u 500 Hz 1000 Hz 2000 Hz

S30 3
oi











SIGNAL DURATION IN MSEC
0c
Figure 16. Acoustic-reflex thresholds of normal (good) and impaired (bad) ears of the
Meniere's group as a function of temporal summation.





83


SH- I Normal ears
40- TI --- Cochlear-good ears
SI 1-----1 Cochlear-bad ears

30- TTr 500 Hz


20 I I

10 i
I0-
: TT1L

II-


40 -rL
e -TI I
0X 40 t
SnI
S 30- T







Im I TT
E- E T 1000 Hz


20 I
N*I 4


10 -



T*
0 10- 1




TTT

S0 2000 Hz



10- -L T T








Figure 17. Inter-subject range of acoustic-reflex
thresholds as a function of temporal integration.
thresholds as a function of temporal integration.





84



would account for the statistically significant findings at

500 Hz. Using discriminant analysis, the more distinct the

groups the less chance of error in classification (Rao, 1952).

It is notable that the point of greatest significant differ-

ence (p<.05 at 500 Hz for normal vs. cochlear-bad) yields the

smallest difference between mean values (Figure 14).

Figure 18 compares the temporal integration slope of

the acoustic reflex of normal-hearing persons in three

studies. The individual mean scores from this study and

Djupesland and Zwislocki (1971) are plotted as well as the

median values. The group mean values from McRobert et al.

(1968) are also included. The average change in slope per

decade in time at 1000 Hz for this study is 6 dB less than

the 21 dB/decade from Djupesland and Zwislocki, as well as

2 dB less than the 25 dB/decade from McRobert et al. In

addition, it appears that the individual mean values for this

study were more scattered than those of Djupesland and

Zwislocki (1971), especially at the 10-msec signal duration.

This large scatter of individual scores may be repre-

sentative of the true population, since no more than nine

normal-hearing subjects were employed in any one study.

These differences may also be a result of the type of

equipment employed, since the two lower values of 19 dB and

21 dB/decade were obtained with a Peters electroacoustic

bridge and a Madsen electroacoustic bridge, respectively.

The steeper slope of 25 dB/decade was obtained with a






85






1000 Hz

O ART for each ear this study
Median for 10 earsj
ART for each ear Djupe
Median for 6 Ss Djupesland
K Mean ART for ca. 9 Ss) McRobert



o 0
0
0



O 30 -
": ao
B Q \0


0 20 1 2




u i0 om i
a\



















hearing ears.
o 20
44 o0
10



oo 6-000\

o -0




10 20 100 200 500

SIGNAL DURATION IN MSEC


Figure 18. A comparison of mean acoustic-reflex
thresholds as a function of temporal summation in normal-
hearing ears.






86



Zwislocki mechanical impedance bridge. There have been

reports indicating that the electroacoustic and mechanical

impedance bridges can vary as much as 35 percent in their

impedance values, depending upon the condition and size of

the tympanic membrane (Lilly, 1970; Wilber, Goodhill, and

Hogue, 1970).

The statistical observations concerning the null

hypothesis of the ART were limited because of unexpected

inter-subject variability causing the sample size of five

normal and five cochlear subjects to be ineffectual statis-

tically. However, descriptive observations can be made,

although these observations may be due to chance.

The time constant of temporal integration of the ART,

that point in time where integration begins,3 is shorter

than the 200 msec reported for threshold audibility. A

shortened time constant has been reported previously by

Small et al. (1962) for temporal integration at supra-

threshold levels. As illustrated in Figure 19, the ART time

constant is most often between the signal durations of

100 msec and 20 msec, or the 100 msec-20 msec interval.

Further inspection of Figure 19 suggests that the time



The time constant was defined in this experiment as
l/e of the maximum threshold change. In practicality, l/e
was 1/2.718, or 32 percent, of the mean integration slope
per decade of stimulus time in normal ears, or 6 dB. The
first 6 dB of threshold change from 500 msec, therefore, was
not considered part of temporal integration. The point at
which 7 dB, or more, of threshold change occurred due to
change in signal duration marked the length of the time
constant.





87










12 63%

z
H 8-
z
0 25%
S4- 112%
I0%
0-
z
H
8-
H 46%
m 27% 27%
4-

U 0%
0 0

67%
2 20-
0
u
a 16

E-4
S12 -


S8 23%
8--

0 10%
4-
S0%
m 0
z 500-200 200-100 100-20 20-10
DURATION INTERNAL IN MSEC


Figure 19. Signal-duration interval in which the time
constant of temporal integration for each test subject occurred
across all frequencies.





88



constant of the two good-ear groups is never shorter than

a duration of 20 msec. Interestingly, the opposite observa-

tion may be made with the cochlear-impaired group. There

were no long time constants in the 500 msec-200 msec inter-

val, but 25 percent of the cochlear-impaired ears, across

all frequencies, demonstrated integration functions that

were in the 20 msec-10 msec interval. That is, temporal

integration for these 4 out of 16 ears may be abnormally

shortened due to the cochlear impairment. This can be most

effectively seen in Figure 20, where three of the cochlear-

impaired ears with a short time constant are compared with

the normal-hearing results at 1000 Hz and 2000 Hz. A short-

ened time constant has been previously reported in cochlear-

impaired ears for temporal integration at threshold of

audibility (Figure 5).

Inspection of Figures 15 and 16 demonstrates that an

average slope value per decade is not necessarily the best

descriptor of temporal integration. The change is not

linear, especially in the cochlear-impaired group. The

cochlear-bad ears at 500 Hz integrate slowly, with a flatter

slope, until 20 msec when the rate of change becomes greater.

For example, four out of five cochlear-bad ears at 500 Hz

change at an average rate per time decade of 8 dB to 20 msec,

and then abruptly change to an approximate slope of 50 dB

to 55 dB per time decade. This change is only distinctive

between good and bad ears at 500 Hz. This difference






89

40 --
0 1000 Hz
Normal ears
m. Cochoear-bad ears

30





20 --







310



2 z
un




E- 0 0
0 20 2000 Hz


S0 T l e co ochlear-bad ears
30





20




10







10 20 100 200 500
SIGNAL DURATION IN MSEC

Figure 20. The delayed time constant in some cochlear-
impaired ears contrasted with the normal time constant of
temporal integration at the acoustic-reflex threshold.





90



probably contributed to the only two rejections of the null

hypothesis which occurred (Table 7).

This abrupt change in slope occurred in the 20 msec-

10 msec interval, the same interval in which the short time

constant of cochlear-imparied ears fell in Figure 20. Since

the definition of the time constant was relatively arbitrary,

the slow-abrupt change in slope and the short time constant

fay reflect the same phenomena in the cochlear-imparied ear.

From the discriminant-analysis data one is able to

identify those subjects consistently classified correctly,

although no statistical significance is attached. Two

goghlear-impaired subjects, FH and BC, were never classified

as normal hearing, while EC was incorrectly classified in

only 2 out of 12 good-bad ear comparisons. Of the five

te@t subjects, the data from these three indicated that one

sh@Oul be able to distinguish a cochlear-impaired ear based

OR a@ggVtic-reflex data. The cochlear abnormality may

eQs8@ a aberration in temporal integration of the ART as

indi@Ba@d by an initial slow growth of the integration func-

ti@n or by an unusually short time constant. One cochlear-

impair@d subject is an appropriate example. LS was incor-

r@@oly classified 9 out of 12 times. Two of the correct

ela@gifiqations were probably based on her very short time

Q8acaant, Her ART changed only 2 dB as the signal duration

wa@ @hortened from 500 msec to 20 msec, then it changed

abuy4ply by 35 dB. She is represented in Figure 20 at 2000

IH by the white hearts.




Full Text
66
was made in SPL at the beginning and conclusion of the experi
ment. The step attenuator and the recording attenuator were
checked for linearity in SPL. A NBS type 9A coupler with a
6-cc cavity attached to c> Bruel and Kjaer 2203 sound level
meter with a one-inch microphone were used to measure SPL
2
(re 0.0002 dyne/cm ). Frequency response of the earphone was
also measured in SPL. Voltage checks were made before and
after each change in test frequency. A General Radio 1900-A
wave analyzer and a General Radio 1521-B graphic level re
corder were used to measure the spectral characteristics of
the test signals. These calibration data may be found in
Appendix D.
The Peters electroacoustic bridge was calibrated accord
ing to the manufacturer's specifications. The probe tone
zeroed the balance meter when the input to the bridge network
was 94 dB SPL and the probe frequency was 276 Hz. Using
four different cavity volumes the compliance dial was cali
brated at 0.2 cc, 1.0 cc, 2.0 cc, and 4.0 cc so that the
balance meter read zero. Finally, the input filter was tuned
to 276 Hz, thereby being most sensitive to the probe-tone
frequency.
Two different earphones were used in this study due to
a malfunction of the TDH-140, 10-ohm phone. A TDH-30, 10-ohm
phone elicited data on four normal-hearing subjects for their
threshold of audibility and on two normal-hearing subjects
for their acousticreflex threshold. Since both phones were


103
Temporal summation of the acoustic reflex does not
function differently in those subjects who have normal
hearing and in those subjects who have a known end-organ
auditory lesion. However, there were statistically signifi
cant differences found between the normal-hearing (good) ear
and the cochlear-impaired (bad) ear comparisons of the 20
msec-200 msec ART difference at 500 Hz. Small sample size,
and unexpected inter-subject variability and flatter inte
gration slopes for normal ears, contributed to the lack of
statistical significance.
Temporal integration at threshold of audibility, in
contrast, did demonstrate statistical differences between
cochlear-impaired ears and normal-hearing ears. Even though
a statistical difference was evidenced, the difference was
not so obvious, in this experiment, as to be a clinically
useful tool. A comparison of temporal integration at thresh
old of audibility and at the ART demonstrated that a short
time constant for the ART may be of diagnostic value in
determining cochlear impairment. However, the question of
whether temporal summation of the acoustic reflex is clini
cally feasible as a diagnostic measure of cochlear abnor
malities was not resolved in this experiment. Any observa
tions made, e.g., the short time constant, were not proven
to be other than chance observations. ART integration,
therefore, could not be safely said to identify an end-organ
impairment.


48
Q
O
a
K
Eh
CJ
SIGNAL DURATION IN MSEC
Figure 6. Temporal summation of the acoustic reflex.
The solid line intersects the median of six individual mean
scores resulting in a slope of 25 dB/decade of signal dura
tion change. The dashed line indicates a similar slope but
a shorter time constant. (Redrawn from Djupesland and
Zwislocki, 1971.)


LIST OF TABLES
Table Page
1. Acoustic-Reflex Threshold in dB SL 16
2. Acoustic-Reflex Threshold in dB HL (ANSI-1969) 17
3. Acoustic-Reflex Threshold in dB SPL (re 0.0002
dyne/cm^) 18
4. Test Stimuli Used to Elicit Threshold of Adui-
bility and Threshold of the Acoustic Reflex 58
5. Conceivable Sound Pressure Level Necessary to
Elicit the Acoustic-Reflex Threshold by a
500-Hz Burst at a Signal Duration of 10 Msec 73
6. Discriminant Analysis of the Acoustic-Reflex
Threshold 79
7. Average Slope Change Per Decade of Time of the
Acoustic-Reflex Threshold 79
8. Discriminant Analysis of the Threshold of Audi
bility Data 93
9. Signal Duration Intervals Within Which Time
Constants of Temporal Integration at Threshold
of Audibility Occur 97
vi


36
SIGNAL DURATION IN MSEC
Figure 4. Average temporal integration slopes as a
function of frequency. In general, as frequency decreases
the slope steepens. (Redrawn from Watson and Gengel, 1969.)


ACOUSTIC-REFLEX THRESHOLD
B RE THRESHOLD AT 500 MSEC
33
Figure 17. Inter-subject range of acoustic-reflex
thresholds as a function of temporal integration.


43
Figure 5. Examples of temporal integration. Cochlear-
impaired integration functions can be flatter, can have a
shorter time constant, or both. (Redrawn from Gengel and
Watson, 1971.)


58
Table 4* Test Stimuli Used to Elicit Threshold of Audi
bility and Threshold of the Acoustic Reflex
Test frequency: 500 Hz, 1000 Hz, 2000 Hz
Signal durations: 500 msec, 200 msec, 100 msec, 20 msec,
10 msec
Rise/fall times: 5 msec
Inter-stimulus interval: 700 msec, 1000 msec, 1100 msec,
1180 msec, 1190 msec
Repetition rate: once every 1200 msec
More specifically, the shaped signals consisted of three
sinusoids at 500 Hz, 1000 Hz, and 2000 Hz. These test fre
quencies were chosen because most of the Meniere's group had
normal hearing in the good ear and poorer hearing in the bad
ear at these frequencies.
Five signal durations were used: 500 msec, 200 msec,
100 msec, 20 msec, and 10 msec. The duration was measured
from the gate onset of a stimulus to the cessation of the
stimulus envelope. Five hundred milliseconds represented,
for practical purposes, an infinitely long signal. The re
maining four durations represented two tenfold, or decade,
changes in time, i.e., 10 msec to 100 msec and 20 msec to
200 msec. These values also included the usually designated
limits of linear temporal integration from the long signal
duration of 200 msec to the short tone pulse of 10 msec.
The repetition rate of the stimulus was held constant by
gating once every 1200 msec. Neural independence was main
tained, because the shortest inter-stimulus interval of the


141
Moller, A. R. Bilateral contraction of the tympanic muscles
in man, examined by measuring acoustic impedance-change.
Annals of Otology, Rhinology and Laryngology, 70:735-
752, 1961a.
. Network model of the middle ear. Journal
of the Acoustical Society of America, 33:168-176,
1961b.
. The sensitivity of contraction of the tympanic
muscles in man. Annals of Otology, Rhinology and
Laryngology, 71:86-95, 1962a.
. Acoustic reflex in man. Journal of the
Acoustical Society of America, 34:1524-1534, 1962b.
. Effect of tympanic muscle activity on movement
of the ear drum, acoustic impedance and cochlear
microphonics. Acta Oto-Laryngologica, 58:525-534,
1964.
. An experimental study of the acoustic impedance
of the middle ear and transmission properties. Acta
Oto-Laryngologica, 60:129-149, 1965.
Munson, W. A. The growth of auditory sensation. Journal
of the Acoustical Society of America, 19:584-591, 1947.
Neergaard, E. B. and Rasmussen, G. Latency of the sta
pedius muscle reflex in man. Archives of Otolaryngology,
84:173-180, 1966.
Nerbonne, M. A. A Comparison of Brief Tone Audiometry
with Other Selected Auditory Tests of Cochlear
Function. Ph.D. dissertation. Michigan State Univer-
sity, 1970.
Niemeyer, W. Relations between the discomfort level and
the reflex threshold of the middle ear muscles.
Audiology, 10:172-176, 1971.
, and Sesterhenn, G. Calculating the hearing
threshold from the stapedius reflex for different
sound stimuli. Presented at International Audiology
Congress, Budapest, 1972.
Nixon, J. and Glorig, A. Reliability of acoustic impedance
measures of the eardrum. Journal of Auditory Re
search, 4:261-276, 1964.


Normal Group
Sub
ject
Ear
Classifi
cation
Frequency
in Hz
125
250
500*
1000*
2000*
3000
4000
JG
L
good
10
10
15
18
18
20
10
R
good
15
10
20
16
17
15
15
CP
L
good
0
0
1
7
8
5
0
R
good
0
0
1
4
11
5
5
GP
L
good
5
5
7
9
15
20
15
R
good
5
10
10
1
2
0
5
JP
L
good
0
0
-8
1
4
5
0
R
good
0
0
-8
2
4
0
0
FS
L
good
0
0
-6
-7
-8
0
0
R
good
0
0
-10
-4
-3
0
10
Actual values obtained by Bekesy tracking
Meniere'
s Group
Sub
ject
Ear
Classifi
cation

Frequency
in Hz
125
250
500*
1000*
2000*
3000
4000
AC
L
bad
65
60
59
68
65
70
70
R
good
20
20
19
16
13
35
35
EC
L
bad
65
60
61
50
57
50
50
R
good
10
10
6
5
8
25
45
FH
L
good
5
5
6
15
14
30
55
R
bad
30
25
30
35
49
50
55
BM
L
bad
25
20
18
22
15
5
10
R
good
20
5
8
11
13
10
10
LS
L
good
40
40
39
19
7
10
15
R
bad
45
40
47
46
40
50
50
Actual values obtained by Bekesy tracking.
108


APPENDIX E
RAX DATA OF THE ACOUSTIC-REFLEX THRESHOLD


114
Table 3. Recording Attenuator Linearity. These Data Were
the Pre- and Post-Experimental Values for the
TDH-140 Earphone at 1000 Hz in dB SPL
Pre
Post
0
114.4
114.0
-10
94.0
93.8
-20
83.4
83.3
-30
73.4
73.2
-40
63.6
63.1
-50
53.6
53.2
-60
43.6
43.2
-70
34.3
33.8


Table 2. AcousticReflex Threshold in dB HL (ANSI-1969)
Sample
Number
Frequency
in Hz
250
500
1000
1500
2000
3000
4000
Anderson and Wedenburg, 1968
200 ears
85
88
86

88

92
Burke et al. 1970a
21 Ss

93
96



92
Burke et al., 1970^
21 SS

95
93


--
89
Harford and Liden, 1967-1968
->
92
95
88

84
82
87
Jerger et al., 1972
382 Ss

89
86

88

86
Lamb et al., 1968
19 Ss
83
88
86

87
86
83
Melcher and Peterson, 1972
30 Ss

83
84
--
83

86
Moller, 1961a
2-5 Ss

85

87

81f

Moller, 1962a
1-3 Ss
7 5C
84d
88e
Peterson and Liden, 1972
88 ears
86
85
85

85

85
^electroacoustic impedance bridge,
mechanical impedance bridge.
.300 Hz.
525 Hz.
^1200 Hz.
3200 Hz.


93
10 out of 12 comparisons. The results are listed in Table 9.
There was no statistical difference between cochlear-good
ears and cochlear-bad ears at the 20 msec-200 msec comparison
in two out of three frequencies. Further, there was no
statistical difference between normal and cochlear good-ears
in all comparisons.
Table 8. Discriminant Analysis of the Threshold of
Audibility Data
500 Hz
1000 Hz
2000 Hz
Normal vs. cochlear-good ears
10 msec-200 msec
ns
ns
ns
20 msec-200 msec
ns
ns
ns
Normal vs. cochlear-bad ears
10 msec-200 msec
.10
.05
.05
20 msec-200 msec
. 05
.05
.05
Cochlear-good vs. cochlear-
bad ears
10 msec-200 msec
.05
.05
.05
20 msec-200 msec
ns
.10
ns
ns = significant.
.01 = significant at the 5 percent level.
.10 = significant at the 10 percent level.
As with the ART, there were no significant differences
among the mean values of the three groups, as presented in
Figure 22. Again, the values for the cochlear-impaired ears
lie below those for the normal-hearing ears. The range of
inter-subject thresholds was not as great as that found with


APPENDIX F
DISCRIMINANT ANALYSIS


29
15 dB more conservative than the other reported studies. As
further support he obtained pre- and post-noise exposure
thresholds, pure tone and reflex, on 11 cats. The results
can be seen in Figure 3. An average permanent threshold shift
of 44 dB resulted with an elevation of the ART by only 1 dB
from the pre-exposure levels. The only noticeable effect upon
the ART was a slight increase in the standard deviation across
the four reflex eliciting frequencies. It appears, therefore,
that the acoustic reflex is a loudness-sensing mechanism that
can be used clinically to indicate the possible presence of
loudness recruitment.
To further authenticate the meaning of the acoustic-re
flex level in cochlearimpaired persons, comparisons have
been made with loudness-balance tests and other audiometric
indicators of cochlear lesions. In most cases a comparison
is made with a unilateral end-organ disease and results of
the ABLB, as the standard of loudness recruitment. The re
sults indicate that the ART level above the threshold of
hearing is as good as the ABLB and slightly better than the
SISI for determining an end-organ lesion (Alberti and
Kristensen, 1970; Ewertsen et_ al. 1958; Kristensen and
Jepsen, 1952; Lamb et al., 1968; Liden, 1970; Thomsen, 1955a;
and others). It is definitely a better and more reliable
predictor of sensory damage than the Monaural Bi-frequency
Loudness Balance, the Difference Limen for Intensity or
Frequency, the Uncomfortable Loudness Level, and Bekesy Types


38
Spectral characteristics of the signal as a function of
duration cannot be divorced from the integrating bandwidth,
or the critical band of the ear (Fletcher, 1940; Scharf,
1970) The theoretical bandwidth of a pulsed tone is pro
portional to the reciprocal of the signal's duration, or
1/t. As long as the spectrum of the test tone remains within
the critical bandwidth of the ear, all of the energy will be
integrated. If not, there should be a loss in sensitivity
(Garner, 1947a; Green et rl. 1957; Olsen and Carhart, 1966;
Sheeley and Bilger, 1964; Wright, 1968a). The spread of
energy by abrupt low frequency signals upward to the more
sensitive frequencies of the ear has been misinterpreted as
an apparent increase of sensitivity within the critical band.
It has resulted in the assumption that the ear does not inte
grate low-frequency tone pips (Campbell and Counter, 1969;
Karlovich, Lane, Smith, Tarlow, Thompson, and Vivion, 1971).
The point is that one must carefully monitor the test stimuli
so that the duration of the tone burst is maintained within
the critical bandwidth (Wright, 1968a).
Some variability in temporal integration has stemmed
from lack of defined criteria of stimulus duration. Goldstein
and Kramer (1960) measured the duration between energy onset
and cessation, while Harris (1957) designated as criteria
the half-power points on the stimulus envelope. An "equiva
lent duration" has also be used (Brahe Pederson and
Elberling, 1972; Dallos and Johnson, 1966; Dallos and Olsen,


15
of values for the ART in 382 normal-hearing persons as being
normally distributed with a mean of 85 dB HL (ANSI-1969).
Ninety-five percent of their population fell within 70 dB to
100 dB HL and 99 percent within 65 dB to 105 dB HL. This
range of thresholds has been supported by others (Anderson
and Wedenberg, 1968; Deutsch, 1968, 1972; Harford and Liden,
1967-1968; Peterson and Liden, 1972) although 95 dB to 100
dB HL is considered the upper limit for normals because
auditory lesions outside the cochlea tend to raise the ART,
e.g., conductive problems or eighth nerve lesions (Brooks,
1971; Anderson, Barr, and Wedenberg, 1970).
As shown in Table 1, the ART in SL is rather uniform
from study to study, except for the technique using manome
try. Weiss, et aL. (1962) show ART levels which are 10 dB
to 20 dB less sensitive than those measured by acoustic
impedance. Inspection of Tables 1 and 2 does not indicate any
systematic effect of frequency in sensation level or in
hearing level in normal subjects. In Table 3 more sound
2
pressure level (SPL re 0.0002 dyne/cm ) is necessary to
elicit the ART at lower frequencies. In this respect, the
acoustic reflex is similar to the threshold of audibility
in response to SPL.
The range of ART levels may be due to various types of
acoustic-impedance instruments (e.g., Burke, Herer, and
McPherson, 1970, as shown in Table 2), threshold criteria
(Jerger et ad., 1972 vs. Moller, 1961a, 1962a), a variety


113
Table 2. Step-Attenuator Linearity. These Data Were
the Pre- and Post-Experimental Values for the
TDH-140 Earphone at 1000 Hz in dB SPL (re 0.0002
dyne/cn\2)
1-dB Increments 10-dB Increments
dB
Pre
Post
dB
Pre
Post
0
139.1
139.0
0
139.1
139.0
-1
138.2
138.0
-10
130.0
129.9
-2
137.4
137.2
-20
120.1
120.0
-3
136.4
136.2
-30
110.1
110.0
-4
135.4
135.3
-40
100.3
100.1
-5
134.6
134.4
-50
90.4
90.2
-6
133.7
133.5
-60
80.6
80.4
-7
132.8
132.6
-70
71.2
71.1
-8
131.9
131.8
-9
130.9
130.8
-10
130.0
129.9


APPENDIX D
CALIBRATION DATA TABLES


44
Nerbonne (1970), who studied temporarily fatigued ears,
and Sanders et. ad. (1971), who studied cochlear lesions,
demonstrated that the SISI and ABLB, as well as brief-tone
audiometry are sensitive indicators of cochlear lesions.
Brief-tone audiometry can yield definitive results when the
ABLB shows partial recruitment and negative SISI scores are
present (Sanders et ad., 1971). There is some correspondence
between amount of recruitment and degree of aberration of
temporal integration, but, for the most part, temporal inte
gration is assessing some other aspect of sensory lesions
(Nerbonne, 1970). Abnormal temporal integration seems to be
part of a syndrome of a distorted time domain of enlarged
critical bands and critical ratios (Northern, 1967; Sheeley,
1963; Simon, 1963) and poorer performance on difference limens
for frequency (Sheeley, 1963). Temporal integration, by way
of its relation to critical bands, seems to have a common
basis with several phenomena related to the tuning of the
auditory system (Licklider, 1951) that are not presently
measured clinically. The method of brief-tone audiometry,
therefore, presents clinicians with another possible test
for end-organ abnormalities, if proper standards are set
forth for the various stimulus and test parameters, and nor
mal response limits are determined.


23
between a 220-Hz and a 800-Hz probe tone frequency. This
latter study does not agree with the findings of Mehmke and
Tegtmeier (1970) in which there should be a 8 dB loss in
transformer efficienty for low frequencies. Lilly and
Shepherd (1964) and Feldman (1967) have also observed that
the acoustic impedance varies with the frequency of the probe
tone.
The choice of probe tone frequency may be dependent upon
the length of the sinusoidal wave and its relation to the
dimensions of the external auditory canal. Acoustic impedance
becomes more sensitive to changes in probe position in the
external auditory meatus and to increased diameter of the
meatus as the probe frequency increases. This may be mini
mized by using a low frequency probe in a larger volume, e.g.,
at the external meatus (Schoel and Arnesen, 1962). This con
tention is not supported by Djupesland et al. (1966), who
see no effects of probe position. Djupesland and his asso
ciates used 5-dB test increments, which may have obscured any
small but significant results of plug position.
It might be possible that the attributed frequency ef
fects of the'probe tone are due to the varied, and often
unreported, intensity levels of the probe tone. It is impor
tant that the probe tone not be intense enough to evoke the
acoustic reflex, since it is there to indicate the existence
of intra-aural muscular contraction and not to elicit it
(see footnote 4). The probe tone level must also be above


330
20
10
0
SIGNAL DURATION IN MSEC
14. Mean acoustic-reflex thresholds as a function of temporal integration.
-0
00


80
Although the group means were not statistically differ
ent, the data from Figure 14 indicate two consistent find
ings. First, the cochlear-impaired (bad) group always had
less mean ART change than the other two. The cochlear-good,
for the most part, had the greatest amount of ART change.
The difference between the cochlear-bad and normal ears at
20 msec increased from between 4 dB and 7 dB to 12 dB for
500 Hz, 1000 Hz,and 2000 Hz, respectively. There were no
other large systematic changes in this respect. Second,
the change from 20 msec to 100 msec (or to 200 msec) became
less as frequency increased for the cochlear-bad ears. This
decrease in mean slope caused the former results of increased
ART difference between normal and cochlear-bad groups.
Figure 15 illustrates the large inter-subject variation
of temporal integration slopes for the normal ears and the
cochlear-bad ears. Similar configurations also appear in
Figure 16, contrasting cochlear-good ears and cochlear-
bad ears. In both figures, the slope of individual sub
jects varies over a wide range within a group, so that the
normal and cochlear-impaired groups overlap. The large inter
subject variability is probably the reason why the mean
values were not statistically different from one another.
Figure 17 contrasts the inter-subject range among
the three groups. This graph also depicts the large amount
of overlap among the three groups, which hampered
statistical significance. The least amount of overlap
occurred at the test frequency of 500 Hz and


LIST OF FIGURES
Figure Page
1. Possible time, strength and direction of the
stapedius and tensor tympani muscles, and
resultant displancement of the tympanic
membrane 8
2. The acoustic-reflex arc 10
3. Pre- and post-noise exposure thresholds ... 30
4. Average temporal integration slopes as a func
tion of frequency 36
5. Examples of temporal integration 43
6. Temporal summation of the acoustic reflex . 48
7. Temporal integration at threshold of audi
bility and of the acoustic reflex for normal
hearing subjects 50
8. Acoustic-reflex contraction 60
9. Spectral content of tone bursts with 5-msec
rise/decay ramp 61
10. A 20-msec signal burst with a smooth 5-msec
rise and decay 63
11. Block diagram of equipment used to elicit and
record the intra-aural reflex and to allow
subject control over the test situation . 64
12. Typical response of the acoustic reflex at
threshold 70
13. Upper limits of acceptable exposure to impulse
noise for 95 percent of the population to
10,000 impulses 72
14. Mean acoustic-reflex thresholds as a function
of temporal integration 78
Vll


Subject Medical History Form
1. Name Age Sex
2. Address
3. Date
4. Subject group: Normal Cochlear
5. History:
A.Conductive
1.Recent middle-ear problems: Yes No
Explain
2.Ear operations: Yes No_
Explain
When Intra-aural muscles involved
B.Sensori-neural hearing problems:
1. Ear: L R Both
2. First noticed Duration
3. Diagnosis By whom
4. Meniere's disease
a. Episodes of true vertigo
b. Feeling of fullness of ears
c. Tinnitus
d. Fluctuating hearing loss
e. Frequency of attacks
f. Time of last attack
g. Present state: Active Quiescent
Unknown
5. Hearing aid: Yes No Don't use
C. Brain Injury: Yes No
Severe concussions: Yes No
CVA: Yes No
D. Medication or drugs currently used: Yes No
106


145
Thomsen, K. A. Employment of impedance measurements in
otologic and otoneurologic diagnostics. Acta Oto-
Laryngologica, 45:159-167, 1955b.
Tillman, T. W. Special hearing tests in otoneurologic
diagnosis. Archives of Otolaryngology, 89:25-30,
1969.
, Dallos, P. J., and Kurvilla, T. Reliability
of measures obtained with the Zwislocki Acoustic
Bridge. Journal of the Acoustical Society of America,
36:582-588, 1963.
Vries, H. de. The minimum audible energy. Acta Oto-
Laryngologica, 36:230-235, 1948.
Ward, W. D. (editor). Proposed damage-risk criterion for
impulse noise (gunfire). Working Group 57 of the
NAS-NRC Committee on Hearing, Bioacoustics, and Bio
mechanics. Washington, D. C., 1968.
Watson, C. S. and Gengel, R. W. Signal duration and signal
frequency in relation to auditory sensitivity. Journal
of the Acoustical Society of America, 46:989-997, 1969.
Weiss, H. S., Mundie, M. R. Cashin, J. L., and Shinabarger,
E. W. The normal human intra-aural muscle reflex in
response to sound. Acta Oto-Laryngologica, 55:505-515,
1962.
Wever, E. G. and Lawrence, M. Physiological Acoustics.
Princeton, J. J.: Princeton University Press, 1954.
Wilber, L. A., Goodhill, V., and Hogue, A. C. Comparative
acoustic impedance measurements. Presented at the
American Speech and Hearing Convention, Chicago, 1970.
Wright, H. N. Switching transients and threshold deter
mination. Journal of Speech and Hearing Research,
1:52-60, 1958"!
. Audibility of switching transients. Journal
of the Acoustical Society of America, 32:138, 1960.
. Clinical measurement of temporal auditory
summation. Journal of Speech and Hearing Research,
11:109-127, 1968a.
The effect of sensori-neural hearing loss on
threshold duration functions. Journal of Speech and
Hearing Research, 11:842-852, 1968b.


CHAPTER V
CONCLUSIONS AND SUMMARY
A concluding comment should be made concerning the
nature of diagnostic tests, as viewed through these results.
If the ART did indeed prove to differ cochlear abnormalities
from normals, what can be considered abnormal? A specific
defect or aberration in auditory processing does not neces
sarily reflect one specific and correlative anatomical lesion,
if, indeed, it is possible to obtain a homogeneously impaired
population. The experimental group is a good case in point,
because they were chosen as a homogeneous population with
classical Meniere's syndrome. Three subjects, EC, FH, and
BC, had Meniere's for less than 6 years. They also had
the most depressed slopes. The other two subjects, LS and
AC, had greater than normal slopes and had had their disease
symptoms for 16 to 35 years, respectively. LS, as well as
FH and BC, demonstrated a short time-constant for at least
one frequency, so, apparently, it does not matter how long
one has had the disease. In addition, these results are not
uniform from one frequency to the next for each subject,
even though the hearing losses were relatively flat.
It did not matter if a clinically significant hearing
loss was present, because BC had practically normal pure-tone
100


88
constant of the two good-ear groups is never shorter than
a duration of 20 msec. Interestingly, the opposite observa
tion may be made with the cochlear-impaired group. There
were no long time constants in the 500 msec-200 msec inter
val, but 25 percent of the cochlear-impaired ears, across
all frequencies, demonstrated integration functions that
were in the 20 msec-10 msec interval. That is, temporal
integration for these 4 out of 16 ears may be abnormally
shortened due to the cochlear impairment. This can be most
effectively seen in Figure 20, where three of the cochlear-
impaired ears with a short time constant are compared with
the normal-hearing results at 1000 Hz and 2000Hz. A short
ened time constant has been previously reported in cochlear-
impaired ears for temporal integration at threshold of
audibility (Figure 5).
Inspection of Figures 15 and 16 demonstrates that an
average slope value per decade is not necessarily the best
descriptor of temporal integration. The change is not
linear, especially in the cochlear-impaired group. The
cochlear-bad ears at 500 Hz integrate slowly, with a flatter
slope, until 20 msec when the rate of change becomes greater.
For example, four out of five cochlear-bad ears at 500 Hz
change at an average rate per time decade of 8 dB to 20 msec,
and then abruptly change to an approximate slope of 50 dB
to 55 dB per time decade. This change is only distinctive
between good and bad ears at 500 Hz. This difference


33
Garner and Miller, 1947), the ear would trade 1 log unit of
intensity (1 bel, 10 dB) for 1 log unit of duration (10 msec
to 100 msec, 20 msec to 200 msec, 15 msec to 150 msec). Even
though there is an increase of power as stimulus time in
creases, the acoustic energy required to maintain the thres
hold of audibility is relatively constant.6 This accumulation
of power over an average tenfold decrease in time may be
inferred from threshold data using a single-slope value (Clack,
1966; Garner and Miller, 1947; Munson, 1947; Northern, 1967;
and others). When this slope value is 10 dB it is considered
perfect temporal integration of energy. Perfect integration
is most often observed for the sinusoid of 1000 Hz.
The time constant of temporal integration (T0) marks the
point at which time and intensity cease to have a linear
relationship; it is thought to occur between 200 msec to 250
msec. Goldstein and Kramer (1960) observe integration occur
ring through 200 msec, but Harris, Haines, and Myers (1958)
report individual time constants ranging from 100 msec to
300 msec with a mean of 200 msec. Plomp and Bouman (1959) and
Hempstock, Bryan, and Tempest (1964) indicate that T0 is
inversely related to frequency; T0 changes from about 375 msec
at 250 Hz to approximately 150 msec at 8000 Hz. Regardless
6Acoustic energy can be stated in the simplest form as
E = PT, where E is acoustic energy, P is acoustic power, and
T is linear for a certain period of time, the time constant.
Further increases in signal duration beyond the time constant
have less and less effect upon threshold level.


139
Lilly, D. J. and Shepherd, D. C. A rebalance technique
for the measurement of absolute changes in acoustic
impedance due to the acoustic reflex. ASHA, 6:
381(a), 1964.
Lindsay, J. R., Kobrak, H. G., and Perlman, H. B. Relation
of the stapedius reflex to hearing sensation in man.
Archives of Otolaryngology, 23:671-678, 1936.
Loeb, M. Psychophysical correlates of intratympanic reflex
action. Psychological Bulletin, 61:140-152, 1964.
Lorente de No, R. The reflex contractions of the muscles
of the middle ear as a hearing test in experimental
animals. Transactions of the American Laryngology,
Rhinology and Otology Society, 39:26-42, 1933.
. The function of the central acoustic nuclei
examined by means of the acoustic reflexes.
Laryngoscope, 45:573-595, 1935.
and Harris, A. S. Experimental studies in
hearing. Laryngoscope, 43:315-326, 193-3.
Madsen. Madsen Model Z070 Electro-Acoustic Impedance
Bridge: Applications and instructions for use.
Copenhagen: Madsen Electronics, n.d.
Martin, F. N. The short increment sensitivity index (SISI) .
In Katz, J. (editor), Handbook of Clinical Audiology.
Baltimore: Williams and Wilkins Company, 1972.
and Wofford, M. J. Temporal summation of brief
tones in normal and cochlear-impaired ears. Journal
of Auditory Research, 10:82-86, 1970.
McRobert, H. The response of the tympanic muscles in
human ears: Possible false inferences from results
of reflex testing on normal and pathological ears.
Sound, 2:71-76, 1968.
, Bryan, M. E., and Tempest, W. The acoustic
stimulation of the middle ear muscles. Sound and
Vibration, 7:129-142, 1968.
Mehmke, S. and Tegtmeir, W. The diagnostic value of
impedance measurements. Fenestra. (An eight page
insert between pp. 12-13,) April 1970.


140
Melcher, J. A. and Peterson, J. L. The effects of age
and hearing impairment on the acoustic reflex decay.
Presented at the American Speech and Hearing Con
vention, San Francisco, 1972.
Mendelson, E. S. A sensitive method for registration of
human intratympanic muscle reflexes. Journal of
Applied Physiology, 11:499-502, 1957.
. Improved method for studying tympanic re
flexes in man. Journal of the Acoustical Society of
America, 44:146-152, 1961.
. The lability of the resting and reflex activity
of the human middle ear muscles. In Fletcher, J. L.
(editor), Middle Ear Function Seminar, U. S. Army
Research Laboratory Report No. 576, Fort Knox,
Kentucky, 1963.
. Acoustic reflexometry. Acta Oto-Laryngologica,
62:125-139, 1966.
Metz, 0. The acoustic impedance measured in normal and
pathological ears. Acta Oto-Laryngologica, Supplement
63:1-254, 1946.
Threshold of reflex contractions of muscles
of middle ear and recruitment of loudness. Archives
of Otolaryngology, 55:536-544, 1952.
Miller, G. A. The perception of short bursts of noise.
Journal of the Acoustical Society of America, 20:
160-170, 1948.
Miskolczy-Fodor, F. Monaural loudness-balance test and
determination of recruitment degree with short sound-
impulses. Acta Oto-Laryngologica, 43:573-595, 1953.
. The relation between hearing loss and recruit
ment and its practical employment in the determination
of receptive hearing loss. Acta Oto-Laryngologica,
46:409-415, 1956.
. Relation between loudness and duration of
tonal pulses. II. Response of normal ears to sounds
with noise sensation. Journal of the Acoustical
Society of America, 32:482-486, 1960.
's
Moller, A. R. Intra-aural muscle contraction in man,
examined by measuring acoustic impedance of the ear.
Laryngoscope, 68:48-62, 1958.


Figure 11. Block diagram of equipment used to elicit and record the intra-aural re
flex and to allow subject control over the test situation.


61
Figure 9. Spectral content of tone bursts with 5-msec
rise/decay ramp. (a) 1000 Hz, 10-msec impulse with half
power width of ca. 120 Hz; (b) 1000 Hz, 20-msec impulse with
half-power width of ca. 50 Hz.


Appendix E(continued)
Sub
ject
BC,
LS
R
LS,
500 Hz
1000 Hz
2000 Hz
Signal Duration in Msec
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
1
118
114
108
106
106
112
110
102
102
104
113
103
97
101
99
2
124
112
110
106
104
112
110
104
104
104
113
107
97
99
99
3
120
112
108
106
104
114
108
104
102
102
111
105
97
97
99
4
120
112
110
106
106
114
108
104
102
104
111
103
95
103
97
5
120
114
108
106
108
116
108
104
104
102
109
105
97
103
97
6
124
112
110
104
106
112
108
104
104
100
109
105
99
95
97
7
122
114
110
106
106
114
110
102
102
104
111
103
97
97
97
8
120
110
112
106
104
114
106
102
104
104
111
105
97
99
97
9
120
112
108
108
106
114
108
102
104
102
117
103
99
103
95
10
124
114
110
106
106
114
110
104
104
100
109
105
97
101
97
1
125
99
99
99
93
124
110
100
92
98
129
96
100
96
94
2
127
101
99
91
91
128
108
102
96
98
131
96
96
96
96
3
127
99
99
97
93
124
106
100
98
100
135
96
96
94
92
4
121
103
101
97
95
122
106
100
96
98
129
96
96
96
92
5
125
109
97
99
95
126
108
100
94
98
129
94
100
96
94
6
125
109
99
99
93
130
110
102
96
94
131
94
96
94
96
7
125
109
95
99
93
128
110
102
92
96
131
96
92
94
92
8
123
109
99
97
93
124
110
100
98
98
133
102
100
98
94
9
125
109
101
99
95
126
108
100
98
98
135
100
94
96
96
10
125
105
99
99
95
130
110
100
98
100
137
96
94
92
94
1
137
123
99
95
91
2
139
121
103
95
89
3
139
119
99
97
93
4
139
121
101
99
91
5
139
121
99
97
93
123


9
It is also surmised that fibers from the accessory
nucleus pass to and from the lateral lemniscus into the
ipsilateral motor nucleus of the trigeminal nerve (n. V)
(Rasmussen, 1946), as the central portion of the tensor tym-
pani-reflex arc. The efferent portion of this reflex arc
constitutes the tensor tympani muscle innervated by the
mandibular branch of the trigeminal nerve. The general schema
for the acoustic-reflex arc as proposed by Rasmussen (1946) is
shown in Figure 2.
Simmons (1963) does not necessarily take exception to
Rasmussen, but suggests that there are several possible re
flex loops, both ipsilateral and crossed. Simmons speculates
that the major differences in latency and response level be
tween the stapedius and tensor tympani muscles are due to the
relative degree of synaptic connections of their respective
reflex arcs. According to Simmons, the stapedial reflex
probably has the more compact, less diffuse, interneuron
connections, thereby explaining its greater sensitivity.
Each of the proposed schema would explain the lower reflex
threshold and shorter latency of the stapedial acoustic-
reflex arc activity in comparison to the tensor tympani
acoustic-reflex responses (as seen in Figure 1). In either
case, the SOC is the most probable reflex center. Neural
activity in the SOC increases with loudness to at least 60
dB sound pressure level (Boudreau, 1965). The acoustic re
flex probably occurs when the neural activity in this


52
In summary, temporal summation is a neural event origi
nating at the cochlea and is essentially complete before the
level of the superior olivary complex (SOC), where the
acoustic reflex is activated. Growth in neural activity in
the SOC parallels the growth in loudness up to 60 dB SPL.
This neural activity decays in an exponential manner, thereby
appearing to trade energy units linearly for about 200 msec.
It is suggested that during acoustic stimulation, neural activ
ity in the reflex center exceeds a certain level and causes
a contraction of the intra-aural muscles. Furthermore, the
mechanism suggested as controlling and/or affecting temporal
summation of threshold and loudness is the filtering charac
teristics of the ear. Lesions in the cochlea can modify the
transmission characteristics through the auditory system in
such a way that the acoustic reflex and temporal integration
are changed in a predictable manner.
The foregoing studies indicate two important clinical
premises: (1) it is not necessary to compare the acoustic-
reflex threshold with the threshold of audibility to obtain
diagnostic information about the state of the cochlea, and
(2) there are factors affecting the acoustic reflex which may
differentiate normal ears from those with end-organ lesions.
Hypothesis
The major hypothesis, formulated for testing in this
study, is based on information concerning temporal summation
of the acoustic reflex: the acoustic reflex changes in a


27
The manometer system in an electroacoustic impedance
bridge is able to approximate the amount of pressure in
the middle-ear. For the most part, 50 mm of equivalent
water pressure (re atmospheric pressure) is considered within
normal limits. Hearing and acoustic-reflex thresholds do not
deteriorate within this pressure range in the middle-ear
(Alberti and Kristensen, 1970; Peterson and Liden, 1970).
Negative middle-ear pressure is more detrimental to thresholds
than positive pressure (Moller, 1965), but the AR may be
stronger with slightly negative canal pressure (Terkildsen,
1964). Terkildsen (1960c) even states that the stapedial
reflex is found around normal atmospheric pressure, while the
tensor reflex is enhanced by negative pressure.
In the absence of middle-ear conductive problems, the
level at which the AR is just elicited above the threshold of
audibility is useful in confirming a cochlear abnormality.
It has been observed that in the majority of mild to moderate
sensorineural hearing losses the acoustic-reflex threshold
occurs at approximately the same hearing-threshold levels (HL)
as the normal ART (Alberti and Kristensen, 1970; Ewertsen,
Filling, Terkildsen, and Thomsen, 1958; Jerger et al^. 1972;
Klockhoff, 1961; Kristensen and Jepsen, 1952; Lamb, Peterson,
and Hansen, 1968; Metz, 1946, 1952; Peterson and Liden, 1972;
Terkildsen, 1960b; Thomsen, 1955a,b; and others). If the
ART occurs at 55 dB to 60 dB SL or less, the hearing loss is
considered to be due to cochlear impairment. The ART is


73
Table 5. Conceivable Sound Pressure Level Necessary to
Elicit the Acoustic-Reflex Threshold by a 500-Hz
Burst at a Signal Duration of 10 Msec
Normal threshold at zero hearing level
(ANSI-1969) in dB SPL 11.5 dB
The ART at the upper limit of normal
sensation level 100.0 dB
Power increase necessary to maintain thres
hold when a tone burst is shortened to
10 msec 25.0 dB
Total SPL 136.5 dB
to 10,000 pulses in some period of time, i.e., one hour to
one day. This limit applies to impulses with a known single
rise-peak decay, not a complex impulse. The Committee
stated that, although there are many unknowns, their data for
upper limits of exposure constituted a conservative estimate.
Only the "weakest ears" of the unaccounted 5 percent might
demonstrate a temporary threshold shift because the acoustic
reflex would afford some protection in some subjects. The
horizontal dashed line in Figure 12 indicates the maximum
output of the earphone within this experimental system at
139 dB SPL and the vertical hatched lines indicate the approxi
mate durations used in this study. As may be seen in Figure
12, the maximum output of the earphone falls below the maxi
mum acceptable exposure limits.
To minimize possible discomfort from loud acoustic
impulses, the subjects were allowed some control over the
test situation. If the subjects felt that the stimulus pulse


86
Zwislocki mechanical impedance bridge. There have been
reports indicating that the electroacoustic and mechanical
impedance bridges can vary as much as 35 percent in their
impedance values, depending upon the condition and size of
the tympanic membrane (Lilly, 1970; Wilber, Goodhill, and
Hogue, 1970).
The statistical observations concerning the null
hypothesis of the ART were limited because of unexpected
inter-subject variability causing the sample size of five
normal and five cochlear subjects to be ineffectual statis
tically. However, descriptive observations can be made,
although these observations may be due to chance.
The time constant of temporal integration of the ART,
3
that point in time where integration begins, is shorter
than the 200 msec reported for threshold audibility. A
shortened time constant has been reported previously by
Small et al. (1962) for temporal integration at supra-
threshold levels. As illustrated in Figure 19, the ART time
constant is most often between the signal durations of
100 msec and 20 msec, or the 100 msec-20 msec interval.
Further inspection of Figure 19 suggests that the time
3
The time constant was defined in this experiment as
1/e of the maximum threshold change. In practicality, 1/e
was 1/2.718, or 32 percent, of the mean integration slope
per decade of stimulus time in normal ears, or 6 dB. The
first 6 dB of threshold change from 500 msec, therefore, was
not considered part of temporal integration. The point at
which 7 dB, or more, of threshold change occurred due to
change in signal duration marked the length of the time
constant.


34
where Tq may fall, integration is essentially complete between
500 msec and 1000 msec (Zwislocki, 1960) .
The phenomenon of integration of acoustic power over
time is attributed to temporal summation in the auditory
system and, most probably, is neural in nature (Zwislocki,
1960)- The term summation is introduced by some authors be
cause it is felt that the response pattern represents tem
poral summation at the synaptic junctions in the neural path
ways. More specifically, this apparent linear function of
log time and log power is due to an exponential decay of the
persisting neural excitation (Plomp and Bouman, 1959; Zwis
locki, 1960) At threshold, there appears to be a direct
proportionality between an increase in the intensity of
acoustic power and the increase in neural excitation, which
is modified somewhat at suprathreshold levels by neural adap
tation and the loudness of the stimuli (Zwislocki, 1960) .
The neural mechanism for temporal integration probably exists
at the neurons above the first and, possibly, the second order,
but before the level of binaural interaction at the superior
olivary complex (Zwislocki, 1960) .
Stimulus parameters
The literature indicates that, in addition to duration,
other various stimulus parameters affect temporal integration.
Those most often cited are stimulus spectrum, rise and decay
time, definition of the signal duration, inter-stimulus


2
test procedures normally employed in the assessment of end-
organ status, in particular those tests considered most
valuable in differential diagnosis, will be reviewed in this
section: Alternate Binaural Loudness Balance, Monaural
Loudness Balance, Short Increment Sensitivity Index, and
Bekesy audiometry.^
The Alternate Binaural Loudness Balance (ABLB) test
(Fowler, 1936) has long been the indicator of loudness recruit-
2
ment. This abnormally rapid growth m loudness is theo
retically regarded as pathognomonic of cochlear lesion
(Dix, Hallpike and Hood, 1948; Fowler, 1936, 1950; Hood, 1969;
Jerger, 1967, among others). In spite of its high validity
in confirming recruitment (Tillman, 1969), this test cannot
be administered to all patients. The patient must have one
normal, non-recruiting ear at each test frequency and at
least a 20 dB threshold difference between ears to obtain
meaningful data. Further, it is estimated that more bilateral
hearing losses than unilateral exist (Carver, 1972) Thus,
It is generally understood that audiometric evaluations
do not measure anatomical lesions, per se. Instead, they
reflect functional alterations in the auditory response caused
by anatomical lesions or physiological changes.
2
Loudness recruitment (Fowler, 1937) is an auditory
phenomenon where the supraliminal loudness growth of one ear
occurs at a faster rate than that of the contralateral ear to
equal intensity units. Whan the threshold differences between
the two ears is more than 20 dB and the supra-threshold
balances are considered equal within 10 dB, the loudness
growth is considered abnormal. This abnormal loudness growth
is called loudness recruitment. (Fowler, 1936; Hood, 1969;
Jerger, 1967.)


56
The subjects had a negative history of middle-ear or
inner-ear surgery and of concussion, brain damage, or
cerebral vascular accidents. Those subjects using drugs or
medication, possibly influencing normal intra-aural muscle
function, were excluded. The medical history form employed is
shown in Appendix A. None of the subjects had conductive
hearing problems. Normal compliance at the tympanic membrane
and normal tympanograms confirmed the absence of a functional
middle-ear problem.
Data from the five control subjects were collected for
each ear. These five subjects demonstrated pure-tone audio
metric measures no worse than 20 dB HL (ANSI-1969) in either
ear for frequencies 125 Hz through 4000 Hz. The normal-hear
ing group ranged from 23 years to 30 years of age.
Only those persons having a classical, unilateral
Meniere's disease, a pure end-organ disorder, with normal
hearing in the contralateral ear, were included in the experi
mental group. The Meniere's disease was medically diagnosed.
The patients had, in the course of their medical history, the
classical symptoms of episodes of true vertigo, fluctuating
hearing loss, "roaring" tinnitus, and feeling of fullness in
the affected ear. These symptoms were supported by the
following objective findings: a sensorineural hearing loss,
loudness recruitment and vestibular canal paresis. This
group will be referred to as the Meniere's group or the
cochlear-impaired group. The Meniere's group ranged from 45
years to 63 years of age.


39
1964; Olsen and Carhart, 1966). An equivalent-duration tone
pip contains the same amount of energy as a rectangular
envelope and allows for the comparison of widely varying
envelope shapes. The use of the equivalent duration fits or
allows the conversion of the overall stimulus envelope to fit
the slope of Garner's model (1947b), which states that not all
energy is used by the ear to summate temporal information
(Dallos and Olsen, 1964). The best measure of the acoustic
waveform is still a moot question because the conversion from
the half-power points to equivalent duration of Harris' data
(1957) by Dallos and Olsen (1964) demonstrate no difference.
The time in which the stimulus envelope rises to and
decays from its effective peak is a critical variable in
auditory measures using pure-tone stimuli. If the stimulus
starts or stops too abruptly, a wide-band transient may be
generated, thereby biasing the results (Wright, 1958, 1968a).
A minimum 5 msec rise/fall time measured on the linear por
tion of the ramp (between 10 percent and 90 percent of maximum
amplitude of the acoustic waveform) is most often suggested
(Harris, 1957; Wright, 1960, 1968a). No differences occur
with rise/fall times varying from 0 msec to 50 msec if the
equivalent duration is held constant (Dallos and Johnson,
1966; Dallos and Olsen, 1964).
The inter-stimulus interval, or the off-time between
successive stimulus envelopes, must be kept long enough to
ensure neural independence between stimulus events. Since


CHAPTER IV
RESULTS AND DISCUSSION
The results of this study indicate that temporal summa
tion of the acoustic-reflex does not function differently in
those subjects who have normal hearing and in those subjects
who have a known end-organ auditory lesion. A statistical
analysis of the data did not support a consistent, significant
difference between normal-hearing (good) ears and cochlear-
impaired (bad) ears, using the acoustic-reflex threshold
(ART). However, there were statistically significant differ
ences between responses of the two groups at threshold of audi
bility. An analysis of these data follows in a comparison of
normal-hearing and hearing-impaired ears at 1) temporal inte
gration at the acoustic-reflex threshold, 2) temporal integra
tion at the threshold of audibility, and 3) temporal integra
tion at the acoustic-reflex compared to temporal integration
at threshold of audibility.
1) Temporal Integration at
the Acoustic-Reflex Threshold1
2
A discriminant analysis (Rao, 1952) was performed on the
ART difference between a 10-msec and a 200-msec duration tone
burst and between a 20-msec and a 200-msec duration tone
^The raw data are reported in Appendix E.
2
Further information concerning discriminant analysis
is contained in Appendix F.
76


89
Figure 20. The delayed time constant in some cochlear-
impaired ears contrasted with the normal time constant of
temporal integration at the acoustic-reflex threshold.


54
6. Is there a significant difference in temporal integra
tion between the normal-hearing ear and the cochlear-
impaired ear of a Meniere's group?
7. If cochlear-impaired ears demonstrate reduced integra
tion at threshold of audibility, do they also demonstrate
reduced temporal integration at the threshold of the
acoustic reflex?


8
TIME 23-
Figure 1. Possible time, strength and direction of
the stapedius and tensor tympani- muscles, and resultant
displacement of the tympanic membrane. (Redrawn from
Mendelson, 1963.)


13
contractions of the intra-aural muscles may not produce a
measurable change in extratympanic air pressure. Instead,
these muscle contractions are reflected in a measurable in
crease in the impedance to acoustic energy within the middle
ear.
The technique used most extensively in the past decade
to indicate intra-aural muscle contraction is acoustic-
impedance measurement. Simply stated, a probe tone is directed
perpendicularly towards the plane of the tympanic membrane.
While most of this acoustic energy is transmitted through the
tympanic membrane and attached ossicular chain to the oval
window, a portion of the acoustical energy wave is reflected
back into the external canal from the tympanic membrane.
Since the probe tone input and the reflected tone pick-up
microphone are connected to the external auditory meatus
with an airtight seal, the reflected probe tone can be
monitored accurately. As the intra-aural muscles contract,
the acoustic impedance increases, causing an increased amount
of the probe tone to be reflected back into the external canal.
It is the increased amplitude and phase change of the reflected
tone which indicate that a change in intra-aural muscle
5
activity has occurred. This type of measurement has passed
^Unless specifically stated otherwise, all acoustic-
reflex data concern the ear in which the reflex is elicited.
The ear at which the reflex is elicited may not necessarily
be the same ear in which the reflex contraction is recorded;
most often the reflex ear and the measurement ear are contra
lateral to each other when using the acoustic impedance


90
probably contributed to the only two rejections of the null
hypothesis which occurred (Table 7).
This abrupt change in slope occurred in the 20 msec-
10 msec interval, the same interval in which the short time
constant of cochlear-imparied ears fell in Figure 20. Since
the definition of the time constant was relatively arbitrary,
the slow-abrupt change in slope and the short time constant
mey reflect the same phenomena in the cochlear-imparied ear.
From the discriminant-analysis data one is able to
identify those subjects consistently classified correctly,
although no statistical significance is attached. Two
ehlear-impaired subjects, FH and BC, were never classified
£ls normal hearing, while EC was incorrectly classified in
only 2 out of 12 good-bad ear comparisons. Of the five
feat subjects, the data from these three indicated that one
§huld feo able to distinguish a cochlear-impaired ear based
R aeystic-reflex data. The cochlear abnormality may
ayse §Lf\ aberration in temporal integration of the ART as
inie^ted by an initial slow growth of the integration func-
fi or by an unusually short time constant. One cochlear-
irogaird subject is an appropriate example. LS was incor
rectly classified 9 out of 12 times. Two of the correct
Classifications were probably based on her very short time
fistanf, Her ART changed only 2 dB as the signal duration
§ §h@rtened from 500 msec to 20 msec, then it changed
afeFUpfely by 35 dB. She is represented in Figure 20 at 2000
S fey %he white hearts.


12
(Kobrak, 1948; Lindsay, Kobrak, and Perlman, 1936; Lorente
de No, 1933; Lorente de No and Harris, 1933; Perlman and
Case, 1939). Since visual observation limits the accuracy
of data quantification, other approaches have been developed
for studying the reflex activity. Electromyography (EMG)
measures the individual muscle fiber action potentials and
is, therefore, the most direct method. EMG activity from
both muscles in humans has been reported in response to both
acoustic and non-acoustic stimulation (Djupesland, 1965;
Fisch and Schulthess, 1963; Salomon and Starr, 1963). This
technique is not clinically feasible.
Extratympanic manometry has been the method employed to
determine direction of tympanic membrane movement. This
technique is made possible by sealing the external auditory
meatus with a probe containing a pressure-sensing device.
If the external meatus is properly sealed, the extratympanic
air pressure should decrease or increase with respective
inward or outward movement of the tympanic membrane (Flottorp
and Djupesland, 1970; Holst et al^., 1963; Liden et al. 1970;
Mendelson, 1961, 1963; Terkildsen, 1957, 1960a,c; Weiss et al.,
1962). McRobert (1968), in a review of this measuring tech
nique, noted that the theory for the tympanic membrane response
to individual muscle contraction is valid, but that there are
many inexplicable results (viz. Mendelson, 1957) suggesting
the need for better instrumentation. Moller (1964) and
Neergaard and Rasmussen (1966) also warn that small


CHAPTER II
STATEMENT OF THE PROBLEM
Innumerable studies detailing the identification and
quantification of cochlear abnormalities indicate that, with
patients who cannot or will not cooperate fully, the present
diagnostic measures are inadequate in obtaining information
on auditory processing. Further understanding of the func
tion of the acousticreflex threshold and the effect of
temporal integration upon threshold measures may fill this
void in diagnostic testing of the auditory system. Such a
measure may be obtained by maintaining a constant reflex
strength as a function of increasing acoustic power while
signal duration is decreased, or, in other words, temporal
integration of the acoustic reflex. Since there are rela
tively few data concerning temporal integration of the acous
tic reflex, the pertinent literature will now be reviewed.
As early as 1935, Lorente de No reported the effects of
varying short signal duration upon the contraction of the
tensor tympani muscle of the cat. He used as much as 140
dB SPL and durations from 3 msec to 100 msec. He explained
his results in terms of neural summation, which occurs at the
synapse.'*' At some synapses one pre-synaptic impulse is not
*"For further information regarding synaptic summation
see the classical treatise of Sherrington (1906, 1947);
also see Brink (1951).
45


50
a
3
o
as
en
g
as
Eh
Figure 7. Temporal integration at threshold of audi
bility and of the acoustic reflex for normal-hearing subjects.
The median slope of the acoustic reflex is about three times
the change in the median slope of audibility. (Based on
Djupesland and Zwislocki, 1971, and Olsen et al., 1973.)


Appendix E--(continued)
500 Hz
1000 Hz
2000 Hz
Sub- Test
Signal Duration in Msec
ject
Run
10
20 100 200 500
10 20 100 200 500
10
20 100 200 500
6
__ __
123
101
99
91
7
137
121
101
97
93
8
139
121
105
97
93
9
137
121
101
95
95
10
139
119
99
97
93
CP =
left
ear of
subject
CP.
II
u
cu
V
right
ear of subject
CP.
f
124


65
Response switch II interrupted the signal presentation at the
electronic switch, while response switch III activated a buz
zer audible to the experimenter. The reflex was measured from
the contralateral external ear canal which was sealed with an
ear olive. The reflected probe tone passed to the Peters
AP61 electroacoustic impedance bridge. The output of the
1
bridge was led through a direct current voltage divider and,
finally, to channel two of the oscilloscope.
To measure threshold of audibility, switches I and II were
closed so that the signal passed from the step attenuator
through the matching transformer to the earphone. Switch III
allowed the signal to be monitored at the oscilloscope and volt
meter without being audible to the subject. The recording at
tenuator was activated by subject response switch I. Thus, a
permanent record of the subject's response was made with a
self-tracking Bekesy technique.
The same basic equipment was utilized to measure the
acoustic-reflex threshold. The recording attenuator was
turned off. The test signal was diverted through the McIntosh
power amplifier to the earphone by using the same switches.
The eliciting signal was stored on the oscilloscope of
channel one, while the acoustic reflex was stored on Channel
two. A photographic record of both oscilloscope channels was
made using a Rolleiflex SL 35 camera with a 50-mm lens.
Calibration was performed at both the output of the
earphone in SPL and across the earphone in volts. Calibration
^The d.c. voltage divider was placed in the system to in
crease the sensitivity of the signal. See Appendix C for the
schematic diagram.


APPENDIX B
PURE TONE HEARING LEVEL (ANSI-1969) OF ALL SUBJECTS


ACOUSTIC-REFLEX THRESHOLD
.B RE THRESHOLD AT 500 MSEC
4 0
1 0
i^Nprmal eaifs 50Q Hz .. Normal ears 1000 Hz .. Normal ears 2000 Hz
i n
11 30-
> <
-a.,, 20 H
111
10 -
0 -
Cochlear-bad ears
1000 Hz
Cochlear-bad ears
2000 Hz
10 20
T- 1 r
n 1 r
100 200 500 10 20 100 200 500 10 20
SIGNAL DURATION IN MSEC
100 200 500
Figure 15. Acoustic-reflex thresholds of normal and cochlear-impaired (bad) ears as
A function of temporal summation.
oo


67
calibrated on the same equipment and in a similar manner,
the data reflect the respective calibration curves. In other
words, the data will be reported as if one earphone were used.
Procedure
The experimental testing was conducted in three sessions
for each individual. The temporal integration at threshold
of audibility was measured for both ears during the first test
session, while temporal integration at threshold of the
acoustic reflex was measured in subsequent sessions for each
of the two ears. The three stimulus frequencies were ran
domized for each ear,and the signal durations were presented
in random order for each test frequency. All auditory measures
were made in a single-walled IAC sound treated chamber.
Temporal Integration at Threshold of Audibility
The subject responses were recorded by using a self
tracking, Bekesy procedure. The recording attenuator was set
at a chart speed of three-fourths of an inch per minute and
1 dB of attenuation every second. The visual mean of a
minimum of 10 threshold crossings was used to establish
threshold. This mean value was located at the nearest whole
dB,and, if the threshold crossings were not stable, the last
10 stable crossings were used to establish the mean threshold
value. The total test time for both ears, including a 15- to
30-minute rest period midway, was two to two and one-half
hours. Since signal crossover to the good ear was not observed


Discriminant Analysis
Discriminant analysis is a form of multivariate
analysis. This type of analysis places the individual
values from two known groups on a continuum. Based upon
the group means and variances, a point on the continuum
divides the individual values into two regions representing
two discrete distributions. Each individual is assigned to
the region to which it has the highest probability of
belonging. This posterior probability of an individual
coming from each group is based on group means, standard
deviations, a group covariance matrix, and a group corre
lation matrix. An approximate F statistic tests equality
of group means.
Once the individuals are assigned to each region,
a contingency table can represent the number of correct
and incorrect classifications, as in Table 1. There were
9 correct and 1 incorrect classifications of 10 normal ears
and 4 correct and 1 incorrect classifications of 5 abnormal
ears.
Table 1. Contingency Table of Discriminant Analysis
N
A
N
9
1
A
1
4
126


68
nor reported by the subjects in the preliminary or experi
mental sessions, a masking noise was not applied to the good
ear of the Meniere's group.
Temporal Integration at the Thres
hold of the Acoustic Reflex
Testing of the acoustic reflex commenced once the exter
nal meatus of the reflex measurement ear was sealed with an
ear olive containing the electroacoustic bridge probe tip.
The seal was considered adequate if a positive 200 mm of
equivalent water pressure in the external auditory canal did
not break the seal. The electroacoustic bridge was periodi
cally checked to determine if the balance dial was within
25 mm of equivalent water pressure of zero. If the balance
meter was out of the target range it was rebalanced by adjust
ing the compliance dial. In order to maintain this criterion,
2
the sensitivity dial was kept on position one. If the
pressure dial was outside the target range, the subject was
asked to perforin a Valsalva or Toynbee maneuver to increase
or decrease, respectively, the middle-ear pressure.
A modified method of limits was employed to determine
the acoustic-reflex threshold. Ten ascending ART trials were
2
Although it is suggested that the higher sensitivity
dial settings be used while determining the acoustic reflex,
i.e., two or three on the Peters electroacoustic bridge and
three or four on the Madsen electroacoustic bridge, prelimi
nary observations suggested an ART difference of no more than
2 dB between sensitivity position one and position three. The
use of sensitivity position one reduced the amount of rebalanc
ing necessary throughout the test procedure. If more sensi
tivity was desired, the vertical multiplier dial or the oscil
loscope was changed.


11
synaptic complex exceeds some critical level of neural ex
citation (Lorente do No, 1933, 1935).
The stapedius muscle is generally considered to be the
most active and dominant muscle in response to acoustic
stimulation in humans, whereas the tensor tympani reacts to
4
acoustic stimulation less often (Flottorp and Djupesland,
1970; Holst, Ingelstedt, and Ortegren, 1963; Jepsen, 1963;
Liden, Peterson, and Harford, 197C; McRobert, 1968; Mendelson
1961, 1966; Terkildsen, 1957, 1960a,c; Weiss, Mundie, Cashin,
and Shinabarger, 1962; and others). It appears that 1
percent to 5 percent of the population with normal hearing
and with no apparent middle-ear pathology have no demonstrable
acoustic-reflex response (Shiffman, 1972; Swannie, 1966;
Terkildsen, 1960c; Weiss et al. 1962; Wright and Btholm,
1973) .
Methods of detecting the intratympanic
muscle activity in man
Some investigators of the intratympanic muscle activity
have attempted to make direct observations of muscle contrac
tions through surgical exposure or via chronic tympanic mem
brane perforations. For the most part, however, only* the
stapedial tendon is directly available for easy visualization
4
There has been considerable discussion of whether the
tensor tympani muscle is active in man to acoustic or non
acoustic stimulation. This is a clinical question of consider
able importance. McRobert (1968), in a critical review of
the literature, and Liden, Peterson and Harford (1970) agree
that tensor activity is present in man but that the stapedius
predominates.


PEAK PRESSURE LEVEL (dB RE 0.0002 DYNE/CM
72
SIGNAL DURATION IN MILLISECONDS
Figure 13. Upper limits of acceptable exposure to impulse
noise for 95 percent of the population to 10,000 impulses.
(Redrawn from Ward, 1968.)


77
burst at each test frequency. In Figure 14, for example,
the mean change for the cochlear-good ear at 500 Hz from
10 msec to 200 msec is 29 dB while the mean change at the
same parameters for the cochlear-bad ears is 22 dB. The
20-msec to 200-msec difference was also chosen for discrimi
nant analysis because the change was large enough to yield
mean differences among groups (see Figure 14) and because
some subjects did not yield an acoustic reflex at 10 msec
with the maximum acoustic power of 139 dB SPL. All thresh
old differences and the data in some figures were subtracted
from, or normalized to, 200 msec. Two-hundred milliseconds
was most often stated as the time constant of temporal inte
gration, and there was little difference between the ART
elicited by a 500-msec and by a 200-msec duration tone. The
mean change of ART between the two durations was 2 dB to
3 dB, depending on frequency and subject group. The figures
were normalized to the signal duration which delineated the
results most clearly.
The results of discriminant analysis are reported as
an "F" ratio. Table 6 gives the statistical results of the
discriminant analysis upon the ART. Two comparisons out of
12 between good and bad ears were statistically significant,
both at 500 Hz and both at the ART difference between 20
msec and 200 msec. Ten of the 12 good-bad ear comparisons
were not significantly different.
The mean ART values of the cochlear-impaired ear, as
demonstrated in Figure 14, were always less than those of the


49
No explanation appears to be offered for the steeper
slope of the acoustic reflex, but the temporal summation ef
fect is there without question as illustrated in Figure 7. The
range of values reported by Olsen, Rose, and Noffsinger (1973)
for temporal integration at threshold of audibility are con
trasted with the data by Djupesland and Zwislocki (1971) for
temporal integration at threshold of the acoustic reflex.
These are both at 1000 Hz in normal ears and have been normal
ized to 200 msec so that the slope has a common reference be
tween the two sets of data. The slope for the acoustic-reflex
threshold at 25 dB/decade is about three times as large as
the median slope of 8 dB/decade at threshold of audibility.
Other authors have reported effects of short stimulus dura
tions upon the acoustic reflex, but their results are not
reported systematically as temporal integration, nor in a form
easily converted to the values presented in Figures 6 and 7
(Johansson, Kylin, and Langfly, 1967; Lilly, 1964; Moller,
1962b; Weiss, Mundie, Cashin, and Shinabarger, 1962).
There have been no published reports dealing with the
effects of temporal summation of the acoustic-reflex thres
hold in hearing-impaired ears, but there are indications of
disturbed spatial summation. Beedle (1970) and Beedle and
Harford (1971) have reported finding an effect of stimulus
intensity upon the growth to the acoustic reflex. The slope
of the reflex growth is much steeper and more rapid for nor
mal ears than for either ear of an unilateral Meniere's group.
Niemeyer and Sesterhenn (1972) have noted the occurrence of


46
enough to initiate a post-synaptic impulse, it is therefore
necessary to sunmate several incoming impulses, i.e., spa
tial and temporal summation. Spatial summation occurs as a
result of neural impulses converging simultaneously at a
synapse but from different neural fibers. Since an increase
in the number of neural fibers transmitting impulses is
thought to be a method of coding increased stimulus intensity,
spatial summation can also be in response to increased inten
sity. Temporal summation results from successive impulses
reaching the synapse through the same fibers. With the tem
poral form of neural summation, the post-synaptic potential
occurs as the signal duration increases. It is the amount of
information that each fiber carries and the number of fibers
invoked which makes the difference in the type of summation.
Lorente de No has suggested, therefore, that varying the
signal duration changes the strength of the acoustic reflex
by virtue of temporal summation.
Simmons (1963) has also demonstrated temporal summation
of the intra-aural muscles in cats to acoustic stimulation.
His results are explained in terms of an on-response in the
auditory system in which there is linear integration of
acoustic power from 5 msec through 50 msec at a rate of 8 dB
per doubling of duration. There is only a 2-dB acoustic-
reflex threshold improvement from 50 msec to 100 msec which
indicates a time constant of 50 msec. In comparison to
temporal integration at threshold of audibility in man, the


19
of stimulus parameters (durations, interstimulus intervals,
intensity increments, and spectral components) and age of
the sample population (Jerger et al. 1972).
The standard deviation of the acoustic-reflex thresholds
in response to pure tones has been reported to have a range
of 6.4 dB (Jerger et al., 1972) to 9 dB (Deutsch, 1968).
Harford and Liden (1967-1968) list high Spearman rank-order
correlations for test-retest reliabilities for 250 Hz, 1000 Hz,
and 2000 Hz and poorer results for 500 Hz, 3000 Hz, and 4000 Hz.
Moller (1962a) kept the acoustic reflex repeatability within
1.2 dB, but used a 10 percent change of maximum reflex con
traction. The reported ART variability of 9 dB may be due,
in part, to the unknown physiological processes causing the
reflex at 250 Hz, 4000 Hz, and 6000 Hz to be absent or in
consistently elicited, even in the presence of normal occur
ring. response to pure tones in the middle frequencies
IJDiutsch, 1972; Fulton and Lamb, 1972; Jerger et al^. 1972).
Xtehas:been:reported.that complex stimuli of narrow band
ore white: noise? elicit- the: ART at- lower: intensity levels than
pore-tone sinusoids.- Along, with. the.increased sensitivity
there is also-improvement of threshold stability (Dallos,
i964seDeutsch*r1968%: 1972; Djupesland, Flottorp, and Winther,
i<966jcLiilyv-.i964j McRo.bert;,: Bryan, and Tempest, 1968; Moller,
i962bv Peterson and Liden, 1970, 1972)%
1554 There is general agreement that:the acoustic reflex is
significantly influenced by the energy level outside the


85
o
o
K
en
§
x
E-<
X
w
3
§
I
u
H
En
w
D
O
U
<
u
w
en
S
o
o
LO
E-i
<
Q
a
o
X
en
K
Eh
03
^3
I 1000 Hz
Figure 18. A comparison of mean acoustic-reflex
thresholds as a function of temporal summation in normal
hearing ears.


Q)
Cn
T3
H
>W U
O ffl
% 0
Cu 4J
*> CO
6 d
8 0
.H n
fcl 4J
O
rH
w
o
0
The d.c. voltage divider was used to cancel the d.c.
output from the electroacoustic bridge while passing the
a.c. singal to be recorded on the oscilloscope.
110
To Input
of Channel 2
of Oscilloscops


Table 1. AcousticReflex Threshold in dB SL
Sample
Number
Frequency in Hz
250
400
500
800
1000 1600
2000
3000
4000
Deutsch, 1968
30 ears
74


--
82

--
81
Jepsen, 1951
98 ears
83

81
--
80
78

76
Jepsen, 1963
88 ears
85

81

75
74

80
Jerger et al., 1972
382 Ss


77

78
77
--
75
Lamb et al., 1968
19 Ss
79
--
82
--
82
82
76
78
Weiss et al., 1962
10 Ss
--
93
--
93
97
--
97*

*3200 Hz


REFERENCES
Alberti, P. W. R. M. and Kristensen, R. The clinical appli
cation of impedance audiometry: A preliminary appraisal
of an electro-acoustic impedance bridge. Laryngoscope,
80:735-746, 1970.
American National Standards Institute. Specifications for
audiometers. ANSI S3.6-1969. American National
Institute, Inc., New York, 1970.
Anderson, H., Barr, B., and Wedenberg, E. Intra-aural
reflexes in retrocochlear lesions. In Hamberger, C. A.
and Wersall, J. (editors), Disorders of the Skull Base
Region. Nobel Symposium, 10th, Stockholm: Almqvist
and Wiksell, 1969.
Bar, B., and Wedenberg, E. The early detection
of acoustic tumors by the stapedius reflex test. In
Wolstenholme, G. E. W. and Knight, J. (editors),
Sensorineural Hearing Loss. A Ciba Foundation Sym
posium, London: J. and A. Churchill, 1970.
and Wedenberg, E. Audiometric identification of
normal hearing carriers of genes for deafness. Acta
Oto-laryngologica, 65:535-554, 1968.
Bates, M. A., Loeb, M., Smith, R. P., and Fletcher, J. L.
Attempts to condition the acoustic reflex. Journal of
Auditory Research, 10:132-135, 1970.
Beedle, R. K. An Investigation of the Relationship Between
the Acoustic Reflex growth and Loudness Growth in
Normal and Pathologic Ears. Ph.D. dissertation.
Northwestern University, Illinois, 1970.
_____ and Harford, E. Acoustic reflex and loudness
growth in normal and pathological ears. Sixth Annual
Report of the Auditory Research Laboratories. North
western University, Illinois, 1971.
Bilger, R. and Feldman, R. M. Frequency dependence in
temporal integration. Journal of the Acoustical Society
of America, 45:293(A), 1969.
131


TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS iii
LIST OF TABLES vi
LIST OF FIGURES vii
ABSTRACT ix
CHAPTER
I. INTRODUCTION AND REVIEW OF THE LITERATURE . 1
Review of the Literature 1
Audiological Tests Commonly Used in
Differential Diagnosis 1
Acoustic Reflex 5
Threshold of the Acoustic Reflex (ART) . 14
Temporal Summation 32
II. STATEMENT OF THE PROBLEM 45
Hypothesis 52
III. METHODS AND PROCEDURES 55
Stimuli 57
Experimental Equipment ..... 62
Procedure 67
Temporal Integration at Threshold of
Audibility 67
Temporal Integration at the Threshold of
the Acoustic Reflex 68
Experimental Safeguards 71
IV. RESULTS AND DISCUSSION 7 6
1) Temporal Integration at the Acoustic-
Reflex Threshold 76
2) Temporal Summation at Threshold of
Audibility 91
3) Temporal Integration at the Acoustic-
Reflex Threshold Compared to Temporal
Integration at Threshold of Audibility 98
IV


95
the ART, and there was not as much overlap between the bad-
ear and good-ear groups. The inter-subject variability is
illustrated in Figure 23.
The data from temporal integration at threshold compared
favorably with those of previous investigators. In Figure
22, the mean normal-hearing integration slopes per decade of
stimulus time were 8 dB to 11 dB. The normal ears demonstrated
a frequency effect whereby the slope flattens as test fre
quency increases. These slopes were 10 dB, 9 dB, and 8 dB
at 500 Hz, 1000 Hz, and 2000 Hz, respectively. No systematic
effect was seen for the cochlear-good ears, but the sample
size was smaller than that of the normal ears.
_ At each frequency in Figure 22, the mean slope for
cochlear-bad ears was flatter than, and below, either of the
normal-hearing groups. Since the mean slopes are between
6 dB and 7 dB for the cochlear-impaired ear, the distinc
tion between normal-hearing and cochlear-impaired ears is
rather subtle. The lack of clear inter-group distinction in
Figure 23 further points to the problem of utilizing
individual results for diagnostic information. Only those
subjects with a slope of less than 5 dB/decade would have
been distinctly classified as cochlear-impaired. Olsen,
Rose, and Noffsinger (1973), with a larger experimental popu
lation, reported normal subjects with slopes as flat as 2 dB
to 3 dB. They concluded that the inter-group "... overlap
was sufficiently great that no characteristic pattern of
behavior defined any group."


57
Since one of the characteristic symptoms of Meniere's
disease is fluctuating hearing in the impaired ear, there
were no minimum requirements of hearing level to qualify the
impaired ear as abnormal. The contralateral ear of each
experimental subject yielded pure-tone thresholds of 20 dB or
less (ANSI-1969) at a minimum of two out of the three test
frequencies. This qualified the contralateral ear as having
normal hearing. To facilitate discussion, ears with normal
hearing were referred to as "good," those ears afflicted with
Meniere's disease were denoted as "bad." This classification
of ears was subsequently used throughout this report.
Thus, there was a total of 15 good ears and 5 bad ears
in 10 subjects. Subject LS of the Meniere's group did not
have normal hearing in her good ear at 500 Hz. Nevertheless,
this particular ear is classified as good because she did
have normal hearing at two of the three test frequencies. The
audiograms of all subjects are contained in Appendix B.
Stimuli
The test stimuli were shaped sinusoids ranging from
500 msec to 10 msec in overall duration with 5-msec rise and
decay times. These signals were gated once every 1200 msec.
The stimuli used in this study are listed in Table 4. The
choice of stimuli was determined by previous temporalinte
gration data, ease and expediency of equipment manipulation,
and maximum utilization of the good and bad ears of the
Meniere1s group.


Appendix E(continued)
500 Hz
1000 Hz
2000 Hz
Sub
ject
ecr
acr
Test
Signal Duration in Msec
nun
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
Cochlear
-Good
Ears
i
119
111
91
89
83
110
96
94
90
86
129
98
88
88
86
2
113
111
93
91
83
114
96
92
90
86
129
100
88
86
82
3
115
113
91
93
85
116
94
90
92
84
129
100
86
88
86
4
113
113
93
91
83
114
98
94
88
86
129
98
88
90
84
5
117
111
91
89
85
112
96
94
92
86
127
102
90
88
84
6
119
113
93
91
85
116
96
92
86
82
129
96
90
88
88
7
117
113
93
91
87
116
96
94
86
84
129
100
90
86
86
8
113
111
93
91
83
120
94
92
88
86
125
100
90
88
86
9
115
113
97
97
85
118
94
94
90
88
127
102
90
86
84
10
117
113
93
91
85
116
94
94
90
86
129
104
92
86
86
1
125
119
111
95
87
114
110
100
98
96
129
119
116
110
110
2
123
121
109
97
89
118
112
98
96
96
129
121
112
116
110
3
127
119
109
97
89
116
110
98
98
94
133
129
116
110
108
4
129
119
109
99
91
120
108
98
100
96
131
125
116
114
112
5
133
117
107
97
87
120
110
98
100
96
137
127
114
112
110
6
125
119
111
97
89
118
110
100
96
94
133
125
118
114
110
7
127
119
109
95
89
120
110
100
94
94
131
125
112
110
108
8
123
119
111
97
89
118
112
96
100
96
131
129
118
110
108
9
129
119
113
93
87
114
110
100
98
90
135
127
118
112
108
10
127
117
107
95
91
120
110
100
98
98
131
103
118
114
112
1
129
123
103
95
95
131
124
104
100
98
134
130
100
96
96
2
127
119
103
99
93
135
126
106
100
98
135
128
100
98
92
3
127
123
101
93
93
131
122
106
102
96
137
130
100
98
94
4
133
125
103
95
93
131
124
104
100
98
132
130
102
96
94
120


31
(Ewertsen et al., 1958; Niemeyer, 1971). Coles (1972), how
ever, argues that the ART, like the Uncomfortable Loudness
Level, is a vague concept and is only diagnostically useful
when absent.
Beedle (1970) doubts that the ART is indicative of end-
organ damage. If, in fact, the ART is a measure of loudness
recruitment, it does not continue to demonstrate this rapid
growth in loudness above the ART. Instead, the reflex growth
function is less steep and slower than in normal ears.
(Also see Beedle and Harford, 1971; Peterson and Liden, 1970,
1972.)
As the hearing loss increases, the level of the acoustic-
reflex threshold also increases but not proportionately (Liden,
1970) Sensorineural hearing losses beyond 70 dB to 80 dB
HL (ISO-1964) do not normally demonstrate an acoustic reflex
at any intensity (Jerger, 1970; Jerger et a^. 1972; Terkild-
sen, 1960b). Jerger makes use of this fact to predict the
probability of the hearing level at the threshold of audi
bility. For example, in the presence of a reflex, there are
5 chances in 10 that the loss does not exceed 85 dB HL (ISO-
1964) and there is only 1 chance in 10 that ir is as much as
100 dB HL. This is a nebulous approach at best, because it
lacks quantification for specific cases.
In summary, the acoustic reflex is a well-defined physio
logical phenomenon which has diagnostic value and gives ob
jective information. Its clinical value and true objectivity


28
seldom seen below 25 dB SL and never less than 5 dB to 10 dB
above puretone threshold (Jerger, 197 0; Jerger et a^l., 1972;
Lamb and Peterson, 1967; Lamb, Peterson, and Hansen, 1968).
If a reflex is obtained at less than 5 dB SL it can be
assumed that the pure-tone hearing loss is due to non-organic
causes (Jepsen, 1953, 1963; Lamb, Peterson, and Hansen, 1968;
Terkildsen, 1964; Thomsen, 1955b).
Liden (1970) demonstrated that the intra-aural reflex
can be used as an objective loudnessrecruitment test. He
compared 52 ears with unilateral Meniere's syndrome, 30 ears
of athetoid children and 9 ears presenting acoustic tumors.
The patients were divided into three groups according to the
level of the reflex thresholds and magnitude of separation
between their threshold of audibility and their ART. If the
ART exceeded 95 dB HL (ISO-1964), or if the intensity span
between the pure-tone threshold and the ART was less than
75 dB, the response was considered abnormal. These two values
correspond, respectively, to the ninetieth and tenth percentiles
on 88 normalhearing control subjects.
Liden maintains that a reduced intensity span between
the ART and the pure tone threshold is a result of a cochlear
lesion, and an elevated ART represents a probable higher
order lesion. A reduced span superimposed on an elevated ART
represents both areas as foci of the lesion. For the most
part, his three pathologic groups support this contention,
even though his span of 75 dB for normal hearing is 10 dB to


74
was uncomfortable, they were instructed to press a button
(see response switch III in Figure 9), which activated a
tone audible only to the experimenter. The experimenter then
reduced the stimulus 10 dB. This situation only occurred
once because of experimenter error. A second control button
(see response switch II in Figure 9) allowed the subject to
discontinue the stimulus presentation at will. This second
button was never employed by the subjects.
Dealing with these high signal levels for long periods
of time was bound to introduce some auditory adaptation or
fatigue. The stimulus presentation rate of once every 1200
msec and a duty cycle of less than 50 percent was used as an
arbitrary compromise between long, fatiguing test sessions
and slow signal presentation rates in an attempt to minimize
neural changes. In addition, at the end of each test run
(10 thresholds per signal duration), the test signal was
turned off in order to allow a two- to five-minute rest inter
val During this interval, the signal duration or frequency
was changed and voltage levels checked for calibration.
Opinions of the subjects indicated that sessions of one and
one-half hours approached the maximum tolerable limit. This
was the time necessary to test the acoustic reflex of one ear.
The possibility that the test signal would have crossed
over to the opposite ear, the recording ear, was considered.
If this had occurred, the data would have been contaminated
in two ways. First, the reflex eliciting tone might have


APPENDIX H
CORRELATION COEFFICIENTS OF THE
TEMPORAL INTEGRATION MEASURES


5
obtained because: (1) the patient cannot or does not under
stand his responsibility in the test situation, i.e., how
to recognize or when to respond to the auditory stimuli;
(2) the motivation of the patient is minimal; or (3) the
patient, because of physical state, is incapable of an appro
priate response (Martin, 1972). Any one of these factors can
reduce the reliability of the test procedure without con
scious intent on the part of the patient, as frequently
occurs with such patients as psychotics, mental retardates,
cerebral palsied persons, preverbal infants, or the comatose
or severely ill individual.
There have been numerous reports on the use of the
acoustic reflex an an indicator of abnormal auditory pro
cessing due to cochlear impairment. It is an involuntary
physiological reaction reflected in the contraction of the
intra-aural muscles to relatively intense acoustic stimuli.
In addition, the auditory system's ability to integrate
acoustic power over an interval of time has been investigated
as a measure of cochlear function and integrity. A combina
tion of these two auditory functions may serve to quantify,
objectively, auditory assessment.
Acoustic Reflex
Anatomy and physiology of the acoustic-reflex arc
The acoustic reflex is a consensual contraction of the
middle-ear muscles producing, in persons with normal auditory


25
Latency, the period from signal onset to contraction,
and contraction time, the interval from stimulus onset to
maximum contraction, were some of the earliest acoustic re
flex characteristics studied (Loeb, 1964) Latency decreases
as the intensity is increased above the ART (Dallos, 1964;
Fisch and Schulthess, 1963; Perlman and Case, 1939). The
diortest latency of the stapedius muscle and tendon is about
10 msec to 15 msec in man (Fisch and Schulthess, 1963; Neer-
gaard and Rasmussen, 1966; and Perlman and Case, 1939), while
the tensor tympani has a comparatively longer latency of 50
to 120 msec. There is considerable intra- and inter-subject
variability in latency (Moller, 1958) This is due, in part,
to the obscurity of the contraction close to threshold, but
the shorter the latency the more reliable the trace (Neergaard
and Rasmussen, 1966). The latencies of the ART are among the
shortest for muscle reflexes in man. Fisch and Schulthess
(1963) conclude from an EGM study that this short latency,
especially for stapedial contraction, is probably due to the
limited number of synapses in the crossed acousticreflex arc.
The longer latency obtained from the tensor tympanireflex
arc may indicate that it has additional synapses (viz. Figure
1) .
The maximum impedance change may be reached within 400
msec to 500 msec after signal onset. Some contraction may be
present as long as one second after signal offset. Djupes-
land and Zwislocki (1971) contend that the growth and decay


132
Blodgett, H. C. Jeffress, L. A., and Taylor, R. W. Rela
tion of masked threshold to signal durations for
various interaural phase-combinations. American
Journal of Psychology, 71:283-290, 1958.
Boudreau, J. C. Stimulus correlates of wave activity in
the superior-olivary complex of the cat. Journal of
the Acoustical Society of America, 35:779-785, 1965.
Brahe Pedersen, C. and Elberling C. Temporal integration
of acoustic energy in normal hearing persons. Acta
Oto-Laryngologica, 74:389-405, 1972.
and Elberling, C. Temporal integration of
acoustic energy in patients with presbyacusis. Acta
Oto-Laryngologica, 75:32-37, 1973.
Brink, F., Jr. Synaptic mechanisms. In Stevens, S. S.
(editor), Handbook of Experimental Psychology. New
York: John Wiley and Sons, 1951.
Brooks, D. N. The use of the electro-acoustic impedance
bridge in the assessment of middle ear function.
Journal of International Audiology, 8:563-569, 1969.
. Electroacoustic impedance studies on normal
ears of children. Journal of Speech and Hearing
Research, 14:247-253, 1971.
Burke, K. S., Herer, G. R., and McPherson, D. L. Middle
ear impedance measurement (Acoustic and electro
acoustic comparisons). Acta Oto-Laryngologica,
70:29-34, 1970.
Campbell, R. S. and Counter, S. A. Temporal integration
and periodicity pitch. Journal of the Acoustical
Society of America, 45:691-693, 1969.
Carhart, R. and Jerger, J. Preferred method for clinical
determination of pure tone thresholds. Journal of
Speech and Hearing Disorders, 24:330-345, 1959.
Carver, W. F. Loudness Balance Procedures. In Katz, J.
(editor), Handbook of Clinical Audiology. Baltimore:
Williams and Wilkins Co., 1972.
Chamberlin, S. C. and Zwislocki, J. J. Threshold of audi
bility as a function of tone duration: Is there a
frequency effect? Journal of the Acoustical Society
of America, 48:71(A), 1970.


24
the background environmental and physiological noise so that
the background noise does not cause spurious results. In
order to meet these two criteria a larger dynamic range is
available with a low frequency probe tone. The fact that low
frequency tones elicit the ART at higher intensity levels can
be seen in Table 3. It is not possible to know whether the dif
ference in ART levels is a real effect of probe frequency or
one due to the intensity level of the probe tone (Lilly and
Shepherd, 1964; Terkildsen, Osterhammel, and Scott Nielsen,
1970).
Acoustic-reflex dynamics of magnitude and latency
The intra-aural muscle response had also been observed in
terms of how the reflex changes as the stimulus intensity is
increased above threshold. The response magnitude of the
acoustic reflex grows as a function of intensity to approxi
mately 30 dB above the acousticreflex threshold (30 dB ART
SL) (Dallos, 1964). Most of this growth occurs within 16 dB
ART SL (Djupesland et rL. 1966) at a near linear rate (Dallos,
1964; Moller, 1962a; Peterson and Liden, 1970; Weiss et al.,
1962) with no observable increases to pure tone stimuli be
yond 120 dB SPL (Hung and Dallos, 1972). Low frequency tones
cause the AR to grow at a faster rate than high frequency
tones (Djupesland, et a_l. 1966; Harford and Liden, 1967-
1968). There does not seem to be agreement whether one par
ticular frequency or noise band causes a larger change in
the reflex (Djupesland et a^l. 1966; Fisch and Schulthess,
1963; Johansson, Kylin, and Langfly, 1967) .


21
to hyperexcitability of the brain-stem structures. These
results might also be ascribed to conditioned "learning,"
but Bates, Loeb, Smith, and Fletcher (1970) were unable to
condition the reflex.
Threshold of the acoustic reflex is normally defined as
the lowest stimulus level at which the reflex can be elicited
(Anderson et_ al_. 1969, 1970; Beedle, 1970; Deutsch, 1968;
Djupesland and Zwislocki, 1971; Peterson and Liden, 1972;
and others). This minimal change in muscle contraction is
unacceptable to Moller (1962a), who uses 10 percent change
of the maximum impedance change. His criterion keeps the
ART reliability within 1.2 dB while reliability deteriorates
as the minimal detectable reflex is approached. In addition,
the ART has been elicited in successive steps of 1 dB
(Djupesland and Zwislocki, 1971), 2 dB (Lamb et al., 1968)
and 5 dB (Jerger et al^. 1972) which might account for some
of the reported threshold level differences.
Stimulus duration, another variable affecting the
acoustic-reflex threshold, is reported by Lorente de No
(1935) as having an effect on the tensor tympani response in
rabbits. The strength of the muscle contraction increases
as a function of increasing stimulus duration with the
stimulus level held constant. Further studies have shown
that the acousticreflex threshold in SPL becomes increasingly
lower (more sensitive) as the signal duration is increased
to about 200 msec (Djupesland and Zwislocki, 1971; McRobert,


This dissertation was submitted to the Department of Speech
in the College of Arts and Sciences and to the Graduate
Council, and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.
August, 1973
Dean, Graduate School


26
of the muscle reflex is symmetrical. Decay from maximum con
traction is apparently independent of intensity and frequency.
Although the importance of the acoustic reflex can be
obscured by the many factors affecting it, consistent thres
hold levels have been obtained in a number of studies. The
consistency of the acoustic-reflex response contributes to
its utility as a clinical tool.
Clinical application of the acoustic reflex
Metz (1946) introduced acoustic impedance measurement as
a clinical tool. He included the presence of the acoustic
reflex and the level necessary to just elicit this reflex as
one of the useful dimensions of acoustic impedance. The
absence of the reflex can support the inference that a middle-
ear problem exists in the ear under test. Terkildsen and
Scott Nielsen (1960) and Klockhoff (1961) presented clinical
cases showing that a relatively normal middle ear is necessary
to elicit an intra-aural reflex. In other words, some authors
feel that the presence of a reflex is indicative of a normal
middle-ear (Feldman, 1967; Klockhoff, 1961). This is not
necessarily true, however. Brooks (1969, 1971) states that
a minor conductive component will not abolish the reflex but
instead elevate the acoustic-reflex threshold. He concludes,
therefore, that only subjects exhibiting a reflex to 95 dB
HL (ISO-1964) or less in the contralateral ear should be
regarded as having normal middle-ear function.


99
Figure 14 are approximately twice those at threshold of
audibility of Figure 22. This gives ART measures the
possible advantage of a greater dynamic range to determine
differences between normal and abnormal ears. Unfortunately,
the large inter-subject variability and group overlap has
eroded this possible advantage.
The time constant of integration was essentially the
same between normal-hearing and cochlear-impaired ears at
threshold of audibility but it was definitely shorter for
cochlear-impaired ears at the ART. The cochlear-impaired
ears at the ART were the only ones to have a time constant
of 20 msec or less. This may be a result of the previously
mentioned larger dynamic range for temporal integration at
the ART. Nevertheless, the shorter time constant is one of
the most distinguishable differences between temporal inte
gration at the ART and at the threshold of audibility.
r.5 ~ It should be mentioned that only one person was
correctly' classified for each parameter across all frequen-
eies:at both thresholds. FH had a flattened slope, as well
as ar.short time constant of less than 2 0 msec.


Figure 10. A 20-msec signal burst with a smooth 5-msec
rise and decay.


37
Wright, 1963). Inter-subject variability may cause the
difference in frequency dependence (Clack, 1966). Another
possibility is that frequency dependence may be observed at
shorter durations but not necessarily at the longer signal
on-times (Olsen and Carhart, 1966). Price (1972) has also
suggested that the external ear may cause a transformation of
the stimulus parameters in some subjects by raising the
energy peak in frequency and by legthening short pulses,
thereby possibly nullifying the effects of frequency.
Frequency dependence is, in part, related to the type
of psychophysical procedure used. Bekesy tracking and forced-
choice tracking show little or no frequency dependence, but
the frequency effect is demonstrated by the method of adjust
ment, method of limits, method of constant stimuli, and
confidence ratings (Bilger and Feldman, 1969; Chamberlin and
Zwislocki, 1970; Watson and Gengel, 1969). In the majority
of studies Bekesy tracking is the method used.
Even though there is large individual variability in the
value of the slope (Clack, 1966; Gengel, 1972; Green et al.,
1957; Hattler and Northern, 1970; Martin and Wofford, 1970),
there is general agreement that the test-retest variability
is good (Doyle, 1970; Hattler and Northern, 1970; Olsen and
Carhart, 1966, Plomp and Bouman, 1959). Gengel and Watson
(1971) suggest at least 12 threshold crossings for each
data point when using Bekesy tracking in order to achieve a
reliable reading.


Table 1. Frequency Response of Earphones in dB SPL
(re 0.0002 dyne/cm2)
500 Hz
1000 Hz
2000 Hz
2 volts input
TDH-39
129
128
127
TDH-140
131
131
131
5 volts input
TDH-140
139
139
139
112


98
3) Temporal Integration at the Acoustic-
Reflex Threshold Compared to Temporal
Integration at Threshold of Audibility
There were no significant correlations occurring con
sistently between the data on temporal integration at the
two thresholds, ART and audibility. There were no corre
lations on direction or amount of change between the two
integration functions, nor were there on classification of
ears. This indicates that abnormal temporal integration at
threshold of audibility does not predict the results of
temporal integration at the ART. It must be remembered,
though, that the threshold criteria for the two thresholds
were not only different but also involved different persons.
The experimenter specified the ART and the subject determined
his own threshold of audibility. The difference in cri
teria may be reflected in many different ways, e.g., a
different point on the intensity function for threshold,
i.e., 25 percent vs. 75 percent of correct responses.
Even though there were no significant comparisons
between the two thresholds, some observations can be made.
The mean temporal-integration slopes of the cochlear-impaired
ears fell below those of the normal-hearing groups with
both thresholds. The ART slopes yielded much larger mean
differences among groups, but also greater inter-subject
variability, than those at threshold of audibility. The
most obvious difference between the two types of thresholds
is the slope per decade of time. The slopes of the ART in
^The correlations are listed in Appendix H.


91
In addition, the acoustic reflex occurred at similar
levels reported in Tables 1, 2, and 3. In Figure 21, the
pure-tone thresholds are separated by 20 dB to 40 dB, but
the acoustic-reflex thresholds are overlapping. The
cochlear-good ear group is depressed somewhat at 500 Hz, but
this is because the sample size was three ears at that
parameter.
The ART was established as the mean value of 10 trials.
This may have been unnecessarily long since there was no
significant difference between the mean of the 10 and the
mean of the first 5 trials. The modified Hughson-Westlake
approach (Carhart and Jerger, 1959) would have been as use
ful as the mean of 10 trials, but for the fact that no single
value could be obtained 50 percent of the time in some
instances. Using 2-dB increments, the average range of
values obtained for threshold in any 10 trials was 4 dB to
6 dB. A range of 10 dB was rarely obtained. For the most
part, the shorter the test duration the more variable the
response became. A mean of 5 trials would seem to yield a
stable ART in half the time used in the present experiment.
2) Temporal Summation at
Threshold of Audibility4
A discriminant analysis of temporal summation at thresh
old of audibility between normal-hearing ears and cochlear-
impaired ears indicated significant good-bad ear differences
4
The raw data are reported in Appendix G.


62
of a 20-msec tons burst in Figure 10. In order to confirm
this, an audible check was made at the earphone. No clicks
were discerned from a 10-msec tone burst at 500 Hz placed
just below threshold audibility. It was concluded, therefore,
that a 10-msec tone pulse with a rise and fall time of 5
msec would produce a relatively undistorted signal in this
test situation, using high intensity levels.
Experimental Equipment
The block diagram in Figure 11 represents the equipment
used to generate the stimulus envelope and to record the
subject's response. A General Radio 1313A oscillator gener
ated the test frequency which was monitored by a Monsanto
Model 100-A electronic counter. The sinusoid passed through
the Grason Stadler 3262A recording attenuator to the Grason
Stadler 829E electronic switch that shaped the signal; then
the signal was step attenuated by an Hewlett Packard 350D
attenuator. By way of switches I, II, and II, the signal was
led through either a McIntosh 162K power amplifier or through
a matching transformer to the earphone, a TDH-140 dynamic
earphone in a MX41/AR cushion. The signal from switch II was
measured on a Ballatine 321 true rms vacuum tube voltmeter
and on channel one of a Tektronix 564 oscilloscope. The
same "clock" triggered both the electronic switch and the
oscilloscope.
Three response switches were under subject control.
Response switch I controlled the recording attenuator.


TEMPORAL SUMMATION OF THE ACOUSTIC-REFLEX THRESHOLD
A POSSIBLE INDICATOR OF COCHLEAR ABNORMALITIES
By
William Lee Parker
A DISSERATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1973


TEMPORAL SUMMATION OF THE ACOUSTIC-REFLEX THRESHOLD
A POSSIBLE INDICATOR OF COCHLEAR ABNORMALITIES
By
William Lee Parker
A DISSERATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1973

To my wife, Christine

ACKNOWLEDGMENTS
It is nice to be able to give thanks to all of those
who helped me complete this paper. My advisory committee,
Drs. F. 0. Black, P. J. Jensen, K. C. Pollock, W. A. Yost,
never failed to extend themselves professionally or per
sonally. Dr. Pollock, my chairman, encouraged my interest
in impedance audiometry and was instrumental in my applica
tion for, and the award of, research funds from the
University of Florida. A special thanks is extended to
Dr. Yost, who took an inordinate interest in this project
and without whose help the project would not have been
completed. To Dr. W. N. Williams goes the award for best
encouragement and prodding, the mark of a true friend.
I also want to thank Drs. E. C. Hutchinson and D. T.
Hughes for their statistical advise and help. Christine
Parker and John Parks contributed their invaluable drawing
skills, and Ginny Parks helped with the mundane chore of
proofreading. And thanks to Sue Kirkpatrick for completing
the final typing.
iii

TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS iii
LIST OF TABLES vi
LIST OF FIGURES vii
ABSTRACT ix
CHAPTER
I. INTRODUCTION AND REVIEW OF THE LITERATURE . 1
Review of the Literature 1
Audiological Tests Commonly Used in
Differential Diagnosis 1
Acoustic Reflex 5
Threshold of the Acoustic Reflex (ART) . 14
Temporal Summation 32
II. STATEMENT OF THE PROBLEM 45
Hypothesis 52
III. METHODS AND PROCEDURES 55
Stimuli 57
Experimental Equipment ..... 62
Procedure 67
Temporal Integration at Threshold of
Audibility 67
Temporal Integration at the Threshold of
the Acoustic Reflex 68
Experimental Safeguards 71
IV. RESULTS AND DISCUSSION 7 6
1) Temporal Integration at the Acoustic-
Reflex Threshold 76
2) Temporal Summation at Threshold of
Audibility 91
3) Temporal Integration at the Acoustic-
Reflex Threshold Compared to Temporal
Integration at Threshold of Audibility 98
IV

CHAPTER Page
V. CONCLUSIONS AND SUMMARY 100
APPENDICES 104
A. MEDICAL HISTORY FORM 105
B. PURE TONE HEARING LEVEL (ANSI-1969) OF ALL
SUBJECTS 107
C. VOLTAGE DIVIDER 109
D. CALIBRATION DATA TABLES Ill
E. RAW DATA OF THE ACOUSTIC-REFLEX THRESHOLD . 115
F. DISCRIMINANT ANALYSIS 125
G. RAW DATA OF THRESHOLD OF AUDIBILITY 127
H. CORRELATION COEFFICIENTS OF THE TEMPORAL
INTEGRATION MEASURES 129
REFERENCES 131
BIOGRAPHICAL SKETCH 147
v

LIST OF TABLES
Table Page
1. Acoustic-Reflex Threshold in dB SL 16
2. Acoustic-Reflex Threshold in dB HL (ANSI-1969) 17
3. Acoustic-Reflex Threshold in dB SPL (re 0.0002
dyne/cm^) 18
4. Test Stimuli Used to Elicit Threshold of Adui-
bility and Threshold of the Acoustic Reflex 58
5. Conceivable Sound Pressure Level Necessary to
Elicit the Acoustic-Reflex Threshold by a
500-Hz Burst at a Signal Duration of 10 Msec 73
6. Discriminant Analysis of the Acoustic-Reflex
Threshold 79
7. Average Slope Change Per Decade of Time of the
Acoustic-Reflex Threshold 79
8. Discriminant Analysis of the Threshold of Audi
bility Data 93
9. Signal Duration Intervals Within Which Time
Constants of Temporal Integration at Threshold
of Audibility Occur 97
vi

LIST OF FIGURES
Figure Page
1. Possible time, strength and direction of the
stapedius and tensor tympani muscles, and
resultant displancement of the tympanic
membrane 8
2. The acoustic-reflex arc 10
3. Pre- and post-noise exposure thresholds ... 30
4. Average temporal integration slopes as a func
tion of frequency 36
5. Examples of temporal integration 43
6. Temporal summation of the acoustic reflex . 48
7. Temporal integration at threshold of audi
bility and of the acoustic reflex for normal
hearing subjects 50
8. Acoustic-reflex contraction 60
9. Spectral content of tone bursts with 5-msec
rise/decay ramp 61
10. A 20-msec signal burst with a smooth 5-msec
rise and decay 63
11. Block diagram of equipment used to elicit and
record the intra-aural reflex and to allow
subject control over the test situation . 64
12. Typical response of the acoustic reflex at
threshold 70
13. Upper limits of acceptable exposure to impulse
noise for 95 percent of the population to
10,000 impulses 72
14. Mean acoustic-reflex thresholds as a function
of temporal integration 78
Vll

Page
Figure
15.
16.
17.
18.
19.
20.
21.
22.
23.
Acoustic-reflex thresholds of normal and
cochlear-impaired (bad) ears as a function
of temporal summation 81
Acoustic-reflex thresholds of normal (good)
and impaired (bad) ears of the Meniere's
group as a function of temporal summation . 82
Inter-subject range of acoustic-reflex thresh
olds as a function of temporal integration 83
A comparison of mean acoustic-reflex thresh
olds as a function of temporal summation in
normal-hearing ears 85
Signal-duration interval in which the time
constant of temporal integration for each
test subject occurred across all frequencies 87
The delayed time constant in some cochlear-
impaired ears contrasted with the normal
time constant of temporal integration at the
acoustic-reflex threshold 89
Thresholds, audibility, and acoustic reflex of
the test subjects at 500-msec signal duration 92
Mean thresholds of audibility as a function of
temporal integration 94
Inter-subject range of thresholds of audibility
as a function of temporal integration .... 96
viii

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of
the Requirements for the Degree of Doctor of Philosophy
TEMPORAL SUMMATION OF THE ACOUSTIC-REFLEX THRESHOLD:
A POSSIBLE INDICATOR OF COCHLEAR ABNORMALITIES
By
William Lee Parker
August, 1973
Chairman: Kenneth C. Pollock, Ph.D.
Major Department: Speech
It has been suggested that temporal summation of the
ART might be a clinical tool distinguishing normal from
cochlear-impaired ears. If the ART measure could yield
differences at least as distinctive as those of temporal
summation at threshold of audibility, the ART could be
used with patients for whom threshold measures are not
appropriate.
In this study, temporal summation of the acoustic-
reflex threshold (ART), as well as of the threshold of
audibility, was measured in five normal-hearing and five
unilaterally hearing-impaired subjects by varying signal
duration of 500 Hz, 1000 Hz, and 2000 Hz tones. The signal
durations employed were 500 msec, 200 msec, 100 msec, 20
msec, and 10 msec.
There were statistically significant differences be
tween normal ears and cochlear-impaired ears at 2 out of
12 comparisons. In addition, there were some descriptive
differences between normal ears and cochlear-impaired ears
IX

for the ART measure. Small sample size, unexpected inter
subject variability, and flatter integration slopes for the
normal ears contributed to the lack of consistent statistical
significance.
Temporal integration at threshold of audibility proved
to be significant in most comparisons between normal ears
and cochlear-impaired ears. Even though temporal integra
tion at threshold of audibility evidenced a statistical
difference, the differences were not obvious enough to make
it an effective clinical tool.
There were some differences between the normal-hearing
and cochlear-impaired ears concerning temporal integration
at the threshold of the acoustic reflex. The mean integra
tion slopes of the cochlear-impaired ears were more depressed
than those of the normal ears.
These data suggest that temporal integration of the
acoustic-reflex threshold may not be used to differentiate
normal ears from cochlear-impaired ears. It may be con
strued that the time constant might provide a statistically
significant difference.
x

CHAPTER I
INTRODUCTION AND REVIEW OF THE LITERATURE
Auditory tests, especially those designed to determine
site of lesion, are complicated by subtle test procedures,
and patient state and sophistication. For the majority of
persons the results of audiometric procedures are of diag
nostic value, but in a few cases even a battery of audiometric
tests gives inconclusive results. The diagnostic picture may
be incomplete because of the inability or unwillingness of the
subject to respond. The literature indicates the status of
current audiological tests and further possibilities offered
by tests of acoustic-reflex and of temporal integration by
the auditory system. Furthermore, these later two auditory
measures may be effectively combined into one diagnostic test.
These conditions, therefore, motivated the present study. It
was designed to investigate temporal summation of the acoustic
reflex as a contribution to the diagnosis of cochlear lesions.
Review of the Literature
Audiological Tests Commonly Used
in Differential Diagnosis
Audiological tests are generally divided into those
most sensitive for the detection of conductive, cochlear,
eighth nerve, and central nervous system lesions. Only those
1

2
test procedures normally employed in the assessment of end-
organ status, in particular those tests considered most
valuable in differential diagnosis, will be reviewed in this
section: Alternate Binaural Loudness Balance, Monaural
Loudness Balance, Short Increment Sensitivity Index, and
Bekesy audiometry.^
The Alternate Binaural Loudness Balance (ABLB) test
(Fowler, 1936) has long been the indicator of loudness recruit-
2
ment. This abnormally rapid growth m loudness is theo
retically regarded as pathognomonic of cochlear lesion
(Dix, Hallpike and Hood, 1948; Fowler, 1936, 1950; Hood, 1969;
Jerger, 1967, among others). In spite of its high validity
in confirming recruitment (Tillman, 1969), this test cannot
be administered to all patients. The patient must have one
normal, non-recruiting ear at each test frequency and at
least a 20 dB threshold difference between ears to obtain
meaningful data. Further, it is estimated that more bilateral
hearing losses than unilateral exist (Carver, 1972) Thus,
It is generally understood that audiometric evaluations
do not measure anatomical lesions, per se. Instead, they
reflect functional alterations in the auditory response caused
by anatomical lesions or physiological changes.
2
Loudness recruitment (Fowler, 1937) is an auditory
phenomenon where the supraliminal loudness growth of one ear
occurs at a faster rate than that of the contralateral ear to
equal intensity units. Whan the threshold differences between
the two ears is more than 20 dB and the supra-threshold
balances are considered equal within 10 dB, the loudness
growth is considered abnormal. This abnormal loudness growth
is called loudness recruitment. (Fowler, 1936; Hood, 1969;
Jerger, 1967.)

3
this standard detector of recruitment cannot be administered
to a majority of hearing-impaired persons to gain diagnostic
information.
The Monaural Bi-frequency Loudness Balance (MBLB) test
(Reger, 1936) compares the loudness above threshold at one
frequency with the loudness of another frequency. Again, the
threshold at one of the test frequencies must be normal with
the other elevated. In this way, recruitment can be tested
unilaterally, thereby allowing a larger segment of the
population to be tested diagnostically for cochlear lesions.
It is perceptually much more difficult for the patient to
perform than the ABLB and is, therefore, less reliable. This
is especially true as the difference between the standard
and comparison frequency increases (Graham, 1967) .
The Short Increment Sensitivity Index (SISI) test
(Jerger, Shedd, and Harford, 1959) indicates the ability of
the cochlear mechanism to respond to small changes in signal
amplitude. It is an indicator of cochlear disorders if a
high percentage of 1 dB increments is detected by the sub
ject, but it apparently is not as reliable a detector of
cochlear impairment as the ABLB measure (Owens, 1971; Tillman,
1969). Its lower reliability may be related to the patient
sophistication necessary to complete the procedure, which
is demonstrated by the difficulty in orienting some patients
to the auditory listening task.
The four Bekesy Types, introduced by Jerger (1960),
were based upon the difference in self-tracked threshold

4
leve]s using both pulsed and continuous puretone sweep fre
quency stimuli. A Type II Bekesy, indicated by a slight
separation between the two tracings and a narrowing of the
tracing for the continuous tone, has been the classic
designator of cochlear pathologies (Jerger, 1960) Unfor
tunately, other patterns such as Type I, IV, or mixed are
also obtained from subjects with confirmed end-organ lesions,
therefore contributing ambiguity to the test results (Owens,
1964). The self-tracking procedure of Bekesy audiometry is,
in addition, a long, fatiguing procedure which may make it
difficult for some patients to complete.
Even though each of these tests has certain drawbacks,
the results from each one can indicate the presence of a
cochlear lesion. Recruitment, as defined operationally by
loudness balancing, is the foremost test of a cochlear
lesion, but abnormal loudness growth is not necessarily
synonymous with recruitment (Hirsh, Palva, and Goodman, 1954;
Simmons and Dixon, 1966). All tests for cochlear lesions
mentioned may represent abnormal auditory processing in a
particular cochlear pathology but may delineate differing
physiological phenomena; otherwise there would be no need
for a battery of audiometric tests.
The four auditory tests discussed above may be valid
indicators of cochlear lesions, but their reliability is
dependent upon the full cooperation of a highly motivated
test subject. In many instances cooperation may not be

5
obtained because: (1) the patient cannot or does not under
stand his responsibility in the test situation, i.e., how
to recognize or when to respond to the auditory stimuli;
(2) the motivation of the patient is minimal; or (3) the
patient, because of physical state, is incapable of an appro
priate response (Martin, 1972). Any one of these factors can
reduce the reliability of the test procedure without con
scious intent on the part of the patient, as frequently
occurs with such patients as psychotics, mental retardates,
cerebral palsied persons, preverbal infants, or the comatose
or severely ill individual.
There have been numerous reports on the use of the
acoustic reflex an an indicator of abnormal auditory pro
cessing due to cochlear impairment. It is an involuntary
physiological reaction reflected in the contraction of the
intra-aural muscles to relatively intense acoustic stimuli.
In addition, the auditory system's ability to integrate
acoustic power over an interval of time has been investigated
as a measure of cochlear function and integrity. A combina
tion of these two auditory functions may serve to quantify,
objectively, auditory assessment.
Acoustic Reflex
Anatomy and physiology of the acoustic-reflex arc
The acoustic reflex is a consensual contraction of the
middle-ear muscles producing, in persons with normal auditory

6
pathways, a bilateral change in acoustic impedance^ in both
ears to acoustic stimuli. Upon simultaneous contraction,
the two intra-aural muscles, the stapedius and tensor tympani,
act in a physiologically antagonistic manner but produce a
synergistic impedance against sound energy.
The stapedius muscle is the smallest muscle in the human
body. It originates from the pyramidal eminence on the
posterior wall of the tympanic cavity and its tendon inserts
on the head, neck, or posterior crus of the stapes. The
tensor tympani muscle arises from the bony semicanal above
the Eustachian tube. Its tendon traverses the tympanum to
insert on the mallar manubrium (Jepsen, 1963; Kobrak, 1959) .
The direction of pull of these two muscles is at right angles
to the axis of their corresponding ossicles, making the func
tional action of the two muscles almost in direct opposition
to one another (Wever and Lawrence, 1954).
If the movement caused by each of the muscles is con
sidered independently of the other, contraction of the
stapedius muscle causes the stapes to be pulled posteriorly
3 .
It is not within the scope of this paper to differ
entiate between the types of impedance present within the
middle ear, nor between the various contributing factors of
increased impedances (see Hung and Dallos, 1972; Lilly,
1972, 1973; Zwislocki, 1961, 1962). It is important, however,
to know that the contraction of one or both of the intra-
aural muscles will produce an opposition to energy flow
through the middle-ear cavity, thereby reducing the acoustic
energy reaching the sensory end-organ.

7
and outward from the oval window, as well as causing the
tympanic membrane to be pushed in a slightly outward direc
tion. The malleus and tympanic membrane swing medially upon
tensor tympani activation (Jepsen, 1963; Kobrak, 1959).
This is not the case, however, upon the contraction of both
muscles simultaneously. The movement of the tympanic membrane
depends upon the relative strength, latency, and contraction
time of each muscle. Figure 1 demonstrates a possible resul
tant of a stronger and faster-acting stapedius muscle con
traction partially opposed by the later and weaker tensor tym
pani muscle contraction. It is important to remember that it
does not matter acoustically which muscle predominates because,
in either case, the acoustic impedance will be increased.
The acoustic-reflex arc consists of an afferent neuron,
a reflex center, and an efferent neuron (Lorente de No, 1933,
1935; Rasmussen, 1946). The afferent portion is the same
afferent pathway for audition starting from the sensory end-
organ but terminating at the level of the superior olivary
complex (SOC). The SOC consists of at least five cellular
groups, but the accessory, or medial, nucleus is felt to be
the central mediator of both reflex arcs, the stapedial and
the tensor tympani. The accessory superior olive gives off
a few fibers to the ipsilateral motor nucleus of the facial
nerve (n. VII) (Rasmussen, 1946), which constitutes the
central portion of the stapedial-reflex arc. The facial
nerve innervates the stapedial muscle to complete the efferent
portion of this reflex arc.

8
TIME 23-
Figure 1. Possible time, strength and direction of
the stapedius and tensor tympani- muscles, and resultant
displacement of the tympanic membrane. (Redrawn from
Mendelson, 1963.)

9
It is also surmised that fibers from the accessory
nucleus pass to and from the lateral lemniscus into the
ipsilateral motor nucleus of the trigeminal nerve (n. V)
(Rasmussen, 1946), as the central portion of the tensor tym-
pani-reflex arc. The efferent portion of this reflex arc
constitutes the tensor tympani muscle innervated by the
mandibular branch of the trigeminal nerve. The general schema
for the acoustic-reflex arc as proposed by Rasmussen (1946) is
shown in Figure 2.
Simmons (1963) does not necessarily take exception to
Rasmussen, but suggests that there are several possible re
flex loops, both ipsilateral and crossed. Simmons speculates
that the major differences in latency and response level be
tween the stapedius and tensor tympani muscles are due to the
relative degree of synaptic connections of their respective
reflex arcs. According to Simmons, the stapedial reflex
probably has the more compact, less diffuse, interneuron
connections, thereby explaining its greater sensitivity.
Each of the proposed schema would explain the lower reflex
threshold and shorter latency of the stapedial acoustic-
reflex arc activity in comparison to the tensor tympani
acoustic-reflex responses (as seen in Figure 1). In either
case, the SOC is the most probable reflex center. Neural
activity in the SOC increases with loudness to at least 60
dB sound pressure level (Boudreau, 1965). The acoustic re
flex probably occurs when the neural activity in this

Facial
Nucleus
SOC
Lateral
Lemniscus
Temporal
Cortex
Trigeminal
Nucleus
Figure 2. The acoustic-reflex arc. The stapedial-reflex arc consists of the end-
organ, the primary nuclei, the SOC, the facial nucleus, and the stapedial muscle. The
tensor tympanireflex arc proceeds from the SOC through the lateral lemniscus to the
trigeminal nucleus, ending in the tensor tympani muscle. (Based on Rasmussen, 1946.)
o

11
synaptic complex exceeds some critical level of neural ex
citation (Lorente do No, 1933, 1935).
The stapedius muscle is generally considered to be the
most active and dominant muscle in response to acoustic
stimulation in humans, whereas the tensor tympani reacts to
4
acoustic stimulation less often (Flottorp and Djupesland,
1970; Holst, Ingelstedt, and Ortegren, 1963; Jepsen, 1963;
Liden, Peterson, and Harford, 197C; McRobert, 1968; Mendelson
1961, 1966; Terkildsen, 1957, 1960a,c; Weiss, Mundie, Cashin,
and Shinabarger, 1962; and others). It appears that 1
percent to 5 percent of the population with normal hearing
and with no apparent middle-ear pathology have no demonstrable
acoustic-reflex response (Shiffman, 1972; Swannie, 1966;
Terkildsen, 1960c; Weiss et al. 1962; Wright and Btholm,
1973) .
Methods of detecting the intratympanic
muscle activity in man
Some investigators of the intratympanic muscle activity
have attempted to make direct observations of muscle contrac
tions through surgical exposure or via chronic tympanic mem
brane perforations. For the most part, however, only* the
stapedial tendon is directly available for easy visualization
4
There has been considerable discussion of whether the
tensor tympani muscle is active in man to acoustic or non
acoustic stimulation. This is a clinical question of consider
able importance. McRobert (1968), in a critical review of
the literature, and Liden, Peterson and Harford (1970) agree
that tensor activity is present in man but that the stapedius
predominates.

12
(Kobrak, 1948; Lindsay, Kobrak, and Perlman, 1936; Lorente
de No, 1933; Lorente de No and Harris, 1933; Perlman and
Case, 1939). Since visual observation limits the accuracy
of data quantification, other approaches have been developed
for studying the reflex activity. Electromyography (EMG)
measures the individual muscle fiber action potentials and
is, therefore, the most direct method. EMG activity from
both muscles in humans has been reported in response to both
acoustic and non-acoustic stimulation (Djupesland, 1965;
Fisch and Schulthess, 1963; Salomon and Starr, 1963). This
technique is not clinically feasible.
Extratympanic manometry has been the method employed to
determine direction of tympanic membrane movement. This
technique is made possible by sealing the external auditory
meatus with a probe containing a pressure-sensing device.
If the external meatus is properly sealed, the extratympanic
air pressure should decrease or increase with respective
inward or outward movement of the tympanic membrane (Flottorp
and Djupesland, 1970; Holst et al^., 1963; Liden et al. 1970;
Mendelson, 1961, 1963; Terkildsen, 1957, 1960a,c; Weiss et al.,
1962). McRobert (1968), in a review of this measuring tech
nique, noted that the theory for the tympanic membrane response
to individual muscle contraction is valid, but that there are
many inexplicable results (viz. Mendelson, 1957) suggesting
the need for better instrumentation. Moller (1964) and
Neergaard and Rasmussen (1966) also warn that small

13
contractions of the intra-aural muscles may not produce a
measurable change in extratympanic air pressure. Instead,
these muscle contractions are reflected in a measurable in
crease in the impedance to acoustic energy within the middle
ear.
The technique used most extensively in the past decade
to indicate intra-aural muscle contraction is acoustic-
impedance measurement. Simply stated, a probe tone is directed
perpendicularly towards the plane of the tympanic membrane.
While most of this acoustic energy is transmitted through the
tympanic membrane and attached ossicular chain to the oval
window, a portion of the acoustical energy wave is reflected
back into the external canal from the tympanic membrane.
Since the probe tone input and the reflected tone pick-up
microphone are connected to the external auditory meatus
with an airtight seal, the reflected probe tone can be
monitored accurately. As the intra-aural muscles contract,
the acoustic impedance increases, causing an increased amount
of the probe tone to be reflected back into the external canal.
It is the increased amplitude and phase change of the reflected
tone which indicate that a change in intra-aural muscle
5
activity has occurred. This type of measurement has passed
^Unless specifically stated otherwise, all acoustic-
reflex data concern the ear in which the reflex is elicited.
The ear at which the reflex is elicited may not necessarily
be the same ear in which the reflex contraction is recorded;
most often the reflex ear and the measurement ear are contra
lateral to each other when using the acoustic impedance

14
from an experimental method (Metz, 1946; Zwislocki, 1957) into
the clinical armamentarium with the introduction of several
commercially available acoustic-impedance meters (Grason-
Stadler, 1973; Madsen, n.d.; Peters, n.d.; Terkildsen and
Scott Nielsen, 1960; Zwislocki, 1963). With the increasing
interest in the United States concerning this technique as
a diagnostic tool, critical evaluations of methods and the
commercially available equipment have been made. These instru
ments have been found to be reliable (Nixon and Glorig, 1964;
Tillman, Dallos, and Kurvilla, 1963), sensitive (Moller,
1964), and also relatively easy to utilize (Brooks, 1971).
Threshold of the Acoustic Reflex (ART)
The acoustic reflex first occurs at a predictable level
above the normal threshold of audibility. The predictability
of this reflex accounts for its clinical usefulness; there
fore, it is forthwith described in detail along with factors
which can affect its performance and measurement.
For pure tones from 250 Hz to 4000 Hz the range of the
acousticreflex threshold (ART) is 70 dB to 90dB above the nor
mal-hearing individual's threshold (sensation level: SL) with
a mean of approximately 80 dB SL (Deutsch, 1968, 1972; Jepsen,
1951). Jerger, Jerger, and Mauldin (1972) report the range
measurement technique. If the reflex threshold concerns the
left ear, the reflex eliciting tone will be delivered to the
left ear. Because both sides will contract to unilateral
stimulation, the reflex action in the above example will be
recorded in the right ear. This eliminates acoustic inter
ference of the eliciting tone with the probe tone, which may
cause misleading results.

15
of values for the ART in 382 normal-hearing persons as being
normally distributed with a mean of 85 dB HL (ANSI-1969).
Ninety-five percent of their population fell within 70 dB to
100 dB HL and 99 percent within 65 dB to 105 dB HL. This
range of thresholds has been supported by others (Anderson
and Wedenberg, 1968; Deutsch, 1968, 1972; Harford and Liden,
1967-1968; Peterson and Liden, 1972) although 95 dB to 100
dB HL is considered the upper limit for normals because
auditory lesions outside the cochlea tend to raise the ART,
e.g., conductive problems or eighth nerve lesions (Brooks,
1971; Anderson, Barr, and Wedenberg, 1970).
As shown in Table 1, the ART in SL is rather uniform
from study to study, except for the technique using manome
try. Weiss, et aL. (1962) show ART levels which are 10 dB
to 20 dB less sensitive than those measured by acoustic
impedance. Inspection of Tables 1 and 2 does not indicate any
systematic effect of frequency in sensation level or in
hearing level in normal subjects. In Table 3 more sound
2
pressure level (SPL re 0.0002 dyne/cm ) is necessary to
elicit the ART at lower frequencies. In this respect, the
acoustic reflex is similar to the threshold of audibility
in response to SPL.
The range of ART levels may be due to various types of
acoustic-impedance instruments (e.g., Burke, Herer, and
McPherson, 1970, as shown in Table 2), threshold criteria
(Jerger et ad., 1972 vs. Moller, 1961a, 1962a), a variety

Table 1. AcousticReflex Threshold in dB SL
Sample
Number
Frequency in Hz
250
400
500
800
1000 1600
2000
3000
4000
Deutsch, 1968
30 ears
74


--
82

--
81
Jepsen, 1951
98 ears
83

81
--
80
78

76
Jepsen, 1963
88 ears
85

81

75
74

80
Jerger et al., 1972
382 Ss


77

78
77
--
75
Lamb et al., 1968
19 Ss
79
--
82
--
82
82
76
78
Weiss et al., 1962
10 Ss
--
93
--
93
97
--
97*

*3200 Hz

Table 2. AcousticReflex Threshold in dB HL (ANSI-1969)
Sample
Number
Frequency
in Hz
250
500
1000
1500
2000
3000
4000
Anderson and Wedenburg, 1968
200 ears
85
88
86

88

92
Burke et al. 1970a
21 Ss

93
96



92
Burke et al., 1970^
21 SS

95
93


--
89
Harford and Liden, 1967-1968
->
92
95
88

84
82
87
Jerger et al., 1972
382 Ss

89
86

88

86
Lamb et al., 1968
19 Ss
83
88
86

87
86
83
Melcher and Peterson, 1972
30 Ss

83
84
--
83

86
Moller, 1961a
2-5 Ss

85

87

81f

Moller, 1962a
1-3 Ss
7 5C
84d
88e
Peterson and Liden, 1972
88 ears
86
85
85

85

85
^electroacoustic impedance bridge,
mechanical impedance bridge.
.300 Hz.
525 Hz.
^1200 Hz.
3200 Hz.

2
Table 3. Acoustic-Reflex Threshold in dB SPL (re 0.0002 dyne/cm )
Sample
Number
Frequency
in Hz
250
500
1000
1500
2000
3000
4000
Harford
and Liden, 1967-1968
?
117
106
95

93
92
96
Jerger
et al., 1972
382 Ss

100
96

97
--
95
Lamb et
al. 1968
19 Ss
109
100
93

96
96
93
Moller,
1961a
2-5 Ss

96

93
--
91

Moller,
1962a
1-3 Ss
iooa
95b
95C
f*300 Hz.
525 Hz.
1200 Hz.
00

19
of stimulus parameters (durations, interstimulus intervals,
intensity increments, and spectral components) and age of
the sample population (Jerger et al. 1972).
The standard deviation of the acoustic-reflex thresholds
in response to pure tones has been reported to have a range
of 6.4 dB (Jerger et al., 1972) to 9 dB (Deutsch, 1968).
Harford and Liden (1967-1968) list high Spearman rank-order
correlations for test-retest reliabilities for 250 Hz, 1000 Hz,
and 2000 Hz and poorer results for 500 Hz, 3000 Hz, and 4000 Hz.
Moller (1962a) kept the acoustic reflex repeatability within
1.2 dB, but used a 10 percent change of maximum reflex con
traction. The reported ART variability of 9 dB may be due,
in part, to the unknown physiological processes causing the
reflex at 250 Hz, 4000 Hz, and 6000 Hz to be absent or in
consistently elicited, even in the presence of normal occur
ring. response to pure tones in the middle frequencies
IJDiutsch, 1972; Fulton and Lamb, 1972; Jerger et al^. 1972).
Xtehas:been:reported.that complex stimuli of narrow band
ore white: noise? elicit- the: ART at- lower: intensity levels than
pore-tone sinusoids.- Along, with. the.increased sensitivity
there is also-improvement of threshold stability (Dallos,
i964seDeutsch*r1968%: 1972; Djupesland, Flottorp, and Winther,
i<966jcLiilyv-.i964j McRo.bert;,: Bryan, and Tempest, 1968; Moller,
i962bv Peterson and Liden, 1970, 1972)%
1554 There is general agreement that:the acoustic reflex is
significantly influenced by the energy level outside the

20
critical bandwidth. The ART is more or less constant as the
bandwidth is increased to a critical value. Beyond this
critical bandwidth the acoustic-reflex threshold will decrease
as the bandwidth increases (Flottorp, Djupesland, and
Winther, 1971; McRobert et al., 1968). Moller (1962b), in
contrast, did not see a consistent change in all subjects as
bandwidth was increased, but it is possible that he did not
extend his bandwidths far enough. Moller did obtain improve
ment in threshold for one subject for which the bandwidth sur
passed the critical width determined by Flottorp et al.
(1971) at 525 Hz fc. The work by McRobert et al. (1968)
established that the lowering of ART with increasing band
width is dependent upon the center frequency, and is more
pronounced at 1000 Hz fc. Apparently, the critical bandwidth
with which the reflex mechanism responds to auditory stimuli
widens with increases in sound pressure level (Hung and
Dallos, 1972).
The acousticreflex threshold does not seem to differ
regardless of ascending or descending stimulus presentation
approach (Beedle, 1970; Harford and Liden, 1967-1968; Peterson
and Liden, 1970, 1972), although Deutsch (1968, 1972) reports
a systematic improvement of threshold over three test trials.
Deutsch attributes the ART improvement of approximately 2 dB
from trial to trial to "auditory sensitization" (Hughes,
1954, 1957). Simmons (1960) labels this response condition
"post-tetanic potentiation" and attributes this sensitivity

21
to hyperexcitability of the brain-stem structures. These
results might also be ascribed to conditioned "learning,"
but Bates, Loeb, Smith, and Fletcher (1970) were unable to
condition the reflex.
Threshold of the acoustic reflex is normally defined as
the lowest stimulus level at which the reflex can be elicited
(Anderson et_ al_. 1969, 1970; Beedle, 1970; Deutsch, 1968;
Djupesland and Zwislocki, 1971; Peterson and Liden, 1972;
and others). This minimal change in muscle contraction is
unacceptable to Moller (1962a), who uses 10 percent change
of the maximum impedance change. His criterion keeps the
ART reliability within 1.2 dB while reliability deteriorates
as the minimal detectable reflex is approached. In addition,
the ART has been elicited in successive steps of 1 dB
(Djupesland and Zwislocki, 1971), 2 dB (Lamb et al., 1968)
and 5 dB (Jerger et al^. 1972) which might account for some
of the reported threshold level differences.
Stimulus duration, another variable affecting the
acoustic-reflex threshold, is reported by Lorente de No
(1935) as having an effect on the tensor tympani response in
rabbits. The strength of the muscle contraction increases
as a function of increasing stimulus duration with the
stimulus level held constant. Further studies have shown
that the acousticreflex threshold in SPL becomes increasingly
lower (more sensitive) as the signal duration is increased
to about 200 msec (Djupesland and Zwislocki, 1971; McRobert,

22
Bryan, and Tempest, 1968; Moller, 1962b; Simmons, 1963;
Weiss, Mundie, Cashin, and Shinabarger, 1962) A 500-msec
duration tone approximates the results of longer stimuli for
eliciting a steady change in impedance at the tympanic
membrane. With regard to the auditory response to long
duration signals, Dallos (1964) and Harford and Liden (1967-
1968) have observed that adaptation takes place only over
longer periods of time, e.g., 15 seconds to 30 seconds. This
adaptation is thought to be an afferent process, because
recovery is instantaneous upon changing to a new test fre
quency. This afferent adaptation is also frequency-dependent,
with higher frequencies demonstrating more response decay
(Anderson et alL. 1969; Melcher and Peterson, 1972).
The ear in which the reflex is recorded can also influ
ence the level of the acousticreflex threshold. Recordings
made in the stimulated ear, instead of the contralateral ear,
improve threshold sensitivity (Moller, 1961b). Bilateral
stimulation improves the threshold value even more (Moller,
1962b).
The frequency of the recording probe tone can in itself
influence the acoustic-reflex values. Peterson and Liden
(1972), averaging the ART for 500 Hz and 4000 Hz, state that
a 220-Hz probe tone is about 6 dB more sensitive than a 625-
Hz probe tone in eliciting the ART, while a 800-Hz tone is
between them. In an earlier study, Harford and Liden (1967-
1968) recorded an unsystematic 2-dB threshold difference

23
between a 220-Hz and a 800-Hz probe tone frequency. This
latter study does not agree with the findings of Mehmke and
Tegtmeier (1970) in which there should be a 8 dB loss in
transformer efficienty for low frequencies. Lilly and
Shepherd (1964) and Feldman (1967) have also observed that
the acoustic impedance varies with the frequency of the probe
tone.
The choice of probe tone frequency may be dependent upon
the length of the sinusoidal wave and its relation to the
dimensions of the external auditory canal. Acoustic impedance
becomes more sensitive to changes in probe position in the
external auditory meatus and to increased diameter of the
meatus as the probe frequency increases. This may be mini
mized by using a low frequency probe in a larger volume, e.g.,
at the external meatus (Schoel and Arnesen, 1962). This con
tention is not supported by Djupesland et al. (1966), who
see no effects of probe position. Djupesland and his asso
ciates used 5-dB test increments, which may have obscured any
small but significant results of plug position.
It might be possible that the attributed frequency ef
fects of the'probe tone are due to the varied, and often
unreported, intensity levels of the probe tone. It is impor
tant that the probe tone not be intense enough to evoke the
acoustic reflex, since it is there to indicate the existence
of intra-aural muscular contraction and not to elicit it
(see footnote 4). The probe tone level must also be above

24
the background environmental and physiological noise so that
the background noise does not cause spurious results. In
order to meet these two criteria a larger dynamic range is
available with a low frequency probe tone. The fact that low
frequency tones elicit the ART at higher intensity levels can
be seen in Table 3. It is not possible to know whether the dif
ference in ART levels is a real effect of probe frequency or
one due to the intensity level of the probe tone (Lilly and
Shepherd, 1964; Terkildsen, Osterhammel, and Scott Nielsen,
1970).
Acoustic-reflex dynamics of magnitude and latency
The intra-aural muscle response had also been observed in
terms of how the reflex changes as the stimulus intensity is
increased above threshold. The response magnitude of the
acoustic reflex grows as a function of intensity to approxi
mately 30 dB above the acousticreflex threshold (30 dB ART
SL) (Dallos, 1964). Most of this growth occurs within 16 dB
ART SL (Djupesland et rL. 1966) at a near linear rate (Dallos,
1964; Moller, 1962a; Peterson and Liden, 1970; Weiss et al.,
1962) with no observable increases to pure tone stimuli be
yond 120 dB SPL (Hung and Dallos, 1972). Low frequency tones
cause the AR to grow at a faster rate than high frequency
tones (Djupesland, et a_l. 1966; Harford and Liden, 1967-
1968). There does not seem to be agreement whether one par
ticular frequency or noise band causes a larger change in
the reflex (Djupesland et a^l. 1966; Fisch and Schulthess,
1963; Johansson, Kylin, and Langfly, 1967) .

25
Latency, the period from signal onset to contraction,
and contraction time, the interval from stimulus onset to
maximum contraction, were some of the earliest acoustic re
flex characteristics studied (Loeb, 1964) Latency decreases
as the intensity is increased above the ART (Dallos, 1964;
Fisch and Schulthess, 1963; Perlman and Case, 1939). The
diortest latency of the stapedius muscle and tendon is about
10 msec to 15 msec in man (Fisch and Schulthess, 1963; Neer-
gaard and Rasmussen, 1966; and Perlman and Case, 1939), while
the tensor tympani has a comparatively longer latency of 50
to 120 msec. There is considerable intra- and inter-subject
variability in latency (Moller, 1958) This is due, in part,
to the obscurity of the contraction close to threshold, but
the shorter the latency the more reliable the trace (Neergaard
and Rasmussen, 1966). The latencies of the ART are among the
shortest for muscle reflexes in man. Fisch and Schulthess
(1963) conclude from an EGM study that this short latency,
especially for stapedial contraction, is probably due to the
limited number of synapses in the crossed acousticreflex arc.
The longer latency obtained from the tensor tympanireflex
arc may indicate that it has additional synapses (viz. Figure
1) .
The maximum impedance change may be reached within 400
msec to 500 msec after signal onset. Some contraction may be
present as long as one second after signal offset. Djupes-
land and Zwislocki (1971) contend that the growth and decay

26
of the muscle reflex is symmetrical. Decay from maximum con
traction is apparently independent of intensity and frequency.
Although the importance of the acoustic reflex can be
obscured by the many factors affecting it, consistent thres
hold levels have been obtained in a number of studies. The
consistency of the acoustic-reflex response contributes to
its utility as a clinical tool.
Clinical application of the acoustic reflex
Metz (1946) introduced acoustic impedance measurement as
a clinical tool. He included the presence of the acoustic
reflex and the level necessary to just elicit this reflex as
one of the useful dimensions of acoustic impedance. The
absence of the reflex can support the inference that a middle-
ear problem exists in the ear under test. Terkildsen and
Scott Nielsen (1960) and Klockhoff (1961) presented clinical
cases showing that a relatively normal middle ear is necessary
to elicit an intra-aural reflex. In other words, some authors
feel that the presence of a reflex is indicative of a normal
middle-ear (Feldman, 1967; Klockhoff, 1961). This is not
necessarily true, however. Brooks (1969, 1971) states that
a minor conductive component will not abolish the reflex but
instead elevate the acoustic-reflex threshold. He concludes,
therefore, that only subjects exhibiting a reflex to 95 dB
HL (ISO-1964) or less in the contralateral ear should be
regarded as having normal middle-ear function.

27
The manometer system in an electroacoustic impedance
bridge is able to approximate the amount of pressure in
the middle-ear. For the most part, 50 mm of equivalent
water pressure (re atmospheric pressure) is considered within
normal limits. Hearing and acoustic-reflex thresholds do not
deteriorate within this pressure range in the middle-ear
(Alberti and Kristensen, 1970; Peterson and Liden, 1970).
Negative middle-ear pressure is more detrimental to thresholds
than positive pressure (Moller, 1965), but the AR may be
stronger with slightly negative canal pressure (Terkildsen,
1964). Terkildsen (1960c) even states that the stapedial
reflex is found around normal atmospheric pressure, while the
tensor reflex is enhanced by negative pressure.
In the absence of middle-ear conductive problems, the
level at which the AR is just elicited above the threshold of
audibility is useful in confirming a cochlear abnormality.
It has been observed that in the majority of mild to moderate
sensorineural hearing losses the acoustic-reflex threshold
occurs at approximately the same hearing-threshold levels (HL)
as the normal ART (Alberti and Kristensen, 1970; Ewertsen,
Filling, Terkildsen, and Thomsen, 1958; Jerger et al^. 1972;
Klockhoff, 1961; Kristensen and Jepsen, 1952; Lamb, Peterson,
and Hansen, 1968; Metz, 1946, 1952; Peterson and Liden, 1972;
Terkildsen, 1960b; Thomsen, 1955a,b; and others). If the
ART occurs at 55 dB to 60 dB SL or less, the hearing loss is
considered to be due to cochlear impairment. The ART is

28
seldom seen below 25 dB SL and never less than 5 dB to 10 dB
above puretone threshold (Jerger, 197 0; Jerger et a^l., 1972;
Lamb and Peterson, 1967; Lamb, Peterson, and Hansen, 1968).
If a reflex is obtained at less than 5 dB SL it can be
assumed that the pure-tone hearing loss is due to non-organic
causes (Jepsen, 1953, 1963; Lamb, Peterson, and Hansen, 1968;
Terkildsen, 1964; Thomsen, 1955b).
Liden (1970) demonstrated that the intra-aural reflex
can be used as an objective loudnessrecruitment test. He
compared 52 ears with unilateral Meniere's syndrome, 30 ears
of athetoid children and 9 ears presenting acoustic tumors.
The patients were divided into three groups according to the
level of the reflex thresholds and magnitude of separation
between their threshold of audibility and their ART. If the
ART exceeded 95 dB HL (ISO-1964), or if the intensity span
between the pure-tone threshold and the ART was less than
75 dB, the response was considered abnormal. These two values
correspond, respectively, to the ninetieth and tenth percentiles
on 88 normalhearing control subjects.
Liden maintains that a reduced intensity span between
the ART and the pure tone threshold is a result of a cochlear
lesion, and an elevated ART represents a probable higher
order lesion. A reduced span superimposed on an elevated ART
represents both areas as foci of the lesion. For the most
part, his three pathologic groups support this contention,
even though his span of 75 dB for normal hearing is 10 dB to

29
15 dB more conservative than the other reported studies. As
further support he obtained pre- and post-noise exposure
thresholds, pure tone and reflex, on 11 cats. The results
can be seen in Figure 3. An average permanent threshold shift
of 44 dB resulted with an elevation of the ART by only 1 dB
from the pre-exposure levels. The only noticeable effect upon
the ART was a slight increase in the standard deviation across
the four reflex eliciting frequencies. It appears, therefore,
that the acoustic reflex is a loudness-sensing mechanism that
can be used clinically to indicate the possible presence of
loudness recruitment.
To further authenticate the meaning of the acoustic-re
flex level in cochlearimpaired persons, comparisons have
been made with loudness-balance tests and other audiometric
indicators of cochlear lesions. In most cases a comparison
is made with a unilateral end-organ disease and results of
the ABLB, as the standard of loudness recruitment. The re
sults indicate that the ART level above the threshold of
hearing is as good as the ABLB and slightly better than the
SISI for determining an end-organ lesion (Alberti and
Kristensen, 1970; Ewertsen et_ al. 1958; Kristensen and
Jepsen, 1952; Lamb et al., 1968; Liden, 1970; Thomsen, 1955a;
and others). It is definitely a better and more reliable
predictor of sensory damage than the Monaural Bi-frequency
Loudness Balance, the Difference Limen for Intensity or
Frequency, the Uncomfortable Loudness Level, and Bekesy Types

30
TEST FREQUENCY
Figure 3. Pre- and post-noise exposure thresholds.
The lower thresholds are audibility and the upper thresh
olds are acoustic reflex in 11 cats. (Redrawn from Liden,
1970.)

31
(Ewertsen et al., 1958; Niemeyer, 1971). Coles (1972), how
ever, argues that the ART, like the Uncomfortable Loudness
Level, is a vague concept and is only diagnostically useful
when absent.
Beedle (1970) doubts that the ART is indicative of end-
organ damage. If, in fact, the ART is a measure of loudness
recruitment, it does not continue to demonstrate this rapid
growth in loudness above the ART. Instead, the reflex growth
function is less steep and slower than in normal ears.
(Also see Beedle and Harford, 1971; Peterson and Liden, 1970,
1972.)
As the hearing loss increases, the level of the acoustic-
reflex threshold also increases but not proportionately (Liden,
1970) Sensorineural hearing losses beyond 70 dB to 80 dB
HL (ISO-1964) do not normally demonstrate an acoustic reflex
at any intensity (Jerger, 1970; Jerger et a^. 1972; Terkild-
sen, 1960b). Jerger makes use of this fact to predict the
probability of the hearing level at the threshold of audi
bility. For example, in the presence of a reflex, there are
5 chances in 10 that the loss does not exceed 85 dB HL (ISO-
1964) and there is only 1 chance in 10 that ir is as much as
100 dB HL. This is a nebulous approach at best, because it
lacks quantification for specific cases.
In summary, the acoustic reflex is a well-defined physio
logical phenomenon which has diagnostic value and gives ob
jective information. Its clinical value and true objectivity

32
occurs when the involuntary acoustic-reflex threshold is
compared with the threshold of audibility. When the dynamic
range between the two thresholds is 55 dB or less, there is
cochlear impairment. By independently varying the signal
duration at threshold of audibility, more diagnostic informa
tion is gained. This manipulation of auditory processing, as
seen in the next section, may be applied to the acoustic re
flex threshold as well.
Temporal Summation
There is a psychophysical assumption that the normal
auditory system can summate, or integrate, acoustic power over
some critical period of time. Since the initial work of Garner
(1947b), Garner and Miller (1947), Hughes (1946), Munson (1947),
and de Vries (1948), considerable interest has been generated
over the concept of "trading" increased acoustic power with
a decreased signal duration to maintain a constant loudness
level. When the on-time of the auditory stimulus is sequen
tially halved, starting from some critical duration, the
signal power is reduced 2 dB to 3 dB with each decrement, i.e.,
from 200 msec to 100 msec, from 100 msec to 50 msec, etc.
The critical long duration is thought to be about 200 msec
and a linear relationship holds to a critical short signal
length of approximately 10 msec. ~
Perfect integration
If one assumes perfect integration (Garner, 1947a,b;

33
Garner and Miller, 1947), the ear would trade 1 log unit of
intensity (1 bel, 10 dB) for 1 log unit of duration (10 msec
to 100 msec, 20 msec to 200 msec, 15 msec to 150 msec). Even
though there is an increase of power as stimulus time in
creases, the acoustic energy required to maintain the thres
hold of audibility is relatively constant.6 This accumulation
of power over an average tenfold decrease in time may be
inferred from threshold data using a single-slope value (Clack,
1966; Garner and Miller, 1947; Munson, 1947; Northern, 1967;
and others). When this slope value is 10 dB it is considered
perfect temporal integration of energy. Perfect integration
is most often observed for the sinusoid of 1000 Hz.
The time constant of temporal integration (T0) marks the
point at which time and intensity cease to have a linear
relationship; it is thought to occur between 200 msec to 250
msec. Goldstein and Kramer (1960) observe integration occur
ring through 200 msec, but Harris, Haines, and Myers (1958)
report individual time constants ranging from 100 msec to
300 msec with a mean of 200 msec. Plomp and Bouman (1959) and
Hempstock, Bryan, and Tempest (1964) indicate that T0 is
inversely related to frequency; T0 changes from about 375 msec
at 250 Hz to approximately 150 msec at 8000 Hz. Regardless
6Acoustic energy can be stated in the simplest form as
E = PT, where E is acoustic energy, P is acoustic power, and
T is linear for a certain period of time, the time constant.
Further increases in signal duration beyond the time constant
have less and less effect upon threshold level.

34
where Tq may fall, integration is essentially complete between
500 msec and 1000 msec (Zwislocki, 1960) .
The phenomenon of integration of acoustic power over
time is attributed to temporal summation in the auditory
system and, most probably, is neural in nature (Zwislocki,
1960)- The term summation is introduced by some authors be
cause it is felt that the response pattern represents tem
poral summation at the synaptic junctions in the neural path
ways. More specifically, this apparent linear function of
log time and log power is due to an exponential decay of the
persisting neural excitation (Plomp and Bouman, 1959; Zwis
locki, 1960) At threshold, there appears to be a direct
proportionality between an increase in the intensity of
acoustic power and the increase in neural excitation, which
is modified somewhat at suprathreshold levels by neural adap
tation and the loudness of the stimuli (Zwislocki, 1960) .
The neural mechanism for temporal integration probably exists
at the neurons above the first and, possibly, the second order,
but before the level of binaural interaction at the superior
olivary complex (Zwislocki, 1960) .
Stimulus parameters
The literature indicates that, in addition to duration,
other various stimulus parameters affect temporal integration.
Those most often cited are stimulus spectrum, rise and decay
time, definition of the signal duration, inter-stimulus

35
interval, attenuation rate, method of threshold calculation,
and psychophysical procedure.
Earlier studies indicate that for a narrow band of
frequencies between 1000 Hz and 4000 Hz a 10 dB/decade slope
is generated (Garner, 1947a; Garner and Miller, 1947; Hughes,
1946; Munson, 1947), while narrow-or wide-band noise bursts
demonstrate about a 8 dB/decade change in threshold inte
gration (Garner, 1947b; Miller, 1948; Small, Brandt, and
Cox, 1962). There are no systematic differences in the slope
between threshold determination in quiet (Garner, 1947b;
Hempstock, Bryan, and Tempest, 1964; and others) and in back
ground noise (Garner and Miller, 1947; Blodgett, Jeffress, and
Taylor, 1958; Hempstock et a_l. 1964; Gengel, 1972; Plomp
and Bouman, 1959; and others).
More recently, a controversy concerning frequency effects
has developed. While 1000 Hz elicits an average slope of 10
dB per decade, lower frequencies have a steeper slope and an
increase in frequency produces a flatter slope, as seen in
Figure 4 (Brahe Pedersen and Elberling, 1972; Elliott, 1963;
Gengel, 1972; Gengel and Watson, 1971; Hattler and Northern,
1970; Hempstock et ad., 1964; Miskolczy-Fodor, 1953; Northern,
1967; Olsen and Carhart, 1966; Sanders, Josey, and Kemer,
1971; Sheeley and Bilger, 1964; Simon, 1963; Tempest and
Bryan, 1971; Watson and Gengel, 1969). Some investigators
feel that there is no frequency dependence (Clack, 1966;
Martin and Wofford, 1970; Wright, 1968a, 1972; Zwicker and

36
SIGNAL DURATION IN MSEC
Figure 4. Average temporal integration slopes as a
function of frequency. In general, as frequency decreases
the slope steepens. (Redrawn from Watson and Gengel, 1969.)

37
Wright, 1963). Inter-subject variability may cause the
difference in frequency dependence (Clack, 1966). Another
possibility is that frequency dependence may be observed at
shorter durations but not necessarily at the longer signal
on-times (Olsen and Carhart, 1966). Price (1972) has also
suggested that the external ear may cause a transformation of
the stimulus parameters in some subjects by raising the
energy peak in frequency and by legthening short pulses,
thereby possibly nullifying the effects of frequency.
Frequency dependence is, in part, related to the type
of psychophysical procedure used. Bekesy tracking and forced-
choice tracking show little or no frequency dependence, but
the frequency effect is demonstrated by the method of adjust
ment, method of limits, method of constant stimuli, and
confidence ratings (Bilger and Feldman, 1969; Chamberlin and
Zwislocki, 1970; Watson and Gengel, 1969). In the majority
of studies Bekesy tracking is the method used.
Even though there is large individual variability in the
value of the slope (Clack, 1966; Gengel, 1972; Green et al.,
1957; Hattler and Northern, 1970; Martin and Wofford, 1970),
there is general agreement that the test-retest variability
is good (Doyle, 1970; Hattler and Northern, 1970; Olsen and
Carhart, 1966, Plomp and Bouman, 1959). Gengel and Watson
(1971) suggest at least 12 threshold crossings for each
data point when using Bekesy tracking in order to achieve a
reliable reading.

38
Spectral characteristics of the signal as a function of
duration cannot be divorced from the integrating bandwidth,
or the critical band of the ear (Fletcher, 1940; Scharf,
1970) The theoretical bandwidth of a pulsed tone is pro
portional to the reciprocal of the signal's duration, or
1/t. As long as the spectrum of the test tone remains within
the critical bandwidth of the ear, all of the energy will be
integrated. If not, there should be a loss in sensitivity
(Garner, 1947a; Green et rl. 1957; Olsen and Carhart, 1966;
Sheeley and Bilger, 1964; Wright, 1968a). The spread of
energy by abrupt low frequency signals upward to the more
sensitive frequencies of the ear has been misinterpreted as
an apparent increase of sensitivity within the critical band.
It has resulted in the assumption that the ear does not inte
grate low-frequency tone pips (Campbell and Counter, 1969;
Karlovich, Lane, Smith, Tarlow, Thompson, and Vivion, 1971).
The point is that one must carefully monitor the test stimuli
so that the duration of the tone burst is maintained within
the critical bandwidth (Wright, 1968a).
Some variability in temporal integration has stemmed
from lack of defined criteria of stimulus duration. Goldstein
and Kramer (1960) measured the duration between energy onset
and cessation, while Harris (1957) designated as criteria
the half-power points on the stimulus envelope. An "equiva
lent duration" has also be used (Brahe Pederson and
Elberling, 1972; Dallos and Johnson, 1966; Dallos and Olsen,

39
1964; Olsen and Carhart, 1966). An equivalent-duration tone
pip contains the same amount of energy as a rectangular
envelope and allows for the comparison of widely varying
envelope shapes. The use of the equivalent duration fits or
allows the conversion of the overall stimulus envelope to fit
the slope of Garner's model (1947b), which states that not all
energy is used by the ear to summate temporal information
(Dallos and Olsen, 1964). The best measure of the acoustic
waveform is still a moot question because the conversion from
the half-power points to equivalent duration of Harris' data
(1957) by Dallos and Olsen (1964) demonstrate no difference.
The time in which the stimulus envelope rises to and
decays from its effective peak is a critical variable in
auditory measures using pure-tone stimuli. If the stimulus
starts or stops too abruptly, a wide-band transient may be
generated, thereby biasing the results (Wright, 1958, 1968a).
A minimum 5 msec rise/fall time measured on the linear por
tion of the ramp (between 10 percent and 90 percent of maximum
amplitude of the acoustic waveform) is most often suggested
(Harris, 1957; Wright, 1960, 1968a). No differences occur
with rise/fall times varying from 0 msec to 50 msec if the
equivalent duration is held constant (Dallos and Johnson,
1966; Dallos and Olsen, 1964).
The inter-stimulus interval, or the off-time between
successive stimulus envelopes, must be kept long enough to
ensure neural independence between stimulus events. Since

40
the theory of temporal summation is based on neural decay,
the minimum off-time should be 200 msec (Zwislocki, 1960).
In addition, a "critical off-time" has been calculated where
by successive stimuli having less than a 200 msec inter
stimulus interval take on the threshold value of one contin
uous tone (Jerger, Jerger, Ainsworth, and Caram, 1966) .
Doyle (1970) in a temporal integration study varied the inter
stimulus interval from 100 msec to greater than 500 msec.
There were no differences in temporalintegration results as
long as the dead time was a minimum of 500 msec, but the
slopes flattened as the off-time was shortened to 100 msec.
In practice, therefore, the inter-stimulus interval should
exceed the theoretical minimum, and probably be no less than
500 msec.
The repetition rate of the signal per unit of time,
usually one second, is also based on consideration of decay
of neural excitation. If one desires to keep the repetition
rate constant, then the rate is determined by the length of
the longestduration test signal and the desired inter-stimulus
interval. Wright (1968a) recommends the use of the same
repetition rate for all test stimuli so that sampling per unit
of time will be uniform.
Some attention has been given to attenuation rate for
those test procedures employing semi-automated equipment, for
example, Bekesy tracking. Hempstock et al. (1964) observed
an increase in the standard deviation of the threshold as the

41
number of pulses presented at each discrete level decreased
from eight pulses to one pulse. There was no significant
difference in the mean threshold. Wright (1968a) suggests
using an attenuation rate of 2.5 dB/second at a repetition
rate of 1/second as a standard.
The choice of ascending, descending, or mean thresholds
can give threshold values differing as much as 3 dB at any
one signal duration (Hempstock et^ a_l. 1964). Therefore,
for the majority of the cases, where applicable, mean values
determine threshold. The vast majority of recent publications
dealing with temporal integration have used the Bekesy track
ing procedure, where the averaged midpoint of the crossings
is considered to be the threshold value.
Effects of end-organ lesions
upon temporal integration
Considerable interest has been generated in the use of
temporal summation as a diagnostic tool for end-organ patholo
gies; this is otherwise known as brief-tone audiometry. A
cochlear lesion seemingly increases the integration ability
of the ear. As the duration is decreased,less acoustic power
is needed to maintain constant energy for threshold (Brahe
Pederden and Elberling, 1972, 1973; Doyle, 1970; Elliot,
1963; Gengel and Watson, 1971; Gengel, 1972; Harris et al.,
1958; Hattler and Northern, 1970; Jerger, 1955; Martin and
Wofford, 1970; Miskolczy-Fodor, 1953, 1956, 1960; Nerbonne,
1970; Olsen, Rose, and Noffsinger, 1973; Sanders and Honig,

42
1967; Sanders, Josey, and Kemer, 1971; Wright, 1968b; Wright
and Cannella, 1969). The integration slope is not affected
by either conductive problems (Miskolczy-Fodor, 1956; Harris
et al. 1958; Wright, 1968b; Wright and Cannella, 1969) or by
retrocochlear problems (Olsen et al., 1973; Sanders et al.,
1971). The critical time at which the temporal integration
begins, or the time constant of integration, also seems to
be shortened to about 100 msec or less depending on the par
ticular frequency (Miskolczy-Fodor, 1953, 1956, 1960; Harris
et al., 1958; Sanders and Honig, 1967; Wright, 1968a). The
abnormal results of cochlear damage to slope and time con
stant can be seen in Figure 5, i.e., the slope is flatter,
the time constant is shorter, or both.
It has also been suggested that brief-tone audiometry
may even be so sensitive as to detect incipient damage to the
cochlea (Campbell and Counter, 1969; Sanders and Honig, 1967;
Wright, 1968a) to detect recruitment (Miskolczy-Fodor, 1956,
1960), and to differentiate between various cochlear
lesions (Harris et al., 1958). Invariably, these attributes
are artifacts of unaccounted frequency effects, energy spread,
and/or inter-subject variability (Gengel and Watson, 1971;
Karlovich et ad., 1971). Olsen et a^. (1973) cast doubt on
the efficacy of temporal integration to differentiate be
tween end-organ and eighth nerve lesions because of the large
overlap between populations.

43
Figure 5. Examples of temporal integration. Cochlear-
impaired integration functions can be flatter, can have a
shorter time constant, or both. (Redrawn from Gengel and
Watson, 1971.)

44
Nerbonne (1970), who studied temporarily fatigued ears,
and Sanders et. ad. (1971), who studied cochlear lesions,
demonstrated that the SISI and ABLB, as well as brief-tone
audiometry are sensitive indicators of cochlear lesions.
Brief-tone audiometry can yield definitive results when the
ABLB shows partial recruitment and negative SISI scores are
present (Sanders et ad., 1971). There is some correspondence
between amount of recruitment and degree of aberration of
temporal integration, but, for the most part, temporal inte
gration is assessing some other aspect of sensory lesions
(Nerbonne, 1970). Abnormal temporal integration seems to be
part of a syndrome of a distorted time domain of enlarged
critical bands and critical ratios (Northern, 1967; Sheeley,
1963; Simon, 1963) and poorer performance on difference limens
for frequency (Sheeley, 1963). Temporal integration, by way
of its relation to critical bands, seems to have a common
basis with several phenomena related to the tuning of the
auditory system (Licklider, 1951) that are not presently
measured clinically. The method of brief-tone audiometry,
therefore, presents clinicians with another possible test
for end-organ abnormalities, if proper standards are set
forth for the various stimulus and test parameters, and nor
mal response limits are determined.

CHAPTER II
STATEMENT OF THE PROBLEM
Innumerable studies detailing the identification and
quantification of cochlear abnormalities indicate that, with
patients who cannot or will not cooperate fully, the present
diagnostic measures are inadequate in obtaining information
on auditory processing. Further understanding of the func
tion of the acousticreflex threshold and the effect of
temporal integration upon threshold measures may fill this
void in diagnostic testing of the auditory system. Such a
measure may be obtained by maintaining a constant reflex
strength as a function of increasing acoustic power while
signal duration is decreased, or, in other words, temporal
integration of the acoustic reflex. Since there are rela
tively few data concerning temporal integration of the acous
tic reflex, the pertinent literature will now be reviewed.
As early as 1935, Lorente de No reported the effects of
varying short signal duration upon the contraction of the
tensor tympani muscle of the cat. He used as much as 140
dB SPL and durations from 3 msec to 100 msec. He explained
his results in terms of neural summation, which occurs at the
synapse.'*' At some synapses one pre-synaptic impulse is not
*"For further information regarding synaptic summation
see the classical treatise of Sherrington (1906, 1947);
also see Brink (1951).
45

46
enough to initiate a post-synaptic impulse, it is therefore
necessary to sunmate several incoming impulses, i.e., spa
tial and temporal summation. Spatial summation occurs as a
result of neural impulses converging simultaneously at a
synapse but from different neural fibers. Since an increase
in the number of neural fibers transmitting impulses is
thought to be a method of coding increased stimulus intensity,
spatial summation can also be in response to increased inten
sity. Temporal summation results from successive impulses
reaching the synapse through the same fibers. With the tem
poral form of neural summation, the post-synaptic potential
occurs as the signal duration increases. It is the amount of
information that each fiber carries and the number of fibers
invoked which makes the difference in the type of summation.
Lorente de No has suggested, therefore, that varying the
signal duration changes the strength of the acoustic reflex
by virtue of temporal summation.
Simmons (1963) has also demonstrated temporal summation
of the intra-aural muscles in cats to acoustic stimulation.
His results are explained in terms of an on-response in the
auditory system in which there is linear integration of
acoustic power from 5 msec through 50 msec at a rate of 8 dB
per doubling of duration. There is only a 2-dB acoustic-
reflex threshold improvement from 50 msec to 100 msec which
indicates a time constant of 50 msec. In comparison to
temporal integration at threshold of audibility in man, the

47
time constant is about 200 msec and the threshold changes
2 dB to 3 db per doubling of signal duration.
There are relatively few data on temporal integration
of the acoustic-reflex threshold in humans. The systematic
published works are by McRobert, Bryan, and Tempest (1968)
and Djupesland and Zwislocki (1971). In both instances data
are plotted in terms of acoustic-power growth as signal dura
tion decreased while maintaining acoustic-reflex threshold.
Their slopes, reported in power to maintain reflex threshold
per average decade change in signal duration, are in the
range reported by Simmons (1963). The slopes range from 14
dB/decade for 300 Hz, to 21dB/decade for 2500 Hz (McRobert
et al., 1968), and to 25 dB/decade for 1000 Hz (Djupesland
and Zwislocki, 1971). Although there are no reports of a
systematic frequency effect (McRobert et a_l. 1968; Simmons,
1963), the middle frequencies apparently have steeper slopes.
Figure 6 represents the effects of temporal integration
upon the threshold of the acoustic reflex in six normal hear
ing subjects by Djupesland and Zwislocki (1971). The median
of the individual threshold means demonstrates a regular
decrease of 5 dB to 7 dB per halving of duration, or a slope
of about 25 dB/decade. The time constant is seen to be, for
the most part, about 200 msec. One subject appears to begin
integration at 100 msec with no obvious change in slope, as
seen by the dashed line connecting the lowest mean ART data.
McRobert et a^. (1968) also present slopes which break
linearity between 100 msec and 200 msec.

48
Q
O
a
K
Eh
CJ
SIGNAL DURATION IN MSEC
Figure 6. Temporal summation of the acoustic reflex.
The solid line intersects the median of six individual mean
scores resulting in a slope of 25 dB/decade of signal dura
tion change. The dashed line indicates a similar slope but
a shorter time constant. (Redrawn from Djupesland and
Zwislocki, 1971.)

49
No explanation appears to be offered for the steeper
slope of the acoustic reflex, but the temporal summation ef
fect is there without question as illustrated in Figure 7. The
range of values reported by Olsen, Rose, and Noffsinger (1973)
for temporal integration at threshold of audibility are con
trasted with the data by Djupesland and Zwislocki (1971) for
temporal integration at threshold of the acoustic reflex.
These are both at 1000 Hz in normal ears and have been normal
ized to 200 msec so that the slope has a common reference be
tween the two sets of data. The slope for the acoustic-reflex
threshold at 25 dB/decade is about three times as large as
the median slope of 8 dB/decade at threshold of audibility.
Other authors have reported effects of short stimulus dura
tions upon the acoustic reflex, but their results are not
reported systematically as temporal integration, nor in a form
easily converted to the values presented in Figures 6 and 7
(Johansson, Kylin, and Langfly, 1967; Lilly, 1964; Moller,
1962b; Weiss, Mundie, Cashin, and Shinabarger, 1962).
There have been no published reports dealing with the
effects of temporal summation of the acoustic-reflex thres
hold in hearing-impaired ears, but there are indications of
disturbed spatial summation. Beedle (1970) and Beedle and
Harford (1971) have reported finding an effect of stimulus
intensity upon the growth to the acoustic reflex. The slope
of the reflex growth is much steeper and more rapid for nor
mal ears than for either ear of an unilateral Meniere's group.
Niemeyer and Sesterhenn (1972) have noted the occurrence of

50
a
3
o
as
en
g
as
Eh
Figure 7. Temporal integration at threshold of audi
bility and of the acoustic reflex for normal-hearing subjects.
The median slope of the acoustic reflex is about three times
the change in the median slope of audibility. (Based on
Djupesland and Zwislocki, 1971, and Olsen et al., 1973.)

51
a difference in reflex eliciting intensity levels for pure
tones versus white noise. The intensity level difference
between the two types of stimuli is much larger in normals
than the range in cochlearimpaired ears.
Norris, Stelmachowicz, and Taylor (1972) have reported
the most significant difference in the acousticreflex action
between normal and pathological cochleas. They measured the
difference in levels necessary to obtain the acoustic-reflex
threshold. The stimulus is pulsed at 2.5 pulses per second,
a 50 percent duty cycle with a 200-msec signal duration.
The resulting modulation of the reflex in response to the
pulsed tone is greater and more regular in normal ears than
in sensorineural hearing losses. This may reflect an in
creased latency of acousticreflex response in cochlear-
impaired ears (Johansson et ad., 1967; Simmons, 1963) or a
distorted spatial summation involving rate of reflex growth
with increased stimulus level (Simmons, 1963).
If effects due to disturbed spatial summation are present
in the acoustic reflex as a result of cochlear pathologies, it
is reasonable to assume that end-organ lesions can also
disturb temporal summation. Cochlear lesions can and do
disturb temporal integration at threshold of audibility, that
is, by flattening the slope of temporal integration and/or
shortening the critical duration. Intuitively, then, tem
poral integration of the acoustic reflex may also be affected
by cochlear lesions, for example, a flattened slope and/or
a shortened critical duration.

52
In summary, temporal summation is a neural event origi
nating at the cochlea and is essentially complete before the
level of the superior olivary complex (SOC), where the
acoustic reflex is activated. Growth in neural activity in
the SOC parallels the growth in loudness up to 60 dB SPL.
This neural activity decays in an exponential manner, thereby
appearing to trade energy units linearly for about 200 msec.
It is suggested that during acoustic stimulation, neural activ
ity in the reflex center exceeds a certain level and causes
a contraction of the intra-aural muscles. Furthermore, the
mechanism suggested as controlling and/or affecting temporal
summation of threshold and loudness is the filtering charac
teristics of the ear. Lesions in the cochlea can modify the
transmission characteristics through the auditory system in
such a way that the acoustic reflex and temporal integration
are changed in a predictable manner.
The foregoing studies indicate two important clinical
premises: (1) it is not necessary to compare the acoustic-
reflex threshold with the threshold of audibility to obtain
diagnostic information about the state of the cochlea, and
(2) there are factors affecting the acoustic reflex which may
differentiate normal ears from those with end-organ lesions.
Hypothesis
The major hypothesis, formulated for testing in this
study, is based on information concerning temporal summation
of the acoustic reflex: the acoustic reflex changes in a

53
systematic fashion in the normal ear as a function of stimu
lus duration, i.e., a tenfold change in time can cause a
median change of 25 dB of acoustic power. The null hypo
thesis is: temporal summation of the acoustic reflex does
not function differently in those subjects who have normal
hearing and in those subjects who have a known end-organ
auditory lesion.
The questions that this study attempted to investigate
are:
1. What is the most sensitive measure of the acoustic-re
flex threshold when using increments of 2 dB: the first
response, the first level to occur 60 percent of the
time, a measure of central tendency of either 5 or 10
trials, or the lowest response of 10 trials?
2. What is the group average and range of acousticreflex
thresholds for a normal-hearing population as a function
of temporal integration?
3. What is the group average and range of acoustic-reflex
thresholds for an unilaterally cochlear impaired popula
tion as a function of temporal integration?
4. Is there a significant difference in temporal integra
tion of the acoustic reflex between normal hearing ears
and cochlear impaired ears?
5. Is there a significant difference in temporal integration
of the acoustic reflex between the ears of normalhearing
subjects and the normalhearing ears of unilaterally
cochlear impaired subjects?

54
6. Is there a significant difference in temporal integra
tion between the normal-hearing ear and the cochlear-
impaired ear of a Meniere's group?
7. If cochlear-impaired ears demonstrate reduced integra
tion at threshold of audibility, do they also demonstrate
reduced temporal integration at the threshold of the
acoustic reflex?

CHAPTER III
METHODS AND PROCEDURES
Temporal summation at both threshold of audibility and
threshold of the acoustic reflex were measured by the follow
ing experimental design. First, a baseline of temporal
integration at threshold of audibility was obtained in order
to compare the results in this study to previously reported
data on normal-hearing and cochlearimpaired subjects. This
information delineated normal integration from reduced inte
gration of cochlear-impaired ears. Second, the results of
temporal summation of the acoustic reflex determined whether
the null hypothesis should be rejected; that is, temporal
summation of the acoustic reflex does not function differ
ently in those subjects who have normal hearing and in those
subjects who have a known end-organ auditory lesion. The
results of this second auditory procedure were obtained with
out the subject's overt cooperation.
Ten cooperative adult subjects were tested. One-half
of the group had normal hearing and were the control subjects.
The other half of the group evidenced unilateral end-organ
hearing impairments and served as the experimental subjects.
These two groups were classified homogeneously and screened
for possible problems that might interfere with normal
acoustic-reflex function.
55

56
The subjects had a negative history of middle-ear or
inner-ear surgery and of concussion, brain damage, or
cerebral vascular accidents. Those subjects using drugs or
medication, possibly influencing normal intra-aural muscle
function, were excluded. The medical history form employed is
shown in Appendix A. None of the subjects had conductive
hearing problems. Normal compliance at the tympanic membrane
and normal tympanograms confirmed the absence of a functional
middle-ear problem.
Data from the five control subjects were collected for
each ear. These five subjects demonstrated pure-tone audio
metric measures no worse than 20 dB HL (ANSI-1969) in either
ear for frequencies 125 Hz through 4000 Hz. The normal-hear
ing group ranged from 23 years to 30 years of age.
Only those persons having a classical, unilateral
Meniere's disease, a pure end-organ disorder, with normal
hearing in the contralateral ear, were included in the experi
mental group. The Meniere's disease was medically diagnosed.
The patients had, in the course of their medical history, the
classical symptoms of episodes of true vertigo, fluctuating
hearing loss, "roaring" tinnitus, and feeling of fullness in
the affected ear. These symptoms were supported by the
following objective findings: a sensorineural hearing loss,
loudness recruitment and vestibular canal paresis. This
group will be referred to as the Meniere's group or the
cochlear-impaired group. The Meniere's group ranged from 45
years to 63 years of age.

57
Since one of the characteristic symptoms of Meniere's
disease is fluctuating hearing in the impaired ear, there
were no minimum requirements of hearing level to qualify the
impaired ear as abnormal. The contralateral ear of each
experimental subject yielded pure-tone thresholds of 20 dB or
less (ANSI-1969) at a minimum of two out of the three test
frequencies. This qualified the contralateral ear as having
normal hearing. To facilitate discussion, ears with normal
hearing were referred to as "good," those ears afflicted with
Meniere's disease were denoted as "bad." This classification
of ears was subsequently used throughout this report.
Thus, there was a total of 15 good ears and 5 bad ears
in 10 subjects. Subject LS of the Meniere's group did not
have normal hearing in her good ear at 500 Hz. Nevertheless,
this particular ear is classified as good because she did
have normal hearing at two of the three test frequencies. The
audiograms of all subjects are contained in Appendix B.
Stimuli
The test stimuli were shaped sinusoids ranging from
500 msec to 10 msec in overall duration with 5-msec rise and
decay times. These signals were gated once every 1200 msec.
The stimuli used in this study are listed in Table 4. The
choice of stimuli was determined by previous temporalinte
gration data, ease and expediency of equipment manipulation,
and maximum utilization of the good and bad ears of the
Meniere1s group.

58
Table 4* Test Stimuli Used to Elicit Threshold of Audi
bility and Threshold of the Acoustic Reflex
Test frequency: 500 Hz, 1000 Hz, 2000 Hz
Signal durations: 500 msec, 200 msec, 100 msec, 20 msec,
10 msec
Rise/fall times: 5 msec
Inter-stimulus interval: 700 msec, 1000 msec, 1100 msec,
1180 msec, 1190 msec
Repetition rate: once every 1200 msec
More specifically, the shaped signals consisted of three
sinusoids at 500 Hz, 1000 Hz, and 2000 Hz. These test fre
quencies were chosen because most of the Meniere's group had
normal hearing in the good ear and poorer hearing in the bad
ear at these frequencies.
Five signal durations were used: 500 msec, 200 msec,
100 msec, 20 msec, and 10 msec. The duration was measured
from the gate onset of a stimulus to the cessation of the
stimulus envelope. Five hundred milliseconds represented,
for practical purposes, an infinitely long signal. The re
maining four durations represented two tenfold, or decade,
changes in time, i.e., 10 msec to 100 msec and 20 msec to
200 msec. These values also included the usually designated
limits of linear temporal integration from the long signal
duration of 200 msec to the short tone pulse of 10 msec.
The repetition rate of the stimulus was held constant by
gating once every 1200 msec. Neural independence was main
tained, because the shortest inter-stimulus interval of the

test series was 700 msec. Preliminary investigation also
indicated that this rate would allow for the muscle contrac
tion to return to the pre-contraction baseline before re
sponding to the next signal burst. Figure 8 illustrates a
tracing of an oscilloscope recording of the acoustic reflex
evoked by a 500-msec signal at 5 dB above the ART. The
muscle activity returned to the pre-contraction baseling
within the 700-msec inter-pulse interval. It should also be
noted that this recording was made towards the end of an
hour and one-half test session, and middle-ear muscle fatigue
was not observed.
The rise and decay times of the stimulus envelope, as
controlled by a pre-set electronic switch, were approximately
5 msec. The 5-msec ramp was verified on an oscilloscope by
determining the duration between the upper 90 percent and
lower 10 percent of the slope.
Figure 9 delineates the spectral characteristics of two
1000-Hz tone burst stimuli measured at the earphone output.
The bandwidth of the 10-msec duration tone pulse is about
120 Hz wide in Figure 9a. This is about 20 Hz wider than
might be expected theoretically, indicating some incidental
spread of energy. When the tone burst is lengthened to 20-
msec duration in Figure 9b, the spectral energy fits nicely
into the theoretical bandwidth of 50 Hz, The spectrum
of these tone bursts suggests that no audible spread of energy
should be detected; this is supported by the smooth envelope

60
Repetition rate
1 stim./1200 msec
Acoustic Muscle
reflex relaxation
Figure 8. Acoustic-reflex contraction. The intra-
aural muscles can contract, relax, and return to the pre
contraction baseline before contracting again. The
repetition rate is once every 1200 msec for a 500-msec tone
burst at 5 dB above the ART.

61
Figure 9. Spectral content of tone bursts with 5-msec
rise/decay ramp. (a) 1000 Hz, 10-msec impulse with half
power width of ca. 120 Hz; (b) 1000 Hz, 20-msec impulse with
half-power width of ca. 50 Hz.

62
of a 20-msec tons burst in Figure 10. In order to confirm
this, an audible check was made at the earphone. No clicks
were discerned from a 10-msec tone burst at 500 Hz placed
just below threshold audibility. It was concluded, therefore,
that a 10-msec tone pulse with a rise and fall time of 5
msec would produce a relatively undistorted signal in this
test situation, using high intensity levels.
Experimental Equipment
The block diagram in Figure 11 represents the equipment
used to generate the stimulus envelope and to record the
subject's response. A General Radio 1313A oscillator gener
ated the test frequency which was monitored by a Monsanto
Model 100-A electronic counter. The sinusoid passed through
the Grason Stadler 3262A recording attenuator to the Grason
Stadler 829E electronic switch that shaped the signal; then
the signal was step attenuated by an Hewlett Packard 350D
attenuator. By way of switches I, II, and II, the signal was
led through either a McIntosh 162K power amplifier or through
a matching transformer to the earphone, a TDH-140 dynamic
earphone in a MX41/AR cushion. The signal from switch II was
measured on a Ballatine 321 true rms vacuum tube voltmeter
and on channel one of a Tektronix 564 oscilloscope. The
same "clock" triggered both the electronic switch and the
oscilloscope.
Three response switches were under subject control.
Response switch I controlled the recording attenuator.

Figure 10. A 20-msec signal burst with a smooth 5-msec
rise and decay.

Figure 11. Block diagram of equipment used to elicit and record the intra-aural re
flex and to allow subject control over the test situation.

65
Response switch II interrupted the signal presentation at the
electronic switch, while response switch III activated a buz
zer audible to the experimenter. The reflex was measured from
the contralateral external ear canal which was sealed with an
ear olive. The reflected probe tone passed to the Peters
AP61 electroacoustic impedance bridge. The output of the
1
bridge was led through a direct current voltage divider and,
finally, to channel two of the oscilloscope.
To measure threshold of audibility, switches I and II were
closed so that the signal passed from the step attenuator
through the matching transformer to the earphone. Switch III
allowed the signal to be monitored at the oscilloscope and volt
meter without being audible to the subject. The recording at
tenuator was activated by subject response switch I. Thus, a
permanent record of the subject's response was made with a
self-tracking Bekesy technique.
The same basic equipment was utilized to measure the
acoustic-reflex threshold. The recording attenuator was
turned off. The test signal was diverted through the McIntosh
power amplifier to the earphone by using the same switches.
The eliciting signal was stored on the oscilloscope of
channel one, while the acoustic reflex was stored on Channel
two. A photographic record of both oscilloscope channels was
made using a Rolleiflex SL 35 camera with a 50-mm lens.
Calibration was performed at both the output of the
earphone in SPL and across the earphone in volts. Calibration
^The d.c. voltage divider was placed in the system to in
crease the sensitivity of the signal. See Appendix C for the
schematic diagram.

66
was made in SPL at the beginning and conclusion of the experi
ment. The step attenuator and the recording attenuator were
checked for linearity in SPL. A NBS type 9A coupler with a
6-cc cavity attached to c> Bruel and Kjaer 2203 sound level
meter with a one-inch microphone were used to measure SPL
2
(re 0.0002 dyne/cm ). Frequency response of the earphone was
also measured in SPL. Voltage checks were made before and
after each change in test frequency. A General Radio 1900-A
wave analyzer and a General Radio 1521-B graphic level re
corder were used to measure the spectral characteristics of
the test signals. These calibration data may be found in
Appendix D.
The Peters electroacoustic bridge was calibrated accord
ing to the manufacturer's specifications. The probe tone
zeroed the balance meter when the input to the bridge network
was 94 dB SPL and the probe frequency was 276 Hz. Using
four different cavity volumes the compliance dial was cali
brated at 0.2 cc, 1.0 cc, 2.0 cc, and 4.0 cc so that the
balance meter read zero. Finally, the input filter was tuned
to 276 Hz, thereby being most sensitive to the probe-tone
frequency.
Two different earphones were used in this study due to
a malfunction of the TDH-140, 10-ohm phone. A TDH-30, 10-ohm
phone elicited data on four normal-hearing subjects for their
threshold of audibility and on two normal-hearing subjects
for their acousticreflex threshold. Since both phones were

67
calibrated on the same equipment and in a similar manner,
the data reflect the respective calibration curves. In other
words, the data will be reported as if one earphone were used.
Procedure
The experimental testing was conducted in three sessions
for each individual. The temporal integration at threshold
of audibility was measured for both ears during the first test
session, while temporal integration at threshold of the
acoustic reflex was measured in subsequent sessions for each
of the two ears. The three stimulus frequencies were ran
domized for each ear,and the signal durations were presented
in random order for each test frequency. All auditory measures
were made in a single-walled IAC sound treated chamber.
Temporal Integration at Threshold of Audibility
The subject responses were recorded by using a self
tracking, Bekesy procedure. The recording attenuator was set
at a chart speed of three-fourths of an inch per minute and
1 dB of attenuation every second. The visual mean of a
minimum of 10 threshold crossings was used to establish
threshold. This mean value was located at the nearest whole
dB,and, if the threshold crossings were not stable, the last
10 stable crossings were used to establish the mean threshold
value. The total test time for both ears, including a 15- to
30-minute rest period midway, was two to two and one-half
hours. Since signal crossover to the good ear was not observed

68
nor reported by the subjects in the preliminary or experi
mental sessions, a masking noise was not applied to the good
ear of the Meniere's group.
Temporal Integration at the Thres
hold of the Acoustic Reflex
Testing of the acoustic reflex commenced once the exter
nal meatus of the reflex measurement ear was sealed with an
ear olive containing the electroacoustic bridge probe tip.
The seal was considered adequate if a positive 200 mm of
equivalent water pressure in the external auditory canal did
not break the seal. The electroacoustic bridge was periodi
cally checked to determine if the balance dial was within
25 mm of equivalent water pressure of zero. If the balance
meter was out of the target range it was rebalanced by adjust
ing the compliance dial. In order to maintain this criterion,
2
the sensitivity dial was kept on position one. If the
pressure dial was outside the target range, the subject was
asked to perforin a Valsalva or Toynbee maneuver to increase
or decrease, respectively, the middle-ear pressure.
A modified method of limits was employed to determine
the acoustic-reflex threshold. Ten ascending ART trials were
2
Although it is suggested that the higher sensitivity
dial settings be used while determining the acoustic reflex,
i.e., two or three on the Peters electroacoustic bridge and
three or four on the Madsen electroacoustic bridge, prelimi
nary observations suggested an ART difference of no more than
2 dB between sensitivity position one and position three. The
use of sensitivity position one reduced the amount of rebalanc
ing necessary throughout the test procedure. If more sensi
tivity was desired, the vertical multiplier dial or the oscil
loscope was changed.

69
made for each signal duration at each frequency in order to
obtain a stable threshold value. The experimenter started
below the acoustic-reflex threshold and increased the
stimulus level in 2-dB steps until a minimum reflex was
detected. At this point, if the response was a definite
change from the ongoing pattern, the stimulus level was re
corded as the ART. The experimenter started the sequence
over again by lowering the intensity 10 dB. If the response
was in doubt, then another response was evoked at the same
signal level or at the next 2-dB higher intensity increment.
If the reflex response was definite at the second response
or the 2-dB higher response, the initially observed sound
pressure level of the reflex was recorded as the ART. Figure
12 illustrates a tracing of a typical acoustic-reflex thres-
old response. In this instance, a 2-dB intensity increase
caused a definite change from the previously recorded pattern.
The tenth ascending threshold response was stored on
channel two of the oscilloscope. The eliciting signal was
increased 5 dB above the reflex threshold and superimposed
upon the ART on channel two of the oscilloscope. The two
superimposed acoustic reflex images on channel two were then
photographed to retain a permanent record.
During the acousticreflex testing procedure, subjects
were requested to glance quietly through some pictorial
magazines in an attempt to equate subject state of alertness
and focus of attention. Subject state and attention have

70
90 92 94 96 96 98 98
Channel 1:
Stimuli of
500-msec
duration
Channel 2:
Acoustic
reflex at
threshold
ART
Figure 12. Typical response of the acoustic reflex
at threshold. Channel 1 illustrates the 500-msec duration
tone burst increasing in 2-dB steps. Channel 2 demonstrates
the response at 96 dB SPL to be significantly different
from the ongoing activity.

71
been reported to affect the acoustic reflex (Durrant and
Shallop, 1969? Gunn, 1967). As long as the subject's atten
tion is generally diffuse and not under intense concentra
tion, e-g., distinguishing visual word choices or auditory
intelligibility, the variability is probably slight (Durrant
and Shallop, 1969).
Experimental Safeguards
In performing temporal integration at the threshold of
the acoustic reflex, sound pressure levels normally asso
ciated with thresholds of discomfort, tickle and pain, and
temporary and permanent threshold shifts were required. It
was necessary to safeguard the subjects against any harm or
unnecessary discomfort. Possible response artifacts result
ing from high intensity signals were also investigated.
It was conceivable that all subjects might be subjected
to stimulus levels in excess of 130 dB SPL, as seen in Table
5. This table demonstrates a possible intensity level neces
sary to elicit a reflex by a 10-msec tone burst at 500 Hz.
For this hypothetical person, the maximum earphone output
would be 136.5 dB SPL.
The Committee on Hearing, Bioacoustics and Biomechanics
of the National Academy of Sciences has published upper limits
of acceptable exposure to impulse noise as a function of
sould pressure level and signal duration as seen in Figure 13
(Ward, 1968). The solid line represents the upper acceptable
limit of exposure for probably 95 percent of the population

PEAK PRESSURE LEVEL (dB RE 0.0002 DYNE/CM
72
SIGNAL DURATION IN MILLISECONDS
Figure 13. Upper limits of acceptable exposure to impulse
noise for 95 percent of the population to 10,000 impulses.
(Redrawn from Ward, 1968.)

73
Table 5. Conceivable Sound Pressure Level Necessary to
Elicit the Acoustic-Reflex Threshold by a 500-Hz
Burst at a Signal Duration of 10 Msec
Normal threshold at zero hearing level
(ANSI-1969) in dB SPL 11.5 dB
The ART at the upper limit of normal
sensation level 100.0 dB
Power increase necessary to maintain thres
hold when a tone burst is shortened to
10 msec 25.0 dB
Total SPL 136.5 dB
to 10,000 pulses in some period of time, i.e., one hour to
one day. This limit applies to impulses with a known single
rise-peak decay, not a complex impulse. The Committee
stated that, although there are many unknowns, their data for
upper limits of exposure constituted a conservative estimate.
Only the "weakest ears" of the unaccounted 5 percent might
demonstrate a temporary threshold shift because the acoustic
reflex would afford some protection in some subjects. The
horizontal dashed line in Figure 12 indicates the maximum
output of the earphone within this experimental system at
139 dB SPL and the vertical hatched lines indicate the approxi
mate durations used in this study. As may be seen in Figure
12, the maximum output of the earphone falls below the maxi
mum acceptable exposure limits.
To minimize possible discomfort from loud acoustic
impulses, the subjects were allowed some control over the
test situation. If the subjects felt that the stimulus pulse

74
was uncomfortable, they were instructed to press a button
(see response switch III in Figure 9), which activated a
tone audible only to the experimenter. The experimenter then
reduced the stimulus 10 dB. This situation only occurred
once because of experimenter error. A second control button
(see response switch II in Figure 9) allowed the subject to
discontinue the stimulus presentation at will. This second
button was never employed by the subjects.
Dealing with these high signal levels for long periods
of time was bound to introduce some auditory adaptation or
fatigue. The stimulus presentation rate of once every 1200
msec and a duty cycle of less than 50 percent was used as an
arbitrary compromise between long, fatiguing test sessions
and slow signal presentation rates in an attempt to minimize
neural changes. In addition, at the end of each test run
(10 thresholds per signal duration), the test signal was
turned off in order to allow a two- to five-minute rest inter
val During this interval, the signal duration or frequency
was changed and voltage levels checked for calibration.
Opinions of the subjects indicated that sessions of one and
one-half hours approached the maximum tolerable limit. This
was the time necessary to test the acoustic reflex of one ear.
The possibility that the test signal would have crossed
over to the opposite ear, the recording ear, was considered.
If this had occurred, the data would have been contaminated
in two ways. First, the reflex eliciting tone might have

75
crossed over and combined with the probe tone, causing a
reflex on the measuring side. This was unlikely, since a
minimum of 40-dB to 50-dB difference in hearing level is
normally necessary for crossover. The effective signal level
reaching the probe ear side would have been 40 dB to 50 dB
less than necessary to evoke the reflex. Since doubling the
signal pressure increases the SPL only 6 dB, the increase in
probe-tone pressure would have been a maximum of 6 dB and
probably only 2 dB to 4 dB. This would not have raised the
probe-tone intensity level enough to cause an acoustic reflex.
Second, the reflex-e1icitng signal that crosses over may
have registered on the pickup microphone of the electroacoustic
bridging network. This network has a filter centered at
276 Hz, and it is at least 20 dB down at 500 Hz. Thus, the
filter should reduce by at least 20 dB any energy crossover
at 500 Hz and possibly even at 1000 Hz and 2000 Hz.
In conclusion, all reasonable safeguards for subject
well-being and integrity of the acoustic reflex data were
taken into account without compromising the experimental
protocol. The subjects' responses, especially from those
exhibiting loudness recruitment, indicated that the subject
safeguards were not actually needed.

CHAPTER IV
RESULTS AND DISCUSSION
The results of this study indicate that temporal summa
tion of the acoustic-reflex does not function differently in
those subjects who have normal hearing and in those subjects
who have a known end-organ auditory lesion. A statistical
analysis of the data did not support a consistent, significant
difference between normal-hearing (good) ears and cochlear-
impaired (bad) ears, using the acoustic-reflex threshold
(ART). However, there were statistically significant differ
ences between responses of the two groups at threshold of audi
bility. An analysis of these data follows in a comparison of
normal-hearing and hearing-impaired ears at 1) temporal inte
gration at the acoustic-reflex threshold, 2) temporal integra
tion at the threshold of audibility, and 3) temporal integra
tion at the acoustic-reflex compared to temporal integration
at threshold of audibility.
1) Temporal Integration at
the Acoustic-Reflex Threshold1
2
A discriminant analysis (Rao, 1952) was performed on the
ART difference between a 10-msec and a 200-msec duration tone
burst and between a 20-msec and a 200-msec duration tone
^The raw data are reported in Appendix E.
2
Further information concerning discriminant analysis
is contained in Appendix F.
76

77
burst at each test frequency. In Figure 14, for example,
the mean change for the cochlear-good ear at 500 Hz from
10 msec to 200 msec is 29 dB while the mean change at the
same parameters for the cochlear-bad ears is 22 dB. The
20-msec to 200-msec difference was also chosen for discrimi
nant analysis because the change was large enough to yield
mean differences among groups (see Figure 14) and because
some subjects did not yield an acoustic reflex at 10 msec
with the maximum acoustic power of 139 dB SPL. All thresh
old differences and the data in some figures were subtracted
from, or normalized to, 200 msec. Two-hundred milliseconds
was most often stated as the time constant of temporal inte
gration, and there was little difference between the ART
elicited by a 500-msec and by a 200-msec duration tone. The
mean change of ART between the two durations was 2 dB to
3 dB, depending on frequency and subject group. The figures
were normalized to the signal duration which delineated the
results most clearly.
The results of discriminant analysis are reported as
an "F" ratio. Table 6 gives the statistical results of the
discriminant analysis upon the ART. Two comparisons out of
12 between good and bad ears were statistically significant,
both at 500 Hz and both at the ART difference between 20
msec and 200 msec. Ten of the 12 good-bad ear comparisons
were not significantly different.
The mean ART values of the cochlear-impaired ear, as
demonstrated in Figure 14, were always less than those of the

330
20
10
0
SIGNAL DURATION IN MSEC
14. Mean acoustic-reflex thresholds as a function of temporal integration.
-0
00

79
good ear of the cochlear-impaired group and the normal
hearing group. However, neither the mean values, nor the
average slope per decade of stimulus time, were significantly
different among the three groups (Table 7 ) .
Table 6. Discriminant Analysis of the Acoustic-Reflex
Threshold *
500 Hz
1000 Hz
2000 Hz
Normal vs. cochlear-good
ears
10 msec-200 msec
ns
ns
ns
20 msec-200 msec
ns
ns
ns
Normal vs. cochlear-bad ears
10 msec-200 msec
ns
ns
ns
20 msec-200 msec
.05
ns
ns
Cochlear-good vs. cochlear-bad
ears
10 msec-200 msec
ns
ns
ns
20 msec-200 msec
.10
ns
ns
ns = not significant
.05 = significant at
.10 = significant at
the 5
the 10
percent
percent
level.
level.
Table 7. Average Slope Change Per Decade of
Acoustic-Reflex Threshold
Time
of the
500 Hz
1000 Hz
2000 Hz
Cochlear-good ears
23 dB
22
dB
25 dB
Normal ears
17 dB
19
dB
19 dB
Cochlear-bad ears
15 dB
14
dB
14 dB

80
Although the group means were not statistically differ
ent, the data from Figure 14 indicate two consistent find
ings. First, the cochlear-impaired (bad) group always had
less mean ART change than the other two. The cochlear-good,
for the most part, had the greatest amount of ART change.
The difference between the cochlear-bad and normal ears at
20 msec increased from between 4 dB and 7 dB to 12 dB for
500 Hz, 1000 Hz,and 2000 Hz, respectively. There were no
other large systematic changes in this respect. Second,
the change from 20 msec to 100 msec (or to 200 msec) became
less as frequency increased for the cochlear-bad ears. This
decrease in mean slope caused the former results of increased
ART difference between normal and cochlear-bad groups.
Figure 15 illustrates the large inter-subject variation
of temporal integration slopes for the normal ears and the
cochlear-bad ears. Similar configurations also appear in
Figure 16, contrasting cochlear-good ears and cochlear-
bad ears. In both figures, the slope of individual sub
jects varies over a wide range within a group, so that the
normal and cochlear-impaired groups overlap. The large inter
subject variability is probably the reason why the mean
values were not statistically different from one another.
Figure 17 contrasts the inter-subject range among
the three groups. This graph also depicts the large amount
of overlap among the three groups, which hampered
statistical significance. The least amount of overlap
occurred at the test frequency of 500 Hz and

ACOUSTIC-REFLEX THRESHOLD
.B RE THRESHOLD AT 500 MSEC
4 0
1 0
i^Nprmal eaifs 50Q Hz .. Normal ears 1000 Hz .. Normal ears 2000 Hz
i n
11 30-
> <
-a.,, 20 H
111
10 -
0 -
Cochlear-bad ears
1000 Hz
Cochlear-bad ears
2000 Hz
10 20
T- 1 r
n 1 r
100 200 500 10 20 100 200 500 10 20
SIGNAL DURATION IN MSEC
100 200 500
Figure 15. Acoustic-reflex thresholds of normal and cochlear-impaired (bad) ears as
A function of temporal summation.
oo

ACOUSTIC-REFLEX THRESHOLD
B RE THRESHOLD AT 500 MSEC
SIGNAL DURATION IN MSEC
Figure 16. Acoustic-reflex thresholds of normal (good) and impaired (bad) ears of the
Meniere's group as a function of temporal summation.

ACOUSTIC-REFLEX THRESHOLD
B RE THRESHOLD AT 500 MSEC
33
Figure 17. Inter-subject range of acoustic-reflex
thresholds as a function of temporal integration.

84
would account for the statistically significant findings at
500 Hz. Using discriminant analysis, the more distinct the
groups the less chance of error in classification (Rao, 1952).
It is notable that the point of greatest significant differ
ence (p<.05 at 500 Hz for normal vs. cochlear-bad) yields the
smallest difference between mean values (Figure 14).
Figure 18 compares the temporal integration slope of
the acoustic reflex of normal-hearing persons in three
studies. The individual mean scores from this study and
Djupesland and Zwislocki (1971) are plotted as well as the
median values. The group mean values from McRobert et al.
(1968) are also included. The average change in slope per
decade in time at 1000 Hz for this study is 6 dB less than
the 21 dB/decade from Djupesland and Zwislocki, as well as
2 dB less than the 25 dB/decade from McRobert et a_l. In
addition, it appears that the individual mean values for this
study were more scattered than those of Djupesland and
Zwislocki (1971), especially at the 10-msec signal duration.
This large scatter of individual scores may be repre
sentative of the true population, since no more than nine
normal-hearing subjects were employed in any one study.
These differences may also be a result of the type of
equipment employed, since the two lower values of 19 dB and
21 dB/decade were obtained with a Peters electroacoustic
bridge and a Madsen electroacoustic bridge, respectively.
The steeper slope of 25 dB/decade was obtained with a

85
o
o
K
en
§
x
E-<
X
w
3
§
I
u
H
En
w
D
O
U
<
u
w
en
S
o
o
LO
E-i
<
Q
a
o
X
en
K
Eh
03
^3
I 1000 Hz
Figure 18. A comparison of mean acoustic-reflex
thresholds as a function of temporal summation in normal
hearing ears.

86
Zwislocki mechanical impedance bridge. There have been
reports indicating that the electroacoustic and mechanical
impedance bridges can vary as much as 35 percent in their
impedance values, depending upon the condition and size of
the tympanic membrane (Lilly, 1970; Wilber, Goodhill, and
Hogue, 1970).
The statistical observations concerning the null
hypothesis of the ART were limited because of unexpected
inter-subject variability causing the sample size of five
normal and five cochlear subjects to be ineffectual statis
tically. However, descriptive observations can be made,
although these observations may be due to chance.
The time constant of temporal integration of the ART,
3
that point in time where integration begins, is shorter
than the 200 msec reported for threshold audibility. A
shortened time constant has been reported previously by
Small et al. (1962) for temporal integration at supra-
threshold levels. As illustrated in Figure 19, the ART time
constant is most often between the signal durations of
100 msec and 20 msec, or the 100 msec-20 msec interval.
Further inspection of Figure 19 suggests that the time
3
The time constant was defined in this experiment as
1/e of the maximum threshold change. In practicality, 1/e
was 1/2.718, or 32 percent, of the mean integration slope
per decade of stimulus time in normal ears, or 6 dB. The
first 6 dB of threshold change from 500 msec, therefore, was
not considered part of temporal integration. The point at
which 7 dB, or more, of threshold change occurred due to
change in signal duration marked the length of the time
constant.

87
<
Figure 19. Signal-duration interval in which the time
constant of temporal integration for each test subject occurred
across all frequencies.

88
constant of the two good-ear groups is never shorter than
a duration of 20 msec. Interestingly, the opposite observa
tion may be made with the cochlear-impaired group. There
were no long time constants in the 500 msec-200 msec inter
val, but 25 percent of the cochlear-impaired ears, across
all frequencies, demonstrated integration functions that
were in the 20 msec-10 msec interval. That is, temporal
integration for these 4 out of 16 ears may be abnormally
shortened due to the cochlear impairment. This can be most
effectively seen in Figure 20, where three of the cochlear-
impaired ears with a short time constant are compared with
the normal-hearing results at 1000 Hz and 2000Hz. A short
ened time constant has been previously reported in cochlear-
impaired ears for temporal integration at threshold of
audibility (Figure 5).
Inspection of Figures 15 and 16 demonstrates that an
average slope value per decade is not necessarily the best
descriptor of temporal integration. The change is not
linear, especially in the cochlear-impaired group. The
cochlear-bad ears at 500 Hz integrate slowly, with a flatter
slope, until 20 msec when the rate of change becomes greater.
For example, four out of five cochlear-bad ears at 500 Hz
change at an average rate per time decade of 8 dB to 20 msec,
and then abruptly change to an approximate slope of 50 dB
to 55 dB per time decade. This change is only distinctive
between good and bad ears at 500 Hz. This difference

89
Figure 20. The delayed time constant in some cochlear-
impaired ears contrasted with the normal time constant of
temporal integration at the acoustic-reflex threshold.

90
probably contributed to the only two rejections of the null
hypothesis which occurred (Table 7).
This abrupt change in slope occurred in the 20 msec-
10 msec interval, the same interval in which the short time
constant of cochlear-imparied ears fell in Figure 20. Since
the definition of the time constant was relatively arbitrary,
the slow-abrupt change in slope and the short time constant
mey reflect the same phenomena in the cochlear-imparied ear.
From the discriminant-analysis data one is able to
identify those subjects consistently classified correctly,
although no statistical significance is attached. Two
ehlear-impaired subjects, FH and BC, were never classified
£ls normal hearing, while EC was incorrectly classified in
only 2 out of 12 good-bad ear comparisons. Of the five
feat subjects, the data from these three indicated that one
§huld feo able to distinguish a cochlear-impaired ear based
R aeystic-reflex data. The cochlear abnormality may
ayse §Lf\ aberration in temporal integration of the ART as
inie^ted by an initial slow growth of the integration func-
fi or by an unusually short time constant. One cochlear-
irogaird subject is an appropriate example. LS was incor
rectly classified 9 out of 12 times. Two of the correct
Classifications were probably based on her very short time
fistanf, Her ART changed only 2 dB as the signal duration
§ §h@rtened from 500 msec to 20 msec, then it changed
afeFUpfely by 35 dB. She is represented in Figure 20 at 2000
S fey %he white hearts.

91
In addition, the acoustic reflex occurred at similar
levels reported in Tables 1, 2, and 3. In Figure 21, the
pure-tone thresholds are separated by 20 dB to 40 dB, but
the acoustic-reflex thresholds are overlapping. The
cochlear-good ear group is depressed somewhat at 500 Hz, but
this is because the sample size was three ears at that
parameter.
The ART was established as the mean value of 10 trials.
This may have been unnecessarily long since there was no
significant difference between the mean of the 10 and the
mean of the first 5 trials. The modified Hughson-Westlake
approach (Carhart and Jerger, 1959) would have been as use
ful as the mean of 10 trials, but for the fact that no single
value could be obtained 50 percent of the time in some
instances. Using 2-dB increments, the average range of
values obtained for threshold in any 10 trials was 4 dB to
6 dB. A range of 10 dB was rarely obtained. For the most
part, the shorter the test duration the more variable the
response became. A mean of 5 trials would seem to yield a
stable ART in half the time used in the present experiment.
2) Temporal Summation at
Threshold of Audibility4
A discriminant analysis of temporal summation at thresh
old of audibility between normal-hearing ears and cochlear-
impaired ears indicated significant good-bad ear differences
4
The raw data are reported in Appendix G.

THRESHOLD dB RE HEARING LEVEL (ANSI-1969)
92
Figure 21. Thresholds, audibility, and acoustic reflex,
of the test subjects at 500-msec signal duration.

93
10 out of 12 comparisons. The results are listed in Table 9.
There was no statistical difference between cochlear-good
ears and cochlear-bad ears at the 20 msec-200 msec comparison
in two out of three frequencies. Further, there was no
statistical difference between normal and cochlear good-ears
in all comparisons.
Table 8. Discriminant Analysis of the Threshold of
Audibility Data
500 Hz
1000 Hz
2000 Hz
Normal vs. cochlear-good ears
10 msec-200 msec
ns
ns
ns
20 msec-200 msec
ns
ns
ns
Normal vs. cochlear-bad ears
10 msec-200 msec
.10
.05
.05
20 msec-200 msec
. 05
.05
.05
Cochlear-good vs. cochlear-
bad ears
10 msec-200 msec
.05
.05
.05
20 msec-200 msec
ns
.10
ns
ns = significant.
.01 = significant at the 5 percent level.
.10 = significant at the 10 percent level.
As with the ART, there were no significant differences
among the mean values of the three groups, as presented in
Figure 22. Again, the values for the cochlear-impaired ears
lie below those for the normal-hearing ears. The range of
inter-subject thresholds was not as great as that found with

15
10
5
0
15
10
5
0
15
10
5
0
94
. a Normal ears
Cochlear-good ears
Cochlear-bad ears
i i 1 r
10 20 100 200
SIGNAL DURATION IN MSEC
22. Mean thresholds of audibility as a function
integration.

95
the ART, and there was not as much overlap between the bad-
ear and good-ear groups. The inter-subject variability is
illustrated in Figure 23.
The data from temporal integration at threshold compared
favorably with those of previous investigators. In Figure
22, the mean normal-hearing integration slopes per decade of
stimulus time were 8 dB to 11 dB. The normal ears demonstrated
a frequency effect whereby the slope flattens as test fre
quency increases. These slopes were 10 dB, 9 dB, and 8 dB
at 500 Hz, 1000 Hz, and 2000 Hz, respectively. No systematic
effect was seen for the cochlear-good ears, but the sample
size was smaller than that of the normal ears.
_ At each frequency in Figure 22, the mean slope for
cochlear-bad ears was flatter than, and below, either of the
normal-hearing groups. Since the mean slopes are between
6 dB and 7 dB for the cochlear-impaired ear, the distinc
tion between normal-hearing and cochlear-impaired ears is
rather subtle. The lack of clear inter-group distinction in
Figure 23 further points to the problem of utilizing
individual results for diagnostic information. Only those
subjects with a slope of less than 5 dB/decade would have
been distinctly classified as cochlear-impaired. Olsen,
Rose, and Noffsinger (1973), with a larger experimental popu
lation, reported normal subjects with slopes as flat as 2 dB
to 3 dB. They concluded that the inter-group "... overlap
was sufficiently great that no characteristic pattern of
behavior defined any group."

40
30
20
10
0
40
30
20
10
0
40
30
20
10
0
96
I 1
l_ H
J 1
hi
J rl
i
i
Normal ears
Cochlear-good ears
Cochlear-bad ears
500 Hz
1000 Hz
i
2000 Hz
i i ¡ 1 r
10 20 100 200 500
SIGNAL DURATION IN MSEC
23. Inter-subject range of thresholds of audi-
function of temporal integration.

97
The time constant of temporal integration at threshold
of audibility in cochlear-impaired ears is said to be
shorter (Figure 5). Table 9 lists the signal duration
intervals within which the time constants occur.^ The re
sults are reported in both the number of times and the per
centage of times a time constant was within a certain inter
val for each group. Even though the cochlear-impaired (bad)
ears had the highest percentage occurring in the 100 msec-
20 msec interval, there was not much difference between
normal and impaired ears. These results could not support
the statement that cochlear-impaired ears have shorter time
constants than normal ears.
Table 9. Signal Duration Intervals Within Which Time
Constants of Temporal Integration at Threshold
of Audibility Occur
Signal Duration Interval in Msec
Total
400-200 200-100 100-20 20-10 Responses
Normal ears
4
(13%)
9
(30%)
17
(57%)

30
(100%)
Cochlear-good
ears
5
(36%)
5
(36%)
4
(28%)

14
(100%)
Cochlear-bad
ears
2
(13%)
3
(19%)
11
(68%)

16
(100%)
The time constant was calculated to be at or above the
first 3 dB of threshold change from 500 msec. The time con
stant criterion is the same as the ART time constant (footnote
3) .

98
3) Temporal Integration at the Acoustic-
Reflex Threshold Compared to Temporal
Integration at Threshold of Audibility
There were no significant correlations occurring con
sistently between the data on temporal integration at the
two thresholds, ART and audibility. There were no corre
lations on direction or amount of change between the two
integration functions, nor were there on classification of
ears. This indicates that abnormal temporal integration at
threshold of audibility does not predict the results of
temporal integration at the ART. It must be remembered,
though, that the threshold criteria for the two thresholds
were not only different but also involved different persons.
The experimenter specified the ART and the subject determined
his own threshold of audibility. The difference in cri
teria may be reflected in many different ways, e.g., a
different point on the intensity function for threshold,
i.e., 25 percent vs. 75 percent of correct responses.
Even though there were no significant comparisons
between the two thresholds, some observations can be made.
The mean temporal-integration slopes of the cochlear-impaired
ears fell below those of the normal-hearing groups with
both thresholds. The ART slopes yielded much larger mean
differences among groups, but also greater inter-subject
variability, than those at threshold of audibility. The
most obvious difference between the two types of thresholds
is the slope per decade of time. The slopes of the ART in
^The correlations are listed in Appendix H.

99
Figure 14 are approximately twice those at threshold of
audibility of Figure 22. This gives ART measures the
possible advantage of a greater dynamic range to determine
differences between normal and abnormal ears. Unfortunately,
the large inter-subject variability and group overlap has
eroded this possible advantage.
The time constant of integration was essentially the
same between normal-hearing and cochlear-impaired ears at
threshold of audibility but it was definitely shorter for
cochlear-impaired ears at the ART. The cochlear-impaired
ears at the ART were the only ones to have a time constant
of 20 msec or less. This may be a result of the previously
mentioned larger dynamic range for temporal integration at
the ART. Nevertheless, the shorter time constant is one of
the most distinguishable differences between temporal inte
gration at the ART and at the threshold of audibility.
r.5 ~ It should be mentioned that only one person was
correctly' classified for each parameter across all frequen-
eies:at both thresholds. FH had a flattened slope, as well
as ar.short time constant of less than 2 0 msec.

CHAPTER V
CONCLUSIONS AND SUMMARY
A concluding comment should be made concerning the
nature of diagnostic tests, as viewed through these results.
If the ART did indeed prove to differ cochlear abnormalities
from normals, what can be considered abnormal? A specific
defect or aberration in auditory processing does not neces
sarily reflect one specific and correlative anatomical lesion,
if, indeed, it is possible to obtain a homogeneously impaired
population. The experimental group is a good case in point,
because they were chosen as a homogeneous population with
classical Meniere's syndrome. Three subjects, EC, FH, and
BC, had Meniere's for less than 6 years. They also had
the most depressed slopes. The other two subjects, LS and
AC, had greater than normal slopes and had had their disease
symptoms for 16 to 35 years, respectively. LS, as well as
FH and BC, demonstrated a short time-constant for at least
one frequency, so, apparently, it does not matter how long
one has had the disease. In addition, these results are not
uniform from one frequency to the next for each subject,
even though the hearing losses were relatively flat.
It did not matter if a clinically significant hearing
loss was present, because BC had practically normal pure-tone
100

101
acuity. This is not to say that temporal integration of
the acoustic-reflex threshold can detect incipient hearing
loss or minor cochlear impairment. Instead, it reflects
the complexity of auditory processing.
In conclusion, there were some apparent, but subtle,
differences between the normal-hearing and hearing-impaired
groups in temporal integration at the ART.
o The mean integration slopes of the cochlear-impaired group
were depressed.
o The time constant of integration was shorter than 20 msec
for some of the cochlear-impaired groups,
o If the time constant is not unusually short, the integra
tion function changed gradually, then abrubtly after 20
msec in the cochlear-impaired ear.
o A cochlear-impaired ear can have both a flattened slope,
as well as a short time constant,
o One did not have to have a hearing loss in the cochlear-
impaired ear to demonstrate either a depressed slope or
a delayed time constant.
Since it has been stated that there are probably differ
ences between normal-hearing and cochlear-impaired groups,
this type of study is open for further investigation. A
similar experiment would require more signal durations,
expecially below 100 msec. Other parameters can be expanded
upon, such as type of stimuli or measuring ipsilaterally
instead of contralaterally to the test ear. Temporal

102
integration measures do not have to take place at threshold,
nor by changing the stimulus durations only. This type of
investigation should be pursued because the acoustic reflex
can reflect a cochlear abnormality without the subject's
cooperation.
In summary, five normal-hearing and five unilaterally
hearing-impaired persons were test subjects. Their results
determined the efficacy of temporal summation of the acoustic
reflex threshold as a possible predictor of cochlear abnor
malities. Temporal summation of the acoustic-reflex thresh
old was obtained by maintaining a constant intra-aural muscle
reflex strength as a function of increased acoustic power
while signal duration was decreased.
A measure of temporal integration was made at both
threshold of audibility and threshold of the acoustic reflex,
the first auditory measure was determined by the subject and
the second by the experimenter. Bekesy tracking was employed
to self-record threshold of audibility, while a modified
method of limits, using 10 ascending trials, determined the
acoustic-reflex threshold. The signal durations used were
500 msec, 200 msec, 100 msec, and 20 msec at the test fre
quencies of 500 Hz, 1000 Hz, and 2000 Hz. As much as 139
dB sound pressure level was employed to maintain constant
energy at threshold. The reflex was monitored from an
oscilloscope connected to a Peters 61AP electroacoustic
bridge.

103
Temporal summation of the acoustic reflex does not
function differently in those subjects who have normal
hearing and in those subjects who have a known end-organ
auditory lesion. However, there were statistically signifi
cant differences found between the normal-hearing (good) ear
and the cochlear-impaired (bad) ear comparisons of the 20
msec-200 msec ART difference at 500 Hz. Small sample size,
and unexpected inter-subject variability and flatter inte
gration slopes for normal ears, contributed to the lack of
statistical significance.
Temporal integration at threshold of audibility, in
contrast, did demonstrate statistical differences between
cochlear-impaired ears and normal-hearing ears. Even though
a statistical difference was evidenced, the difference was
not so obvious, in this experiment, as to be a clinically
useful tool. A comparison of temporal integration at thresh
old of audibility and at the ART demonstrated that a short
time constant for the ART may be of diagnostic value in
determining cochlear impairment. However, the question of
whether temporal summation of the acoustic reflex is clini
cally feasible as a diagnostic measure of cochlear abnor
malities was not resolved in this experiment. Any observa
tions made, e.g., the short time constant, were not proven
to be other than chance observations. ART integration,
therefore, could not be safely said to identify an end-organ
impairment.

APPENDICES

APPENDIX A
MEDICAL HISTORY FORM

Subject Medical History Form
1. Name Age Sex
2. Address
3. Date
4. Subject group: Normal Cochlear
5. History:
A.Conductive
1.Recent middle-ear problems: Yes No
Explain
2.Ear operations: Yes No_
Explain
When Intra-aural muscles involved
B.Sensori-neural hearing problems:
1. Ear: L R Both
2. First noticed Duration
3. Diagnosis By whom
4. Meniere's disease
a. Episodes of true vertigo
b. Feeling of fullness of ears
c. Tinnitus
d. Fluctuating hearing loss
e. Frequency of attacks
f. Time of last attack
g. Present state: Active Quiescent
Unknown
5. Hearing aid: Yes No Don't use
C. Brain Injury: Yes No
Severe concussions: Yes No
CVA: Yes No
D. Medication or drugs currently used: Yes No
106

APPENDIX B
PURE TONE HEARING LEVEL (ANSI-1969) OF ALL SUBJECTS

Normal Group
Sub
ject
Ear
Classifi
cation
Frequency
in Hz
125
250
500*
1000*
2000*
3000
4000
JG
L
good
10
10
15
18
18
20
10
R
good
15
10
20
16
17
15
15
CP
L
good
0
0
1
7
8
5
0
R
good
0
0
1
4
11
5
5
GP
L
good
5
5
7
9
15
20
15
R
good
5
10
10
1
2
0
5
JP
L
good
0
0
-8
1
4
5
0
R
good
0
0
-8
2
4
0
0
FS
L
good
0
0
-6
-7
-8
0
0
R
good
0
0
-10
-4
-3
0
10
Actual values obtained by Bekesy tracking
Meniere'
s Group
Sub
ject
Ear
Classifi
cation

Frequency
in Hz
125
250
500*
1000*
2000*
3000
4000
AC
L
bad
65
60
59
68
65
70
70
R
good
20
20
19
16
13
35
35
EC
L
bad
65
60
61
50
57
50
50
R
good
10
10
6
5
8
25
45
FH
L
good
5
5
6
15
14
30
55
R
bad
30
25
30
35
49
50
55
BM
L
bad
25
20
18
22
15
5
10
R
good
20
5
8
11
13
10
10
LS
L
good
40
40
39
19
7
10
15
R
bad
45
40
47
46
40
50
50
Actual values obtained by Bekesy tracking.
108

APPENDIX C
VOLTAGE DIVIDER

Q)
Cn
T3
H
>W U
O ffl
% 0
Cu 4J
*> CO
6 d
8 0
.H n
fcl 4J
O
rH
w
o
0
The d.c. voltage divider was used to cancel the d.c.
output from the electroacoustic bridge while passing the
a.c. singal to be recorded on the oscilloscope.
110
To Input
of Channel 2
of Oscilloscops

APPENDIX D
CALIBRATION DATA TABLES

Table 1. Frequency Response of Earphones in dB SPL
(re 0.0002 dyne/cm2)
500 Hz
1000 Hz
2000 Hz
2 volts input
TDH-39
129
128
127
TDH-140
131
131
131
5 volts input
TDH-140
139
139
139
112

113
Table 2. Step-Attenuator Linearity. These Data Were
the Pre- and Post-Experimental Values for the
TDH-140 Earphone at 1000 Hz in dB SPL (re 0.0002
dyne/cn\2)
1-dB Increments 10-dB Increments
dB
Pre
Post
dB
Pre
Post
0
139.1
139.0
0
139.1
139.0
-1
138.2
138.0
-10
130.0
129.9
-2
137.4
137.2
-20
120.1
120.0
-3
136.4
136.2
-30
110.1
110.0
-4
135.4
135.3
-40
100.3
100.1
-5
134.6
134.4
-50
90.4
90.2
-6
133.7
133.5
-60
80.6
80.4
-7
132.8
132.6
-70
71.2
71.1
-8
131.9
131.8
-9
130.9
130.8
-10
130.0
129.9

114
Table 3. Recording Attenuator Linearity. These Data Were
the Pre- and Post-Experimental Values for the
TDH-140 Earphone at 1000 Hz in dB SPL
Pre
Post
0
114.4
114.0
-10
94.0
93.8
-20
83.4
83.3
-30
73.4
73.2
-40
63.6
63.1
-50
53.6
53.2
-60
43.6
43.2
-70
34.3
33.8

APPENDIX E
RAX DATA OF THE ACOUSTIC-REFLEX THRESHOLD

Appendix E. Raw Data of the Acoustic Reflex Threshold
CTl
Sub
ject
JGL
500 Hz
1000 Hz
2000 Hz
Test
Signal Duration in Msec
nun
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
Normal
Ears
i
123
108
98
96
94
126
124
100
98
94
127
120
99
97
95
2
119
111
98
96
94
124
116
100
98
96
127
125
97
97
93
3
119
111
101
98
96
126
116
100
96
96
125
122
97
97
97
4
126
113
98
96
96
124
120
100
100
96
127
125
97
95
95
5
121
112
101
94
94
126
114
101
96
96
127
123
97
93
95
6
128
111
100
98
94

116
101
98
96
127
123
97
97
97
7
123
111
100
98
94
126
116
101
96
96
127
127
95
95
97
8
121
111
100
98
96
126
112
101
98
96

120
97
97
95
9
126
113
100
96
96
126
116
103
100
96
127
118
97
97
93
10
126
117
101
96
96

116
101
98
96
125
123
97
95
95
1
126
116
98
90
84
126
105
94
88
84
121
110
97
95
93
2
124
110
98
88
84
124
107
92
88
86
125
113
97
93
91
3
126
112
98
88
86
128
107
92
90
84
127
110
95
93
93
4
126
116
100
92
86
126
105
94
90
86
125
115
97
95
91
5
124
114
98
92
88
124
109
92
84
86
'
115
95
95
91
6
126
116
98
94
84
126
107
92
86
86
125
110
97
95
91
7
124
116
100
92
86
124
107
94
84
84

113
95
95
91
8

114
98
92
84
128
109
92
84
88

115
97
93
93
9
128
114
98
94
86
126
114
92
88
88
125
110
97
93
91
10
128
114
98
96
86
128
109
92
86
86
125
115
97
95
93
1
NR
115
108
106
104
139
128
110
106
102
NR
118
106
102
97
2
117
108
104
106

124
108
106
102
119
106
100
100
3
120
104
104
108
139
126
110
104
100
118
106
102
97
4
118
108
108
108
139
128
110
106
102
123
102
100
97

ApRsn<3:j.?c R(pqn^iRVi^d))
50Q Hz 1000 Hz 2000 Hz
Sulp- Test Signal Duration in Msec
ject
Run,
10,
20,
100
200
500
10
20
100
200
500
10
20
100
200
500
5,
117,
108
106
108
130
108
106
102
120
104
102
97
6
| | '1
117
108
106
108
T
130
108
108
102
120
106
100
97
7,
1?4
108
100,
106
r-
130
110
104
100
119
106
102
97
8,
1 1
118,
108
108
108
130
110
106
104
117
106
102
97
9
l l i
117
108
104,
108
128
108
108
102

106
102
98
10
1 1 !
123
108
108
108

130
108
106
100
119
104
104
97
JGr
1
1 1 !
NR,
} i >
127,
115
110
115
NR
130
110
106
102
NR
139
114
112
103
2
i l '
125
113
110
113
12 8
110
106
102
139
114
114
103
3
127
113
111
113
130
112
106
100
137
116
112
101
4
127
113
111
115
130
112
108
102
139
114
112
101
i :
5,
131
113
113
115
130
110
108
104
135
116
112
101
i.
6
1 I l
131
117
111
113
134
112
108
102
137
114
114
103
7
1 1 1
129
115
111
113
! 1 '
130
114
106
102
137
114
112
101
8
||ft
129
115
111
115
1 *
128
110
108
104
137
114
114
103
9
1 l 1
129
115
113
115
j \ 1
130
110
110
102
139
118
112
105
10
127
113
113
113
1 '
132
110
108
100
139
114
112
103
JPL
1
122
110
101
99
99
121
116
96
100
96
138
133
117
115
111
2
122
113
101
101
99
120
116
98
102
96

131
119
115
111
3
126
115
99
101
99
122
116
98
102
96
138
133
115
115
111
4
127
115
101
101
97
120
118
96
100
96
138
131
117
113
111
i :
5,
125
113
101
103
99
123
118
98
102
96
138
133
117
115
109
6
12 3
115
101
103
97
122
119
98
100
94

135
117
115
115
7
126
115;
99
101
99
125
118
94
98
98

135
119
117
111
8
126
117
103
103
97
125
118
96
98
96
138
135
117
119
113
9
127
115
103
101
99
123
120
98
98
96
138
131
117
113
111
10
126
115
103
101
97
122
119
98
98
98
138
129
117
115
111
117

Appendix E(continued)
500 iHz
1000 Hz
2000 Hz
Sub
ject
Signal Duration in Msec
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
1
117
104
i 93
93
91
111
100
94
96
94
120
110
104
106
102
2
119
103
93
93
91
112
102
96
98
94
123
111
106
101
101
3
112
104
93
91
91
112
102
96
98
92
124
111
106
101
101
4
113
104
95
93
91
114
103
96
98
90
126
111
106
102
102
5
117
103
95
91
93
112
103
96
98
92
127
111
108
102
102
6
113
108
95
91
95
112
103
94
96
92
124
110
110
102
101
7
113
103
93
91
93
112
105
96
98
94
124
110
108
104
101
8
117
104
93
93
95
112
103
96
96
92
124
111
106
106
99
9
115
103
95
93
91
114
103
96
98
92
125
111
106
106
101
10
113
104
95
93
91
112
103
96
98
96
125
110
106
104
102
1
113
108
99
93
83
105
102
96
92
90
108
102
91
91
85
2
111
106
97
95
83
109
100
96
94
88
106
102
95
91
87
3
111
106
99
99
85
109
100
98
96
90
110
101
95
93
87
4
110
106
101
99
89
109
102
96
90
88
111
101
95
91
85
5
111
108
99
97
87
109
102
94
94
92
111
102
95
91
87
6
113
111
101
99
89
109
102
96
94
88
113
102
93
93
85
7
108
110
101
99
91
109
103
98
94
88
110
99
93
93
89
8
115
110
99
101
91
107
98
98
94
90
111
102
93
91
87
9
113
108
101
97
89
107
102
96
96
90
110
101
95
89
87
10
115
100
101
99
93
107
102
96
94
88
111
102
93
89
91
1
117
111
101
99
99
107
98
94
90
88
119
112
110
104
102
2
125
111
105
101
99
111
100
92
88
88
123
112
106
104
100
3
119
111
105
101
99
111
100
92
90
88
119
115
108
106
104
4
119
111
101
101
97
111
98
92
92
88
119
114
110
104
102
5
117
113
103
103
99
111
98
94
90
88
119
112
110
104
104
6
123
111
103
101
97
112
100
92
92
90
119
115
110
106
104
JP,
FS,
FS
R
\
118

Appendix E(continued)
Sub
ject
500 Hz
1000 Hz
2000 Hz
Signal Duration in Msec
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
7
119
109
103
99
97
111
102
92
90
92
117
114
108
104
102
8
121
111
105
101
97
111
100
90
94
90
119
114
108
106
102
9
121
109
103
101
99
109
102
92
92
92
121
114
108
104
104
10
119
111
103
103
99
107
102
90
92
88
119
114
108
104
102
1
137
123
109
107
107
132
116
106
102
100
NR
131
112
106
98
2
139
121
109
109
107
130
120
108
104
100
135
108
104
102
3
137
123
111
107
103
134
118
108
102
100
1 :1
108
104
102
4
135
123
109
107
105
136
120
106
102
98
129
108
104
102
5
133
125
111
107
109
130
118
106
102
100
133
106
102
100
6
139
121
111
107
107
134
120
108
100
98
135
110
102
100
7
139
125
111
103
109
134
118
110
102
98
133
108
102
102
8
139
123
111
105
107
134
116
108
102
98
133
110
102
102
9
137
121
111
107
107
136
118
108
100
100
133
108
104
102
10
135
121
111
109
109
136
118
106
102
100
133
108
104
102
1
138
119
107
105
103
130
114
106
104
102
123
115
98
94
94
2
132
119
109
107
105
128
116
104
102
102
125
115
98
94
96
3
138
121
109
107
103
126
116
106
102
102
129
113
100
94
94
4
134
119
109
107
103
126
112
108
102
100
133
115
100
92
94
5
134
123
109
105
103
128
114
106
102
102
129
111
100
96
96
6
134
123
109
107
105
128
114
108
102
100
131
111
100
96
92
7
136
121
109
105
101
130
116
106
102
102
131
113
100
94
94
8
134
123
109
105
103
124
114
106
102
102
135
123
100
96
94
9
136
121
109
105
103
130
114
104
104
102
131
119
98
96
94
10
138
121
109
105
105
132
114
108
102
102
131
119
100
96
96
GP,
GP,
119

Appendix E(continued)
500 Hz
1000 Hz
2000 Hz
Sub
ject
ecr
acr
Test
Signal Duration in Msec
nun
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
Cochlear
-Good
Ears
i
119
111
91
89
83
110
96
94
90
86
129
98
88
88
86
2
113
111
93
91
83
114
96
92
90
86
129
100
88
86
82
3
115
113
91
93
85
116
94
90
92
84
129
100
86
88
86
4
113
113
93
91
83
114
98
94
88
86
129
98
88
90
84
5
117
111
91
89
85
112
96
94
92
86
127
102
90
88
84
6
119
113
93
91
85
116
96
92
86
82
129
96
90
88
88
7
117
113
93
91
87
116
96
94
86
84
129
100
90
86
86
8
113
111
93
91
83
120
94
92
88
86
125
100
90
88
86
9
115
113
97
97
85
118
94
94
90
88
127
102
90
86
84
10
117
113
93
91
85
116
94
94
90
86
129
104
92
86
86
1
125
119
111
95
87
114
110
100
98
96
129
119
116
110
110
2
123
121
109
97
89
118
112
98
96
96
129
121
112
116
110
3
127
119
109
97
89
116
110
98
98
94
133
129
116
110
108
4
129
119
109
99
91
120
108
98
100
96
131
125
116
114
112
5
133
117
107
97
87
120
110
98
100
96
137
127
114
112
110
6
125
119
111
97
89
118
110
100
96
94
133
125
118
114
110
7
127
119
109
95
89
120
110
100
94
94
131
125
112
110
108
8
123
119
111
97
89
118
112
96
100
96
131
129
118
110
108
9
129
119
113
93
87
114
110
100
98
90
135
127
118
112
108
10
127
117
107
95
91
120
110
100
98
98
131
103
118
114
112
1
129
123
103
95
95
131
124
104
100
98
134
130
100
96
96
2
127
119
103
99
93
135
126
106
100
98
135
128
100
98
92
3
127
123
101
93
93
131
122
106
102
96
137
130
100
98
94
4
133
125
103
95
93
131
124
104
100
98
132
130
102
96
94
120

Appendix E(continued)
500 Hz 1000 Hz 2000 Hz
Signal Duration in Msec
ject
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
5
131
125
97
95
95
129
124
104
98
96
134
132
98
98
92
6
133
123
101
99
93
135
124
104
100
96
134
130
102
100
94
7
131
125
101
95
91
135
126
106
100
100
139
132
100
102
94
8
129
125
103
99
93
137
124
108
100
98
132
128
100
98
94
9
133
121
101
97
93
133
126
104
98
98
135
130
102
98
94
10
127
121
99
99
93
133
122
106
100
98
135
130
100
100
94
bcr
No Reflex
lsl
1
Cochlear
Impaired Fre-
135
126
104
102
100
NR
133
106
100
96
2
quency
137
130
102
104
100
133
108
100
94
3
135
132
102
104
98
133
110
104
94
4
134
130
104
100
100
133
110
100
92
5
139
132
104
104
100
129
110
102
96
6
135
132
102
104
98
129
106
102
94
7
137
130
104
104
102
139
110
100
94
8
135
126
102
106
100
137
108
100
94
9
137
130
102
104
100
135
110
102
94
10
137
132
100
102
100
139
110
100
94
Cochlear
-Bad
Ears
ecl
1
127
119
111
111
109
116
104
106
96
94
125
110
98
96
96
2
127
117
111
109
109
120
104
98
96
92
123
110
100
100
96
3
125
119
113
111
109
120
106
98
96
90
129
108
100
98
94
4
127
119
113
109
109
120
106
100
96
94
125
108
100
102
96
5
127
119
113
111
107
118
106
100
96
94
127
110
100
98
96
121

Appendix E(continued)
Sub
ject
500 Hz
1000 Hz
2000 Hz
Signal Duration in Msec
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
6
129
119
113
109
109
116
106
96
94
94
123
104
102
98
94
7
129
117
111
109
107
120
106
98
96
94
129
110
100
98
94
8
129
115
115
113
105
120
110
98
96
94
129
110
102
98
96
9
127
119
113
111
105
120
108
98
96
92
131
108
104
96
94
10
129
119
113
111
107
118
102
98
96
92
129
110
102
100
96
1
105
107
97
103
95
100
90
90
88
90
108
98
106
102
98
2
107
105
99
99
99
100
90
92
90
90
108
98
100
100
100
3
109
107
99
103
97
98
90
92
90
90
108
102
104
104
100
4
109
109
99
99
97
100
90
90
88
90
106
100
100
102
100
5
109
105
95
97
97
100
92
90
86
88
104
102
108
102
96
6
107
105
101
97
95
100
90
90
90
90
110
104
98
102
100
7
109
101
103
103
95
100
90
90
90
90
104
102
100
100
102
8
107
105
101
101
95
102
90
88
90
90
108
102
100
104
104
9
107
107
103
101
97
102
90
90
88
90
110
102
102
102
102
10
107
107
101
99
95
100
88
90
90
90
108
102
100
102
100
1
117
103
93
93
91
135
122
110
106
104
135
128
116
104
104
2
117
103
93
95
91
135
124
110
106
106
135
130
120
104
104
3
119
101
93
95
93
134
124
114
104
106
135
130
112
104
102
4
119
103
93
95
93
135
126
112
106
106
137
130
116
104
102
5
119
101
93
93
91
139
130
110
110
104
134
135
116
102
104
6
119
101
95
93
91
139
126
110
112
104
137
130
116
104
102
7
115
103
95
93
93
139
124
110
112
106
134
130
116
102
102
8
113
101
95
93
93
139
126
108
110
104
137
132
114
104
104
9
119
103
95
95
93
137
122
110
110
104
135
134
116
102
104
10
117
101
95
95
93
139
128
108
110
106
137
134
120
104
102
FH
R
AC,
122

Appendix E(continued)
Sub
ject
BC,
LS
R
LS,
500 Hz
1000 Hz
2000 Hz
Signal Duration in Msec
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
1
118
114
108
106
106
112
110
102
102
104
113
103
97
101
99
2
124
112
110
106
104
112
110
104
104
104
113
107
97
99
99
3
120
112
108
106
104
114
108
104
102
102
111
105
97
97
99
4
120
112
110
106
106
114
108
104
102
104
111
103
95
103
97
5
120
114
108
106
108
116
108
104
104
102
109
105
97
103
97
6
124
112
110
104
106
112
108
104
104
100
109
105
99
95
97
7
122
114
110
106
106
114
110
102
102
104
111
103
97
97
97
8
120
110
112
106
104
114
106
102
104
104
111
105
97
99
97
9
120
112
108
108
106
114
108
102
104
102
117
103
99
103
95
10
124
114
110
106
106
114
110
104
104
100
109
105
97
101
97
1
125
99
99
99
93
124
110
100
92
98
129
96
100
96
94
2
127
101
99
91
91
128
108
102
96
98
131
96
96
96
96
3
127
99
99
97
93
124
106
100
98
100
135
96
96
94
92
4
121
103
101
97
95
122
106
100
96
98
129
96
96
96
92
5
125
109
97
99
95
126
108
100
94
98
129
94
100
96
94
6
125
109
99
99
93
130
110
102
96
94
131
94
96
94
96
7
125
109
95
99
93
128
110
102
92
96
131
96
92
94
92
8
123
109
99
97
93
124
110
100
98
98
133
102
100
98
94
9
125
109
101
99
95
126
108
100
98
98
135
100
94
96
96
10
125
105
99
99
95
130
110
100
98
100
137
96
94
92
94
1
137
123
99
95
91
2
139
121
103
95
89
3
139
119
99
97
93
4
139
121
101
99
91
5
139
121
99
97
93
123

Appendix E--(continued)
500 Hz
1000 Hz
2000 Hz
Sub- Test
Signal Duration in Msec
ject
Run
10
20 100 200 500
10 20 100 200 500
10
20 100 200 500
6
__ __
123
101
99
91
7
137
121
101
97
93
8
139
121
105
97
93
9
137
121
101
95
95
10
139
119
99
97
93
CP =
left
ear of
subject
CP.
II
u
cu
V
right
ear of subject
CP.
f
124

APPENDIX F
DISCRIMINANT ANALYSIS

Discriminant Analysis
Discriminant analysis is a form of multivariate
analysis. This type of analysis places the individual
values from two known groups on a continuum. Based upon
the group means and variances, a point on the continuum
divides the individual values into two regions representing
two discrete distributions. Each individual is assigned to
the region to which it has the highest probability of
belonging. This posterior probability of an individual
coming from each group is based on group means, standard
deviations, a group covariance matrix, and a group corre
lation matrix. An approximate F statistic tests equality
of group means.
Once the individuals are assigned to each region,
a contingency table can represent the number of correct
and incorrect classifications, as in Table 1. There were
9 correct and 1 incorrect classifications of 10 normal ears
and 4 correct and 1 incorrect classifications of 5 abnormal
ears.
Table 1. Contingency Table of Discriminant Analysis
N
A
N
9
1
A
1
4
126

APPENDIX G
RAW DATA OF THRESHOLD OF AUDIBILITY

128
Appendix G. Raw Data of Threshold of Audibility
500 Hz 1000 Hz 2000 Hz
Signal Duration in Msec
ject
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
Normal
Ears
CPL
31
25
16
17
12
26
24
14
15
14
29
25
20
19
17
CPR
27
25
13
13
12
31
26
19
13
11
33
30
23
20
20
jgl
42
37
27
27
26
38
33
25
26
25
39
34
28
27
27
jgr
42
38
35
31
31
36
36
26
24
23
38
33
27
26
26
JPL
26
16
9
6
3
20
18
10
10
8
30
26
18
14
13
JPR
24
17
7
6
3
24
16
12
11
9
24
21
14
15
13
fsl
21
13
6
6
5
10
8
0
1
0
14
9
4
3
1
fsr
15
10
3
2
1
26
15
7
6
2
18
12
11
8
6
gpl
32
29
19
18
18
31
24
18
16
16
35
32
23
22
24
gpr
40
36
27
25
21
29
22
12
8
8
22
19
11
11
11
Cochlear
-Good
Ears
ecr
35
27
21
19
17
30
25
20
15
12
32
26
22
21
17
fhl
37
29
23
21
17
36
34
26
25
22
42
34
28
24
23
acr
43
38
32
29
30
39
33
27
25
23
36
28
24
22
22
BCS
37
30
21
19
19
38
35
28
23
18
36
29
25
23
22
LSl Hearing Impaired Fre- 39 34 28 27 26 34 27 20 18 16
quency
Cochlear-Bad Ears
lcl
80
77
72
73
72
72
67
59
57
57
72
70
67
66
66
fiir
56
51
46
45
41
53
48
40
43
42
63
61
59
58
58
acl
85
78
72
72
70
85
81
76
77
75
87
82
77
75
74
bcl !
45
40
34
29
29
37
37
31
30
29
35
31
28
28
24
lsr
68
63
59
58
58
64
61
55
55
53
61
56
53
50
49
lsl
61
57
52
51
50
f

APPENDIX H
CORRELATION COEFFICIENTS OF THE
TEMPORAL INTEGRATION MEASURES

Correlation Coefficients of the Measures Between
Temporal Integration of the Acoustic-Reflex Threshold and
Temporal Integration of the Threshold of Audibility.
Frequency
Correlation
Signifi
cance
20 msec-200 msec measure
Normal ears
500
Hz
0.3763
ns
1000
Hz
0.6926
ns
2000
Hz
0,6349
ns
Cochlear-good ears
500
Hz
0.9177
.05
1000
Hz
-0.9768
.01
2000
Hz
0.0186
ns
Cochlear-bad ears
500
Hz
-0.2654
ns
1000
Hz
-0.0995
ns
2000
Hz
0,6880
ns
10 msec-200 msec measure
Normal ears
500
Hz
0,0943
ns
1000
Hz
0.3887
ns
2000
Hz
0.6397
ns
Cochlear-good ears
500
Hz
-0.7857
ns
1000
Hz
0.3493
ns
2000
Hz
-0.9661
.01
Cochlear-bad ears
500
Hz
-0.1461
ns
1000
Hz
0.1460
ns
2000
Hz
0.7719
ns
ns = not statistically significant.
.05 = significant at the 5 percent level.
.01 = significant at the 1 percent level.
130

REFERENCES
Alberti, P. W. R. M. and Kristensen, R. The clinical appli
cation of impedance audiometry: A preliminary appraisal
of an electro-acoustic impedance bridge. Laryngoscope,
80:735-746, 1970.
American National Standards Institute. Specifications for
audiometers. ANSI S3.6-1969. American National
Institute, Inc., New York, 1970.
Anderson, H., Barr, B., and Wedenberg, E. Intra-aural
reflexes in retrocochlear lesions. In Hamberger, C. A.
and Wersall, J. (editors), Disorders of the Skull Base
Region. Nobel Symposium, 10th, Stockholm: Almqvist
and Wiksell, 1969.
Bar, B., and Wedenberg, E. The early detection
of acoustic tumors by the stapedius reflex test. In
Wolstenholme, G. E. W. and Knight, J. (editors),
Sensorineural Hearing Loss. A Ciba Foundation Sym
posium, London: J. and A. Churchill, 1970.
and Wedenberg, E. Audiometric identification of
normal hearing carriers of genes for deafness. Acta
Oto-laryngologica, 65:535-554, 1968.
Bates, M. A., Loeb, M., Smith, R. P., and Fletcher, J. L.
Attempts to condition the acoustic reflex. Journal of
Auditory Research, 10:132-135, 1970.
Beedle, R. K. An Investigation of the Relationship Between
the Acoustic Reflex growth and Loudness Growth in
Normal and Pathologic Ears. Ph.D. dissertation.
Northwestern University, Illinois, 1970.
_____ and Harford, E. Acoustic reflex and loudness
growth in normal and pathological ears. Sixth Annual
Report of the Auditory Research Laboratories. North
western University, Illinois, 1971.
Bilger, R. and Feldman, R. M. Frequency dependence in
temporal integration. Journal of the Acoustical Society
of America, 45:293(A), 1969.
131

132
Blodgett, H. C. Jeffress, L. A., and Taylor, R. W. Rela
tion of masked threshold to signal durations for
various interaural phase-combinations. American
Journal of Psychology, 71:283-290, 1958.
Boudreau, J. C. Stimulus correlates of wave activity in
the superior-olivary complex of the cat. Journal of
the Acoustical Society of America, 35:779-785, 1965.
Brahe Pedersen, C. and Elberling C. Temporal integration
of acoustic energy in normal hearing persons. Acta
Oto-Laryngologica, 74:389-405, 1972.
and Elberling, C. Temporal integration of
acoustic energy in patients with presbyacusis. Acta
Oto-Laryngologica, 75:32-37, 1973.
Brink, F., Jr. Synaptic mechanisms. In Stevens, S. S.
(editor), Handbook of Experimental Psychology. New
York: John Wiley and Sons, 1951.
Brooks, D. N. The use of the electro-acoustic impedance
bridge in the assessment of middle ear function.
Journal of International Audiology, 8:563-569, 1969.
. Electroacoustic impedance studies on normal
ears of children. Journal of Speech and Hearing
Research, 14:247-253, 1971.
Burke, K. S., Herer, G. R., and McPherson, D. L. Middle
ear impedance measurement (Acoustic and electro
acoustic comparisons). Acta Oto-Laryngologica,
70:29-34, 1970.
Campbell, R. S. and Counter, S. A. Temporal integration
and periodicity pitch. Journal of the Acoustical
Society of America, 45:691-693, 1969.
Carhart, R. and Jerger, J. Preferred method for clinical
determination of pure tone thresholds. Journal of
Speech and Hearing Disorders, 24:330-345, 1959.
Carver, W. F. Loudness Balance Procedures. In Katz, J.
(editor), Handbook of Clinical Audiology. Baltimore:
Williams and Wilkins Co., 1972.
Chamberlin, S. C. and Zwislocki, J. J. Threshold of audi
bility as a function of tone duration: Is there a
frequency effect? Journal of the Acoustical Society
of America, 48:71(A), 1970.

133
Clack, T. D. Effect of signal duration on the auditory-
sensitivity of humans and monkeys (Macaca mulatta).
Journal of the Acoustical Society of America, 40:1140-
1146, 1966.
Coles, R. R. A. Can present day audiology really help in
diagnosis? An otologist's question. Journal of
Laryngology and Otology, 86:191-224, 1972.
Dallos, P. J. Dynamics of the acoustic reflex: Phenomeno
logical aspects. Journal of the Acoustical Society of
America, 36:2175-2183, 1964.
and Johnson, K. R. Influence of rise-fall time
upon short tone thresholds. Journal of the Acoustical
Society of America, 40:1160-1163, 1966.
anc^ Olsen, W. Integration of energy at threshold
with gradual rise-fall tone pips. Journal of the
Acoustical Society of America, 36:741-751, 1964.
Deutsch, L. J. The threshold of the stapedius reflex to
selected acoustic stimuli in normal human ears.
U. S. Naval Submarine Medical Center Research Report
No. 546:1-27, 1968.
. The threshold of the stapedius reflex for pure
tone and noise stimuli. Acta Oto-Laryngologica,
74:248-251, 1972.
Dix, M. R., Hallpike, C. S., and Hood, J. D. Observations
upon the loudness recruitment phenomenon with special
reference to the differential diagnosis of disorders
of the internal ear and VIII nerve. Proceedings of the
Royal Society of Medicine, 41:516-526, 1948.
Djupesland, G. Electromyography of the tympanic muscles
in man. Journal International Audiology, 4: 34-41,
1965.
/ Flottorp, G., and Winther, F. Size and duration
of acoustically elicited impedance changes in man.
Acta Oto-Laryngologica, 224:220-228, 1966.
and Zwislocki, J. J. Effect of temporal
summation on the human stapedius reflex. Acta Oto-
Laryngologica, 71:262-265, 1971.
Doyle, T. N. Auditory Temporal Summation with Variable Inter
signal Intervals in Normal and Non-normal Subjects.
Ph.D. dissertation. University of Minnesota, 1970.

134
Durrant, J. D. and Shallop, J. K. Effects of differing
states of attention on acoustic reflex and temporary
threshold shift. Journal of the Acoustical Society
of America, 46:904-913, 1969.
Elliott, L. L. Tonal thresholds for short-duration stimuli
as related to subject hearing level. Journal of the
Acoustical Society of America, 35:578-580, 1963.
Ewertsen, H., Filling, S., Terkildsen, K., and Thomsen, K. A.
Comparative recruitment testing. Acta Oto-Laryngologica,
Supplement 140:116-122, 1958.
Feldman, A. S. Acoustic Impedance studies of the normal
ear. Journal of Speech and Hearing Disorders, 10:165-
176, 1967.
Fisch, U. and Schulthess, G. V. Electromyographic studies
on the human stapedial muscle. Acta Oto-Laryngologica,
56:287-297, 1963.
Fletcher, H. Auditory patterns. Review of Modern Physics.
12:47-65, 1940.
Flottorp, G. and Djupesland, G. Diphasic impedance change
and its applicability in clinical work. Acta Oto-
Laryngologica Supplement 263:200-204, 1970.
, Djupesland, G., and Winther, F. The acoustic
stapedius reflex in relation to critical bandwidth.
Journal of the Acoustical Society of America, 49:457-
461, 1971.
Fowler, E. P. A method for the early detection of oto
sclerosis. Archives of Otolaryngology, 24:731-741,
1936.
. The diagnosis of diseases of the neural
mechanism of hearing by the aid of sounds well above
threshold. Transactions of the American Otological
Society, 27:207-219, 1937.
. The recruitment of loudness phenomenon.
Laryngoscope, 60:680-695, 1950.
Fulton, R. T. and Lamb, L. E. Acoustic impedance and
tympanometry with the retarded: A normative study.
Audiology, 11:199-208, 1972.
Garner, W. R. The effect of frequency spectrum on temporal
integration of energy in the ear. Journal of the
Acoustical Society of America, 19:808-815, 1947a.

135
Garner, W. R. Auditory thresholds of short tones as a
function of repetition rates. Journal of the Acous
tical Society of America, 19:600-608, 1947b.
and Miller, G. A. The masked threshold of pure
tones as a function of duration. Journal of Experi
mental Psychology, 37:293-303, 1947.
Gengel, R. W. Auditory temporal integration at relatively
high masked threshold levels. Journal of the Acoustical
Society of America, 51:1849-1851, 1972.
and Watson, C. S. Temporal integration: I.
Clinical implications of a laboratory study. II.
Additional data from hearing-impaired subjects.
Journal of Speech and Hearing Disorders, 36:213-224,
1971.
Goldstein, R. and Kramer, J. C. Factors affecting thresh
old for short tones. Journal of Speech and Hearing
Research, 3:249-256, I960.
Graham, A. B. (editor). Alternate loudness balance
techniques. In Sensorineural Hearing Processes and
Disorders. Boston: Little, Brown and Company, 1967.
Grason-Stadler. Otoadmittance Handbook 2: A guide to
users of the Grason-Stadler Model 1720 Otoadmittance
Meter. Concord, Mass: Grason-Stadler Company, 1973.
Green, D. M., Birdsall, T. G., and Tanner, W. P. Jr. Signal
detection as a function of signal intensity and
duration. Journal of the Acoustical Society of
America, 29:523-531, 1957.
Gunn, W. J. Effects of attending to auditory signals on
the magnitude of the acoustic reflex. U. S. Army
Research Laboratory Report No. 751, Fort Knox,
Kentucky, 1967.
Harford, E. and Liden, G. Acoustic impedance, the middle-
ear muscle reflex, tympanometry and extratympanic
manmet.ry. Annual Report of the Auditory Research
Labortories. Northwestern University, 1967-1968.
Harris, J. D. Peak vs. total energy in threshold for very
short tones. Acta Oto-Laryngologica, 47:134-140, 1957.
, Haines, H. L., and Myers, C. K. Brief-tone
audiometry: Temporal Integration in the hypacusic.
Archives of Otolaryngology, 67:699-713, 1958.

136
Hattler, K. W. and Northern, J. L. Clinical application
of temporal summation. Journal of Auditory Research,
10:72-78, 1970.
Hempstock, T. J., Bryan, M. E., and Tempest, W. A. A
redetermination of quiet thresholds as a function of
stimulus duration. Sound and Vibration, 1:365-380,
1964.
Hirsh, I. J., Palva, T., and Goodman, A. Difference limen
and recruitment. Archives of Otolaryngology, 60:525-
540, 1954.
Holst, H., Ingelstedt, S., and Ortegren, U. Ear drum move
ments following stimulation of the middle ear muscles.
Acta Oto-Laryngologica, Supplement 182:73-83, 1963.
Hood, J. D. Basic audiological requirements in neuro
otology. Journal of Laryngology and Otology, 83:
695-711, 1969.
Hughes, J. R. Auditory sensitization. Journal to the
Acoustical Society of America, 26:1064-1070, 1954.
. Electrophysical evidence or auditory sensiti
zation. Journal of the Acoustical Society of America,
29:275-280, 1957.
Hughes, J. W. The threshold of audition for short periods
of stimulation. Proceedings of the Royal Society of
Medicine, B133 : 486-490, 1946.
Hung, I. J. and Dallos, P. Study of the acoustic reflex
in human beings. I. Dynamic characteristics.
Journal of the Acoustical Society of America, 52:
1168-1180, 1972.
International Organization for Standardization. Standard
reference zero for the calibration of pure tone
audiometers. ISO R 389-1964. Geneva, Switzerland.
Jepsen, O. The threshold of the reflexes of the intra-
tympanic muscles in a normal material examined by
means of the impedance method. Acta Oto-Laryngologica,
39:406-408, 1951.
. Intratympanic muscle reflexes in psychogenic
deafness. Acta Oto-Laryngologica, Supplement 109:
61-69, 1953.

137
Jepsen, 0. Middle ear muscle reflexes in man. In Jerger,
J. (editor), Modern Developments in Audiology.
New York: Academic Press, 1963.
Jerger, J. Influence of stimulus duration on the pure-tone
threshold during recovery from auditory fatigue.
Journal of the Acoustical Society of America, 27:121-
124, 1955.
. Bekesy audiometry in analysis of auditory dis
orders. Journal of Speech and Hearing Research, 3:
275-237, 1960.
. The audiological examination as an aid in
diagnosis. Archives of Otolaryngology, 85: 552-
554, 1967.
. Clinical experience with impedance audiometry.
Archives of Otolaryngology, 92:311-324, 1970.
f Jerger, S., Ainsworth, J., and Caram, P.
Recovery of auditory function after surgical removal
of cerebellar tumors. Journal of Speech and Hearing
Disorders, 31:377-382, 1966.
, Jerger, S., and Mauldin, L. Studies in im
pedance autiometry: I. Normal and sensorineural
ears. Archives of Otolaryngology, 96: 513-523, 1972.
, Shedd, J., and Harford, E. On the detection of
extremely small changes in sound intensity. Archives
of Otolaryngology, 69:200-211, 1959.
I
Johansson, B., Kylin, B., and Langfly, M. Acoustic reflex
as a test of individual susceptibility to noise.
Acta Oto-Laryngologica, 64:256-262, 1967.
Karlovich, R. S., Lane, R. H., Smith, L. L., Tarlow, A. J.,
Thompson, D., and Vivion, M. C. Auditory threshold
at 125 Hz as a function of signal duration and signal
filtering. Journal of the Acoustical Society of
America, 49 :1897-1899, 1971.
Klockhoff, I. Middle ear muscle reflexes in man: A
clinical and experimental study with special reference
to diagnostic problems in hearing impairment. Acta
Oto-Laryngologica, Supplement 164:1-92, 1961
Kobrak, H. G. The present status of objective hearing tests.
Annals of Otology, Rhinology and Laryngology, 57:1018-
1026, 1948.

138
Kobrak, H. G. The Middle Ear. Chicago: University of
Chicago Press, 1959.
Kristensen, H. K. and Jepsen, 0. Recruitment in otoneuralo-
gical diagnostics. Acta Oto-Laryngologica, 42:553-
560, 1952.
Lamb, L. E. and Peterson, J. Middle ear reflex measure
ments in pseudohypacusis. Journal of Speech and
Hearing Disorders, 32:42-51, 1967.
, Peterson, J., and Hansen, S. Application of
stapedius muscle reflex measures to diagnosis of
auditory problems. Journal of International Audiology,
7:188-199, 1968.
Licklider, J. C. R. The perception of speech. In Stevens,
S. S. (editor), Handbook of Experimental Psychology.
New York: John Wiley and Sons, 1951.
Liden, G. The stapedius muscle reflex used as an objective
recruitment test: A clinical and experimental study.
In Wolstenholme, G. E. W. and Knight, J. (editors),
Sensorineural Hearing Loss. A Ciba Foundation
..aa£ symposium, London: J~. and A. Churchill, 1970.
~ Peterson, J. L. and Harford, E. R. Simul
taneous recording of changes in relative impedance in
:--r--air pressure during acoustic and non-acoustic
elicitation of the middle-ear reflexes. Acta Oto-
Lryngologica Supplement 263:208-217, 1970.
Lilly, D. J. Some properties of the acoustic reflex in
man* Journal of the Acoustical Society of America,
!ll2QQ7 (A) 1964.
l.zr.zz^:: a comparison of acoustic impedance data obtained
with Madsen and Zwislocki instruments. Presented at
the-American Speech and Hearing Convention, Chicago,
I37J1.
. Acoustic impedance at the tympanic membrane.
In Katz, J. (editor), Handbook of Clinical Audiology.
Baltimore: Williams and Wilkins Company, 1972.
. Measurement of acoustic impedance at the
tympanic membrane. In Jerger, J. (editor), Modern
Bvelopments in Audiology, 2nd Edition. New York:
Academic Press, 1973.
i

139
Lilly, D. J. and Shepherd, D. C. A rebalance technique
for the measurement of absolute changes in acoustic
impedance due to the acoustic reflex. ASHA, 6:
381(a), 1964.
Lindsay, J. R., Kobrak, H. G., and Perlman, H. B. Relation
of the stapedius reflex to hearing sensation in man.
Archives of Otolaryngology, 23:671-678, 1936.
Loeb, M. Psychophysical correlates of intratympanic reflex
action. Psychological Bulletin, 61:140-152, 1964.
Lorente de No, R. The reflex contractions of the muscles
of the middle ear as a hearing test in experimental
animals. Transactions of the American Laryngology,
Rhinology and Otology Society, 39:26-42, 1933.
. The function of the central acoustic nuclei
examined by means of the acoustic reflexes.
Laryngoscope, 45:573-595, 1935.
and Harris, A. S. Experimental studies in
hearing. Laryngoscope, 43:315-326, 193-3.
Madsen. Madsen Model Z070 Electro-Acoustic Impedance
Bridge: Applications and instructions for use.
Copenhagen: Madsen Electronics, n.d.
Martin, F. N. The short increment sensitivity index (SISI) .
In Katz, J. (editor), Handbook of Clinical Audiology.
Baltimore: Williams and Wilkins Company, 1972.
and Wofford, M. J. Temporal summation of brief
tones in normal and cochlear-impaired ears. Journal
of Auditory Research, 10:82-86, 1970.
McRobert, H. The response of the tympanic muscles in
human ears: Possible false inferences from results
of reflex testing on normal and pathological ears.
Sound, 2:71-76, 1968.
, Bryan, M. E., and Tempest, W. The acoustic
stimulation of the middle ear muscles. Sound and
Vibration, 7:129-142, 1968.
Mehmke, S. and Tegtmeir, W. The diagnostic value of
impedance measurements. Fenestra. (An eight page
insert between pp. 12-13,) April 1970.

140
Melcher, J. A. and Peterson, J. L. The effects of age
and hearing impairment on the acoustic reflex decay.
Presented at the American Speech and Hearing Con
vention, San Francisco, 1972.
Mendelson, E. S. A sensitive method for registration of
human intratympanic muscle reflexes. Journal of
Applied Physiology, 11:499-502, 1957.
. Improved method for studying tympanic re
flexes in man. Journal of the Acoustical Society of
America, 44:146-152, 1961.
. The lability of the resting and reflex activity
of the human middle ear muscles. In Fletcher, J. L.
(editor), Middle Ear Function Seminar, U. S. Army
Research Laboratory Report No. 576, Fort Knox,
Kentucky, 1963.
. Acoustic reflexometry. Acta Oto-Laryngologica,
62:125-139, 1966.
Metz, 0. The acoustic impedance measured in normal and
pathological ears. Acta Oto-Laryngologica, Supplement
63:1-254, 1946.
Threshold of reflex contractions of muscles
of middle ear and recruitment of loudness. Archives
of Otolaryngology, 55:536-544, 1952.
Miller, G. A. The perception of short bursts of noise.
Journal of the Acoustical Society of America, 20:
160-170, 1948.
Miskolczy-Fodor, F. Monaural loudness-balance test and
determination of recruitment degree with short sound-
impulses. Acta Oto-Laryngologica, 43:573-595, 1953.
. The relation between hearing loss and recruit
ment and its practical employment in the determination
of receptive hearing loss. Acta Oto-Laryngologica,
46:409-415, 1956.
. Relation between loudness and duration of
tonal pulses. II. Response of normal ears to sounds
with noise sensation. Journal of the Acoustical
Society of America, 32:482-486, 1960.
's
Moller, A. R. Intra-aural muscle contraction in man,
examined by measuring acoustic impedance of the ear.
Laryngoscope, 68:48-62, 1958.

141
Moller, A. R. Bilateral contraction of the tympanic muscles
in man, examined by measuring acoustic impedance-change.
Annals of Otology, Rhinology and Laryngology, 70:735-
752, 1961a.
. Network model of the middle ear. Journal
of the Acoustical Society of America, 33:168-176,
1961b.
. The sensitivity of contraction of the tympanic
muscles in man. Annals of Otology, Rhinology and
Laryngology, 71:86-95, 1962a.
. Acoustic reflex in man. Journal of the
Acoustical Society of America, 34:1524-1534, 1962b.
. Effect of tympanic muscle activity on movement
of the ear drum, acoustic impedance and cochlear
microphonics. Acta Oto-Laryngologica, 58:525-534,
1964.
. An experimental study of the acoustic impedance
of the middle ear and transmission properties. Acta
Oto-Laryngologica, 60:129-149, 1965.
Munson, W. A. The growth of auditory sensation. Journal
of the Acoustical Society of America, 19:584-591, 1947.
Neergaard, E. B. and Rasmussen, G. Latency of the sta
pedius muscle reflex in man. Archives of Otolaryngology,
84:173-180, 1966.
Nerbonne, M. A. A Comparison of Brief Tone Audiometry
with Other Selected Auditory Tests of Cochlear
Function. Ph.D. dissertation. Michigan State Univer-
sity, 1970.
Niemeyer, W. Relations between the discomfort level and
the reflex threshold of the middle ear muscles.
Audiology, 10:172-176, 1971.
, and Sesterhenn, G. Calculating the hearing
threshold from the stapedius reflex for different
sound stimuli. Presented at International Audiology
Congress, Budapest, 1972.
Nixon, J. and Glorig, A. Reliability of acoustic impedance
measures of the eardrum. Journal of Auditory Re
search, 4:261-276, 1964.

142
Norris, T. W., Stelmachowicz, P., and Taylor, D. Effects
of stimulus variations on acoustic reflex patterns.
Presented at International Audiology Congress, Budapest,
1972.
Northern, J. L. Temporal summation for critical bandwidth
signals. Journal of the Acoustical Society of America,
42:456-461, 1967.
Olsen, W. 0. and Carhart, R. Integration of acoustic power
at threshold by normal listeners. Journal of the
Acoustical Society of America, 40:591-599, 1966.
, Rose, D. E., and Noffsinger, D. Brief tone
audiometry with normal, cochlear, and VIII nerve
tumor patients. A paper accepted for publication
Archives of Otolaryngology, 1973.
Owens, E. Bekesy tracing and site of lesion. Journal of
Speech and Hearing Disorders, 29:456-468, 1964.
. Audiologic evaluation in cochlear versus
retrocochlear lesions. Acta Oto-Laryngologica,
Supplement 283:1-45, 1971.
Perlman, H. B. and Case, T. J. Latent period of the crossed
stapedius reflex in man. Annals of Otology, Rhinology
and Laryngology, 48:663-675, 1939.
Peters Ltd. Peters AP 61 Acoustic Impedance Meter: Oper
ating instructions. New York: Lehr Instrument
Corporation, n.d.
Peterson, J. L. and Liden, G. Dynamics of the stapedial
muscle reflex. Presented at the 10th International
Audiology Congress, Dallas, Texas, 1970.
and Liden, G. Some static characteristics of
the stapedial muscle reflex. Audiology, 11:97-114, 1972.
Plomp, R. and Bouman, M. A. Relation between hearing
threshold and duration for tone pulses. Journal of
the Acoustical Society of America, 31:749-758, 1959.
Price, G. R. Influence of external ear acoustics on im
pulse arriving at the ear drum. Journal of the
Acoustical Society of America, 52:129 (a), 1972.
Rao, C. R. Advanced Statistical Methods in Biometric
Research. New York: John Wiley and Sons, 1952.

143
Rasmussen, G. L. The olivary peduncle and other fiber
projections of the superior olivary complex. Journal
of Comparative Neurology, 84:141-219, 1946.
Reger, S. N. Differences in loudness response of normal
and hard-of-hearing ears at intensity levels slightly
above threshold. Annals of Otology, Rhinology and
Laryngology, 45:1029-1039, 1936.
Salomon, G. and Starr, A. Electromyography of middle ear
in man during motor activities. Acta Neurologica
Scandinavia, 39:161-168, 1963.
Sanders, J. W. and Honig, E. A. Brief-tone audiometry:
Results in normal and impaired ears. Archives of
Otolaryngology, 85:640-647, 1967.
, Josey, A. F., and Kemer, F. J. Brief-tone
audiometry in patients with VUIth nerve tumor.
Journal of Speech and Hearing Research, 14:172-178,
1971.
Scharf, B. Critical bands. In Tobias, J. B. (editor),
Foundations of Modern Auditory Theory, Volume I.
New York: Academic Press, 1970.
Schoel, T. and Arnesen, G. The choice of probe-tube
position and test frequency in determining the intra-
aural reflexes. Acta Oto-Laryngologica, 54:233-238,
1962.
Sheeley, E. C. and Bilger, R. C. Temporal integration as
a function of frequency. Journal of the Acoustical
Society of America, 36:1850-1857, 1964.
Sherrington, C. S. The integrative action of the nervous
system. London: Constable, 1906.
. The integrative action of the nervous system.
New Haven, Conn.: Yale University Press, 1947.
Shiftman, F. Bridge clinic. Impedance Newsletter, 1:11,
1972. (Madsen Electronics Corp., New York.)
Simmons, F. B. Post-tetanic potentiation in the middle
ear muscle acoustic reflex. Journal of the Acoustical
Society of America, 32:1589-1591, 1960.
. An analysis of the middle-ear muscle acoustic
reflex of the cat. In Fletcher, J. L. (editor),
Middle Ear Function Seminar. U. S. Army Medical Re
search Laboratory Report No. 576, Fort Know, Kentucky,
1963.

144
Simmons, F. B. and Dixon, R. F. Clinical implications of
loudness balancing. Archives of Otolaryngology,
83:449-454, 1966.
Simon, G. R. The critical bandwidth level in recruiting
ears and its relation to temporal summation. Journal
of Auditory Research, 3:109-119, 1963.
Small, A. M., Jr., Brandt, J. F., and Cox, P. G. Loudness
as a function of signal duration. Journal of the
Acoustical Society of America, 34:513-514, 1962.
Swannie, E. M. Impedance audiometry in clinical practice.
Proceedings of the Royal Society of Medicine, 59:
971-974, 1966.
Tempest, W. and Bryan, M. E. The auditory threshold for
short-duration pulses. Journal of the Acoustical
Society of America, 49:1901-1902, 1971.
Terkildsen, K. Movements of the eardrums following intra-
aural muscle reflexes. Archives of Otolaryngology,
66:484-488, 1957.
. The intra-aural muscle reflexes in normal
persons and in workers exposed in intense industrial
noise. Acta Oto-Laryngologica, 52:384-396, 1960a.
. An evaluation of perceptive hearing losses in
children, based on recruitment determinations. Acta
Oto-Laryngologica, 51:476-484, 1960b.
. Acoustic reflexes of the human muscle tensor
tympani. Acta Oto-Laryngologica, Supplement 158:
230-238, 1960c.
. Clinical application of impedance measurements
with a fixed frequency technique. Journal of Inter
national Audiology, 3:147-155, 1964.
/ Osterhammel, P., and Scott Nielsen, S.
Impedance measurements: Probe-tone intensity and
middle-ear reflexes. Acta Oto-Laryngologica, Supplement
263:205-207, 1970.
and Scott Nielsen, S. Electroacoustic impedance
measuring bridge for clinical use. Archives of
Otolaryngology, 72:339-346, 1960.
Thomsen, K. A. The Metz recruitment test. Acta Oto-
Laryngologica, 45:544-552, 1955a.

145
Thomsen, K. A. Employment of impedance measurements in
otologic and otoneurologic diagnostics. Acta Oto-
Laryngologica, 45:159-167, 1955b.
Tillman, T. W. Special hearing tests in otoneurologic
diagnosis. Archives of Otolaryngology, 89:25-30,
1969.
, Dallos, P. J., and Kurvilla, T. Reliability
of measures obtained with the Zwislocki Acoustic
Bridge. Journal of the Acoustical Society of America,
36:582-588, 1963.
Vries, H. de. The minimum audible energy. Acta Oto-
Laryngologica, 36:230-235, 1948.
Ward, W. D. (editor). Proposed damage-risk criterion for
impulse noise (gunfire). Working Group 57 of the
NAS-NRC Committee on Hearing, Bioacoustics, and Bio
mechanics. Washington, D. C., 1968.
Watson, C. S. and Gengel, R. W. Signal duration and signal
frequency in relation to auditory sensitivity. Journal
of the Acoustical Society of America, 46:989-997, 1969.
Weiss, H. S., Mundie, M. R. Cashin, J. L., and Shinabarger,
E. W. The normal human intra-aural muscle reflex in
response to sound. Acta Oto-Laryngologica, 55:505-515,
1962.
Wever, E. G. and Lawrence, M. Physiological Acoustics.
Princeton, J. J.: Princeton University Press, 1954.
Wilber, L. A., Goodhill, V., and Hogue, A. C. Comparative
acoustic impedance measurements. Presented at the
American Speech and Hearing Convention, Chicago, 1970.
Wright, H. N. Switching transients and threshold deter
mination. Journal of Speech and Hearing Research,
1:52-60, 1958"!
. Audibility of switching transients. Journal
of the Acoustical Society of America, 32:138, 1960.
. Clinical measurement of temporal auditory
summation. Journal of Speech and Hearing Research,
11:109-127, 1968a.
The effect of sensori-neural hearing loss on
threshold duration functions. Journal of Speech and
Hearing Research, 11:842-852, 1968b.

146
Wright, H. N. Temporal summation for tones at threshold.
Presented at the 84th meeting of the Acoustical Society
of America, Miami Beach, 1972.
and Cannella, F. Differential effect of con
ductive hearing loss on the threshold-duration function
Journal of Speech and Hearing Research, 12:607-615,
1969.
Wright, J. and Btholm, B. Anomalies of the middle-ear
muscles. Journal of Laryngology and Otology, 87:
281-288, 1973.
Zwicker, E. and Wright, H. N. Temporal summation for tones
in narrow-band noise. Journal of the Acoustical
Society of America, 35:691-699, 1963.
Zwislocki, J. Some measurements of the impedance at the
eardrum. Journal of the Acoustical Society of America,
29:349-356, 1957.
. Theory of temporal auditory summation. Journal
of the Acoustical Society of America, 32:1046-1060,
1960.
. Acoustic measurement of the middle ear function
Annals of Otology, Rhinology and Laryngology, 70:
599-606, 1961.
. Analysis of the middle-ear function. I.
Input impedance. Journal of the Acoustical Society
of America, 34:1514-1523, 1962.
. An acoustic method for clinical examination
of the ear. Journal of Speech and Hearing Research,
6:303-314, 196T:

BIOGRAPHICAL SKETCH
William Lee Parker was born in 1945, earned his first
dollar in 1954, and was married in 1966 to the delightful
Cristine Mary Sisson. He received his academic degrees
of Bachelor and Master of Arts with a major in speech
pathology at California State College at Long Beach. In
1973, he received the degree Doctor of Philosophy with a
major in speech, the field of audiology, from the University
of Florida, Gainesville, Florida. At the time of his gradua
tion from California State College at Long Beach, he was
the father of Andrea, Mark, and Joshuaall of whose help
he could not have done without if he had wanted to get out
of school any slower, but whose presence made every moment
a memorable one.
147

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Kenneth C.
Associate
Pollock,
Professor
Pn:D., Chairman
of Speech
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
7
F. Owen Black, M.D.
Associate Professor of Surgery
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Assistant Professor of Speech

This dissertation was submitted to the Department of Speech
in the College of Arts and Sciences and to the Graduate
Council, and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.
August, 1973
Dean, Graduate School



4
leve]s using both pulsed and continuous puretone sweep fre
quency stimuli. A Type II Bekesy, indicated by a slight
separation between the two tracings and a narrowing of the
tracing for the continuous tone, has been the classic
designator of cochlear pathologies (Jerger, 1960) Unfor
tunately, other patterns such as Type I, IV, or mixed are
also obtained from subjects with confirmed end-organ lesions,
therefore contributing ambiguity to the test results (Owens,
1964). The self-tracking procedure of Bekesy audiometry is,
in addition, a long, fatiguing procedure which may make it
difficult for some patients to complete.
Even though each of these tests has certain drawbacks,
the results from each one can indicate the presence of a
cochlear lesion. Recruitment, as defined operationally by
loudness balancing, is the foremost test of a cochlear
lesion, but abnormal loudness growth is not necessarily
synonymous with recruitment (Hirsh, Palva, and Goodman, 1954;
Simmons and Dixon, 1966). All tests for cochlear lesions
mentioned may represent abnormal auditory processing in a
particular cochlear pathology but may delineate differing
physiological phenomena; otherwise there would be no need
for a battery of audiometric tests.
The four auditory tests discussed above may be valid
indicators of cochlear lesions, but their reliability is
dependent upon the full cooperation of a highly motivated
test subject. In many instances cooperation may not be


70
90 92 94 96 96 98 98
Channel 1:
Stimuli of
500-msec
duration
Channel 2:
Acoustic
reflex at
threshold
ART
Figure 12. Typical response of the acoustic reflex
at threshold. Channel 1 illustrates the 500-msec duration
tone burst increasing in 2-dB steps. Channel 2 demonstrates
the response at 96 dB SPL to be significantly different
from the ongoing activity.


APPENDIX C
VOLTAGE DIVIDER


APPENDIX G
RAW DATA OF THRESHOLD OF AUDIBILITY


133
Clack, T. D. Effect of signal duration on the auditory-
sensitivity of humans and monkeys (Macaca mulatta).
Journal of the Acoustical Society of America, 40:1140-
1146, 1966.
Coles, R. R. A. Can present day audiology really help in
diagnosis? An otologist's question. Journal of
Laryngology and Otology, 86:191-224, 1972.
Dallos, P. J. Dynamics of the acoustic reflex: Phenomeno
logical aspects. Journal of the Acoustical Society of
America, 36:2175-2183, 1964.
and Johnson, K. R. Influence of rise-fall time
upon short tone thresholds. Journal of the Acoustical
Society of America, 40:1160-1163, 1966.
anc^ Olsen, W. Integration of energy at threshold
with gradual rise-fall tone pips. Journal of the
Acoustical Society of America, 36:741-751, 1964.
Deutsch, L. J. The threshold of the stapedius reflex to
selected acoustic stimuli in normal human ears.
U. S. Naval Submarine Medical Center Research Report
No. 546:1-27, 1968.
. The threshold of the stapedius reflex for pure
tone and noise stimuli. Acta Oto-Laryngologica,
74:248-251, 1972.
Dix, M. R., Hallpike, C. S., and Hood, J. D. Observations
upon the loudness recruitment phenomenon with special
reference to the differential diagnosis of disorders
of the internal ear and VIII nerve. Proceedings of the
Royal Society of Medicine, 41:516-526, 1948.
Djupesland, G. Electromyography of the tympanic muscles
in man. Journal International Audiology, 4: 34-41,
1965.
/ Flottorp, G., and Winther, F. Size and duration
of acoustically elicited impedance changes in man.
Acta Oto-Laryngologica, 224:220-228, 1966.
and Zwislocki, J. J. Effect of temporal
summation on the human stapedius reflex. Acta Oto-
Laryngologica, 71:262-265, 1971.
Doyle, T. N. Auditory Temporal Summation with Variable Inter
signal Intervals in Normal and Non-normal Subjects.
Ph.D. dissertation. University of Minnesota, 1970.


for the ART measure. Small sample size, unexpected inter
subject variability, and flatter integration slopes for the
normal ears contributed to the lack of consistent statistical
significance.
Temporal integration at threshold of audibility proved
to be significant in most comparisons between normal ears
and cochlear-impaired ears. Even though temporal integra
tion at threshold of audibility evidenced a statistical
difference, the differences were not obvious enough to make
it an effective clinical tool.
There were some differences between the normal-hearing
and cochlear-impaired ears concerning temporal integration
at the threshold of the acoustic reflex. The mean integra
tion slopes of the cochlear-impaired ears were more depressed
than those of the normal ears.
These data suggest that temporal integration of the
acoustic-reflex threshold may not be used to differentiate
normal ears from cochlear-impaired ears. It may be con
strued that the time constant might provide a statistically
significant difference.
x


146
Wright, H. N. Temporal summation for tones at threshold.
Presented at the 84th meeting of the Acoustical Society
of America, Miami Beach, 1972.
and Cannella, F. Differential effect of con
ductive hearing loss on the threshold-duration function
Journal of Speech and Hearing Research, 12:607-615,
1969.
Wright, J. and Btholm, B. Anomalies of the middle-ear
muscles. Journal of Laryngology and Otology, 87:
281-288, 1973.
Zwicker, E. and Wright, H. N. Temporal summation for tones
in narrow-band noise. Journal of the Acoustical
Society of America, 35:691-699, 1963.
Zwislocki, J. Some measurements of the impedance at the
eardrum. Journal of the Acoustical Society of America,
29:349-356, 1957.
. Theory of temporal auditory summation. Journal
of the Acoustical Society of America, 32:1046-1060,
1960.
. Acoustic measurement of the middle ear function
Annals of Otology, Rhinology and Laryngology, 70:
599-606, 1961.
. Analysis of the middle-ear function. I.
Input impedance. Journal of the Acoustical Society
of America, 34:1514-1523, 1962.
. An acoustic method for clinical examination
of the ear. Journal of Speech and Hearing Research,
6:303-314, 196T:


75
crossed over and combined with the probe tone, causing a
reflex on the measuring side. This was unlikely, since a
minimum of 40-dB to 50-dB difference in hearing level is
normally necessary for crossover. The effective signal level
reaching the probe ear side would have been 40 dB to 50 dB
less than necessary to evoke the reflex. Since doubling the
signal pressure increases the SPL only 6 dB, the increase in
probe-tone pressure would have been a maximum of 6 dB and
probably only 2 dB to 4 dB. This would not have raised the
probe-tone intensity level enough to cause an acoustic reflex.
Second, the reflex-e1icitng signal that crosses over may
have registered on the pickup microphone of the electroacoustic
bridging network. This network has a filter centered at
276 Hz, and it is at least 20 dB down at 500 Hz. Thus, the
filter should reduce by at least 20 dB any energy crossover
at 500 Hz and possibly even at 1000 Hz and 2000 Hz.
In conclusion, all reasonable safeguards for subject
well-being and integrity of the acoustic reflex data were
taken into account without compromising the experimental
protocol. The subjects' responses, especially from those
exhibiting loudness recruitment, indicated that the subject
safeguards were not actually needed.


CHAPTER I
INTRODUCTION AND REVIEW OF THE LITERATURE
Auditory tests, especially those designed to determine
site of lesion, are complicated by subtle test procedures,
and patient state and sophistication. For the majority of
persons the results of audiometric procedures are of diag
nostic value, but in a few cases even a battery of audiometric
tests gives inconclusive results. The diagnostic picture may
be incomplete because of the inability or unwillingness of the
subject to respond. The literature indicates the status of
current audiological tests and further possibilities offered
by tests of acoustic-reflex and of temporal integration by
the auditory system. Furthermore, these later two auditory
measures may be effectively combined into one diagnostic test.
These conditions, therefore, motivated the present study. It
was designed to investigate temporal summation of the acoustic
reflex as a contribution to the diagnosis of cochlear lesions.
Review of the Literature
Audiological Tests Commonly Used
in Differential Diagnosis
Audiological tests are generally divided into those
most sensitive for the detection of conductive, cochlear,
eighth nerve, and central nervous system lesions. Only those
1


ACKNOWLEDGMENTS
It is nice to be able to give thanks to all of those
who helped me complete this paper. My advisory committee,
Drs. F. 0. Black, P. J. Jensen, K. C. Pollock, W. A. Yost,
never failed to extend themselves professionally or per
sonally. Dr. Pollock, my chairman, encouraged my interest
in impedance audiometry and was instrumental in my applica
tion for, and the award of, research funds from the
University of Florida. A special thanks is extended to
Dr. Yost, who took an inordinate interest in this project
and without whose help the project would not have been
completed. To Dr. W. N. Williams goes the award for best
encouragement and prodding, the mark of a true friend.
I also want to thank Drs. E. C. Hutchinson and D. T.
Hughes for their statistical advise and help. Christine
Parker and John Parks contributed their invaluable drawing
skills, and Ginny Parks helped with the mundane chore of
proofreading. And thanks to Sue Kirkpatrick for completing
the final typing.
iii


97
The time constant of temporal integration at threshold
of audibility in cochlear-impaired ears is said to be
shorter (Figure 5). Table 9 lists the signal duration
intervals within which the time constants occur.^ The re
sults are reported in both the number of times and the per
centage of times a time constant was within a certain inter
val for each group. Even though the cochlear-impaired (bad)
ears had the highest percentage occurring in the 100 msec-
20 msec interval, there was not much difference between
normal and impaired ears. These results could not support
the statement that cochlear-impaired ears have shorter time
constants than normal ears.
Table 9. Signal Duration Intervals Within Which Time
Constants of Temporal Integration at Threshold
of Audibility Occur
Signal Duration Interval in Msec
Total
400-200 200-100 100-20 20-10 Responses
Normal ears
4
(13%)
9
(30%)
17
(57%)

30
(100%)
Cochlear-good
ears
5
(36%)
5
(36%)
4
(28%)

14
(100%)
Cochlear-bad
ears
2
(13%)
3
(19%)
11
(68%)

16
(100%)
The time constant was calculated to be at or above the
first 3 dB of threshold change from 500 msec. The time con
stant criterion is the same as the ART time constant (footnote
3) .


Appendix E(continued)
500 Hz 1000 Hz 2000 Hz
Signal Duration in Msec
ject
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
5
131
125
97
95
95
129
124
104
98
96
134
132
98
98
92
6
133
123
101
99
93
135
124
104
100
96
134
130
102
100
94
7
131
125
101
95
91
135
126
106
100
100
139
132
100
102
94
8
129
125
103
99
93
137
124
108
100
98
132
128
100
98
94
9
133
121
101
97
93
133
126
104
98
98
135
130
102
98
94
10
127
121
99
99
93
133
122
106
100
98
135
130
100
100
94
bcr
No Reflex
lsl
1
Cochlear
Impaired Fre-
135
126
104
102
100
NR
133
106
100
96
2
quency
137
130
102
104
100
133
108
100
94
3
135
132
102
104
98
133
110
104
94
4
134
130
104
100
100
133
110
100
92
5
139
132
104
104
100
129
110
102
96
6
135
132
102
104
98
129
106
102
94
7
137
130
104
104
102
139
110
100
94
8
135
126
102
106
100
137
108
100
94
9
137
130
102
104
100
135
110
102
94
10
137
132
100
102
100
139
110
100
94
Cochlear
-Bad
Ears
ecl
1
127
119
111
111
109
116
104
106
96
94
125
110
98
96
96
2
127
117
111
109
109
120
104
98
96
92
123
110
100
100
96
3
125
119
113
111
109
120
106
98
96
90
129
108
100
98
94
4
127
119
113
109
109
120
106
100
96
94
125
108
100
102
96
5
127
119
113
111
107
118
106
100
96
94
127
110
100
98
96
121


2
Table 3. Acoustic-Reflex Threshold in dB SPL (re 0.0002 dyne/cm )
Sample
Number
Frequency
in Hz
250
500
1000
1500
2000
3000
4000
Harford
and Liden, 1967-1968
?
117
106
95

93
92
96
Jerger
et al., 1972
382 Ss

100
96

97
--
95
Lamb et
al. 1968
19 Ss
109
100
93

96
96
93
Moller,
1961a
2-5 Ss

96

93
--
91

Moller,
1962a
1-3 Ss
iooa
95b
95C
f*300 Hz.
525 Hz.
1200 Hz.
00


ACOUSTIC-REFLEX THRESHOLD
B RE THRESHOLD AT 500 MSEC
SIGNAL DURATION IN MSEC
Figure 16. Acoustic-reflex thresholds of normal (good) and impaired (bad) ears of the
Meniere's group as a function of temporal summation.


41
number of pulses presented at each discrete level decreased
from eight pulses to one pulse. There was no significant
difference in the mean threshold. Wright (1968a) suggests
using an attenuation rate of 2.5 dB/second at a repetition
rate of 1/second as a standard.
The choice of ascending, descending, or mean thresholds
can give threshold values differing as much as 3 dB at any
one signal duration (Hempstock et^ a_l. 1964). Therefore,
for the majority of the cases, where applicable, mean values
determine threshold. The vast majority of recent publications
dealing with temporal integration have used the Bekesy track
ing procedure, where the averaged midpoint of the crossings
is considered to be the threshold value.
Effects of end-organ lesions
upon temporal integration
Considerable interest has been generated in the use of
temporal summation as a diagnostic tool for end-organ patholo
gies; this is otherwise known as brief-tone audiometry. A
cochlear lesion seemingly increases the integration ability
of the ear. As the duration is decreased,less acoustic power
is needed to maintain constant energy for threshold (Brahe
Pederden and Elberling, 1972, 1973; Doyle, 1970; Elliot,
1963; Gengel and Watson, 1971; Gengel, 1972; Harris et al.,
1958; Hattler and Northern, 1970; Jerger, 1955; Martin and
Wofford, 1970; Miskolczy-Fodor, 1953, 1956, 1960; Nerbonne,
1970; Olsen, Rose, and Noffsinger, 1973; Sanders and Honig,


22
Bryan, and Tempest, 1968; Moller, 1962b; Simmons, 1963;
Weiss, Mundie, Cashin, and Shinabarger, 1962) A 500-msec
duration tone approximates the results of longer stimuli for
eliciting a steady change in impedance at the tympanic
membrane. With regard to the auditory response to long
duration signals, Dallos (1964) and Harford and Liden (1967-
1968) have observed that adaptation takes place only over
longer periods of time, e.g., 15 seconds to 30 seconds. This
adaptation is thought to be an afferent process, because
recovery is instantaneous upon changing to a new test fre
quency. This afferent adaptation is also frequency-dependent,
with higher frequencies demonstrating more response decay
(Anderson et alL. 1969; Melcher and Peterson, 1972).
The ear in which the reflex is recorded can also influ
ence the level of the acousticreflex threshold. Recordings
made in the stimulated ear, instead of the contralateral ear,
improve threshold sensitivity (Moller, 1961b). Bilateral
stimulation improves the threshold value even more (Moller,
1962b).
The frequency of the recording probe tone can in itself
influence the acoustic-reflex values. Peterson and Liden
(1972), averaging the ART for 500 Hz and 4000 Hz, state that
a 220-Hz probe tone is about 6 dB more sensitive than a 625-
Hz probe tone in eliciting the ART, while a 800-Hz tone is
between them. In an earlier study, Harford and Liden (1967-
1968) recorded an unsystematic 2-dB threshold difference



PAGE 1

TEMPORAL SUMMATION OF THE ACOUSTIC-REFLEX THRESHOLD A POSSIBLE INDICATOR OF COCHLEAR ABNORMALITIES By William Lee Parker A DISSERATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1973

PAGE 2

To my wife, Christine

PAGE 3

ACKNOWLEDGMENTS It is nice to be able to give thanks to all of those who helped me complete this paper. My advisory committee, Drs. F. 0. Black, P. J. Jensen, K. C. Pollock, W. A. Yost, never failed to extend themselves professionally or personally. Dr. Pollock, my chairman, encouraged my interest in impedance audiometry and was instrumental in my application for, and the award of, research funds from the University of Florida. A special thanks is extended to Dr. Yost, who took an inordinate interest in this project and without whose help the project would not have been completed. To Dr. W. N. Williams goes the award for best encouragement and prodding, the mark of a true friend. I also want to thank Drs. E. C. Hutchinson and D. T. Hughes for their statistical advise and help. Christine Parker and John Parks contributed their invaluable drawing skills, and Ginny Parks helped with the mundane chore of proofreading. And thanks to Sue Kirkpatrick for completing the final typing. iii

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGMENTS iii LIST OF TABLES vi LIST OF FIGURES vii ABSTRACT ix CHAPTER I. INTRODUCTION AND REVIEW OF THE LITERATURE 1 Review of the Literature 1 Audiological Tests Commonly Used in Differential Diagnosis 1 Acoustic Reflex 5 Threshold of the Acoustic Reflex (ART) 14 Temporal Summation 32 II. STATEMENT OF THE PROBLEM 45 Hypothesis 52 III. METHODS AND PROCEDURES 55 Stimuli 57 Experimental Equipment 62 Procedure 67 Temporal Integration at Threshold of Audibility • 67 Temporal Integration at the Threshold of the Acoustic Reflex 68 Experimental Safeguards 71 IV. RESULTS AND DISCUSSION 76 1) Temporal Integration at the AcousticReflex Threshold 76 2) Temporal Summation at Threshold of Audibility 91 3) Temporal Integration at the AcousticReflex Threshold Compared to Temporal Integration at Threshold of Audibility 98 iv

PAGE 5

CHAPTER Page V. CONCLUSIONS AND SUMMARY 100 APPENDICES 104 A. MEDICAL HISTORY FORM 105 B. PURE TONE HEARING LEVEL (ANSI-1969) OF ALL SUBJECTS 107 C. VOLTAGE DIVIDER 109 D. CALIBRATION DATA TABLES Ill E. RAW DATA OF THE ACOUSTIC-REFLEX THRESHOLD ... 115 F. DISCRIMINANT ANALYSIS 125 G. RAW DATA OF THRESHOLD OF AUDIBILITY 127 H. CORRELATION COEFFICIENTS OF THE TEMPORAL INTEGRATION MEASURES 129 REFERENCES ,,,,,,,,, 131 BIOGRAPHICAL SKETCH 147 v

PAGE 6

LIST OF TABLES Table Page 1. Acoustic-Reflex Threshold in dB SL 16 2. Acoustic-Reflex Threshold in dB HL (ANSI-1969) 17 3. Acoustic-Reflex Threshold in dB SPL (re 0.0002 dyne/cm^) 18 4. Test Stimuli Used to Elicit Threshold of Aduibility and Threshold of the Acoustic Reflex 58 5. Conceivable Sound Pressure Level Necessary to Elicit the Acoustic-Reflex Threshold by a 500-Hz Burst at a Signal Duration of 10 Msec 73 6. Discriminant Analysis of the Acoustic-Reflex Threshold 79 7. Average Slope Change Per Decade of Time of the Acoustic-Reflex Threshold 79 8. Discriminant Analysis of the Threshold of Audibility Data 93 9. Signal Duration Intervals Within Which Time Constants of Temporal Integration at Threshold of Audibility Occur 97 vi

PAGE 7

LIST OF FIGURES Figure Page 1. Possible time, strength and direction of the stapedius and tensor tympani muscles, and resultant displancement of the tympanic membrane 8 2. The acoustic-reflex arc 10 3. Preand post-noise exposure thresholds ... 30 4. Average temporal integration slopes as a function of frequency 36 5. Examples of temporal integration 43 6. Temporal summation of the acoustic reflex 48 7. Temporal integration at threshold of audibility and of the acoustic reflex for normalhearing subjects 50 8. Acoustic-reflex contraction 60 9. Spectral content of tone bursts with 5-msec rise/decay ramp 61 10. A 20-msec signal burst with a smooth 5-msec rise and decay 63 11. Block diagram of equipment used to elicit and record the intra-aural reflex and to allow subject control over the test situation 64 12. Typical response of the acoustic reflex at threshold 70 13. Upper limits of acceptable exposure to impulse noise for 95 percent of the population to 10,000 impulses 72 14. Mean acoustic-reflex thresholds as a function of temporal integration 78 vii

PAGE 8

Figure Page 15. Acoustic-reflex thresholds of normal and cochlearimpaired (bad) ears as a function of temporal summation 81 16. Acoustic-reflex thresholds of normal (good) and impaired (bad) ears of the Meniere's group as a function of temporal summation 82 17. Inter-subject range of acoustic-reflex thresholds as a function of temporal integration 83 18. A comparison of mean acoustic-reflex thresholds as a function of temporal summation in normal-hearing ears 85 19. Signal -duration interval in which the time constant of temporal integration for each test subject occurred across all frequencies 87 20. The delayed time constant in some cochlearimpaired ears contrasted with the normal time constant of temporal integration at the acoustic-reflex threshold 89 21. Thresholds, audibility, and acoustic reflex of the test subjects at 500-msec signal duration 92 22. Mean thresholds of audibility as a function of temporal integration 94 23. Inter-subject range of thresholds of audibility as a function of temporal integration .... 96 viii

PAGE 9

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TEMPORAL SUMMATION OF THE ACOUSTIC-REFLEX THRESHOLD: A POSSIBLE INDICATOR OF COCHLEAR ABNORMALITIES By William Lee Parker August, 1973 Chairman: Kenneth C. Pollock, Ph.D. Major Department: Speech It has been suggested that temporal summation of the ART might be a clinical tool distinguishing normal from cochlear-impaired ears. If the ART measure could yield differences at least as distinctive as those of temporal summation at threshold of audibility, the ART could be used with patients for whom threshold measures are not appropriate In this study, temporal summation of the acousticreflex threshold (ART) as well as of the threshold of audibility, was measured in five normal-hearing and five unilaterally hearing-impaired subjects by varying signal duration of 500 Hz, 1000 Hz, and 2000 Hz tones. The signal durations employed were 500 msec, 200 msec, 100 msec, 20 msec, and 10 msec. There were statistically significant differences between normal ears and cochlear-impaired ears at 2 out of 12 comparisons. In addition, there were some descriptive differences between normal ears and cochlear-impaired ears ix

PAGE 10

for the ART measure. Small sample size, unexpected intersubject variability, and flatter integration slopes for the normal ears contributed to the lack of consistent statistical significance Temporal integration at threshold of audibility proved to be significant in most comparisons between normal ears and cochlear-impaired ears. Even though temporal integration at threshold of audibility evidenced a statistical difference, the differences were not obvious enough to make it an effective clinical tool. There were some differences between the normal-hearing and cochlear-impaired ears concerning temporal integration at the threshold of the acoustic reflex. The mean integration slopes of the cochlear-impaired ears were more depressed than those of the normal ears. These data suggest that temporal integration of the acoustic-reflex threshold may not be used to differentiate normal ears from cochlear-impaired ears. It may be construed that the time constant might provide a statistically significant difference. x

PAGE 11

CHAPTER I INTRODUCTION AND REVIEW OF THE LITERATURE Auditory tests, especially those designed to determine site of lesion, are complicated by subtle test procedures, and patient state and sophistication. For the majority of persons the results of audiometric procedures are of diagnostic value, but in a few cases even a battery of audiometric tests gives inconclusive results. The diagnostic picture may be incomplete because of the inability or unwillingness of the subject to respond. The literature indicates the status of current audiological tests and further possibilities offered by tests of acoustic-reflex and of temporal integration by the auditory system. Furthermore, these later two auditory measures may be effectively combined into one diagnostic test. These conditions, therefore, motivated the present study. It was designed to investigate temporal summation of the acoustic reflex as a contribution to the diagnosis of cochlear lesions. Review of the Literature Audiological Tests Commonly Used in Differential Diagnosis Audiological tests are generally divided into those most sensitive for the detection of conductive, cochlear, eighth nerve, and central nervous system lesions. Only those 1

PAGE 12

2 test procedures normally employed in the assessment of endorgan status, in particular those tests considered most valuable in differential diagnosis, will be reviewed in this section: Alternate Binaural Loudness Balance, Monaural Loudness Balance, Short Increment Sensitivity Index, and Bekesy audiometry.^" The Alternate Binaural Loudness Balance (ABLB) test (Fowler, 1936) has long been the indicator of loudness recruit 2 ment. This abnormally rapid growth in loudness is theoretically regarded as pathognomonic of cochlear lesion (Dix, Hallpike and Hood, 1948; Fowler, 1936, 1950; Hood, 1969; Jerger, 1967, among others) In spite of its high validity in confirming recruitment (Tillman, 1969) this test cannot be administered to all patients. The patient must have one normal, non-recruiting ear at each test frequency and at least a 20 dB threshold difference between ears to obtain meaningful data. Further, it is estimated that more bilateral hearing losses than unilateral exist (Carver, 1972) Thus, It is generally understood that audiometric evaluations do not measure anatomical lesions, per se. Instead, they reflect functional alterations in the auditory response caused by anatomical lesions or physiological changes. 2 Loudness recruitment (Fowler, 1937) is an auditory phenomenon where the supraliminal loudness growth of one ear occurs at a faster rate than that of the contralateral ear to equal intensity units. When the threshold differences between the two ears is more than 20 dB and the suprathreshold balances are considered equal within 10 dB the loudness growth is considered abnormal. This abnormal loudness growth is called loudness recruitment. (Fowler, 1936; Hood, 1969; Jerger, 1967.)

PAGE 13

3 this standard detector of recruitment cannot be administered to a majority of hearing-impaired persons to gain diagnostic information. The Monaural Bi-frequency Loudness Balance (MBLB) test (Reger, 1936) compares the loudness above threshold at one frequency with the loudness of another frequency. Again, the threshold at one of the test frequencies must be normal with the other elevated. In this way, recruitment can be tested unilaterally, thereby allowing a larger segment of the population to be tested diagnostically for cochlear lesions. It is perceptually much more difficult for the patient to perform than the ABLB and is, therefore, less reliable. This is especially true as the difference between the standard and comparison frequency increases (Graham, 1967) The Short Increment Sensitivity Index (SISI) test (Jerger, Shedd, and Harford, 1959) indicates the ability of the cochlear mechanism to respond to small changes in signal amplitude. It is an indicator of cochlear disorders if a high percentage of 1 dB increments is detected by the subject, but it apparently is not as reliable a detector of cochlear impairment as the ABLB measure (Owens, 1971; Tillman, 1969) Its lower reliability may be related to the patient sophistication necessary to complete the procedure, which is demonstrated by the difficulty in orienting some patients to the auditory listening task. The four Bekesy Types, introduced by Jerger (1960), were based upon the difference in self-tracked threshold

PAGE 14

4 leveHs using both pulsed and continuous pure-tone sweep frequency stimuli. A Type II Bekesy, indicated by a slight separation between the two tracings and a narrowing of the tracing for the continuous tone, has been the classic designator of cochlear pathologies (Jerger, 1960) Unfortunately, other patterns such as Type I, IV, or mixed are also obtained from subjects with confirmed end-organ lesions, therefore contributing ambiguity to the test results (Owens, 1964). The self-tracking procedure of Bekesy audiometry is, in addition, a long, fatiguing procedure which may make it difficult for some patients to complete. Even though each of these tests has certain drawbacks, the results from each one can indicate the presence of a cochlear lesion. Recruitment, as defined operationally by loudness balancing, is the foremost test of a cochlear lesion, but abnormal loudness growth is not necessarily synonymous with recruitment (Hirsh, Palva, and Goodman, 1954; Simmons and Dixon, 1966) All tests for cochlear lesions mentioned may represent abnormal auditory processing in a particular cochlear pathology but may delineate differing physiological phenomena; otherwise there would be no need for a battery of audiometric tests. The four auditory tests discussed above may be valid indicators of cochlear lesions, but their reliability is dependent upon the full cooperation of a highly motivated test subject. In many instances cooperation may not be

PAGE 15

5 obtained because: (1) the patient cannot or does not understand his responsibility in the test situation, i.e., how to recognize or when to respond to the auditory stimuli; (2) the motivation of the patient is minimal; or (3) the patient, because of physical state, is incapable of an appropriate response (Martin, 1972) Any one of these factors can reduce the reliability of the test procedure without conscious intent on the part of the patient, as frequently occurs with such patients as psychotics, mental retardates, cerebral palsied persons, preverbal infants, or the comatose or severely ill individual. There have been numerous reports on the use of the acoustic reflex an an indicator of abnormal auditory processing due to cochlear impairment. It is an involuntary physiological reaction reflected in the contraction of the intra-aural muscles to relatively intense acoustic stimuli. In addition, the auditory system's ability to integrate acoustic power over an interval of time has been investigated as a measure of cochlear function and integrity. A combination of these two auditory functions may serve to quantify, objectively, auditory assessment. Acoustic Reflex Anatomy and physiology of the acoustic-reflex arc The acoustic reflex is a consensual contraction of the middle-ear muscles producing, in persons with normal auditory

PAGE 16

6 pathways, a bilateral change in acoustic impedance in both ears to acoustic stimuli. Upon simultaneous contraction, the two intra-aural muscles, the stapedius and tensor tympani, act in a physiologically antagonistic manner but produce a synergistic impedance against sound energy. The stapedius muscle is the smallest muscle in the human body. It originates from the pyramidal eminence on the posterior wall of the tympanic cavity and its tendon inserts on the head, neck, or posterior crus of the stapes. The tensor tympani muscle arises from the bony semicanal above the Eustachian tube. Its tendon traverses the tympanum to insert on the mallar manubrium (Jepsen, 1963; Kobrak, 1959). The direction of pull of these two muscles is at right angles to the axis of their corresponding ossicles, making the functional action of the two muscles almost in direct opposition to one another (Wever and Lawrence, 1954) If the movement caused by each of the muscles is considered independently of the other, contraction of the stapedius muscle causes the stapes to be pulled posteriorly It is not within the scope of this paper to differentiate between the types of impedance present within the middle ear, nor between the various contributing factors of increased impedances (see Hung and Dallos, 1972; Lilly, 1972, 1973; Zwislocki, 1961, 1962). It is important, however, to know that the contraction of one or both of the intraaural muscles will produce an opposition to energy flow through the middle-ear cavity, thereby reducing the acoustic energy reaching the sensory end-organ.

PAGE 17

7 and outward from the oval window, as well as causing the tympanic membrane to be pushed in a slightly outward direction. The malleus and tympanic membrane swing medially upon tensor tympani activation (Jepsen, 1963; Kobrak, 1959). This is not the case, however, upon the contraction of both muscles simultaneously. The movement of the tympanic membrane depends upon the relative strength, latency, and contraction time of each muscle. Figure 1 demonstrates a possible resultant of a stronger and faster-acting stapedius muscle contraction partially opposed by the later and weaker tensor tympani muscle contraction. It is important to remember that it does not matter acoustically which muscle predominates because in either case, the acoustic impedance will be increased. The acoustic-reflex arc consists of an afferent neuron, a reflex center, and an efferent neuron (Lorente de No, 1933, 1935; Rasmussen, 1946) The afferent portion is the same afferent pathway for audition starting from the sensory endorgan but terminating at the level of the superior olivary complex (SOC) The SOC consists of at least five cellular groups, but the accessory, or medial, nucleus is felt to be the central mediator of both reflex arcs, the stapedial and the tensor tympani. The accessory superior olive gives off a few fibers to the ipsilateral motor nucleus of the facial nerve (n. VII) (Rasmussen, 1946) which constitutes the central portion of the stapedial-ref lex arc. The facial nerve innervates the stapedial muscle to complete the efferent portion of this reflex arc.

PAGE 18

8 TIME Figure 1. Possible time, strength and direction of the stapedius and tensor tympani. muscles, and resultant displacement of the tympanic membrane. (Redrawn from Mendelson, 19 63.)

PAGE 19

9 It is also surmised that fibers from the accessory nucleus pass to and from the lateral lemniscus into the ipsilateral motor nucleus of the trigeminal nerve (n. V) (Rasmussen, 1946) as the central portion of the tensor tympani-reflex arc. The efferent portion of this reflex arc constitutes the tensor tympani muscle innervated by the mandibular branch of the trigeminal nerve. The general schema for the acoustic-reflex arc as proposed by Rasmussen (1946) is shown in Figure 2. Simmons (1963) does not necessarily take exception to Rasmussen, but suggests that there are several possible reflex loops, both ipsilateral and crossed. Simmons speculates that the major differences in latency and response level between the stapedius and tensor tympani muscles are due to the relative degree of synaptic connections of their respective reflex arcs. According to Simmons, the stapedial reflex probably has the more compact, less diffuse, interneuron connections, thereby explaining its greater sensitivity. Each of the proposed schema would explain the lower reflex threshold and shorter latency of the stapedial acousticreflex arc activity in comparison to the tensor tympani acoustic-reflex responses (as seen in Figure 1) In either case, the SOC is the most probable reflex center. Neural activity in the SOC increases with loudness to at least 60 dB sound pressure level (Boudreau, 1965) The acoustic reflex probably occurs when the neural activity in this

PAGE 20

10 H cd X 5-i • O "H CD -P O U W C a) P co e 3 CD U CO 01 CO H 3 E to cd m — i a to -p CO •H rH CO Cd C -H O t5 U CD 0 id us o 5h -P 5H (d CO CD P CD X CD cd CO CD 45 H rd HP m MH CD CD 'd 45 M C -P 1 cd • H 45 CD (d co tnr-i 3 O M •rH 3 CD CD & U 45 fd 3 -P •P C CO rH O h m U £5 5h -H cd -a c •rH P CO -rH 3 CD O rH O U CD EH X CD CD >ii4-l CO CN CD fa CD 3 5H CD I. H O a, CD e H d 45 >i cd tn-P -P £5 M C O cd co cr> CT> C5 -H SH CD 5-1 O -P -P H E 0

PAGE 21

11 synaptic complex exceeds some critical level of neural excitation (Lorente de No, 1933, 1935). The stapedius muscle is generally considered to be the most active and dominant muscle in response to acoustic stimulation in humans, whereas the tensor tympani reacts to acoustic stimulation less often 4 (Flottorp and Djupesland, 1970; Hoist, Ingelstedt, and Ortegren, 1963; Jepsen, 1963; Liden, Peterson, and Harford, 197C; McRobert, 1968; Mendelson 1961, 1966; Terkildsen, 1957, 1960a, c; Weiss, Mundie, Cashin, and Shinabarger, 1962; and others). It appears that 1 percent to 5 percent of the population with normal hearing and with no apparent middle-ear pathology have no demonstrable acoustic-reflex response (Shiffman, 1972; Swannie, 1966; Terkildsen, 1960c; Weiss et a_l 1962; Wright and Btholm, 1973) Methods of detecting the intra tympanic muscle activity in man Some investigators of the intratympanic muscle activity have attempted to make direct observations of muscle contractions through surgical exposure or via chronic tympanic membrane perforations. For the most part, however, only the stapedial tendon is directly available for easy visualization 4 There has been considerable discussion of whether the tensor tympani muscle is active in man to acoustic or nonacoustic stimulation. This is a clinical question of consider able importance. McRobert (1968), in a critical review of the literature, and Liden, Peterson and Harford (1970) agree that tensor activity is present in man but that the stapedius predominates

PAGE 22

12 (Kobrak, 1948; Lindsay, Kobrak, and Perlman, 1936; Lorente de No, 1933; Lorente de No and Harris, 19 33; Perlman and Case, 1939) Since visual observation limits the accuracy of data quantification, other approaches have been developed for studying the reflex activity. Electromyography (EMG) measures the individual muscle fiber action potentials and is, therefore, the most direct method. EMG activity from both muscles in humans has been reported in response to both acoustic and non-acoustic stimulation (Djupesland, 1965; Fisch and Schulthess, 1963; Salomon and Starr, 1963). This technique is not clinically feasible. Extratympanic manometry has been the method employed to determine direction of tympanic membrane movement. This technique is made possible by sealing the external auditory meatus with a probe containing a pressure-sensing device. If the external meatus is properly sealed, the extratympanic air pressure should decrease or increase with respective inward or outward movement of the tympanic membrane (Flottorp and Djupesland, 1970; Hoist et al. 1963; Liden et al. 1970; Mendelson, 1961, 1963; Terkildsen, 1957, 1960a, c; Weiss et al. 1962). McRobert (1968), in a review of this measuring technique, noted that the theory for the tympanic membrane response to individual muscle contraction is valid, but that there are many inexplicable results (viz. Mendelson, 1957) suggesting the need for better instrumentation. Moller (1964) and Neergaard and Rasmussen (1966) also warn that small

PAGE 23

13 contractions of the intra-aural muscles may not produce a measurable change in extratympanic air pressure. Instead, these muscle contractions are reflected in a measurable increase in the impedance to acoustic energy within the middle ear The technique used most extensively in the past decade to indicate intra-aural muscle contraction is acousticimpedance measurement. Simply stated, a probe tone is directed perpendicularly towards the plane of the tympanic membrane. While most of this acoustic energy is transmitted through the tympanic membrane and attached ossicular chain to the oval window, a portion of the acoustical energy wave is reflected back into the external canal from the tympanic membrane. Since the probe tone input and the reflected tone pick-up microphone are connected to the external auditory meatus with an airtight seal, the reflected probe tone can be monitored accurately. As the intra-aural muscles contract, the acoustic impedance increases, causing an increased amount of the probe tone to be reflected back into the external canal. It is the increased amplitude and phase change of the reflected tone which indicate that a change in intra-aural muscle • 5 activity has occurred. This type of measurement has passed ^Unless specifically stated otherwise, all acousticreflex data concern the ear in which the reflex is elicited. The ear at which the reflex is elicited may not necessarily be the same ear in which the reflex contraction is recorded; most often the reflex ear and the measurement ear are contralateral to each other when using the acoustic impedance

PAGE 24

14 from an experimental method (Metz, 194 6; Zwislocki, 1957) into the clinical armamentarium with the introduction of several commercially available acoustic-impedance meters (GrasonStadler, 1973; Madsen, n.d.; Peters, n.d.; Terkildsen and Scott Nielsen, 1960; Zwislocki, 1963). With the increasing interest in the United States concerning this technique as a diagnostic tool, critical evaluations of methods and the commercially available equipment have been made. These instruments have been found to be reliable (Nixon and Glorig, 1964; Tillman, Dallos, and Kurvilla, 1963), sensitive (Moller, 1964) and also relatively easy to utilize (Brooks, 1971) Threshold of the Acoustic Reflex (ART ) The acoustic reflex first occurs at a predictable level above the normal threshold of audibility. The predictability of this reflex accounts for its clinical usefulness; therefore, it is forthwith described in detail along with factors which can affect its performance and measurement. For pure tones from 250 Hz to 4000 Hz the range of the acoustic-reflex threshold (ART) is 70 dB to 90dB above the normal-hearing individual's threshold (sensation level: SL) with a mean of approximately 80 dB SL (Deutsch, 1968, 1972; Jepsen, 1951) Jerger, Jerger, and Mauldin (1972) report the range measurement technique. If the reflex threshold concerns the left ear, the reflex eliciting tone will be delivered to the left ear. Because both sides will contract to unilateral stimulation, the reflex action in the above example will be recorded in the right ear. This eliminates acoustic interference of the eliciting tone with the probe tone, which may cause misleading results.

PAGE 25

15 of values for the ART in 382 normalhearing persons as being normally distributed with a mean of 85 dB HL (ANSI-1969) Ninety-five percent of their population fell within 7 0 dB to 100 dB HL and 99 percent within 65 dB to 105 dB HL. This range of thresholds has been supported by others (Anderson and Wedenberg, 1968; Deutsch, 1968, 1972; Harford and Liden, 1967-1968; Peterson and Liden, 1972) although 95 dB to 100 dB HL is considered the upper limit for normals because auditory lesions outside the cochlea tend to raise the ART, e.g., conductive problems or eighth nerve lesions (Brooks, 1971; Anderson, Barr and Wedenberg, 1970). As shown in Table 1, the ART in SL is rather uniform from study to study, except for the technique using manometry. Weiss, et a_l. (1962) show ART levels which are 10 dB to 20 dB less sensitive than those measured by acoustic impedance. Inspection of Tables 1 and 2 does not indicate any systematic effect of frequency in sensation level or in hearing level in normal subjects. In Table 3 more sound 2 pressure level (SPL re 0.0002 dyne/cm ) is necessary to elicit the ART at lower frequencies. In this respect, the acoustic reflex is similar to the threshold of audibility in response to SPL. The range of ART levels may be due to various types of acoustic -impedance instruments (e.g., Burke, Herer, and McPherson, 1970, as shown in Table 2) threshold criteria (Jerger et al., 1972 vs. Moller, 1961a, 1962a), a variety

PAGE 26

16 o o rH O m 00 1 o 00 00 r* 1 o o i o i 1 1 1 i 0^ ro o o 1 CO rCN 1 o 1 t> r00 1 CN N o IE CD i i i i I 1 i i i 1 c rH >1 O u IN if} CO CN 1 o 00 1 00 1 /It 1 T 1 D 1 CI) 1 I i 1 1 CO 1 1 l | 1 CT> 00 >H o [ rH i ri i 1 U) o CO L CO 1 1 in _^ PA 'O c o 1 1 1 1 1 1 CO •H o 1 1 1 i I 1 "3rH 0 Ah O n U 1 1 I CTi i CO in r CO CO 1 i — i At CD CN 5-1 v2 rn U J Efl r i tH CU H (j M M U) rH CU CD IU rH IU \j j 01 CO KA PS ft^ vl) u J rn CU 3 g Cn rH (0 3 o CO u Q IH CO 53 ro CO CO 1 1 (1) K 1 CN u CN •H CO •P rH VD CTi CO a~i rH rH u o vo H ro H (Ti in rH rH 0^ cr^ f3 H rH •
PAGE 27

17 O O CM in r>. CO 1 o m CO CO CO CO co 1 o H | o 1 1 CN 1 >x> 1 r— | o 1 CO 1 00 CO CO N id o o 00 1 1 CO ( — CO 1 c o 00 1 CO CO CO CO 1 H CM >i U o c o 1 1 1 in 1 1 1 1 I CO iH tr G n o o VD lO ro CO V£> L£> 1 o 00 C\ CTi CO CO CO CO ) rH o CO M LP, LP, cn CO ro LP) o CO cr, CO CO CO CO in o m 1 1 CM I ro 1 1 in CO 1 1 I CO 1 1 CM in CO CO CO CO 0 in W (H 0 M m rH Q) 0 CO CO CO CO e e o o H rH W 2 CM CM CM LP) CO in CO in CO CO to CO to to M L0 CO to CO CO ft! CO CO Q) CM in co CO CTi o 1 1 00 m H CO CM rH 00 CO CO en CM H rH 1 rcn CM r> rH I*tT io a> H eft rH C XI CM 0 c o o to c OJ rrc m CO M a o> (1) rH c Id ml rd 1 c Hi C m a id CCt 1 id rH rH c +*l 01 c o VI 01 **\ M 0 CO u M 01 QJ M M CO n 0) o 0 x! 0 0 M a) M CM tn o rH rH
PAGE 28

o o o >* o o o N X o o c o •H CN >1 o O c O 0) in 3 rH C7* 0) M o Cn o o H o o LT) o in CN 0) u i— I rH 1 r~en rH CN c cn CO 0) rH \£> T3 •H rH rd 41 rH CN a G cn cn id ail •31 rH rH SI M u U o
PAGE 29

19 of stimulus parameters (durations, inter stimulus intervals, intensity increments, and spectral components) and age of the sample population (Jerger et a_l 1972). The standard deviation of the acoustic-reflex thresholds in response to pure tones has been reported to have a range of 6.4 dB (Jerger et al 1972) to 9 dB (Deutsch, 1968). Harford and Liden (1967-1968) list high Spearman rank-order correlations for test-retest reliabilities for 250 Hz, 1000 Hz, and 2000 Hz and poorer results for 500 Hz, 3000 Hz, and 4000 Hz. Moller (1962a) kept the acoustic reflex repeatability within 1.2 dB, but used a 10 percent change of maximum reflex contraction. The reported ART variability of 9 dB may be due, in part, to the unknown physiological processes causing the arefiex at 250 Hz, 4000 Hz, and 6000 Hz to be absent or inconsistently elicited, even in the presence of normal occurxing. response to pure tones in the middle frequencies ZjDeutsch,, 1972; Fulton and Lamb, 1972; Jerger et al 1972). Xte has; beenreported. that complex stimuli of narrow band Offc whiter noiseelicitthe: ART: atlowers intensity levels than ptare-r-tone sinusoids." ; Along: with_ the_ increased sensitivity tiaerBiis also" improvement, of threshold stability (Dallos, 3.964; -_ Deutsch,: 1968, 1.972; Djupesland : Flottorp and Winther i#66jc Lillys _19 64 j Mc Roberta Bryan, and Tempest, 1968; Moller, i96-2b-; Peterson and Liden,: 1970, 1972)%: : : 1?:4 There is general agreement that. the acoustic reflex is significantly influenced by the energy level outside the

PAGE 30

20 critical bandwidth. The ART is more or less constant as the bandwidth is increased to a critical value. Beyond this critical bandwidth the acoustic-reflex threshold will decrease as the bandwidth increases (Flottorp, Djupesland, and Winther, 1971; McRobert et al. 1968). Moller (1962b), in contrast, did not see a consistent change in all subjects as bandwidth was increased, but it is possible that he did not extend his bandwidths far enough. Moller did obtain improvement in threshold for one subject for which the bandwidth surpassed the critical width determined by Flottorp et al (1971) at 525 Hz f c The work by McRobert et al. (1968) established that the lowering of ART with increasing bandwidth is dependent upon the center frequency, and is more pronounced at 1000 Hz f Apparently, the critical bandwidth with which the reflex mechanism responds to auditory stimuli widens with increases in sound pressure level (Hung and Dallos, 1972) The acoustic— reflex threshold does not seem to differ regardless of ascending or descending stimulus presentation approach (Beedle, 1970; Harford and Liden, 1967-1968; Peterson and Liden, 1970, 1972), although Deutsch (1968, 1972) reports a systematic improvement of threshold over three test trials. Beutsch attributes the ART improvement of approximately 2 dB from trial to trial to "auditory sensitization" (Hughes, 1954, 1957) Simmons (1960) labels this response condition "post-tetanic potentiation" and attributes this sensitivity

PAGE 31

21 to hyperexcitability of the brainstem structures. These results might also be ascribed to conditioned "learning," but Bates, Loeb, Smith, and Fletcher (1970) were unable to condition the reflex. Threshold of the acoustic reflex is normally defined as the lowest stimulus level at which the reflex can be elicited (Anderson et al. ; 1969, 1970; Beedle, 1970; Deutsch, 1968; Djupesland and Zwislocki, 1971; Peterson and Liden, 1972; and others) This minimal change in muscle contraction is unacceptable to Moller (1962a) who uses 10 percent change of the maximum impedance change. His criterion keeps the ART reliability within 1.2 dB while reliability deteriorates as the minimal detectable reflex is approached. In addition, the ART has been elicited in successive steps of 1 dB (Djupesland and Zwislocki, 1971), 2 dB (Lamb et al 1968) and 5 dB (Jerger et al 1972) which might account for some of the reported threshold level differences. Stimulus duration, another variable affecting the acoustic-reflex threshold, is reported by Lorente de No (1935) as having an effect on the tensor tympani response in rabbits. The strength of the muscle contraction increases as a function of increasing stimulus duration with the stimulus level held constant. Further studies have shown that the acoustic— reflex threshold in SPL becomes increasingly lower (more sensitive) as the signal duration is increased to about 200 msec (Djupesland and Zwislocki, 1971; McRobert,

PAGE 32

22 Bryan, and Tempest, 1968; Moller, 1962b; Simmons, 1963; Weiss, Mundie, Cashin, and Shinabarger, 1962) A 500-msec duration tone approximates the results of longer stimuli for eliciting a steady change in impedance at the tympanic membrane. With regard to the auditory response to long duration signals, Dallos (1964) and Harford and Liden (19671968) have observed that adaptation takes place only over longer periods of time, e.g., 15 seconds to 30 seconds. This adaptation is thought to be an afferent process, because recovery is instantaneous upon changing to a new test frequency. This afferent adaptation is also frequency-dependent, with higher frequencies demonstrating more response decay (Anderson et a_l 1969; Melcher and Peterson, 1972). The ear in which the reflex is recorded can also influence the level of the acoustic— reflex threshold. Recordings made in the stimulated ear, instead of the contralateral ear, improve threshold sensitivity (Moller, 1961b) Bilateral stimulation improves the threshold value even more (Moller, 1962b) The frequency of the recording probe tone can in itself influence the acoustic-reflex values. Peterson and Liden (1972), averaging the ART for 500 Hz and 4000 Hz, state that a 220-Hz probe tone is about 6 dB more sensitive than a 625Hz probe tone in eliciting the ART, while a 800-Hz tone is between them. In an earlier study, Harford and Liden (19671968) recorded an unsystematic 2-dB threshold difference

PAGE 33

23 between a 2 20Hz and a 800Hz probe tone frequency. This latter study does not agree with the findings of Mehmke and Tegtmeier (197 0) in which there should be a 8 dB loss in transformer efficienty for low frequencies. Lilly and Shepherd (1964) and Feldman (1967) have also observed that the acoustic impedance varies with the frequency of the probe tone. The choice of probe tone frequency may be dependent upon the length of the sinusoidal wave and its relation to the dimensions of the external auditory canal. Acoustic impedanc becomes more sensitive to changes in probe position in the external auditory meatus and to increased diameter of the meatus as the probe frequency increases. This may be minimized by using a low frequency probe in a larger volume, e.g. at the external meatus (Schoel and Arnesen, 1962) This contention is not supported by Djupesland et al. (1966) who see no effects of probe position. Djupesland and his associates used 5-dB test increments, which may have obscured any small but significant results of plug position. It might be possible that the attributed frequency effects of the 'probe tone are due to the varied, and often unreported, intensity levels of the probe tone. It is important that the probe tone not be intense enough to evoke the acoustic reflex, since it is there to indicate the existence of intra-aural muscular contraction and not to elicit it (see footnote 4) The probe tone level must also be above

PAGE 34

24 the background environmental and physiological noise so that the background noise does not cause spurious results. In order to meet these two criteria a larger dynamic range is available with a low frequency probe tone. The fact that low frequency tones elicit the ART at higher intensity levels can be seen in Table 3. It is not possible to know whether the difference in ART levels is a real effect of probe frequency or one due to the intensity level of the probe tone (Lilly and Shepherd 1964 ; Terkildsen, Osterhammel, and Scott Nielsen, 1970) Acoustic -reflex dynamics of magnitude and latency The intra-aural muscle response had also been observed in terms of how the reflex changes as the stimulus intensity is increased above threshold. The response magnitude of the acoustic reflex grows as a function of intensity to approximately 30 dB above the acoustic-reflex threshold (30 dB ART SL) (Dallos, 1964) Most of this growth occurs within 16 dB ART SL (Djupesland et a_l 1966) at a near linear rate (Dallos, 1964; Moller, 1962a; Peterson and Liden, 1970; Weiss et al 1962) with no observable increases to pure tone stimuli beyond 120 dB SPL (Hung and Dallos, 1972) Low frequency tones cause the AR to grow at a faster rate than high frequency tones (Djupesland, et al 1966; Harford and Liden, 19671968) There does not seem to be agreement whether one particular frequency or noise band causes a larger change in the reflex (Djupesland et al 1966; Fisch and Schulthess, 1963; Johansson, Kylin, and Langfly, 1967)

PAGE 35

25 Latency, the period from signal onset to contraction, and contraction time, the interval from stimulus onset to maximum contraction, were some of the earliest acoustic reflex characteristics studied (Loeb, 1964) Latency decreases as the intensity is increased above the ART (Dallos, 1964; Fisch and Schulthess, 1963; Perlman and Case, 1939) The diortest latency of the stapedius muscle and tendon is about 10 msec to 15 msec in man (Fisch and Schulthess, 1963; Neergaard and Rasmussen, 1966; and Perlman and Case, 1939) while the tensor tympani has a comparatively longer latency of 50 to 120 msec. There is considerable intraand inter-subject variability in latency (Moller, 1958) This is due, in part, to the obscurity of the contraction close to threshold, but the shorter the latency the more reliable the trace (Neergaard and Rasmussen, 1966) The latencies of the ART are among the shortest for muscle reflexes in man. Fisch and Schulthess (1963) conclude from an EGM study that this short latency, especially for stapedial contraction, is probably due to the limited number of synapses in the crossed acoustic— reflex arc. The longer latency obtained from the tensor tympani— reflex arc may indicate that it has additional synapses (viz. Figure 1) • The maximum impedance change may be reached within 400 msec to 500 msec after signal onset. Some contraction may be present as long as one second after signal offset. Djupesland and Zwislocki (1971) contend that the growth and decay

PAGE 36

26 of the muscle reflex is symmetrical. Decay from maximum contraction is apparently independent of intensity and frequency. Although the importance of the acoustic reflex can be obscured by the many factors affecting it, consistent threshold levels have been obtained in a number of studies. The consistency of the acoustic-reflex response contributes to its utility as a clinical tool. Clinical application of the acoustic reflex Metz (1946) introduced acoustic impedance measurement as a clinical tool. He included the presence of the acoustic reflex and the level necessary to just elicit this reflex as one of the useful dimensions of acoustic impedance. The absence of the reflex can support the inference that a middleear problem exists in the ear under test. Terkildsen and Scott Nielsen (1960) and Klockhoff (1961) presented clinical cases showing that a relatively normal middle ear is necessary to elicit an intra-aural reflex. In other words, some authors feel that the presence of a reflex is indicative of a normal middle-ear (Feldman, 1967; Klockhoff, 1961) This is not necessarily true, however. Brooks (1969, 1971) states that a minor conductive component will not abolish the reflex but instead elevate the acoustic -reflex threshold. He concludes, therefore, that only subjects exhibiting a reflex to 95 dB HL (ISO-1964) or less in the contralateral ear should be regarded as having normal middle-ear function.

PAGE 37

27 The manometer system in an electroacoustic impedance bridge is able to approximate the amount of pressure in the middle-ear. For the most part, 50 mm of equivalent water pressure (re atmospheric pressure) is considered within normal limits. Hearing and acoustic-reflex thresholds do not deteriorate within this pressure range in the middle-ear (Alberti and Kristensen, 1970; Peterson and Liden, 1970). Negative middle-ear pressure is more detrimental to thresholds than positive pressure (Moller, 1965), but the AR may be stronger with slightly negative canal pressure (Terkildsen, 1964). Terkildsen (1960c) even states that the stapedial reflex is found around normal atmospheric pressure, while the tensor reflex is enhanced by negative pressure. In the absence of middle-ear conductive problems, the level at which the AR is just elicited above the threshold of audibility is useful in confirming a cochlear abnormality. It has been observed that in the majority of mild to moderate sensorineural hearing losses the acoustic-reflex threshold occurs at approximately the same hearing-threshold levels (HL) as the normal ART (Alberti and Kristensen, 1970; Ewertsen, Filling, Terkildsen, and Thomsen, 1958; Jerger et a_l. 1972; Klockhoff, 1961; Kristensen and Jepsen, 1952; Lamb, Peterson, and Hansen, 1968; Metz, 1946, 1952; Peterson and Liden, 1972; Terkildsen, 1960b; Thomsen, 1955a,b; and others) If the ART occurs at 55 dB to 60 dB SL or less, the hearing loss is considered to be due to cochlear impairment. The ART is

PAGE 38

28 seldom seen below 25 dB SL and never less than 5 dB to 10 dB above pure-tone threshold (Jerger, 1970; Jerger et al 1972; Lamb and Peterson, 1967; Lamb, Peterson, and Hansen, 1968). If a reflex is obtained at less than 5 dB SL it can be assumed that the pure-tone hearing loss is due to non-organic causes (Jepsen, 1953, 1963; Lamb, Peterson, and Hansen, 1968; Terkildsen, 1964; Thomsen, 1955b) Liden (1970) demonstrated that the intra-aural reflex can be used as an objective loudness-recruitment test. He compared 52 ears with unilateral Meniere's syndrome, 30 ears of athetoid children and 9 ears presenting acoustic tumors The patients were divided into three groups according to the level of the reflex thresholds and magnitude of separation between their threshold of audibility and their ART. If the ART exceeded 95 dB HL (ISO-1964) or if the intensity span between the pure-tone threshold and the ART was less than 75 dB, the response was considered abnormal. These two value correspond, respectively to the ninetieth and tenth percentiles on 88 normal— hearing control subjects. Liden maintains that a reduced intensity span between the ART and the pure tone threshold is a result of a cochlear lesion, and an elevated ART represents a probable higher order lesion. A reduced span superimposed on an elevated ART represents both areas as foci of the lesion. For the most part, his three pathologic groups support this contention, even though his span of 75 dB for normal hearing is 10 dB to

PAGE 39

29 15 dB more conservative than the other reported studies. As further support he obtained preand post-noise exposure thresholds, purs tone and reflex, on 11 cats. The results can be seen in Figure 3. An average permanent threshold shift of 44 dB resulted with an elevation of the ART by only 1 dB from the pre-exposure levels. The only noticeable effect upon the ART was a slight increase in the standard deviation across the four reflex eliciting frequencies. It appears, therefore, that the acoustic reflex is a loudness-sensing mechanism that can be used clinically to indicate the possible presence of loudness recruitment. To further authenticate the meaning of the acoustic-reflex level in cochlear— impaired persons, comparisons have been made with loudness-balance tests and other audiometric indicators of cochlear lesions. In most cases a comparison is made with a unilateral end-organ disease and results of the ABLB, as the standard of loudness recruitment. The results indicate that the ART level above the threshold of hearing is as good as the ABLB and slightly better than the SISI for determining an end-organ lesion (Alberti and Kris ten s en 1970; Ewertsen et al 1958; Kristensen and Jepsen, 1952; Lamb et al., 1968; Liden, 1970; Thomsen, 1955a; and others) It is definitely a better and more reliable predictor of sensory damage than the Monaural Bi-frequency Loudness Balance, the Difference Limen for Intensity or Frequency, the Uncomfortable Loudness Level, and Bekesy Types

PAGE 40

30 (N Q O a co w OS Eh a o w 2 X Q CM O o o CO w CQ H U w Q 120 110 100 90 80 70 60 50 40 30 20 10 0 -10 -20 -30 PRE-NOISE POST-NOISE 0.1 kHz 1 kHz 10 kHz TEST FREQUENCY Figure 3. Preand post-noise exposure thresholds. The lower thresholds are audibility and the upper thresholds are acoustic reflex in 11 cats. (Redrawn from Liden, 1970.)

PAGE 41

31 (Ewertsen et al. 1958; Niemeyer, 1971). Coles (1972), however, argues that the ART, like the Uncomfortable Loudness Level, is a vague concept and is only diagnostically useful when absent. Beedle (1970) doubts that the ART is indicative of endorgan damage. If, in fact, the ART is a measure of loudness recruitment, it does not continue to demonstrate this rapid growth in loudness above the ART. Instead, the reflex growth function is less steep and slower than in normal ears. (Also see Beedle and Harford, 1971; Peterson and Liden, 1970, 1972.) As the hearing loss increases, the level of the acousticreflex threshold also increases but not proportionately (Liden, 1970) • Sensorineural hearing losses beyond 70 dB to 80 dB HL (ISC— 1964) do not normally demonstrate an acoustic reflex at any intensity (Jerger, 1970; Jerger et al., 1972; Terkildsen, 1960b) Jerger makes use of this fact to predict the probability of the hearing level at the threshold of audibility. For example, in the presence of a reflex, there are 5 chances in 10 that the loss does not exceed 85 dB HL (ISO1964) and there is only 1 chance in 10 that it is as much as 100 dB HL. This is a nebulous approach at best, because it lacks quantification for specific cases. In summary, the acoustic reflex is a well-defined physiological phenomenon which has diagnostic value and gives objective information, its clinical value and true objectivity

PAGE 42

32 occurs when the involuntary acoustic-reflex threshold is compared with the threshold of audibility. When the dynamic range between the two thresholds is 55 dB or less, there is cochlear impairment. By independently varying the signal duration at threshold of audibility, more diagnostic information is gained. This manipulation of auditory processing, as seen in the next section, may be applied to the acoustic reflex threshold as well. Temporal Summation There is a psychophysical assumption that the normal auditory system can summate, or integrate, acoustic power over some critical period of time. Since the initial work of Garner (1947b), Garner and Miller (1947), Hughes (1946), Munson (1947), and de Vries (1948) considerable interest has been generated over the concept of "trading" increased acoustic power with a decreased signal duration to maintain a constant loudness level. When the on-time of the auditory stimulus is sequentially halved, starting from some critical duration, the signal power is reduced 2 dB to 3 dB with each decrement, i.e., from 200 msec to 100 msec, from 100 msec to 50 msec, etc. The critical long duration is thought to be about 200 msec and a linear relationship holds to a critical short signal length of approximately 10 msec. Perfect integration If one assumes perfect integration (Garner, 1947a, b;

PAGE 43

33 Garner and Miller, 1947) the ear would trade 1 log unit of intensity (1 bel, 10 dB) for 1 log unit of duration (10 msec to 100 msec, 20 msec to 200 msec, 15 msec to 150 msec) Even though there is an increase of power as stimulus time increases, the acoustic energy required to maintain the threshold of audibility is relatively constant. This accumulation of power over an average tenfold decrease in time may be inferred from threshold data using a single-slope value (Clack, 1966; Garner and Miller, 1947; Munson, 1947; Northern, 1967; and others) When this slope value is 10 dB it is considered perfect temporal integration of energy. Perfect integration is most often observed for the sinusoid of 1000 Hz. The time constant of temporal integration (T Q ) marks the point at which time and intensity cease to have a linear relationship; it is thought to occur between 200 msec to 250 msec. Goldstein and Kramer (1960) observe integration occurring through 200 msec, but Harris, Haines, and Myers (1958) report individual time constants ranging from 100 msec to 300 msec with a mean of 200 msec. Plomp and Bouman (1959) and Hempstock, Bryan, and Tempest (1964) indicate that T Q is inversely related to frequency; T Q changes from about 375 msec at 250 Hz to approximately 150 msec at 8000 Hz. Regardless 6 Acoustic energy can be stated in the simplest form as E = PT, where E is acoustic energy, P is acoustic power, and T is linear for a certain period of time, the time constant. Further increases in signal duration beyond the time constant have less and less effect upon threshold level.

PAGE 44

34 where T Q may fall, integration is essentially complete between 500 msec and 1000 msec (Zwislocki, 1960) The phenomenon of integration of acoustic power over time is attributed to temporal summation in the auditory system and, most probably, is neural in nature (Zwislocki, 1960) The term summation is introduced by some authors because it is felt that the response pattern represents temporal summation at the synaptic junctions in the neural pathways. More specifically, this apparent linear function of log time and log power is due to an exponential decay of the persisting neural excitation (Plomp and Bouman, 19 59; Zwislocki, 1960) At threshold, there appears to be a direct proportionality between an increase in the intensity of acoustic power and the increase in neural excitation, which is modified somewhat at suprathreshold levels by neural adaptation and the loudness of the stimuli (Zwislocki, 1960) The neural mechanism for temporal integration probably exists at the neurons above the first and, possibly, the second order, but before the level of binaural interaction at the superior olivary complex (Zwislocki, 1960) Stimulus parameters The literature indicates that, in addition to duration, other various stimulus parameters affect temporal integration. Those most often cited are stimulus spectrum, rise and decay time, definition of the signal duration, inter-stimulus

PAGE 45

35 interval, attenuation rate, method of threshold calculation, and psychophysical procedure. Earlier studies indicate that for a narrow band of frequencies between 1000 Hz and 4000 Hz a 10 dB/decade slope is generated (Garner, 1947a; Garner and Miller, 1947; Hughes, 1946; Munson, 1947) while narrowor wide-band noise bursts demonstrate about a 8 dB/decade change in threshold integration (Garner, 1947b; Miller, 1948; Small, Brandt, and Cox, 1962) There are no systematic differences in the slope between threshold determination in quiet (Garner, 1947b; Hempstock, Bryan, and Tempest, 1964; and others) and in background noise (Garner and Miller, 1947; Blodgett, Jeffress, and Taylor, 1958; Hempstock et al., 1964; Gengel, 1972; Plomp and Bouman, 1959; and others). More recently, a controversy concerning frequency effects has developed. While 1000 Hz elicits an average slope of 10 dB per decade, lower frequencies have a steeper slope and an increase in frequency produces a flatter slope, as seen in Figure 4 (Brahe Pedersen and Elberling, 1972; Elliott, 1963; Gengel, 1972; Gengel and Watson, 1971; Hattler and Northern, 1970; Hempstock et al. 1964; Miskolczy-Fodor 1953; Northern, 19 67; Olsen and Carhart, 1966; Sanders, Josey, and Kemer, 1971; Sheeley and Bilger, 1964; Simon, 1963; Tempest and Bryan, 1971; Watson and Gengel, 1969). Some investigators feel that there is no frequency dependence (Clack, 1966; Martin and Wofford, 1970; Wright, 1968a, 1972; Zwicker and

PAGE 46

36 16 32 64 128 256 512 1024 SIGNAL DURATION IN MSEC Figure 4. Average temporal integration slopes as a functxon of frequency. In general, as frequency decreases the slope steepens. (Redrawn from Watson and Gengel, 1969.)

PAGE 47

37 Wright, 1963) Inter-subject variability may cause the difference in frequency dependence (Clack, 1966) Another possibility is that frequency dependence may be observed at shorter durations but not necessarily at the longer signal on-times (Olsen and Carhart, 1966) Price (1972) has also suggested that the external ear may cause a transformation of the stimulus parameters in some subjects by raising the energy peak in frequency and by legthening short pulses, thereby possibly nullifying the effects of frequency. Frequency dependence is, in part, related to the type of psychophysical procedure used. Bekesy tracking and forcedchoice tracking show little or no frequency dependence, but the frequency effect is demonstrated by the method of adjustment, method of limits, method of constant stimuli, and confidence ratings (Bilger and Feldman, 1969; Chamberlin and Zwislocki, 1970; Watson and Gengel, 1969). In the majority of studies Bekesy tracking is the method used. Even though there is large individual variability in the value of the slope (Clack, 1966; Gengel, 1972; Green et al. 1957; Hattler and Northern, 1970; Martin and Wofford, 1970) there is general agreement that the test-retest variability is good (Doyle, 1970; Hattler and Northern, 1970; Olsen and Carhart, 1966, Plomp and Bouman, 1959) Gengel and Watson (1971) suggest at least 12 threshold crossings for each data point when using Bekesy tracking in order to achieve a reliable reading.

PAGE 48

38 Spectral characteristics of the signal as a function of duration cannot be divorced from the integrating bandwidth, or the critical band of the ear (Fletcher, 1940; Scharf, 1970) The theoretical bandwidth of a pulsed tone is proportional to the reciprocal of the signal's duration, or 1/t. As long as the spectrum of the test tone remains within the critical bandwidth of the ear, all of the energy will be integrated. If not, there should be a loss in sensitivity (Garner, 1947a; Green et al 1957; Olsen and Carhart, 1966; Sheeley and Bilger, 1964; Wright, 1968a). The spread of energy by abrupt low frequency signals upward to the more sensitive frequencies of the ear has been misinterpreted as an apparent increase of sensitivity within the critical band. It has resulted in the assumption that the ear does not integrate low-frequency tone pips (Campbell and Counter, 1969; Karlovich, Lane, Smith, Tar low, Thompson, and Vivion, 1971) The point is that one must carefully monitor the test stimuli so that the duration of the tone burst is maintained within the critical bandwidth (Wright, 1968a) Some variability in temporal integration has stemmed from lack of defined criteria of stimulus duration. Goldstein and Kramer (1960) measured the duration between energy onset and cessation, while Harris (1957) designated as criteria the half -power points on the stimulus envelope. An "equivalent duration" has also be used (Brahe Pederson and Elberling, 1972; Dallos and Johnson, 1966; Dallos and Olsen,

PAGE 49

39 1964; Olsen and Carhart, 1966) An equivalent-duration tone pip contains the same amount of energy as a rectangular envelope and allows for the comparison of widely varying envelope shapes. The use of the equivalent duration fits or allows the conversion of the overall stimulus envelope to fit the slope of Garner's model (1947b), which states that not all energy is used by the ear to summate temporal information (Dallos and Olsen, 1964) The best measure of the acoustic waveform is still a moot question because the conversion from the half-power points to equivalent duration of Harris 1 data (1957) by Dallos and Olsen (1964) demonstrate no difference. The time in which the stimulus envelope rises to and decays from its effective peak is a critical variable in auditory measures using pure-tone stimuli. If the stimulus starts or stops too abruptly, a wide-band transient may be generated, thereby biasing the results (Wright, 1958, 1968a). A minimum 5 msec rise/fall time measured on the linear portion of the ramp (between 10 percent and 90 percent of maximum amplitude of the acoustic waveform) is most often suggested (Harris, 1957; Wright, 1960, 1968a). No differences occur with rise/fall times varying from 0 msec to 50 msec if the equivalent duration is held constant (Dallos and Johnson, 1966; Dallos and Olsen, 1964) The inter-stimulus interval, or the off -time between successive stimulus envelopes, must be kept long enough to ensure neural independence between stimulus events. Since

PAGE 50

40 the theory of temporal summation is based on neural decay, the minimum off-time should be 200 msec (Zwislocki, 1960) In addition, a "critical off-time" has been calculated whereby successive stimuli having less than a 200 msec interstimulus interval take on the threshold value of one continuous tone (Jerger, Jerger, Ainsworth, and Caram, 1966) Doyle (1970) in a temporal integration study varied the interstimulus interval from 100 msec to greater than 500 msec. There were no differences in temporal— integration results as long as the dead time was a minimum of 500 msec, but the slopes flattened as the off-time was shortened to 100 msec. In practice, therefore, the inter-stimulus interval should exceed the theoretical minimum, and probably be no less than 500 msec. The repetition rate of the signal per unit of time, usually one second, is also based on consideration of decay of neural excitation. If one desires to keep the repetition rate constant, then the rate is determined by the length of the longest-duration test signal and the desired inter-stimulus interval. Wright (1968a) recommends the use of the same repetition rate for all test stimuli so that sampling per unit of time will be uniform. Some attention has been given to attenuation rate for those test procedures employing semi -automated equipment, for example, Bekesy tracking. Hempstock et al. (1964) observed an increase in the standard deviation of the threshold as the

PAGE 51

41 number of pulses presented at each discrete level decreased from eight pulses to one pulse. There was no significant difference in the mean threshold. Wright (1968a) suggests using an attenuation rate of 2.5 dB/second at a repetition rate of 1/second as a standard. The choice of ascending, descending, or mean thresholds can give threshold values differing as much as 3 dB at any one signal duration (Hempstock et al 1964). Therefore, for the majority of the cases, where applicable, mean values determine threshold. The vast majority of recent publications dealing with temporal integration have used the Bekesy tracking procedure, where the averaged midpoint of the crossings is considered to be the threshold value. Effects of end-organ lesions upon temporal integration Considerable interest has been generated in the use of temporal summation as a diagnostic tool for end-organ pathologies; this is otherwise known as brief -tone audiometry. A cochlear lesion seemingly increases the integration ability of the ear. As the duration is decreased, less acoustic power is needed to maintain constant energy for threshold (Brahe Pederden and Elberling, 1972, 1973; Doyle, 1970; Elliot, 1963; Gengel and Watson, 1971; Gengel, 1972; Harris et al. 1958; Hattler and Northern, 1970; Jerger, 1955; Martin and Wofford, 1970; Miskolc zy-Fodor 1953, 1956, 1960; Nerbonne, 1970; Olsen, Rose, and Noffsinger, 1973; Sanders and Honig,

PAGE 52

42 1967; Sanders, Josey, and Kemer 1971; Wright, 1968b; Wright and Cannella, 1969) The integration slope is not affected by either conductive problems (Miskolczy-Fodor 1956; Harris et al 1958; Wright, 1968b; Wright and Cannella, 1969) or by retrocochlear problems (Olsen et al 1973; Sanders et al. 1971) The critical time at which the temporal integration begins, or the time constant of integration, also seems to be shortened to about 100 msec or less depending on the particular frequency (Miskolczy-Fodor, 1953, 1956, 1960; Harris et al. / 1958; Sanders and Honig, 1967; Wright, 1968a). The abnormal results of cochlear damage to slope and time constant can be seen in Figure 5, i.e., the slope is flatter, the time constant is shorter, or both. It has also been suggested that brief-tone audiometry may even be so sensitive as to detect incipient damage to the cochlea (Campbell and Counter, 1969; Sanders and Honig, 1967; Wright, 1968a) to detect recruitment (Miskolczy-Fodor, 1956, 1960) and to differentiate between various cochlear lesions (Harris et al., 1958). Invariably, these attributes are artifacts of unaccounted frequency effects, energy spread, and/or inter-subject variability (Gengel and Watson, 1971; Karlovich et al 1971). Olsen et al. (1973) cast doubt on the efficacy of temporal integration to differentiate between end-organ and eighth nerve lesions because of the large overlap between populations.

PAGE 53

43 Figure 5. Examples of temporal integration. Cochlearimpaired integration functions can be flatter, can have a shorter time constant, or both. (Redrawn from Gengel and Watson, 1971.)

PAGE 54

44 Nerbonne (1970) who studied temporarily fatigued ears, and Sanders et a_l. (1971), who studied cochlear lesions, demonstrated that the SISI and ABLB, as well as brief-tone audiometry are sensitive indicators of cochlear lesions. Brief-tone audiometry can yield definitive results when the ABLB shows partial recruitment and negative SISI scores are present (Sanders et a_l. 1971) There is some correspondence between amount of recruitment and degree of aberration of temporal integration, but, for the most part, temporal integration is assessing some other aspect of sensory lesions (Nerbonne, 1970) Abnormal temporal integration seems to be part of a syndrome of a distorted time domain of enlarged critical bands and critical ratios (Northern, 1967; Sheeley, 1963; Simon, 1963) and poorer performance on difference limens for frequency (Sheeley, 1963) Temporal integration, by way of its relation to critical bands, seems to have a common basis with several phenomena related to the tuning of the auditory system (Licklider, 1951) that are not presently measured clinically. The method of brief -tone audiometry, therefore, presents clinicians with another possible test for end-organ abnormalities, if proper standards are set forth for the various stimulus and test parameters, and normal response limits are determined.

PAGE 55

CHAPTER II STATEMENT OF THE PROBLEM Innumerable studies detailing the identification and quantification of cochlear abnormalities indicate that, with patients who cannot or will not cooperate fully, the present diagnostic measures are inadequate in obtaining information on auditory processing. Further understanding of the function of the acoustic-reflex threshold and the effect of temporal integration upon threshold measures may fill this void in diagnostic testing of the auditory system. Such a measure may be obtained by maintaining a constant reflex strength as a function of increasing acoustic power while signal duration is decreased, or, in other words, temporal integration of the acoustic reflex. Since there are relatively few data concerning temporal integration of the acoustic reflex, the pertinent literature will now be reviewed. As early as 1935, Lorente de No reported the effects of varying short signal duration upon the contraction of the tensor tympani muscle of the cat. He used as much as 140 dB SPL and durations from 3 msec to 100 msec. He explained his results in terms of neural summation, which occurs at the synapse. 1 At some synapses one pre-synaptic impulse is not 1 For further information regarding synaptic summation see the classical treatise of Sherrington (1906, 1947); also see Brink (1951) 45

PAGE 56

46 enough to initiate a post-synaptic impulse, it is therefore necessary to summate several incoming impulses, i.e., spatial and temporal summation. Spatial summation occurs as a result of neural impulses converging simultaneously at a synapse but from different neural fibers. Since an increase in the number of neural fibers transmitting impulses is thought to be a method of coding increased stimulus intensity, spatial summation can also be in response to increased intensity. Temporal summation results from successive impulses reaching the synapse through the same fibers. With the temporal form of neural summation, the post-synaptic potential occurs as the signal duration increases. It is the amount of information that each fiber carries and the number of fibers invoked which makes the difference in the type of summation. Lorente de No has suggested, therefore, that varying the signal duration changes the strength of the acoustic reflex by virtue of temporal summation. Simmons (1963) has also demonstrated temporal summation of the intra-aural muscles in cats to acoustic stimulation. His results are explained in terms of an on-response in the auditory system in which there is linear integration of acoustic power from 5 msec through 50 msec at a rate of 8 dB per doubling of duration. There is only a 2-dB acousticreflex threshold improvement from 50 msec to 100 msec which indicates a time constant of 50 msec. In comparison to temporal integration at threshold of audibility in man, the

PAGE 57

47 time constant is about 200 msec and the threshold changes 2 dB to 3 db per doubling of signal duration. There are relatively few data on temporal integration of the acoustic-reflex threshold in humans. The systematic published works are by McRobert, Bryan, and Tempest (1968) and Djupesland and Zwislocki (1971) In both instances data are plotted in terms of acoustic-power growth as signal duration decreased while maintaining acoustic-reflex threshold. Their slopes, reported in power to maintain reflex threshold per average decade change in signal duration, are in the range reported by Simmons (1963) The slopes range from 14 dB/decade for 300 Hz, to 21dB/decade for 2500 Hz (McRobert et al 1968), and to 25 dB/decade for 1000 Hz (Djupesland and Zwislocki, 1971) Although there are no reports of a systematic frequency effect (McRobert et al 1968; Simmons, 1963) the middle frequencies apparently have steeper slopes. Figure 6 represents the effects of temporal integration upon the threshold of the acoustic reflex in six normal hearing subjects by Djupesland and Zwislocki (1971) The median of the individual threshold means demonstrates a regular decrease of 5 dB to 7 dB per halving of duration, or a slope of about 25 dB/decade. The time constant is seen to be, for the most part, about 20 0 msec. One subject appears to begin integration at 100 msec with no obvious change in slope, as seen by the dashed line connecting the lowest mean ART data. McRobert et al. (1968) also present slopes which break linearity between 100 msec and 200 msec.

PAGE 58

48 u g 60 o o o m Eh < B 50 40 30 Q O 20 Q O CO W BJ K U) K W in Eh ^ rc Eh n I i i "iij 1 — r i 1 1 mi j 1 — i i i rrrq — i rrm Median of 6 Ss Lowest Mean ART 1 1 1 1 in 1 1 1 1 H ii 10 100 1000 10000 SIGNAL DURATION IN MSEC Figure 6. Temporal summation of the acoustic reflex. The solid line intersects the median of six individual mean scores resulting in a slope of 25 dB/decade of signal duration change. The dashed line indicates a similar slope but a shorter time constant. (Redrawn from Djupesland and Zwislocki, 1971.)

PAGE 59

49 No explanation appears to be offered for the steeper slope of the acoustic reflex, but the temporal summation effect is there without question as illustrated in Figure 7. The range of values reported by Olsen, Rose, and Noffsinger (1973) for temporal integration at threshold of audibility are contrasted with the data by Djupesland and Zwislocki (1971) for temporal integration at threshold of the acoustic reflex. These are both at 1000 Hz in normal ears and have been normalized to 200 msec so that the slope has a common reference between the two sets of data. The slope for the acoustic-reflex threshold at 25 dB/decade is about three times as large as the median slope of 8 dB/decade at threshold of audibility. Other authors have reported effects of short stimulus durations upon the acoustic reflex, but their results are not reported systematically as temporal integration, nor in a form easily converted to the values presented in Figures 6 and 7 (Johansson, Kylin, and Langfly, 1967; Lilly, 1964; Moller, 1962b; Weiss, Mundie, Cashin, and Shinabarger, 1962) There have been no published reports dealing with the effects of temporal summation of the acoustic-reflex threshold in hearing-impaired ears, but there are indications of disturbed spatial summation. Beedle (1970) and Beedle and Harford (1971) have reported finding an effect of stimulus intensity upon the growth to the acoustic reflex. The slope of the reflex growth is much steeper and more rapid for normal ears than for either ear of an unilateral Meniere's group. Niemeyer and Sesterhenn (1972) have noted the occurrence of

PAGE 60

50 • Acoustic Reflex Threshold — Median for Acoustic Reflex Threshold o — 1 1 L_ 10 20 100 200 SIGNAL DURATION IN MSEC .... Fl 9 u re 7. Temporal integration at threshold of audimC lity nd f the acoustic reflex for normal-hearing subjects The median slope of the acoustic reflex is about three times the change in the median slope of audibility. (Based on Djupesland and Zwislocki, 1971, and Olsen et al., 1973 )

PAGE 61

51 a difference in reflex eliciting intensity levels for pure tones versus white noise. The intensity level difference between the two types of stimuli is much larger in normals than the range in cochlear— impaired ears. Norris, Stelmachowicz and Taylor (1972) have reported the most significant difference in the acoustic— reflex action between normal and pathological cochleas. They measured the difference in levels necessary to obtain the acoustic-reflex threshold. The stimulus is pulsed at 2.5 pulses per second, a 50 percent duty cycle with a 200-msec signal duration. The resulting modulation of the reflex in response to the pulsed tone is greater and more regular in normal ears than in sensorineural hearing losses. This may reflect an increased latency of acoustic— reflex response in cochlear— impaired ears (Johansson et al 1967; Simmons, 1963) or a distorted spatial summation involving rate of reflex growth with increased stimulus level (Simmons, 1963). If effects due to disturbed spatial summation are present in the acoustic reflex as a result of cochlear pathologies, it is reasonable to assume that end-organ lesions can also disturb temporal summation. Cochlear lesions can and do disturb temporal integration at threshold of audibility, that is, by flattening the slope of temporal integration and/or shortening the critical duration. Intuitively, then, temporal integration of the acoustic reflex may also be affected by cochlear lesions, for example, a flattened slope and/or a shortened critical duration.

PAGE 62

52 In summary, temporal summation is a neural event originating at the cochlea and is essentially complete before the level of the superior olivary complex (SOC) where the acoustic reflex is activated. Growth in neural activity in the SOC parallels the growth in loudness up to 60 dB SPL. This neural activity decays in an exponential manner, thereby appearing to trade energy units linearly for about 200 msec. It is suggested that during acoustic stimulation, neural activity in the reflex center exceeds a certain level and causes a contraction of the intra-aural muscles. Furthermore, the mechanism suggested as controlling and/or affecting temporal summation of threshold and loudness is the filtering characteristics of the ear. Lesions in the cochlea can modify the transmission characteristics through the auditory system in such a way that the acoustic reflex and temporal integration are changed in a predictable manner. The foregoing studies indicate two important clinical premises: (1) it is not necessary to compare the acoustic reflex threshold with the threshold of audibility to obtain diagnostic information about the state of the cochlea, and (2) there are factors affecting the acoustic reflex which may differentiate normal ears from those with end-organ lesions. Hypothesis The major hypothesis, formulated for testing in this study, is based on information concerning temporal summation of the acoustic reflex: the acoustic reflex changes in a

PAGE 63

53 systematic fashion in the normal ear as a function of stimulus duration, i.e., a tenfold change in time can cause a median change of 25 dB of acoustic power. The null hypothesis is: temporal summation of the acoustic refle x does not function differently in those subjects who have normal hearing and in those subjects who have a known end-organ auditory lesion The questions that this study attempted to investigate are: 1. What is the most sensitive measure of the acoustic-reflex threshold when using increments of 2 dB : the first response, the first level to occur 60 percent of the time, a measure of central tendency of either 5 or 10 trials, or the lowest response of 10 trials? 2. What is the group average and range of acoustic-reflex thresholds for a normal-hearing population as a function of temporal integration? 3. What is the group average and range of acoustic-reflex thresholds for an unilaterally cochlear impaired population as a function of temporal integration? 4. Is there a significant difference in temporal integration of the acoustic reflex between normal hearing ears and cochlear impaired ears? 5. Is there a significant difference in temporal integration of the acoustic reflex between the ears of normal-hearing subjects and the normal-hearing ears of unilaterally cochlear impaired subjects?

PAGE 64

54 Is there a significant difference in temporal integration between the normal-hearing ear and the cochlearimpaired ear of a Meniere's group? If cochlear-impaired ears demonstrate reduced integration at threshold of audibility, do they also demonstrate reduced temporal integration at the threshold of the acoustic reflex?

PAGE 65

CHAPTER III METHODS AND PROCEDURES Temporal summation at both threshold of audibility and threshold of the acoustic reflex were measured by the following experimental design. First, a baseline of temporal integration at threshold of audibility was obtained in order to compare the results in this study to previously reported data on normal-hearing and cochlear— impaired subjects. This information delineated normal integration from reduced integration of cochlear-impaired ears. Second, the results of temporal summation of the acoustic reflex determined whether the null hypothesis should be rejected; that is, temporal summation of the acoustic reflex does not function differently in those subjects who have normal hearing and in those subjects who have a known end-organ auditory lesion. The results of this second auditory procedure were obtained without the subject's overt cooperation. Ten cooperative adult subjects were tested. One-half of the group had normal hearing and were the control subjects. The other half of the group evidenced unilateral end-organ hearing impairments and served as the experimental subjects. These two groups were classified homogeneously and screened for possible problems that might interfere with normal acoustic-reflex function. 55

PAGE 66

56 The subjects had a negative history of middle-ear or inner-ear surgery and of concussion, brain damage, or cerebral vascular accidents. Those subjects using drugs or medication, possibly influencing normal intra-aural muscle function, were excluded. The medical history form employed i shown in Appendix A. None of the subjects had conductive hearing problems. Normal compliance at the tympanic membrane and normal tympanograms confirmed the absence of a functional middle-ear problem. Data from the five control subjects were collected for each ear. These five subjects demonstrated pure-tone audiometric measures no worse than 20 dB HL (ANSI-1969) in either ear for frequencies 125 Hz through 4000 Hz. The normal-hearing group ranged from 23 years to 30 years of age. Only those persons having a classical, unilateral Meniere's disease, a pure end-organ disorder, with normal hearing in the contralateral ear, were included in the experimental group. The Meniere's disease was medically diagnosed. The patients had, in the course of their medical history, the classical symptoms of episodes of true vertigo, fluctuating hearing loss, "roaring" tinnitus, and feeling of fullness in the affected ear. These symptoms were supported by the following objective findings: a sensorineural hearing loss, loudness recruitment and vestibular canal paresis. This group will be referred to as the Meniere's group or the cochlear-impaired group. The Meniere's group ranged from 45 years to 63 years of age.

PAGE 67

57 Since one of the characteristic symptoms of Meniere's disease is fluctuating hearing in the impaired ear, there were no minimum requirements of hearing level to qualify the impaired ear as abnormal. The contralateral ear of each experimental subject yielded pure-tone thresholds of 20 dB or less (ANSI-1969) at a minimum of two out of the three test frequencies. This qualified the contralateral ear as having normal hearing. To facilitate discussion, ears with normal hearing were referred to as "good," those ears afflicted with Meniere's disease were denoted as "bad." This classification of ears was subsequently used throughout this report. Thus, there was a total of 15 good ears and 5 bad ears in 10 subjects. Subject LS of the Meniere's group did not have normal hearing in her good ear at 500 Hz. Nevertheless, this particular ear is classified as good because she did have normal hearing at two of the three test frequencies. The audiograms of all subjects are contained in Appendix B. Stimuli The test stimuli were shaped sinusoids ranging from 500 msec to 10 msec in overall duration with 5-msec rise and decay times. These signals were gated once every 1200 msec. The stimuli used in this study are listed in Table 4. The choice of stimuli was determined by previous temporal-integration data, ease and expediency of equipment manipulation, and maximum utilization of the good and bad ears of the Meniere 1 s group

PAGE 68

58 Table 4. Test Stimuli Used to Elicit Threshold of Audibility and Threshold of the Acoustic Reflex Test frequency: 500 Hz, 1000 Hz, 2000 Hz Signal durations: 500 msec, 200 msec, 100 msec, 20 msec, 10 msec Rise/fall times: 5 msec Inter-stimulus interval: 700 msec, 1000 msec, 1100 msec, 1180 msec, 1190 msec Repetition rate: once every 1200 msec More specifically, the shaped signals consisted of three sinusoids at 500 Hz, 1000 Hz, and 2000 Hz. These test frequencies were chosen because most of the Meniere's group had normal hearing in the good ear and poorer hearing in the bad ear at these frequencies. Five signal durations were used: 500 msec, 200 msec, 100 msec, 20 msec, and 10 msec. The duration was measured from the gate onset of a stimulus to the cessation of the stimulus envelope. Five hundred milliseconds represented, for practical purposes, an infinitely long signal. The remaining four durations represented two tenfold, or decade, changes in time, i.e., 10 msec to 100 msec and 20 msec to 200 msec. These values also included the usually designated limits of linear temporal integration from the long signal duration of 200 msec to the short tone pulse of 10 msec. The repetition rate of the stimulus was held constant by gating once every 1200 msec. Neural independence was maintained, because the shortest inter-stimulus interval of the

PAGE 69

5 9 test series was 700 msec. Preliminary investigation also indicated that this rate would allow for the muscle contraction to return to the pre-contraction baseline before responding to the next signal burst. Figure 8 illustrates a tracing of an oscilloscope recording of the acoustic reflex evoked by a 500-msec signal at 5 dB above the ART. The muscle activity returned to the pre-contraction baseling within the 700-msec inter-pulse interval. It should also be noted that this recording was made towards the end of an hour and one-half test session, and middle-ear muscle fatigue was not observed. The rise and decay times of the stimulus envelope, as controlled by a pre-set electronic switch, were approximately 5 msec. The 5-msec ramp was verified on an oscilloscope by determining the duration between the upper 90 percent and lower 10 percent of the slope. Figure 9 delineates the spectral characteristics of two 1000-Hz tone burst stimuli measured at the earphone output. The bandwidth of the 10-msec duration tone pulse is about 120 Hz wide in Figure 9a. This is about 20 Hz wider than might be expected theoretically, indicating some incidental spread of energy. When the tone burst is lengthened to 20msec duration in Figure 9b, the spectral energy fits nicely into the theoretical bandwidth of 50 Hz, ./ •.. The spectrum of these tone bursts suggests that no audible spread of energy should be detected; this is supported by the smooth envelope

PAGE 70

60 Repetition rate 1 stim./1200 msec Acoustic Muscle reflex relaxation Figure 8. Acoustic-reflex contraction. The intraaural muscles can contract, relax, and return to the precontraction baseline before contracting again. The repetition rate is once every 1200 msec for a 500-msec tone burst at 5 dB above the ART.

PAGE 71

61 1 1 r a b Figure 9. Spectral content of tone bursts with 5-msec rxse/decay ramp. (a) 1000 Hz, 10-msec impulse with halfpower width of ca. 120 Hz; (b) 1000 Hz, 20-msec impulse with half -power width of ca. 50 Hz.

PAGE 72

62 of a 20-msec ton 3 burst in Figure 10. In order to confirm this, an audible check was made at the earphone. No clicks were discerned from a 10-msec tone burst at 500 Hz placed just below threshold audibility. It was concluded, therefore, that a 10-msec tone pulse with a rise and fall time of 5 msec would produce a relatively undistorted signal in this test situation, using high intensity levels. Experimental Equipment The block diagram in Figure 11 represents the equipment used to generate the stimulus envelope and to record the subject's response. A General Radio 1313A oscillator generated the test frequency which was monitored by a Monsanto Model 100-A electronic counter. The sinusoid passed through the Grason Stadler 3262A recording attenuator to the Grason Stadler 829E electronic switch that shaped the signal; then the signal was step attenuated by an Hewlett Packard 350D attenuator. By way of switches I, II, and II, the signal was led through either a Mcintosh 162K power amplifier or through a matching transformer to the earphone, a TDH-14 0 dynamic earphone in a MX41/AR cushion. The signal from switch II was measured on a Ballatine 321 true rms vacuum tube voltmeter and on channel one of a Tektronix 564 oscilloscope. The same "clock" triggered both the electronic switch and the oscilloscope Three response switches were under subject control. Response switch I controlled the recording attenuator.

PAGE 73

63 Figure 10. A 20-msec signal burst with a smooth 5-msec rise and decay.

PAGE 74

w H Cm H CM < w u z w u < u o Q Q W H o On pi £ m H 0) H H-l

PAGE 75

65 Response switch II .interrupted the signal presentation at the electronic switch, while response switch III activated a buzzer audible to the sxperimenter The reflex was measured from the contralateral external ear canal which was sealed with an ear olive. The reflected probe tone passed to the Peters AP61 electroacoustic impedance bridge. The output of the bridge was led through a direct current voltage divider"'" and, finally, to channel two of the oscilloscope. To measure threshold of audibility, switches I and II were closed so that the signal passed from the step attenuator through the matching transformer to the earphone. Switch III allowed the signal to be monitored at the oscilloscope and volt meter without being audible to the subject. The recording attenuator was activated by subject response switch I. Thus, a permanent record of the subject's response was made with a self -tracking Bekesy technique. The same basic equipment was utilized to measure the acoustic-reflex threshold. The recording attenuator was turned off. The test signal was diverted through the Mcintosh power amplifier to the earphone by using the same switches. The eliciting signal was stored on the oscilloscope of channel one, while the acoustic reflex was stored on Channel two. A photographic record of both oscilloscope channels was made using a Rolleiflex SL 3 5 camera with a 50-mm lens. Calibration was performed at both the output of the earphone in SPL and across the earphone in volts. Calibration 1 The d.c. voltage divider was placed in the system to increase the sensitivity of the signal. See Appendix C for the schematic diagram.

PAGE 76

66 was made in SPL at the beginning and conclusion of the experi ment. The step attenuator and the recording attenuator were checked for linearity in SPL. A NBS type 9A coupler with a 6-cc cavity attached to v Bruel and Kjaer 2203 sound level meter with a one-inch microphone were used to measure SPL (re 0.0002 dyne/cm ). Frequency response of the earphone was also measured in SPL. Voltage checks were made before and after each change in test frequency. A General Radio 1900-A wave analyzer and a General Radio 1521-B graphic level recorder were used to measure the spectral characteristics of the test signals. These calibration data may be found in Appendix D. The Peters electroacoustic bridge was calibrated according to the manufacturer's specifications. The probe tone zeroed the balance meter when the input to the bridge network was 94 dB SPL and the probe frequency was 27 6 Hz. Using four different cavity volumes the compliance dial was calibrated at 0.2 cc, 1.0 cc, 2.0 cc, and 4.0 cc so that the balance meter read zero. Finally, the input filter was tuned to 276 Hz, thereby being most sensitive to the probe-tone frequency. Two different earphones were used in this study due to a malfunction of the TDH-140, 10-ohm phone. A TDH-30, 10-ohm phone elicited data on four normalhearing subjects for their threshold of audibility and on two normalhearing subjects for their acoustic-reflex threshold. Since both phones were

PAGE 77

67 calibrated on the same equipment and in a similar manner, the data reflect the respective calibration curves. In other words, the data will be reported as if one earphone were used. Procedure The experimental testing was conducted in three sessions for each individual. The temporal integration at threshold of audibility was measured for both ears during the first test session, while temporal integration at threshold of the acoustic reflex was measured in subsequent sessions for each of the two ears. The three stimulus frequencies were randomized for each ear, and the signal durations were presented in random order for each test frequency. All auditory measures were made in a single-walled IAC sound treated chamber. Temporal Integration at Threshold of Audibility The subject responses were recorded by using a selftracking, Bekesy procedure. The recording attenuator was set at a chart speed of three-fourths of an inch per minute and 1 dB of attenuation every second. The visual mean of a minimum of 10 threshold crossings was used to establish threshold. This mean value was located at the nearest whole dB, and, if the threshold crossings were not stable, the last 10 stable crossings were used to establish the mean threshold value. The total test time for both ears, including a 15to 30-minute rest period midway, was two to two and one-half hours. Since signal crossover to the good ear was not observed

PAGE 78

68 nor reported by the subjects in the preliminary or experimental sessions, a masking noise was not applied to the good ear of the Meniere's group. Temporal Integration at the Thres hold of the Acoustic Reflex Testing of the acoustic reflex commenced once the external meatus of the reflex measurement ear was sealed with an ear olive containing the electroacoustic bridge probe tip. The seal was considered adequate if a positive 200 mm of equivalent water pressure in the external auditory canal did not break the seal. The electroacoustic bridge was periodically checked to determine if the balance dial was within 25 mm of equivalent water pressure of zero. If the balance meter was out of the target range it was rebalanced by adjusting the compliance dial. In order to maintain this criterion, the sensitivity dial was kept on position one. 2 If the pressure dial was outside the target range, the subject was asked to perform a Valsalva or Toynbee maneuver to increase or decrease, respectively, the middle-ear pressure. A modified method of limits was employed to determine the acoustic-reflex threshold. Ten ascending ART trials were 2 Although it is suggested that the higher sensitivity dial settings be used while determining the acoustic reflex, i.e., two or three on the Peters electroacoustic bridge and three or four on the Madsen electroacoustic bridge, preliminary observations suggested an ART difference of no more than 2 dB between sensitivity position one and position three. The use of sensitivity position one reduced the amount of rebalancing necessary throughout the test procedure. If more sensitivity was desired, the vertical multiplier dial or the oscilloscope was changed.

PAGE 79

69 made for each signal duration at each frequency in order to obtain a stable threshold value. The experimenter started below the acoustic-reflex threshold and increased the stimulus level in 2-dB steps until a minimum reflex was detected. At this point, if the response was a definite change from the ongoing pattern, the stimulus level was recorded as the ART The experimenter started the sequence over again by lowering the intensity 10 dB. If the response was in doubt, then another response was evoked at the same signal level or at the next 2-dB higher intensity increment. If the reflex response was definite at the second response or the 2-dB higher response, the initially observed sound pressure level of the reflex was recorded as the ART. Figure 12 illustrates a tracing of a typical acoustic-reflex thresold response. In this instance, a 2-dB intensity increase caused a definite change from the previously recorded pattern. The tenth ascending threshold response was stored on channel two of the oscilloscope. The eliciting signal was increased 5 dB above the reflex threshold and superimposed upon the ART on channel two of the oscilloscope. The two superimposed acoustic reflex images on channel two were then photographed to retain a permanent record. During the acoustic -reflex testing procedure, subjects were requested to glance quietly through some pictorial magazines in an attempt to equate subject state of alertness and focus of attention. Subject state and attention have

PAGE 80

70 90 92 94 96 96 98 98 Channel 1: Stimuli of 500-msec duration Channel 2: Acoustic reflex at threshold Figure 12. Typical response of the acoustic reflex at threshold. Channel 1 illustrates the 500-msec duration tone burst increasing in 2-dB steps. Channel 2 demonstrates the response at 96 dB SPL to be significantly different from the ongoing activity.

PAGE 81

71 been reported to affect the acoustic reflex (Durrant and Shallop, 1969; Gunn, 1967). As long as the subject's attention is generally diffuse and not under intense concentration, e.g., distinguishing visual word choices or auditory intelligibility, the variability is probably slight (Durrant and Shallop, 1969) Experimental Safeguards In performing temporal integration at the threshold of the acoustic reflex, sound pressure levels normally associated with thresholds of discomfort, tickle and pain, and temporary and permanent threshold shifts were required. It was necessary to safeguard the subjects against any harm or unnecessary discomfort. Possible response artifacts resulting from high intensity signals were also investigated. It was conceivable that all subjects might be subjected to stimulus levels in excess of 130 dB SPL, as seen in Table 5. This table demonstrates a possible intensity level necessary to elicit a reflex by a 10-msec tone burst at 500 Hz. For this hypothetical person, the maximum earphone output would be 136.5 dB SPL. The Committee on Hearing, Bioacoustics and Biomechanics of the National Academy of Sciences has published upper limits of acceptable exposure to impulse noise as a function of sould pressure level and signal duration as seen in Figure 13 (Ward, 1968) The solid line represents the upper acceptable limit of exposure for probably 95 percent of the population

PAGE 82

72 Figure 13. Upper limits of acceptable exposure to impulse noise for 95 percent of the population to 10,000 impulses. (Redrawn from Ward, 1968.)

PAGE 83

73 Table 5. Conceivable Sound Pressure Level Necessary to Elicit the Acoustic-Reflex Threshold by a 500-Hz Burst at a Signal Duration of 10 Msec Normal threshold at zero hearing level (ANSI-1969) in dB SPL 11.5 dB The ART at the upper limit of normal sensation level 100.0 dB Power increase necessary to maintain threshold when a tone burst is shortened to 10 msec 25.0 dB Total SPL 13 6 .5 dB to 10,000 pulses in some period of time, i.e., one hour to one day. This limit applies to impulses with a known single rise-peak decay, not a complex impulse. The Committee stated that, although there are many unknowns, their data for upper limits of exposure constituted a conservative estimate. Only the "weakest ears" of the unaccounted 5 percent might demonstrate a temporary threshold shift because the acoustic reflex would afford some protection in some subjects. The horizontal dashed line in Figure 12 indicates the maximum output of the earphone within this experimental system at 139 dB SPL and the vertical hatched lines indicate the approximate durations used in this study. As may be seen in Figure 12, the maximum output of the earphone falls below the maximum acceptable exposure limits. To minimize possible discomfort from loud acoustic impulses, the subjects were allowed some control over the test situation. If the subjects felt that the stimulus pulse

PAGE 84

74 was uncomfortable, they were instructed to press a button (see response switch III in Figure 9) which activated a tone audible only to the experimenter. The experimenter then reduced the stimulus 10 dB. This situation only occurred once because of experimenter error. A second control button (see response switch II in Figure 9) allowed the subject to discontinue the stimulus presentation at will. This second button was never employed by the subjects. Dealing with these high signal levels for long periods of time was bound to introduce some auditory adaptation or fatigue. The stimulus presentation rate of once every 1200 msec and a duty cycle of less than 50 percent was used as an arbitrary compromise between long, fatiguing test sessions and slow signal presentation rates in an attempt to minimize neural changes. In addition, at the end of each test run (10 thresholds per signal duration) the test signal was turned off in order to allow a twoto five-minute rest interval. During this interval, the signal duration or frequency was changed and voltage levels checked for calibration. Opinions of the subjects indicated that sessions of one and one-half hours approached the maximum tolerable limit. This was the time necessary to test the acoustic reflex of one ear. The possibility that the test signal would have crossed over to the opposite ear, the recording ear, was considered. If this had occurred, the data would have been contaminated in two ways. First, the reflex eliciting tone might have

PAGE 85

75 crossed over and combined with the probe tone, causing a reflex on the measuring side. This was unlikely, since a minimum of 4 0-dB to 50-dB difference in hearing level is normally necessary for crossover. The effective signal level reaching the probe ear side would have been 40 dB to 50 dB less than necessary to evoke the reflex. Since doubling the signal pressure increases the SPL only 6 dB, the increase in probe-tone pressure would have been a maximum of 6 dB and probably only 2 dB to 4 dB. This would not have raised the probe-tone intensity level enough to cause an acoustic reflex. Second, the reflexe 1 icitng signal that crosses over may have registered on the pickup microphone of the electroacoustic bridging network. This network has a filter centered at 276 Hz, and it is at least 20 dB down at 500 Hz. Thus, the filter should reduce by at least 20 dB any energy crossover at 500 Hz and possibly even at 1000 Hz and 2000 Hz. In conclusion, all reasonable safeguards for subject well-being and integrity of the acoustic reflex data were taken into account without compromising the experimental protocol. The subjects' responses, especially from those exhibiting loudness recruitment, indicated that the subject safeguards were not actually needed.

PAGE 86

CHAPTER IV RESULTS AND DISCUSSION The results of this study indicate that temporal summation of the acoustic-reflex does not function differently in those subjects who have normal hearing and in those subjects who have a known end -organ auditory lesion. A statistical analysis of the data did not support a consistent, significant difference between normal-hearing (good) ears and cochlearimpaired (bad) ears, using the acoustic-reflex threshold (ART) However, there were statistically significant differences between responses of the two groups at threshold of audibility. An analysis of these data follows in a comparison of normal-hearing and hearing-impaired ears at 1) temporal inte gration at the acous tic-reflex threshold 2) temporal integra tion at the threshol d of audibility and 3) temporal integra tion at the acous tic-reflex compared to temporal integration at threshold of audibility 1) Temporal Integration at the Acoustic-Reflex Threshold A discriminant analysis 2 (Rao, 1952) was performed on the ART difference between a 10-msec and a 200-msec duration tone burst and between a 20-msec and a 200-msec duration tone The raw data are reported in Appendix E. 2 Further information concerning discriminant analysis is contained m Appendix F. 76

PAGE 87

77 burst at each test frequency. In Figure 14, for example, the mean change for the cochlear-good ear at 500 Hz from 10 msec to 200 msec is 29 dB while the mean change at the same parameters for the cochlear-bad ears is 22 dB. The 20-msec to 200-msec difference was also chosen for discriminant analysis because the change was large enough to yield mean differences among groups (see Figure 14) and because some subjects did not yield an acoustic reflex at 10 msec with the maximum acoustic power of 139 dB SPL. All threshold differences and the data in some figures were subtracted from, or normalized to, 200 msec. Two-hundred milliseconds was most often stated as the time constant of temporal integration, and there was little difference between the ART elicited by a 500-msec and by a 2 00-msec duration tone. The mean change of ART between the two durations was 2 dB to 3 dB, depending on frequency and subject group. The figures were normalized to the signal duration which delineated the results most clearly. The results of discriminant analysis are reported as an "P" ratio. Table 6 gives the statistical results of the discriminant analysis upon the ART. Two comparisons out of 12 between good and bad ears were statistically significant, both at 500 Hz and both at the ART difference between 20 msec and 200 msec. Ten of the 12 good-bad ear comparisons were not significantly different. The mean ART values of the cochlear-impaired ear, as demonstrated in Figure 14, were always less than those of the

PAGE 88

D3SW 00 3 IV CTI0HS3HH.L 3H 9P criOHsaHHi xauaa-oiisnoov

PAGE 89

79 good ear of the cochlear-impaired group and the normalhearing group. However, neither the mean values, nor the average slope per decade of stimulus time, were significantly different among the three groups (Table 7 ) Table 6. Discriminant Analysis of the Acoustic-Reflex Threshold 500 Hz 1000 Hz 2000 Hz Normal vs. cochlear-good ears 10 msec-200 msec ns ns ns 20 msec-200 msec ns ns ns Normal vs. cochlear-bad ears 10 msec-200 msec ns ns ns 20 msec-200 msec .05 ns ns Cochlear-good vs. cochlear-bad ears 10 msec-200 msec ns ns ns 20 msec-200 msec .10 ns ns ns = not significant. .05 = significant at the 5 .10 = significant at the 10 percent percent level level Table 7 Average Slope Change Per Decade of Time of the Acoustic-Reflex Threshold 500 Hz 1000 Hz 2000 Hz Cochlear-good ears Normal ears Cochlear-bad ears 23 dB 17 dB 15 dB 22 dB 19 dB 14 dB 25 dB 19 dB 14 dB

PAGE 90

80 Although the group means were not statistically different, the data from Figure 14 indicate two consistent findings. First, the cochlear-impaired (bad) group always had less mean ART change than the other two. The cochlear-good, for the most part, had the greatest amount of ART change. The difference between the cochlear-bad and normal ears at 20 msec increased from between 4 dB and 7 dB to 12 dB for 500 Hz, 1000 Hz, and 2000 Hz, respectively. There were no other large systematic changes in this respect. Second, the change from 20 msec to 100 msec (or to 200 msec) became less as frequency increased for the cochlear-bad ears. This decrease in mean slope caused the former results of increased ART difference between normal and cochlear-bad groups. Figure 15 illustrates the large inter-subject variation of temporal integration slopes for the normal ears and the cochlear-bad ears. Similar configurations also appear in Figure 16, contrasting cochlear-good ears and cochlearbad ears. In both figures, the slope of individual subjects varies over a wide range within a group, so that the normal and cochlear-impaired groups overlap. The large intersubject variability is probably the reason why the mean values were not statistically different from one another. Figure 17 contrasts the inter-subject range among the three groups. This graph also depicts the large amount of overlap among the three groups, which hampered statistical significance. The least amount of overlap occurred at the test frequency of 500 Hz and

PAGE 91

81

PAGE 92

82 o o in o o CN O o o CN o o o in o w M o o CN O O o CM o o in o o CN o o o CN 2 H z o i D Q J < Z U H CO 0 +J o CO n rd (U Q) M H rd s H T) C rd 0 0 rd • 6 C M O O -H a -p rd o i CO CO a o CO +1 CD O H U-i CD M I C o •H -P o o H fi -P 3 CO 4-1 3 O rd O < CO rd SO o n a) CO D3SW 00 S IV aT!0HS3HHtI 3H HP aiOHsaHHi xauaa-oiisnoov M •H CD h P CD H C

PAGE 93

83 T Q O 33 CO as Eh X w a fa ? u § H S Eh Eh CO D W o 2 u < CQ o w CO a o o m Eh < P O CO 40 30 20 10 0 40 30 20 10 T I I o — 40 -\ 30 20 100 rl TT i! ir i i i I 10 TTit i -i-i T i ; 20 H Normal ears —i Cochlear-good ears H Cochlear-bad ears 500 Hz i I J. 1000 Hz ItI 2000 Hz TT T l 100 if _L I 200 500 SIGNAL DURATION IN MSEC Figure 17. Inter-subject range of acoustic-reflex thresholds as a function of temporal integration.

PAGE 94

84 would account for the statistically significant findings at 500 Hz. Using discriminant analysis, the more distinct the groups the less chance of error in classification (Rao, 1952) It is notable that the point of greatest significant difference (p< .05 at 500 Hz for normal vs. cochlear-bad) yields the smallest difference between mean values (Figure 14) Figure 18 compares the temporal integration slope of the acoustic reflex of normal-hearing persons in three studies. The individual mean scores from this study and Djupesland and Zwislocki (1971) are plotted as well as the median values. The group mean values from McRobert et al. (1968) are also included. The average change in slope per decade in time at 1000 Hz for this study is 6 dB less than the 21 dB/decade from Djupesland and Zwislocki, as well as 2 dB less than the 25 dB/decade from McRobert et al. In addition, it appears that the individual mean values for this study were more scattered than those of Djupesland and Zwislocki (1971) especially at the 10-msec signal duration. This large scatter of individual scores may be representative of the true population, since no more than nine normal-hearing subjects were employed in any one study. These differences may also be a result of the type of equipment employed, since the two lower values of 19 dB and 21 dB/decade were obtained with a Peters electroacoustic bridge and a Madsen electroacoustic bridge, respectively. The steeper slope of 25 dB/decade was obtained with a

PAGE 95

85 1000 Hz 3 this study ART for each ear Median for 10 ears ART for each ear-> Median for 6 Ss J D 3upesland Mean ART for ca. 9 Ss^ McRobert 40 u w £ Q O K Ul o o m X & X w fa I u M Eh U) o u < Eh < D i—l O X CO Eh CQ T3 30 20 10 -0 _l 1 1 H 10 20 100 200 500 SIGNAL DURATION IN MSEC Figure 18. A comparison of mean acoustic-reflex thresholds as a function of temporal summation in normalhearing ears.

PAGE 96

86 Zwislocki mechanical impedance bridge. There have been reports indicating that the electroacoustic and mechanical impedance bridges can vary as much as 35 percent in their impedance values, depending upon the condition and size of the tympanic membrane (Lilly, 1970; Wilber, Goodhill, and Hogue, 1970) The statistical observations concerning the null hypothesis of the ART were limited because of unexpected inter-subject variability causing the sample size of five normal and five cochlear subjects to be ineffectual statistically. However, descriptive observations can be made, although these observations may be due to chance. The time constant of temporal integration of the ART, that point in time where integration begins, 3 is shorter than the 2 00 msec reported for threshold audibility. A shortened time constant has been reported previously by Small et al. (1962) for temporal integration at suprathreshold levels. As illustrated in Figure 19, the ART time constant is most often between the signal durations of 100 msec and 20 msec, or the 100 msec-20 msec interval. Further inspection of Figure 19 suggests that the time The time constant was defined in this experiment as 1/e of the maximum threshold change. In practicality, 1/e was 1/2.718, or 32 percent, of the mean integration slope per decade of stimulus time in normal ears, or 6 dB. The first 6 dB of threshold change from 500 msec, therefore, was not considered part of temporal integration. The point at which 7 dB, or more, of threshold change occurred due to change in signal duration marked the length of the time constant.

PAGE 97

87 I w Eh 2 2 O fa o D 2 12-1 8 Eh 4 Q < 2 H Eh H CD U O o Eh § 20 o u 8-i 4 16 12 8 4 63% .12% 0% 27% 27% 46% 0% 67% 500-200 200-100 100-20 20-10 DURATION INTERNAL IN MSEC Figure 19. Signal-duration interval in which the time constant of temporal integration for each test subject occurred across all frequencies.

PAGE 98

88 constant of the two good-ear groups is never shorter than a duration of 20 msec. Interestingly, the opposite observation may be made with the cochlear-impaired group. There were no long time constants in the 500 msec-200 msec interval, but 25 percent of the cochlear-impaired ears, across all frequencies, demonstrated integration functions that were in the 20 msec-10 msec interval. That is, temporal integration for these 4 out of 16 ears may be abnormally shortened due to the cochlear impairment. This can be most effectively seen in Figure 20, where three of the cochlearimpaired ears with a short time constant are compared with the normal-hearing results at 1000 Hz and 2000Hz. A shortened time constant has been previously reported in cochlearimpaired ears for temporal integration at threshold of audibility (Figure 5) Inspection of Figures 15 and 16 demonstrates that an average slope value per decade is not necessarily the best descriptor of temporal integration. The change is not linear, especially in the cochlear-impaired group. The cochlear-bad ears at 500 Hz integrate slowly, with a flatter slope, until 20 msec when the rate of change becomes greater. For example, four out of five cochlear-bad ears at 500 Hz change at an average rate per time decade of 8 dB to 2 0 msec, and then abruptly change to an approximate slope of 50 dB to 55 dB per time decade. This change is only distinctive between good and bad ears at 500 Hz. This difference

PAGE 99

Figure 20. The delayed time constant in some cochlearimpaired ears contrasted with the normal time constant of temporal integration at the acoustic-reflex threshold.

PAGE 100

90 probably contributed to the only two rejections of the null hypothesis which occurred (Table 7) This abrupt change in slope occurred in the 2 0 msec10 msec interval, the same interval in which the short time constant of cochlear-imparied ears fell in Figure 20. Since the definition of the time constant was relatively arbitrary, the slow-abrupt change in slope and the short time constant may reflect the same phenomena in the cochlear-imparied ear. From the discriminant-analysis data one is able to identify those subjects consistently classified correctly, although no statistical significance is attached. Two §oehlearimpaired subjects, FH and BC, were never classified as normal hearing, while EC was incorrectly classified in only 2 out of 12 good-bad ear comparisons. Of the five test subjects, the data from these three indicated that one §huld be able to distinguish a cochlear-impaired ear based or ggeystic-ref lex data. The cochlear abnormality may g§\i§e aberration in temporal integration of the ART as ifliie^ted by an initial slow growth of the integration functiR @r by an unusually short time constant. One cochlearimpai?§d subject is an appropriate example. LS was incorrectly classified 9 out of 12 times. Two of the correct elgsgif ieations were probably based on her very short time @n,a%a.nt, Her ART changed only 2 dB as the signal duration W3§ §h@tened from 500 msec to 20 msec, then it changed afe?\jptiy by 35 dB. She is represented in Figure 20 at 2000 Hs fey the white hearts.

PAGE 101

91 In addition, the acoustic reflex occurred at similar levels reported in Tables 1, 2, and 3. In Figure 21, the pure-tone thresholds are separated by 20 dB to 4 0 dB, but the acoustic-reflex thresholds are overlapping. The cochlear-good ear group is depressed somewhat at 500 Hz, but this is because the sample size was three ears at that parameter. The ART was established as the mean value of 10 trials. This may have been unnecessarily long since there was no significant difference between the mean of the 10 and the mean of the first 5 trials. The modified Hughson-Westlake approach (Carhart and Jerger, 1959) would have been as useful as the mean of 10 trials, but for the fact that no singl value could be obtained 50 percent of the time in some instances. Using 2-dB increments, the average range of values obtained for threshold in any 10 trials was 4 dB to 6 dB. A range of 10 dB was rarely obtained. For the most part, the shorter the test duration the more variable the response became. A mean of 5 trials would seem to yield a stable ART in half the time used in the present experiment. 2) Temporal Summation at Threshold of Audibility ^ A discriminant analysis of temporal summation at thresh old of audibility between normal-hearing ears and cochlearimpaired ears indicated significant good-bad ear differences 4 The raw data are reported in Appendix G.

PAGE 102

92 CM e o H Q o EC W aa Eh 100 90 80 70 60 50 A O Normal ears Cochlear-good ears Cochlear-bad ears Audibility threshold Acoustic-reflex threshold „-•250 500 1000 2000 TEST FREQUENCY 4000 Figure 21. Thresholds, audibility, and acoustic reflex, of the test subjects at 500-msec signal duration.

PAGE 103

93 10 out of 12 comparisons. The results are listed in Table 9. There was no statistical difference between cochlear-good ears and cochlear-bad ears at the 20 msec-200 msec comparison in two out of three frequencies. Further, there was no statistical difference between normal and cochlear good-ears in all comparisons. Table 8. Discriminant Analysis of the Threshold of Audibility Data 500 Hz 1000 Hz 2000 Hz Normal vs. cochlear-good ears 10 msec-200 msec ns ns ns 20 msec-200 msec ns ns ns Normal vs. cochlear-bad ears 10 msec-200 msec .10 .05 .05 20 msec-200 msec .05 .05 .05 Cochlear-good vs. cochlearbad ears 10 msec-200 msec .05 .05 .05 20 msec-200 msec ns .10 ns ns = significant. .01 = significant at the 5 percent level. .10 = significant at the 10 percent level. As with the ART, there were no significant differences among the mean values of the three groups, as presented in Figure 22. Again, the values for the cochlear-impaired ears lie below those for the normal-hearing ears. The range of inter-subject thresholds was not as great as that found with

PAGE 104

94 a a Normal ears o— o Cochlear-good ears • • Cochlear-bad ears 10 20 100 200 SIGNAL DURATION IN MSEC Figure 22. Mean thresholds of audibility as a function of temporal integration.

PAGE 105

95 the ART, and there was not as much overlap between the badear and good-ear groups. The inter-subject variability is illustrated in Figure 23. The data from temporal integration at threshold compared favorably with those of previous investigators. In Figure 22, the mean normal-hearing integration slopes per decade of stimulus time were 8 dB to 11 dB. The normal ears demonstrated a frequency effect whereby the slope flattens as test frequency increases. These slopes were 10 dB, 9 dB, and 8 dB at 500 Hz, 1000 Hz, and 2000 Hz, respectively. No systematic effect was seen for the cochlear-good ears, but the sample size was smaller than that of the normal ears. At each frequency in Figure 22, the mean slope for cochlear-bad ears was flatter than, and below, either of the normal-hearing groups. Since the mean slopes are between 6 dB and 7 dB for the cochlear-impaired ear, the distinction between normal-hearing and cochlear-impaired ears is rather subtle. The lack of clear inter-group distinction in Figure 23 further points to the problem of utilizing individual results for diagnostic information. Only those subjects with a slope of less than 5 dB/decade would have been distinctly classified as cochlear-impaired. Olsen, Rose, and Noff singer (1973), with a larger experimental population, reported normal subjects with slopes as flat as 2 dB to 3 dB. They concluded that the inter-group "... overlap was sufficiently great that no characteristic pattern of behavior defined any group."

PAGE 106

96 40 30 20 10 1 JtI i -L — I Normal ears — I Cochlear-good ears -I Cochlear-bad ears 500 Hz III TTT u w CO s o o Eh H H D D < O E-t < D O CO D O P K K CO E-t PS w w £-• CO 40 30 20 10 0 Irl i Irl 1000 Hz IT x iil i 40 30 H 20 10 0 H i It 10 hi 2000 Hz [II ill 20 100 200 SIGNAL DURATION IN MSEC 500 Figure 23. Inter-subject range of thresholds of audibility as a function of temporal integration.

PAGE 107

97 The time constant of temporal integration at threshold of audibility in cochlear-impaired ears is said to be shorter (Figure 5) Table 9 lists the signal duration 5 intervals within which the time constants occur. The results are reported in both the number of times and the percentage of times a time constant was within a certain interval for each group. Even though the cochlear-impaired (bad) ears had the highest percentage occurring in the 100 msec20 msec interval, there was not much difference between normal and impaired ears. These results could not support the statement that cochlear-impaired ears have shorter time constants than normal ears. Table 9 Signal Duration Intervals Within Which Time Constants of Temporal Integration at Threshold of Audibility Occur Si gnal Duration Interval in Msec 400-200 200-100 100-20 20-10 Total Responses Normal ears 4 (13%) 9 (30%) 17 (57%) 30 (100%) Cochlear-good ears 5 (36%) 5 (36%) 4 (28%) 14 (100%) Cochlear-bad ears 2 (13%) 3 (19%) 11 (68%) 16 (100%) The time constant was calculated to be at or above the first 3 dB of threshold change from 500 msec. The time constant criterion is the same as the ART time constant (footnote 3)

PAGE 108

98 3) Temporal Integration at the Ac ousticReflex Threshold Compared to Tempor al Integration at Threshold of Audibility There were no significant correlations occurring consistently between the data on temporal integration at the two thresholds, ART and audibility. There were no correlations on direction or amount of change between the two integration functions, nor were there on classification of ears. This indicates that abnormal temporal integration at threshold of audibility does not predict the results of temporal integration at the ART. It must be remembered, though, that the threshold criteria for the two thresholds were not only different but also involved different persons. The experimenter specified the ART and the subject determined his own threshold of audibility. The difference in criteria may be reflected in many different ways, e.g., a different point on the intensity function for threshold, i.e., 25 percent vs. 75 percent of correct responses. Even though there were no significant comparisons between the two thresholds, some observations can be made. The mean temporal-integration slopes of the cochlear-impaired ears fell below those of the normal-hearing groups with both thresholds. The ART slopes yielded much larger mean differences among groups, but also greater inter-subject variability, than those at threshold of audibility. The most obvious difference between the two types of thresholds is the slope per decade of time. The slopes of the ART in 6 The correlations are listed in Appendix H.

PAGE 109

99 Figure 14 are approximately twice those at threshold of audibility of Figure 22. This gives ART measures the possible advantage of a greater dynamic range to determine differences between normal and abnormal ears. Unfortunately, the large inter-subject variability and group overlap has eroded this possible advantage. The time constant of integration was essentially the same between normal-hearing and cochlear-impaired ears at threshold of audibility but it was definitely shorter for cochlear-impaired ears at the ART. The cochlear-impaired ears at the ART were the only ones to have a time constant of 20 msec or less. This may be a result of the previously mentioned larger dynamic range for temporal integration at the ART. Nevertheless, the shorter time constant is one of the most distinguishable differences between temporal integration at the ART and at the threshold of audibility. r~e It should be mentioned that only one person was correctly classified for each parameter across all frequencies-^ at: both thresholds. FH had a flattened slope, as well as ar. short time constant of less than 2 0 msec.

PAGE 110

CHAPTER V CONCLUSIONS AND SUMMARY A concluding comment should be made concerning the nature of diagnostic tests, as viewed through these results. If the ART did indeed prove to differ cochlear abnormalities from normals, what can be considered abnormal? A specific defect or aberration in auditory processing does not necessarily reflect one specific and correlative anatomical lesion, if, indeed, it is possible to obtain a homogeneously impaired population. The experimental group is a good case in point, because they were chosen as a homogeneous population with classical Meniere's syndrome. Three subjects, EC, FH, and BC, had Meniere's for less than 6 years. They also had the most depressed slopes. The other two subjects, LS and AC, had greater than normal slopes and had had their disease symptoms for 16 to 35 years, respectively. LS, as well as FH and BC, demonstrated a short time-constant for at least one frequency, so, apparently, it does not matter how long one has had the disease. In addition, these results are not uniform from one frequency to the next for each subject, even though the hearing losses were relatively flat. It did not matter if a clinically significant hearing loss was present, because BC had practically normal pure-tone 100

PAGE 111

101 acuity. This is not to say that temporal integration of the acoustic-reflex threshold can detect incipient hearing loss or minor cochlear impairment. Instead, it reflects the complexity of auditory processing. In conclusion, there vere some apparent, but subtle, differences between the normal-hearing and hearing-impaired groups in temporal integration at the ART o The mean integration slopes of the cochlear-impaired group were depressed. o The time constant of integration was shorter than 2 0 msec for some of the cochlear-impaired groups, o If the time constant is not unusually short, the integration function changed gradually, then abrubtly after 20 msec in the cochlear-impaired ear. o A cochlear-impaired ear can have both a flattened slope, as well as a short time constant, o One did not have to have a hearing loss in the cochlearimpaired ear to demonstrate either a depressed slope or a delayed time constant. Since it has been stated that there are probably differences between normal-hearing and cochlear-impaired groups, this type of study is open for further investigation. A similar experiment would require more signal durations, expecially below 100 msec. Other parameters can be expanded upon, such as type of stimuli or measuring ipsilaterally instead of contralaterally to the test ear. Temporal

PAGE 112

102 integration measures do not have to take place at threshold, nor by changing the stimulus durations only. This type of investigation should be pursued because the acoustic reflex can reflect a cochlear abnormality without the subject's cooperation. In summary, five normal-hearing and five unilaterally hearing-impaired persons were test subjects. Their results determined the efficacy of temporal summation of the acousticreflex threshold as a possible predictor of cochlear abnormalities. Temporal summation of the acoustic-reflex threshold was obtained by maintaining a constant intra-aural musclereflex strength as a function of increased acoustic power while signal duration was decreased. A measure of temporal integration was made at both threshold of audibility and threshold of the acoustic reflex, the first auditory measure was determined by the subject and the second by the experimenter. Bekesy tracking was employed to self-record threshold of audibility, while a modified method of limits, using 10 ascending trials, determined the acoustic-reflex threshold. The signal durations used were 500 msec, 200 msec, 100 msec, and 20 msec at the test frequencies of 500 Hz, 1000 Hz, and 2000 Hz. As much as 139 dB sound pressure level was employed to maintain constant energy at threshold. The reflex was monitored from an oscilloscope connected to a Peters 61AP electroacoustic bridge.

PAGE 113

103 Temporal summation of the acoustic reflex does not function differently in those subjects who have normal hearing and in those subjects who have a known end-organ auditory lesion. However, there were statistically significant differences found between the normal-hearing (good) ear and the cochlear-impaired (bad) ear comparisons of the 20 msec-200 msec ART difference at 500 Hz. Small sample size, and unexpected inter-subject variability and flatter integration slopes for normal ears, contributed to the lack of statistical significance. Temporal integration at threshold of audibility, in contrast, did demonstrate statistical differences between cochlear-impaired ears and normal-hearing ears. Even though a statistical difference was evidenced, the difference was not so obvious, in this experiment, as to be a clinically useful tool. A comparison of temporal integration at threshold of audibility and at the ART demonstrated that a short time constant for the ART may be of diagnostic value in determining cochlear impairment. However, the question of whether temporal summation of the acoustic reflex is clinically feasible as a diagnostic measure of cochlear abnormalities was not resolved in this experiment. Any observations made, e.g., the short time constant, were not proven to be other than chance observations. ART integration, therefore, could not be safely said to identify an end-organ impairment.

PAGE 114

APPENDICES

PAGE 115

APPENDIX A MEDICAL HISTORY FORM

PAGE 116

Subject Medical History Form 1. Name Age Sex 2. Address 3 Date__ 4. Subject group: Normal Cochlear 5. History: A. Conductive 1. Recent middle-ear problems: Yes No Explain 2. Ear operations: Yes No_ Explain_ When Intra-aural muscles involved B. Sensori-neural hearing problems: 1. Ear: L R Both 2. First noticed Duration 3. Diagnosis By whom 4. Meniere's disease a. Episodes of true vertigo b. Feeling of fullness of ears c. Tinnitus d. Fluctuating hearing loss e. Frequency of attacks f. Time of last attack g. Present state: Active Quiescent Unknown 5. Hearing aid: Yes No Don't use C. Brain Injury: Yes No Severe concussions: Yes No CVA: Yes No D. Medication or drugs currently used: Yes No 106

PAGE 117

APPENDIX B PURE TONE HEARING LEVEL (ANSI-1969) OF ALL SUBJECTS

PAGE 118

Normal Group SubJ CivU Juul Classif iLa L XvJIl Frequency in Hz 1 TC J Ten 500* 1000* 9 n n n JUUU A n n n t u u u JG L good 10 10 15 18 18 20 10 R good 15 10 20 16 17 15 15 CP L good 0 0 1 7 8 5 0 R good 0 0 1 4 11 5 5 GP L good 5 5 7 9 15 20 15 R good 5 10 10 1 2 0 5 JP L good 0 0 -8 1 4 5 0 R good 0 0 -8 2 4 0 0 FS L good 0 0 -6 -7 -8 0 0 R good 0 0 -10 -4 -3 0 10 *Actual values obtained by Bekesy tracking. Meniere's Group Subject Ear Classification Frequency in Hz 125 250 500* 1000* 2000* 3000 4000 AC L bad 65 60 59 68 65 70 70 R good 20 20 19 16 13 35 35 EC L bad 65 60 61 50 57 50 50 R good 10 10 6 5 8 25 45 FH L good 5 5 6 15 14 30 55 R bad 30 25 30 35 49 50 55 BM L bad 25 20 18 22 15 5 10 R good 20 5 8 11 13 10 10 LS L good 40 40 39 19 7 10 15 R bad 45 40 47 46 40 50 50 Actual values obtained by Bekesy tracking. 108

PAGE 119

APPENDIX C VOLTAGE DIVIDER

PAGE 120

01 -H o o O O 6 as 8 0 fri 4J O to 3 6V d.c. 3K ohm to cn a, o +> rH o 3
PAGE 121

APPENDIX D CALIBRATION DATA TABLES

PAGE 122

Table 1. Frequency Response of Earphones in dB SPL (re 0.0002 dyne/cm 2 ) 500 Hz 1000 Hz 2000 Hz 2 volts input TDH-39 129 128 127 TDH-14 0 131 131 131 5 volts input TDH-140 139 139 139 112

PAGE 123

113 Table 2. Step-Attenuator Linearity. These Data Were the Preand Post-Experimental Values for the TDH-140 Earphone at 1000 Hz in dB SPL (re 0.0002 dyne/cm2) 1-dB Increments 10-dB Increments dB Pre Post dB Pre Post 0 139.1 139.0 0 139.1 139.0 -1 138.2 138.0 -10 130.0 129.9 — 2. 137 4 137.2 -20 120 1 120. 0 -3 136.4 136.2 -30 110.1 110.0 -4 135.4 135.3 -40 100.3 100.1 -5 134.6 134.4 -50 90.4 90.2 -6 133.7 133.5 -60 80.6 80.4 -7 132.8 132.6 -70 71.2 71.1 -8 131.9 131.8 -9 130.9 130.8 -10 130. 0 129.9

PAGE 124

114 Table 3. Recording Attenuator Linearity. These Data Were the Preand Post-Experimental Values for the TDH-140 Earphone at 1000 Hz in dB SPL Pre Post 0 114 4 114 0 -10 94.0 93.8 -20 83.4 83.3 -30 73.4 73.2 -40 63.6 63.1 -50 53.6 53.2 -60 43.6 43.2 -70 34.3 33.8

PAGE 125

APPENDIX E RAX DATA OF THE ACOUSTIC-REFLEX THRESHOLD

PAGE 126

in rLi in i-r> in n in CTiC7\0")J*. O CT> C"> CTi C\ r— crvcicyicricyicyicricACftcrv cr> o cti cr> infirominininnciin (mono t^r-"inr-inr-mr^r~~r-vo 10 ko cn CTiCriC^0 Cr\OCriCr>CT>0 i OOOO oincMLomror-ooon CNCNCNCNCNCNCNCNrHCN i— Ir— IrHi— (rHf— liHi— Ir-li— ( omominominoin co cti oo ci H H H (N r~ in r~ rt-~ r> I r> in OJ (N (N (N (N CN I CN OV0C0*£>C0O00 c^c^c^oo^cr\c^c^ocri COOOOOTW^frfCDlO COOOC^CTiCOOOOOCOOOOO ID VO -cr OOOO OOOOrHrHrHiHrOrH OOOOOOOOOO rHrHHr-HrHrHrHrHrHrH o oo o o HOHH (NHrHlNHHrlHHH rHrHrHrHrHrHrHrHrHrH inr^t > -ino^r^r^ai'* l£) I VD I CNCNCNCNCN I CN CN CN I vDTrookD'j'^orroo^Doo CNCNCNCNCNCNCNCNCNCN m i en n C^CTiCTiC^CnCACrtCTiO^CTi ^rsrkOVOC0TjkO^jkOkD ^ CO CO 00 00 00 00 00 00 00 00 00 00 OOOO vovoooio^roococovovo CTiCT>CTiCriCTiCr>Cr\CTiO>0^ OCOCOCNCNrrCNCN^VO U) ^ CO cricococricr\cr\cr\CT\cT\cri oooo rH rH rH rH COCOHCOHOOOOrl cncrioooooooo rH rl rl rl rl H H cocoooococoocococo CO CO ^ CO cncr>cr>ocricriocr>cricri oooo •H rH rH rH rH rH 00HHCOMHHHfn> OHrlHHHrfHHrH HHHrlHHHHrlrl vDOfMVD^rvDVD^r'3'rr in i — o oo rHrHrHiHiHrHrHHrHH H H M H HrlHHHHHHrlH i— I rH i— I rH fOCnCTSlXIrHCOCOrHVOVO CNrHrHCNCNCNCNCNCNCN HHHHrlHHHHH vD^rvoMj-cjivDrr I oooo CN CN CN CN CN CN CN I CN CN PS rH rH rH rH rH rH rH rH rH 2 rHCNro^rinior^cocno HCNm^rinioooocno HCNcrt CJ u 4 116

PAGE 127

117 c % a o a i i w 5 O in o o CM ti tr" nm ^> o O rH o o CN CN O r4 o O o o CO in S c o -H o N CN PC c o o -H O o -P o o td rH rH rH Q o CN H cd C D> O •iH rH o in o o CN N o o rH o in O CN O rH r-rco nririHHnHninn HHriHCM/iHciHH CT\ O CTv CT> CTi CPi OOOOOOOOOO HrlrlrlOrlHHHH 0) £J CD 3 I -P •O o CN O CN CN i COCOCOCOCOfOCOCOCOfO rHrHrHrHrHrHrHrHrHrH CNCNOCN^rCNCN^CNO OOOOOOOOOO V0V0V000C0 00VD00O00 OOOOOOOOrHO Lnininnintn^-cfiroin rHrHrHrHrHrHrHrHrHrH rHrHrHrHrHrHrHrHrHrH r~~cninr--r~-r--crir-r-r~rHrHrHrHrHrHrHrHrHrH rHrHrHrHrHrHrHrHrHrH fOiHcorHcoinininrHCTi fOCOCOCOCOCOCOCOCOCN rHrHrHrHrHrHrHrHrHrH CO I CO 00 00 I I CO 00 00 co I co co co I I n co co vovoj3U3V£>Troovov0oo OCNCNOCNO00 00 00 00 OOOOOOCf\Cftfl\Cfi OOCNCNOCN^OOO kDOOOO^COCO^rVDCOCO rHrHrHrHrHrHrHrHrHrH (TlC^C^O^C^CACT\CX>G>Crt OC0OOOTO00OCN VOVDVOOOOOCTiCOOOOCn COCNCOCOCOCOCOCNCOCO iHrHrHrHiHrHrHrHCNrH 2 rHiHrHrHrHrHrHiH rHOCNOCOCNinincOCN CNCNCNCNCNCNCNCNCNCN rHrHrHrHrHrHrHrHrHrH incocoinincocoininco cricriO~\r^crir~~cr\r~-o^r~^ HHH rHrHrHrHrHiHrHrHrH OOrHrHCOrHrHrHCOCO CTlHHHCOCOHCOHH rH rH rH rH rH rH CTiOOOOOOOOO incocococot-^mininco HrimHHHCftMnn rHrHrHrHrHrHrHrHrHrH OOCTlOOOCTlOOO rHrHrHrHrHrHrHrHrHrH H H HHH HHH r^-inr^r^HHcricricrir^ ocoinincoininr-^inin CNCNCNCNfOCOCNCNCNCN H H H H H HHH in vo r— co cn o H S5 HCNco^rinvor-oocno H H H H H H cNCNvor~incovD^or--vo CNCNCNCNCNCNCNCNCS1CN HHHHHHHHHH HCNC0TTinvDr~00CT\O

PAGE 128

118 i w •H a SB o o o CN N X CD O O N K o o in u 0) cn £ C •H c o •H -P cd a id c •H CO o o m o o CN O O O CN O O iri o o CN o o o CN o o in o o CN O o o CN O H -P crcncTio>CTicTicricr> nnHooHHrironn o~\CTvcri CXi CTi OOHHHHHHHH rH CN rH rH rH rH HHHHHH O00O00CN0000OO00 CO CO CO CO CO o chcocTicocricococricrioo cooocooooocri (N'fl'inO'J^'^'rrVD'J O CO O CN O CN VOlOCDlO'l'lDCDCOlDlD ^ CN CN CN CN G\CT\ ci d rcri r~~ cooocooooooo cr> cr> a\ coincricnr-cricriHrcnHffiHH(riHH h in in h n n (J\
PAGE 129

119 o c •H -P C o I I w X H T3 C QJ < o o U I O t — i < N E o o Q o o CN o CN O rH r \ KJ C; \L> tf\ UJ <— • O — N N r-rH U O o *H o 4-* CD c — 1 frt *U H I— 1 CN p| frt c /— N •H iH UJ O o U 1 o o N E o o i — i o 1 fS Li ) o CN O rH -P C 1 +) X! O 0) CO •n CN CN CA 00 O ^* CN CN CTl CTi CTl C7i CN O CN O CT\ CTv CTi CTi CN O CN CN O O O O rH rH rH rH rH rH o o rH rH r~roi cri a\ C0O0000V£> OOOOOOrHOOO HHHrlHHHrlHH V0O00OC0OC0<£>C000 rHCNrHCNrHCNrHrHrHrH rHrHrHrHrHrHrHrHrHrH nroMnnrocofOfon r-ir-it-{r-{r-it—{t-{r-{r-tr-i r~r--nmcTir~~c7icrio icr\o*\cTicT\a> ^^•^TCNVOVDTTUDVDVr) cTic?s(TicriOAcricricricri ooooooooooooo cficnoooooomo rH rH rH rH rH rH rH ininroLDHHcommtri rHrHrHrHrHrHrHCNrHrH roinaincriHHinHH CNCNCNfOCNnnnnm CNCNCNOCNOCNCNCNCN OOOOOOOOOO t-trHr-tT-\T-\r-\r-iiH^-ir-{ ^CNCNCNCNCNCNCNrrCN OOOOOOOOOO r-ir-ir-fr^r-ir^r-tr-if-if-i OOOOOOOOOO r-tr-Ai-tr^r-ir-ir-it-i^i-i *3VOVOCN*3 ,, <3 V£)Ta'3 T}< rocNCNCNCNCNncNnn t-ir-it-{r-ir-\<-tt-{r-{r-tr-i OOOOOOOOOO rHrHrHrHrHrHrHrHrHrH OOOOOOOOOO rHrHrHrHrHrHrHrHrHrH f-CTiCXiCiCriCTiCXvCriCTiCri OOOOOOOOOO rHrHrHrHrHrHrHrHrHrH CTiCTlrHCnClrOrHdrHrH rHrHCNrHCNCNCNCNCNCN rHrHrHrHrHrHrHrHrHrH COCNCO^t^VDrrVDOD rHrHrHrHrHrHrHrHrHrH r-cocfto rHCNrn^ririMDr-cocTiO rHCNn^invDr^cocrio Pa a

PAGE 130

120 oooococooocooococooo OOOOCMOOOOOOOOCvJ VO CM T t} i-HrHOrHrHrHOOOrH CTiCTiO^CTi rHr— IrHrHrHrHi— li— IrHrH CCDCOOOOCDVDCn3VD 0OO00OCTN00COCOCOOO0O OUOfN^fOO(N CTi (Ti HrlrlrlHHHHHH COCOVOCOOOOOOCN cococococricricricricricr* VDfM^DVDTOOfMOOOOOO O O O iH0oocriOO oooo WU)TTCOVOVDU5'rHr^H inr~r~crir~r--inr^min in ca m in oocr> en r--crir~ n 01 n in HrlHHHHHHHH HfMHHHHHHHH CNrHCNCN HHrlrlrlHrlHrlH HHHHHrlHHHH H H H H aininni^oM^nint^ innr^a\rninr~mcAr~ en r*r m HHrlrlrlrlHHHH (NCNCNCNCOCNCNCNCNCJ CN (N t^oo
PAGE 131

121 N E o o o CM >0 dJ c •H -P c o I I w X -H •d a a, a < a o o o N E O O in Q r\l -t1 ^ff ^t* *Tt CN o on on on on on on cy. cy\ Ol o> cn on cn in • — rri t — i rxi rrt m < — ^ m m n< / — \ ^ — t -J r*j t — s — 1 o CT O O (Ti O o o o o o o o c o O CN J I _j rn m m _J __1 __l rn j r— J 1 r-1 (Yi fNl ( — > r\l rn CO O o o UD o CO o O O CXi O O O O O O O rH rH rH O o rH rH rH r~i Pn m m H HH H H rH rH rH rH rH m ro n c> rLO on CN n n n fN ro ro CO CM CN ro CO CO CO __j _i __j _j _j __j ~ ~ 1 rn rn rn rn rHi—tr-Hi— IrHl— ll—ll—li— IrH O a. cn in m rH CO CO CO CO CO CO Pi 1 __i __l _J __I __J r~i rn rn r~i r~i ri lz V > lO Ift m y\ r"Y*l ff\ O O CO o O CO CM o c o rt\ m /T\ /"V\ rp\ o o cn o O cn o c o o rn 1 r~i rH rH rH H rH rH rH rH H /-v-\ /-^i /~-\ /"y~> cn n< o CM H fT\ (~^t m CT\ (* — *l o o o o o o o o o O CN rH rH rH rH HHH rH rH rH rH rH rH rH H *-> u •H <• \D UJ \L> CN CN CM CM CN O -t— rn o o o o (3 Q ^ rH i r i rn m m rn rH rH rH rH H rH rH rH rH r-i M I— J rn ^-fi IA *-M I O rvl VJ Vl^ ^J" \J l N o CM o IfJ o CM CM CN CN CN CN CN CN cn co co CO ro ro m CN CO ro j rn i i m i i rn rn rH rH rH rH rH rH rH rH rH rH rrt c tr -H — j CM CO CO CO CO CO co co co ro ro rn co ro CO CO rn j _i _j __j _j _j rn rn rn m rn rn r-\ H r-i rH rH rH rH rH rH rH lCi rr* __i /-n #-n ro Cj^ C7\ 1 U 1 (1) u h o io in a\ cr o 0\ i rn i i j m v_j i o fT\ CTj rn rn rr\ rH r—i | p— j p-j E H <-n UJ r* ; IT) LT) rn rn M rvi /-,! <>N.l /-St iM /SI a) CN v N CN CN CN CN rd rH rH rH rH rH i—j rH O y. o rH CO rH CTt CO r-* o u cy rH CO CO CO CN CO CN H 0 3 r-i t-i r-\ r-\ <-\ r-\ U-l Eh H 1 +> fl o PS o CO •n CO !h id w l iH rj O rH x; 0 o u VO VD ^ V£> on on on O 00 CN 00 on o on o on oo o o o o on o o o o r-\ r-\ <-\ r-\ O O 00 00 o rH rH O O rH rH rH rH rH rH lo ro on lo rCN CM CN CN CN >J (N O ^ on o C\ g\ on VO VD on on cri on ucocooo o on on o o 'S* vo VO o o o o o rH rH rH rH rH <£> O O O 00 rH CN CN CN rH on on on on o o o o o rH rH rH rH rH rH On rH On rH rH O rH O rH rH rH CO CO CO r-i t-t r-\ r-i r-t r-{ r-\ r-i cr* on on on rH rH rH rH rH rH rH i— I rH rH I s in CN CN CN CN CN rH rH rH rH rH iH CM CO -rf LO CJ w

PAGE 132

122 T3 Q *> — > Q frt rH rH M f~i *-* Q o rH rl r* *&> o w o o U) c — > O rvi IN N K o o o rH Q in V J CN o rH i-f UJ P 1 4J £3 O rrinininrin CTiCTiCTiCTiCTiCTlCTiCTtCTiCTl ^•^J'CNCN^J'CNCN^r^rCN OOOOOOOOOO rlrlrlrlrlrlrlrHrlrH T^r^TrcNTCNrrcN^f oooooooooo rHrlrHrlrHt-Hrlrlrlrl COOCNCOCDCOCOTCOO rlCNi-HrHrlrli-lrHrlCN raoooinooiNf f (Ndcnnnroronron HrlrlrlHHrlHrlrl ininiflrTr-1'r-mr^ COCOCO3"'3 C0^1 ^1 CO OOOOOOOOOO r-HrtrlrlrlrlrlrHrHrl COCO^COOCNCNOOO OOOOrlrHrHrlrlrH i-HrlrlrHrlr-lrli-HrHrl oo^rcNooocoooo rlrl HHHHrlOHO rlHrlHHHHHHrl CNTj.'3'COOCOTrCOCNOO CNCNCNCNCnCNCNCNCNCN rl rl rlrlrlrHrHrlrlrH mLnTincTiCTiCTiCTir^CTi roennfoconmnrnn rlrlrHrlrHrlrli-HrHrl rHi-H nciriHcnnnn CT CTl CTiCTiCTiCTiCTiCTiCTiCTi rocTirocTir^-r^rorHrHcTi ninininnnroroinin OCTiOCTiCTlCTiOOOCTi CTCTCTiCTiCTiCTiCTiCTiCTiCTi rH i— I HHH o din ro n r^mcricriiriHrnHrnH nnrofOfiininini/nn rl rl rl rl rl CTiCTiCTiCTiCTiOOOOO CTl (Ti CTlCTiCTlCTlCTCTCTlCTl rl rl rl rl rl rl rl rl rl rl cti rld ct cti r^inr^cTimini-Hinr-r~mm HroHHmHroH rl rl rl rl rl OOOOOOOOOO OOOOOOOOOO rl rl rl rl rl rlrlrlrlrlrHrlrlrlrl ,-|,Hr-li-lrlrlrl,-HrlrH cti cti cT cti Lor^cTicTicrir-~cTir^r~rr-^i^-cTiCTicTiCTiinrocTirCNCNCNCNCN OOOOOOOOOO rlrlrlrlrHrlrlrHrlrl rHrlrlrlrl r-^r-t^^r-ir-ir-ifir-i^ rlrlrlrHrlrHrlrlrlrl CD r~CO (Ti o rl ricNcn^
PAGE 133

123 o o^cria>r-r^r^r~-rinro C^^C^iCyCXtCr>(T\(TtC">cri U ] o f— i r*^ ro ro lo r~~ ct ro i — i o o cr> o*> o o ct> C*\ CTi CT~i CJ^ G> CJ^ o rH / — \ — <• — > CN O oi M/i ci m in mm n in t? O G> O iHiHiHiHrHrHrHiHrHrH o r — 1 r— 1 t — 1 r— I i— i O O i — 1 i — 1 i — \ O iHrHrHrHrHr— Ir— IrHr— Ir— 1 o O ^ CN CN O ^ "^J* CN O 0) o OOOOOOOOOO Ms u > rl r—i r-i i— 1 i — 1 i—i r-i i— f r— \ t—\ c o CN ^* CN CN CNJ tj 1 •H o OOOOOOOOOO M CNI i i i i i i i i i— 1 i— i i—i 11 i—l f"H i— 1 i— i t—i r— I M Uh c o V — J •H o CN ^ TJ* Tj* Tj 1 ^J* CN CN CN ^J 1 o 1 1 o OOOOOOOOOO i—l i— 1 i—l H i— 1 i—l i—l i-n i— ( r — 1 __J u 3 Q o OOC00000C0OUDC0O N i~n i— t <— ) C_J C_> O i — 1 O O r-H H iHiHiHiHi— Ir— IrHrHi— li— 1 (T3 C tn o •H ^1 i—1 t~l t~ f r— 1 i—l r— 1 r 1 r— 1 r— 1 CO rHrHi— li— IrHi— IrHrHrHrH o o OOOOOOOOOO LO Hi— ti— IiHi— 1 r Ir— 1 r| H r 1 o vO vO vo vo ^* CO vo o OOOOOOOOOO CN rH i— 1 i— 1 r— iiHrHrHiHt— 1 rH N o OOOOOOOOOOfMOOO f— \ 1 — f~n O r— 1 v — J f j t — ( t — 1 O i — I o rH i— liHrHrHrHi— liHrHrHiH Q ID O CN rHi— Ir— Ir-Hrli— li-Hi— ),— (iH iHiHiHi— IrHrHHiHiHrH O OO^OOO^rCNOO^J' rH H(N(NfM(N(N{N(NCN(N HHHHHrlHHHH 4n 3 Q) 3 HlNCTTlflVDI^OOfflO E-< P X! O 3 CD U CO •n CQ 01D10 1DO*C(MO^^f 00>Oi010CTiOfTi(T>0^CT>OfJ iH rH (Jl H CTl CI (3\ in CTiOCTiOOOOOOO rH ^\ r-{ r-i ^ ^ r-t ^ Lnr^r^rHinmmmtnin OJ(N(N m m
PAGE 134

124 o o in o o o o o CM o rH o o in o o CN O o o CM o o in o o CM O O O CN O -P W C CD 3 CT\ 0> CT\ <7i CT\ o\ r in t- CTi CT\ CTi r— I r— I ID iH
PAGE 135

APPENDIX F DISCRIMINANT ANALYSIS

PAGE 136

Discriminant Analysis Discriminant analysis is a form of multivariate analysis. This type of analysis places the individual values from two known groups on a continuum. Based upon the group means and variances, a point on the continuum divides the individual values into two regions representing two discrete distributions. Each individual is assigned to the region to which it has the highest probability of belonging. This posterior probability of an individual coming from each group is based on group means, standard deviations, a group covariance matrix, and a group correlation matrix. An approximate F statistic tests equality of group means. Once the individuals are assigned to each region, a contingency table can represent the number of correct and incorrect classifications, as in Table 1. There were 9 correct and 1 incorrect classifications of 10 normal ears and 4 correct and 1 incorrect classifications of 5 abnormal ears Table 1. Contingency Table of Discriminant Analysis N A N 9 1 A 1 4 126

PAGE 137

APPENDIX G RAW DATA OF THRESHOLD OF AUDIBILITY

PAGE 138

C5 X -H c 1 o iH o to TCniDVDOCNOr~-00CN p o -p o u rH rH CN CN rH rH rH rH -H o cd rH as rH H W •H ,Q Q o i-i •^rvorovooovoootDTrcN •H CN ns CN CN CO CO rH rH rHCNCN TJ ; rH e d i ^ *-i C3 o CP o vOHOO^DO'J'OVOHai MH •H rH CMCOfOrOCNCNrHCNCOCN 0 co "0 rH O (NfMVOHClClinHOJH 0 o rH rH CN CO rH CN Xi ID CO a) M O O H rl M Cl rH CN E-i CN n N 0 SB O voror^iDcnr^vorocnr^ O rH rH CN CO rH CN (O o rH | 1 o ID Q O inmr-oovor-roocnvo CN NCJOClHHHHCNn O Hr^NCNvo^HincMO • rH CIINV^MNMHCIT I +J XI o CO -n cn iH id w T3 0 o u I id cd rH 0 o u rro cn cn vo rH CN CN CN rH H t (N CI CO CN CN CN CN rH CN 00 ^ ID O CN CN CN CN CN vo co cn r-~ CN CO CN CN CN CN CN VO VO *T ro ^* ro co ro CN CN CO CO VO rH CN CN rH CN ID ID ID CO f~ rH CN CN CN CN o vo rco co CN CN CN CN CN 1/1 T O 1/1 rf cn ro ro ro ro O VO Cl CO CTi ro ro ro ro ro n iH W -a id m i M rd CD rH U 0 u ro cn rH rH CO rH I CD M h CTl rH cn cn rH CN CN rH TJ CD •H rH ro cn rH rd CN CN ro CN 04 g H r* cn 00 o tr> cn cn ro ro c >i •H U M C CO o id r co rdj 3 ci o t ci K cp VC 00 "tf *T 1 VO ID f-~ CN T VO CO ID CO O VO ID [ — CN ID r~ cn r* 00 ro VO ID I — CN ID 0 rH CN rH VO 1 VO CO Cl ID CN Cl I ID rH r~ VO CO CO VO CN ID 01 Cl ID tJ* I CN ID r ci r O if) id r~ CO ID Cn O VO rH ID ID r*CO ID rCO rH rH vo 00 co vo cn ci in r> t id co co vo CN rH O Ol CO O t cn id id ro id cn cn co rH r^* [ (N ID ID CN VO CN T CT\
PAGE 139

APPENDIX H CORRELATION COEFFICIENTS OF THE TEMPORAL INTEGRATION MEASURES

PAGE 140

Correlation Coefficients of the Measures Between Temporal Integration of the Acoustic-Reflex Threshold and Temporal Integration of the Threshold of Audibility. Signif iFrequency Correlation cance 20 msec-200 msec measure Normal ears 500 Hz 0.3763 ns 1000 Hz 0,6926 ns 2000 Hz 0,6349 ns Cochlear-good ears 500 Hz 0.9177 .05 1000 Hz -0.9768 .01 2000 Hz 0.0186 ns Cochlear-bad ears 500 Hz -0.2654 ns 1000 Hz -0.0995 ns 2000 Hz 0,6880 ns 10 msec-200 msec measure Normal ears 500 Hz 0,0943 ns 1000 Hz 0.3887 ns 2000 Hz 0.6397 ns Cochlear-good ears 500 Hz -0.7857 ns 1000 Hz 0.3493 ns 2000 Hz -0.9661 .01 Cochlear-bad ears 500 Hz -0.1461 ns 1000 Hz 0.1460 ns 2000 Hz 0.7719 ns ns = not statistically significant. .05 = significant at the 5 percent level. .01 = significant at the 1 percent level. 130

PAGE 141

REFERENCES Alberti, P. W. R. M. and Kristensen, R. The clinical application of impedance audiometry: A preliminary appraisal of an electro-acoustic impedance bridge. Laryngoscop e 80:735-746, 1970. American National Standards Institute. Specifications for audiometers. ANSI S3. 6-1969. American National Institute, Inc., New York, 1970. Anderson, H., Barr, B., and Wedenberg, E. Intra-aural reflexes in retrocochlear lesions. In Hamberger, C. A. and Wersall, J. (editors) Disorders of the Skull Base Region Nobel Symposium, 10th, Stockholm: Almqvist and Wiksell, 1969. Bar, B., and Wedenberg, E. The early detection of acoustic tumors by the stapedius reflex test. In Wolstenholme, G. E. W. and Knight, J. (editors) Sensorineural Hearing Loss A Ciba Foundation Symposium, London: J. and A. Churchill, 1970. and Wedenberg, E. Audiometric identification of normal hearing carriers of genes for deafness. Acta Oto-laryngologica 65:535-554, 1968. Bates, M. A., Loeb, M. Smith, R. P., and Fletcher, J. L. Attempts to condition the acoustic reflex. Journal of Auditory Research 10:132-135, 1970. Beedle, R. K. An Investigation of the Relationship Between the Acoustic Reflex growth and Loudness Growth in Normal and Pathologic Ears Ph.D. dissertation. Northwestern University, Illinois, 1970. and Harford, E. Acoustic reflex and loudness growth in normal and pathological ears. Sixth Annual Report of the Auditory Research Laboratories Northwestern University, Illinois, 1971. Bilger, R. and Feldman, R. M. Frequency dependence in temporal integration. Journal of the Acoustical Society of America 45:293(A), 1969. 131

PAGE 142

132 Blodgett, H. C. Jeffress, L. A., and Taylor, R. W. Relation of masked threshold to signal durations for various interaural phase-combinations. American Journal of Psychology 71:283-290, 1958. Boudreau, J. C. Stimulus correlates of wave activity in the superior-olivary complex of the cat. Journal of the Acoustical Society of America 35:779-785, 1965. Brahe Pedersen, C. and Elberling C. Temporal integration of acoustic energy in normal hearing persons. Acta Oto-Laryngologica 74:389-405, 1972. and Elberling, C. Temporal integration of acoustic energy in patients with presbyacusis Acta Oto-Laryngologica 75:32-37, 1973. Brink, F., Jr. Synaptic mechanisms. In Stevens, S. S. (editor) Handbook of Experimental Psychology New York: John Wiley and Sons, 1951. Brooks, D. N. The use of the electro-acoustic impedance bridge in the assessment of middle ear function. Journal of International Audiology 8:563-569, 1969. • Electroacoustic impedance studies on normal ears of children. Journal of Speech and Hearing Research 14:247-253, 1971. Burke, K. S., Herer, G. R. and McPherson, D. L. Middle ear impedance measurement (Acoustic and electroacoustic comparisons). Acta Oto-Lar yngologica, 70:29-34, 1970. Campbell, R. S. and Counter, S. A. Temporal integration and periodicity pitch. Journal of the Acoustical Society of America 45:691-693, 1969. Carhart, R. and Jerger, J. Preferred method for clinical determination of pure tone thresholds. Journal of Speech and Hearing Disorders 24:330-345, 1959. Carver, W. F. Loudness Balance Procedures. In Katz J. (editor) Handbook of Clinical Audiology Baltimore: Williams and Wilkins Co., 1972. Chamberlin, S. C. and Zwislocki, J. J. Threshold of audibility as a function of tone duration: Is there a frequency effect? Journal of the Acoustical Society of America 48:71(A), 1970.

PAGE 143

133 Clack, T. D. Effect of signal duration on the auditorysensitivity of humans and monkeys (Macaca mulatta) Journal of the Acoustical Society of America 40:11401146, 1966. Coles, R. R. A. Can present day audiology really help in diagnosis? An otologist's question. Journal of Laryngology and Otology 86:191-224, 1972. Dallos, P. J. Dynamics of the acoustic reflex: Phenomenological aspects. Journal of the Acoustical Society of America 36:2175-2183, 1964. and Johnson, K. R. Influence of rise-fall time upon short tone thresholds. Journal of the Acoustical Society of America 40:1160-1163, 1966. ^ and Olsen, W. Integration of energy at threshold wxth gradual rise-fall tone pips. Journal of the Acoustical Society of America 36:741-751, 1964. Deutsch, L. J. The threshold of the stapedius reflex to selected acoustic stimuli in normal human ears. U. S. Naval Submarine Medical Center Rese arch Report No. 546:1-27, 1968. The threshold of the stapedius reflex for pure tone and noise stimuli. Acta Oto-Laryngologica 74:248-251, 1972. Dix, M. R. Hallpike, C. S., and Hood, J. D. Observations upon the loudness recruitment phenomenon with special reference to the differential diagnosis of disorders of the internal ear and VIII nerve. Proceedings of the Royal Society of Medicine 41:516-526, 1948. Djupesland, G. Electromyography of the tympanic muscles in man. Journal I nternational Audiologv/ 4: 34-41, 1965. — Flottorp, G., and Winther, F. Size and duration of acoustically elicited impedance changes in man. Acta Oto-Laryngologica 224:220-228, 1966. and Zwislocki, J. J. Effect of temporal summation on the human stapedius reflex. Acta Oto Laryngologica 71:262-265, 1971. Doyle, T. N. Auditory Temporal Summation with Variable Inter signal Intervals in Normal and Non -normal Subjects Ph.D. dissertation. University of Minnesota, 1970.

PAGE 144

134 Durrant, J. D. and Shallop, J. K. Effects of differing states of attention on acoustic reflex and temporary threshold shift. Journal of th e Acoustical Society of America 46:904-913, 1969. Elliott, L. L. Tonal thresholds for short-duration stimuli as related to subject hearing level. Journal of the Acoustical Society of America 35:578-580, 1963. Ewertsen, H. Filling, S., Terkildsen, K. and Thomsen, K. A. Comparative recruitment testing. Acta Oto-Laryngologica Supplement 140:116-122, 1958. Feldman, A. S. Acoustic Impedance studies of the normal ear. Journal of Speech and Hearing Disorders 10:165176, 1967"; Fisch, U. and Schulthess, G. V. Electromyographic studies on the human stapedial muscle. Acta Oto-Laryngologica 56:287-297, 1963. Fletcher, H. Auditory patterns. Review of Modern Physics 12:47-65, 1940. Flottorp, G. and Djupesland, G. Diphasic impedance change and its applicability in clinical work. Acta Oto Laryngologica Supplement 263 :200-204 1970"; Djupesland, G., and Winther, F. The acoustic stapedius reflex in relation to critical bandwidth. Journal of the Acoustical Society of America, 49:457461, 1971. Fowler, E. P. A method for the early detection of otosclerosis. Archives of Otolaryngology 24:731-741, 1936. The diagnosis of diseases of the neural mechanism of hearing by the aid of sounds well above threshold. Transactions of the American Otological Society 27:207-219, 1937. ~ The recruitment of loudness phenomenon. Laryngoscope 60:680-695, 1950. Fulton, R. T. and Lamb, L. E. Acoustic impedance and tympanometry with the retarded: A normative study. Audiology 11:199-208, 1972. Garner, W. R. The effect of frequency spectrum on temporal integration of energy in the ear. Journal of the Acoustical Society of America 19:808-815, 1947a.

PAGE 145

135 Garner, W. R. Auditory thresholds of short tones as a function of repetition rates. Journal of the Acoustical Society of America 19:600-608, 1947b. and Miller, G. A. The masked threshold of pure tones as a function of duration. Journal of Experi mental Psychology 37:293-303, 19477 Gengel, R. W. Auditory temporal integration at relatively high masked threshold levels. Journal of the Acoustical Society of America 51:1849-1851, 1972. ~~ and Watson, C. S. Temporal integration: I. Clinical implications of a laboratory study. II. Additional data from hearing-impaired subjects. Journal of Speech and Hearing Disorders, 36:213-224, 1971. Goldstein, R. and Kramer, J. C. Factors affecting threshold for short tones. Journal of Speech and Hearing Research 3:249-256, 19607 Graham, A. B. (editor). Alternate loudness balance techniques. In Sensorineural Hearing Processes and Disorders Boston! Little, Brown and Company, 1967. Grason-Stadler. Otoadmittance Handbook 2: A guide to users of the Grason-Stadler Model 1720 Otoadmittance Meter Concord, Mass: Grason-Stadler Company, 1973. Green, D. M. Birdsall, T. G. and Tanner, W. P. Jr. Signal detection as a function of signal intensity and duration. Journal of the Acoustical Socie ty of America 29:523-531, 1957. Gunn, W. J. Effects of attending to auditory signals on the magnitude of the acoustic reflex U. S. Army Research Laboratory Report No. 751, Fort Knox, Kentucky, 19 67. Harford, E. and Liden, G. Acoustic impedance, the middleear muscle reflex, tympanometry and extratympanic manometry. Annual Report of the Auditory Research Labortories Northwestern University, 1967-1968. Harris, J. D. Peak vs. total energy in threshold for very short tones. Acta Oto-Laryngologica 47:134-140, 1957. Haines, H. L. and Myers, C. K. Brief-tone audiometry: Temporal Integration in the hypacusic. Archives of Otolaryngology 67:699-713, 1958.

PAGE 146

136 Hattler, K. W. and Northern, J. L. Clinical application of temooral summation. Journal of Auditory Research 10:72-78, 1970. Hempstock, T. J., Bryan, M. E., and Tempest, W. A. A redetermination of quiet thresholds as a function of stimulus duration. Sound and Vibration 1:365-380, 1964. Hirsh, I. J., Palva, T. and Goodman, A. Difference limen and recruitment. Archives of Otolaryngology 60:525540, 1954. Hoist, H Ingelstedt, S., and Ortegren, U. Ear drum movements following stimulation of the middle ear muscles. Acta Oto-Laryngologica Supplement 182:73-83, 1963. Hood, J. D. Basic audiological requirements in neurootology. Journal of Laryngology and Otology 83: 695-711, 19€9~. Hughes, J. R. Auditory sensitization. Journal to the Acoustical Society of America 26:1064-1070, 1954. Electrophysical evidence or auditory sensitization. Journal of the Acoustical Society of America 29:275-280, 1957. Hughes, J. W. The threshold of audition for short periods of stimulation. Proceedings of the Royal Society of Medicine B133 : 486-490 1946. Hung, I. J. and Dallos, P. Study of the acoustic reflex in human beings. I. Dynamic characteristics. Journal of the Acoustical Society of Ameri ca, 52: 1168-1180, 1972. International Organization for Standardization. Standard reference zero for the calibration of pure tone audiometers. ISO R 389-19 64. Geneva, Switzerland. Jepsen, O. The threshold of the reflexes of the intratympanic muscles in a normal material examined by means of the impedance method. Acta O to-Laryngologica, 39:406-408, 1951. • Infcratympanic muscle reflexes in psychogenic deafness. Acta Oto-Laryngologica, Supplement 109: 61-69, 1953:

PAGE 147

137 Jepsen, 0. Middle ear muscle reflexes in man. In Jerger, J. (editor) Modern Developments in Audiology New York: Academic Press, 1963. Jerger, J. Influence of stimulus duration on the pure-tone threshold during recovery from auditory fatigue. Journal of the Acoustical Society of Americ a, 27:121124, 1955T • Bekesy audiometry in analysis of auditory disorders. Journal of Speech and Hearing Research 3: 275-287, 1960. The audiological examination as an aid in diagnosis. Archives of Otolaryngology 85: 552554, 1967. Clinical experience with impedance audiometry. Archives of Otolaryngology 92:311-324, 1970. t Jerger, S., Ainsworth, J., and Caram, P. Recovery of auditory function after surgical removal of cerebellar tumors. Journal of Speech and Hearing Disorders 31:377-382, 1966. Jerger, S., and Mauldin, L. Studies in impedance autiometry: I. Normal and sensorineural ears. Archives of Otolaryngology 96: 513-523, 1972. / Shedd, J., and Harford, E. On the detection of extremely small changes in sound intensity. Archives of Otolaryngolog y, 69:200-211, 1959. Johansson, B., Kylin, B. and Langfly, M. Acoustic reflex as a test of individual susceptibility to noise. Acta Oto-Laryngologica 64:256-262, 1967. Karlovich, R. S., Lane, R. H. Smith, L. L. Tarlow, A. J., Thompson, D. and Vivion, M. C. Auditory threshold at 125 Hz as a function of signal duration and signal filtering. Journal of the Acoustical Society of America 49:1897-1899, 1971. Klockhoff, I. Middle ear muscle reflexes in man: A clinical and experimental study with special reference to diagnostic problems in hearing impairment. Acta Oto-Laryngologica Supplement 164:1-92, 1961 Kobrak, H. G. The present status of objective hearing tests. Annals of Otology Rhinology and Laryngology, 57:10181026, 1948. ^~

PAGE 148

138 Kobrak, H. G. The Middle Ear Chicago: University of Chicago Press, 1959. Kristensen, H. K. and Jepsen, 0. Recruitment in otoneuralogical diagnostics. Acta Oto-Laryngologica 42:553560, 1952. Lamb, L. E. and Peterson, J. Middle ear reflex measurements in pseudohypacusis Journal of Speech and Hearing Disorders 32:42-51, 1967. Peterson, J., and Hansen, S. Application of stapedius muscle reflex measures to diagnosis of auditory problems. Journal of International Audiology 7:188-199, 1968. Licklider, J. C. R. The perception of speech. In Stevens, S. S. (editor) Handbook of Experimental Psychology New York: John Wiley and Sons, 1951. Liden, G. The stapedius muscle reflex used as an objective recruitment test: A clinical and experimental study. In Wolstenholme, G. E. W. and Knight, J. (editors), Sensorineural Hearing Loss A Ciba Foundation Symposium, London: J. and A. Churchill, 1970. Peterson, J. L. and Harford, E. R. Simultaneous recording of changes in relative impedance in -r --air pressure during acoustic and non-acoustic elicitation of the middle-ear reflexes. Acta OtoLaryngologica Supplement 263:208-217, 1970. Lilly.* _Eu J. Some properties of the acoustic reflex in raah i Journal of the Acoustical Society of America 15^2007 (A) 19 64. -— -'— r --A comparison of acoustic impedance data obtained with Madsen and Zwislocki instruments. Presented at €he American Speech and Hearing Convention, Chicago, — — • Acoustic impedance at the tympanic membrane. In-Katz, J. (editor), Handbook of Clinical Audiology Baltimore: Williams and Wilkins Company, 1972. Ker '~' : Measurement of acoustic impedance at the tympanic membrane. In Jerger, J. (editor), Modern Sevelopments in Audiology 2nd Edition. New York: Academic Press, 19 73.

PAGE 149

139 Lilly, D. J. and Shepherd, D. C. A rebalance technique for the measurement of absolute changes in acoustic impedance due to the acoustic reflex. ASHA 6: 381 (a) 1964. Lindsay, J. R. Kobrak, H. G. and Perlman, H. B. Relation of the stapedius reflex to hearing sensation in man. Archives of Otolaryng ology, 23:671-678, 1936. Loeb, M. Psychophysical correlates of intratympanic reflex action. Psychological Bulletin 61:140-152, 1964. Lorente de No, R. The reflex contractions of the muscles of the middle ear as a hearing test in experimental animals. Transactions of the American Laryngology Rhinology and Otology Society 39:26-42, 1933. The function of the central acoustic nuclei examined by means of the acoustic reflexes. Laryngoscope 45:573-595, 1935. and Harris, A. S. Experimental studies in hearing. Laryngoscope 43:315-326, 193-3. Madsen. Madsen Model Z070 Electro-Acoustic Impedance Bridge: Applications and instructions for use. Copenhagen: Madsen Electronics, n.d. Martin, F. N. The short increment sensitivity index (SISI) In Katz, J. (editor) Handbook of Clinical Audiology Baltimore: Williams and Wilkins Company, 1972. and Wofford, M. J. Temporal summation of brief tones in normal and cochlear-impaired ears. Journal of Auditory Research 10:82-86, 1970. McRobert, H. The response of the tympanic muscles in human ears: Possible false inferences from results of reflex testing on normal and pathological ears. Sound 2:71-76, 1968. 0 Bryan, M. E. and Tempest, W. The acoustic stimulation of the middle ear muscles. Sound and Vibration 7:129-142, 1968. Mehmke, S. and Tegtmeir, W. The diagnostic value of impedance measurements. Fenestra (An eight page insert between pp. 12-13,) April 1970.

PAGE 150

140 Melcher, J. A. and Peterson, J. L. The effects of age and hearing impairment on the acoustic reflex decay. Presented at the American Speech and Hearing Convention, San Francisco, 1972. Mendelson, E. S. A sensitive method for registration of human intratympanic muscle reflexes. Journal of Applied Physiology 11:499-502, 1957. Improved method for studying tympanic reflexes in man. Journal of the Acoustical Society of America 44:146-152, 1961. The lability of the resting and reflex activity of the human middle ear muscles. In Fletcher, J. L. (editor) Middle Ear Function Seminar U. S. Army Research Laboratory Report No. 576, Fort Knox, Kentucky, 1963. Acoustic ref lexometry Acta Oto-Laryngologica 62:125-139, 1966. Metz, 0. The acoustic impedance measured in normal and pathological ears. Acta Oto-Laryngologica Supplement 63:1-254, 1946. Threshold of reflex contractions of muscles of middle ear and recruitment of loudness. Archives of Otolaryngology 55:536-544, 1952. Miller, G. A. The perception of short bursts of noise. Journal of the Acoustical Socie ty of America, 20: 160-170, 1948. Miskolczy-Fodor, F. Monaural loudness-balance test and determination of recruitment degree with short soundimpulses. Acta Oto-Laryngologica 43:573-595, 1953. • The relation between hearing loss and recruitment and its practical employment in the determination of receptive hearing loss. Acta Oto-Laryngologica, 46:409-415, 1956. 2 • Relation between loudness and duration of tonal pulses. II. Response of normal ears to sounds with noise sensation. Journal of the Acoustical Society of America 32 :482-486 19~60~T Moller, A. R. Intra-aural muscle contraction in man, examined by measuring acoustic impedance of the ear. Laryngoscope 68:48-62, 1958.

PAGE 151

141 Moller, A. R. Bilateral contraction of the tympanic muscles in man, examined by measuring acoustic impedance-change. Annals of Otology, Rhinology and Laryngolo gy, 70:735752, 1961a. Network model of the middle ear. Journal of the Acoustical Society of America 33 : 168-176 1961b. The sensitivity of contraction of the tympanic muscles in man. Annals of Otology, Rhinology and Laryngology 71:86-95, 1962a. Acoustic reflex in man. Journal of the Acoustical Society of America 34:1524-1534, 1962b. Effect of tympanic muscle activity on movement of the ear drum, acoustic impedance and cochlear microphonics. Acta Oto-Laryn gologica 58:525-534, 1964. An experimental study of the acoustic impedance of the middle ear and transmission properties. Acta Oto-Laryngologica 60:129-149, 1965. Munson, W. A. The growth of auditory sensation. Journal of the Acoustical Society of America 19:584-591, 1947. Neergaard, E. B. and Rasmussen, G. Latency of the stapedius muscle reflex in man. Archives of Otolaryngology, 84:173-180, 1966. — — 2JL Nerbonne, M. A. A Comparison of Brief Tone Audiometry with Other SeTected Auditory Tests of Cochle ar Function" Ph.D. dissertation. Michigan State University, 1970. Niemeyer, W. Relations between the discomfort level and the reflex threshold of the middle ear muscles. Audiology 10:172-176, 1971. and Sesterhenn, G. Calculating the hearing threshold from the stapedius reflex for different sound stimuli. Presented at International Audiology Congress, Budapest, 1972. Nixon, J. and Glorig A. Reliability of acoustic impedance measures of the eardrum. Journal of Auditor y Research, 4:261-276, 1964.

PAGE 152

142 Norris, T. W. Stelmachowicz P., and Taylor, D. Effects of stimulus variations on acoustic reflex patterns. Presented at International Audiology Congress, BudaDest, 1972. Northern, J. L. Temporal summation for critical bandwidth signals. Journal of the Acoustical Society of America 42:456-461, 1967. Olsen, W. 0. and Carhart R. Integration of acoustic power at threshold by normal listeners. Journal of the Acoustical Society of America 40:591-599, 1966. Rose, D. E and Noffsinger, D. Brief tone audiometry with normal, cochlear, and VIII nerve tumor patients. A paper accepted for publication Archives of Otolaryngology 1973. Owens, E. Bekesy tracing and site of lesion. Journal of Speech and Hearing Disorders 29:456-468, 1964. • Audiologic evaluation in cochlear versus retrocochlear lesions. Acta Oto-Laryngologica Supplement 283:1-45, 1971. Perlman, H. B. and Case, T. J. Latent period of the crossed stapedius reflex in man. Annals of Otology, Rhinology and Laryngology 48:663-675, 1939. Peters Ltd. Peters AP 61 Acoustic Impedance Meter: Oper ating instructions New York: Lehr Instrument Corporation, n.d. Peterson, J. L. and Liden, G. Dynamics of the stapedial muscle reflex. Presented at the 10th International Audiology Congress, Dallas, Texas, 1970. and Liden, G. Some static characteristics of the stapedial muscle reflex. Audiology 11:97-114, 1972. Plomp, R. and Bouman, M. A. Relation between hearing threshold and duration for tone pulses. Journal of the Acoustical Society of America 31:749-758, 1959. Price, G. R. Influence of external ear acoustics on impulse arriving at the ear drum. Journal of the Acoustical Society of America 52:129 (a), 1972. Rao, C. R. Advanced Statistical Methods in Biometric Research New York: John Wiley and Sons, 1952.

PAGE 153

143 Rasmussen, G. L. The olivary peduncle and other fiber projections of the superior olivary complex. Journal of Comparative Neurology 84:141-219, 1946. Reger, S. N. Differences in loudness response of normal and hard-of -hearing ears at intensity levels slightly above threshold. Annals of Otology, Rhinology and Laryngology 45:1029-1039, 1936. Salomon, G. and Starr, A. Electromyography of middle ear in man during motor activities. Acta Neurologica Scandinavia 39:161-168, 1963. Sanders, J. W. and Honig, E. A. Brief -tone audiometry: Results in normal and impaired ears. Archives of Otolaryngology 85:640-647, 1967. / Josey, A. F. and Kemer F. J. Brief -tone audiometry in patients with Vlllth nerve tumor. Journal of Speech and Hearing Research 14:172-178, 1971. Scharf, B. Critical bands. In Tobias, J. B. (editor), Foundations of Modern Auditory Theory Volume I New York: Academic Press, 1970. Schoel, T. and Arnesen, G. The choice of probe-tube position and test frequency in determining the intraaural reflexes. Acta Oto-Laryngologica 54:233-238, 1962. Sheeley, E. C. and Bilger, R. C. Temporal integration as a function of frequency. Journal of the Acoustical Society of America 36:1850-1857, 1964. Sherrington, C. S. The integrative action of the nervous system London: Constable, 1906. • The integrative action of the nervous system New Haven, Conn. : Yale University Press, 1947. Shiffman, F. Bridge clinic. Impedance Newsletter 1:11, 1972. (Madsen Electronics Corp., New York.) Simmons, F. B. Post-tetanic potentiation in the middle ear muscle acoustic reflex Journal of the Acoustical Society of America 32:1589-1591, 1960. • An analysis of the middle-ear muscle acoustic reflex of the cat. In Fletcher, J. L. (editor) Middle Ear Function Seminar U. S. Army Medical Research Laboratory Report No. 57 6, Fort Know, Kentucky, 1963.

PAGE 154

144 Simmons, F. B. and Dixon, R. F. Clinical implications of loudness balancing. Archives of Otolaryngol ogy, 83:449-454, 1966. Simon, G. R. The critical bandwidth level in recruiting ears and its relation to temporal summation. Journal of Auditory Research 3:109-119, 1963. Small, A. M. Jr., Brandt, J. F., and Cox, P. G. Loudness as a function of signal duration. Journal of the Acoustical Society of America 34:513-514, 1962. Swannie, E. M. Impedance audiometry in clinical practice. Proceedings of the Royal Society of Medicine 59: 971-974, 1966. Tempest, W. and Bryan, M. E. The auditory threshold for short-duration pulses. Journal of the Acoustical Society of America 49:1901-1902, 1971. Terkildsen, K. Movements of the eardrums following intraaural muscle reflexes. Archives of Otolaryngo logy, 66:484-488, 1957. The intra-aural muscle reflexes in normal persons and in workers exposed in intense industrial noise. Acta Oto-Laryngologica 52:384-396, 1960a. An evaluation of perceptive hearing losses in children, based on recruitment determinations. Acta Oto-Laryngologica 51:476-484, 1960b. Acoustic reflexes of the human muscle tensor tympani. Acta Oto-Laryngologica Supplement 158: 230-238, 1960c. w Clinical application of impedance measurements with a fixed frequency technique. Journal of International Audiology 3:147-155, 1964T _, Osterhammel, P., and Scott Nielsen, S Impedance measurements: Probe-tone intensity and middle-ear reflexes. Acta Oto-Laryngologica, Supplement 263:205-207, 1970. 2 and Scott Nielsen, S. Electroacoustic impedance measuring bridge for clinical use. Archives of Otolaryngology 72:339-346, 1960. Thomsen, K. A. The Metz recruitment test. Acta OtoLaryngologica 45:544-552, 1955a.

PAGE 155

145 Thomsen, K. A. Employment of impedance measurements in otologic and otoneurologic diagnostics. Acta Oto Lar yngoJ oq ica 45:159-167, 1955b. Tillman, T. W. Special hearing tests in otoneurologic diagnosis. Archives of Otolaryngology 89:25-30, 1969. Dallos, P. J., and Kurvilla, T. Reliability of measures obtained with the Zwislocki Acoustic Bridge. Journal of the Acoustical Society of America 36:582-588, 1963. Vries, H. de. The minimum audible energy. Acta Oto Laryngologica 36:230-235, 1948. Ward, W. D. (editor) Proposed damage-risk criterion for impulse noise (gunfire) Working Group 57 of the NAS-NRC Committee on Hearing, Bioacoustics and Biomechanics. Washington, D. C, 1968. Watson, C. S. and Gengel, R. W. Signal duration and signal frequency in relation to auditory sensitivity. Journal of the Acoustical Society of America 46:989-997, 1969. Weiss, H. S., Mundie, M. R. Cashin, J. L. and Shinabarger, E. W. The normal human intra-aural muscle reflex in response to sound. Acta Oto-Laryng ologica, 55:505-515, 1962. Wever, E. G. and Lawrence, M. Physiological Acoustics Princeton, J. J.: Princeton University Press, 1954. Wilber, L. A., Goodhill, V., and Hogue, A. C. Comparative acoustic impedance measurements. Presented at the American Speech and Hearing Convention, Chicago, 1970. Wright, H. N. Switching transients and threshold determination. Journal of Speech and Hearing Research 1:52-60, 1958": • Audibility of switching transients. Journal of the Acoustical Society of America 32:138, 1960. • Clinical measurement of temporal auditory summation. Journal of Speech and Hearing Researc h, 11:109-127, 1968a. • The effect of sensori-neural hearing loss on threshold duration functions. Journal of Speech and Hearing Research 11:842-852, 1968b.

PAGE 156

146 Wright, H. N. Temporal summation for tones at threshold. Presented at the 84th meeting of the Acoustical Society of America, Miami Beach, 1972. and Cannella, F. Differential effect of conductive hearing loss on the threshold-duration function. Journal of Speech and Hearing Research 12:607-615, 1969. Wright, J. and Btholm, B. Anomalies of the middle-ear muscles. Journal of Laryngology and Otology 87: 281-288, 1973. Zwicker, E. and Wright, H. N. Temporal summation for tones in narrow-band noise. Journal of the Acoustical Society of America 35:691-699, 1963. Zwislocki, J. Some measurements of the impedance at the eardrum. Journal of the Acoustical Society of America 29:349-356, 1957. • Theory of temporal auditory summation. Journal of the Acoustical Society of America 32:1046-1060"^ 1960. Acoustic measurement of the middle ear function. Annals of Otology, Rhinology an d Laryngology, 70: 599-606, 1961. • Analysis of the middle-ear function. I. Input impedance. Journal of the Acoustical Society of America 34:1514-1523, 1962. • An acoustic method for clinical examination of the ear. Journal of Speech and Hea ring Research, 6:303-314, 1963": ~~

PAGE 157

BIOGRAPHICAL SKETCH William Lee Parker was born in 1945, earned his first dollar in 1954, and was married in 1966 to the delightful Cristine Mary Sisson. He received his academic degrees of Bachelor and Master of Arts with a major in speech pathology at California State College at Long Beach. In 1973, he received the degree Doctor of Philosophy with a major in speech, the field of audiology, from the University of Florida, Gainesville, Florida. At the time of his graduation from California State College at Long Beach, he was the father of Andrea, Mark, and Joshua — all of whose help he could not have done without if he had wanted to get out of school any slower, but whose presence made every moment a memorable one. 147

PAGE 158

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. / / Kenneth C. "Pollock, Ph:D., Chairman Associate Professor of Speech I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. F. Owen Black, M.D. Associate Professor of Surgery I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^9 i T. (J J 3 sen, Ph.D. Associate Professor of Linguistics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. W. A. Yost, Assistant Professc/r of Speech

PAGE 159

This dissertation was submitted to the Department of Speech in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 197 3 Dean, Graduate School


138
Kobrak, H. G. The Middle Ear. Chicago: University of
Chicago Press, 1959.
Kristensen, H. K. and Jepsen, 0. Recruitment in otoneuralo-
gical diagnostics. Acta Oto-Laryngologica, 42:553-
560, 1952.
Lamb, L. E. and Peterson, J. Middle ear reflex measure
ments in pseudohypacusis. Journal of Speech and
Hearing Disorders, 32:42-51, 1967.
, Peterson, J., and Hansen, S. Application of
stapedius muscle reflex measures to diagnosis of
auditory problems. Journal of International Audiology,
7:188-199, 1968.
Licklider, J. C. R. The perception of speech. In Stevens,
S. S. (editor), Handbook of Experimental Psychology.
New York: John Wiley and Sons, 1951.
Liden, G. The stapedius muscle reflex used as an objective
recruitment test: A clinical and experimental study.
In Wolstenholme, G. E. W. and Knight, J. (editors),
Sensorineural Hearing Loss. A Ciba Foundation
..aa£ symposium, London: J~. and A. Churchill, 1970.
~ Peterson, J. L. and Harford, E. R. Simul
taneous recording of changes in relative impedance in
:--r--air pressure during acoustic and non-acoustic
elicitation of the middle-ear reflexes. Acta Oto-
Lryngologica Supplement 263:208-217, 1970.
Lilly, D. J. Some properties of the acoustic reflex in
man* Journal of the Acoustical Society of America,
!ll2QQ7 (A) 1964.
l.zr.zz^:: a comparison of acoustic impedance data obtained
with Madsen and Zwislocki instruments. Presented at
the-American Speech and Hearing Convention, Chicago,
I37J1.
. Acoustic impedance at the tympanic membrane.
In Katz, J. (editor), Handbook of Clinical Audiology.
Baltimore: Williams and Wilkins Company, 1972.
. Measurement of acoustic impedance at the
tympanic membrane. In Jerger, J. (editor), Modern
Bvelopments in Audiology, 2nd Edition. New York:
Academic Press, 1973.
i


30
TEST FREQUENCY
Figure 3. Pre- and post-noise exposure thresholds.
The lower thresholds are audibility and the upper thresh
olds are acoustic reflex in 11 cats. (Redrawn from Liden,
1970.)


143
Rasmussen, G. L. The olivary peduncle and other fiber
projections of the superior olivary complex. Journal
of Comparative Neurology, 84:141-219, 1946.
Reger, S. N. Differences in loudness response of normal
and hard-of-hearing ears at intensity levels slightly
above threshold. Annals of Otology, Rhinology and
Laryngology, 45:1029-1039, 1936.
Salomon, G. and Starr, A. Electromyography of middle ear
in man during motor activities. Acta Neurologica
Scandinavia, 39:161-168, 1963.
Sanders, J. W. and Honig, E. A. Brief-tone audiometry:
Results in normal and impaired ears. Archives of
Otolaryngology, 85:640-647, 1967.
, Josey, A. F., and Kemer, F. J. Brief-tone
audiometry in patients with VUIth nerve tumor.
Journal of Speech and Hearing Research, 14:172-178,
1971.
Scharf, B. Critical bands. In Tobias, J. B. (editor),
Foundations of Modern Auditory Theory, Volume I.
New York: Academic Press, 1970.
Schoel, T. and Arnesen, G. The choice of probe-tube
position and test frequency in determining the intra-
aural reflexes. Acta Oto-Laryngologica, 54:233-238,
1962.
Sheeley, E. C. and Bilger, R. C. Temporal integration as
a function of frequency. Journal of the Acoustical
Society of America, 36:1850-1857, 1964.
Sherrington, C. S. The integrative action of the nervous
system. London: Constable, 1906.
. The integrative action of the nervous system.
New Haven, Conn.: Yale University Press, 1947.
Shiftman, F. Bridge clinic. Impedance Newsletter, 1:11,
1972. (Madsen Electronics Corp., New York.)
Simmons, F. B. Post-tetanic potentiation in the middle
ear muscle acoustic reflex. Journal of the Acoustical
Society of America, 32:1589-1591, 1960.
. An analysis of the middle-ear muscle acoustic
reflex of the cat. In Fletcher, J. L. (editor),
Middle Ear Function Seminar. U. S. Army Medical Re
search Laboratory Report No. 576, Fort Know, Kentucky,
1963.


134
Durrant, J. D. and Shallop, J. K. Effects of differing
states of attention on acoustic reflex and temporary
threshold shift. Journal of the Acoustical Society
of America, 46:904-913, 1969.
Elliott, L. L. Tonal thresholds for short-duration stimuli
as related to subject hearing level. Journal of the
Acoustical Society of America, 35:578-580, 1963.
Ewertsen, H., Filling, S., Terkildsen, K., and Thomsen, K. A.
Comparative recruitment testing. Acta Oto-Laryngologica,
Supplement 140:116-122, 1958.
Feldman, A. S. Acoustic Impedance studies of the normal
ear. Journal of Speech and Hearing Disorders, 10:165-
176, 1967.
Fisch, U. and Schulthess, G. V. Electromyographic studies
on the human stapedial muscle. Acta Oto-Laryngologica,
56:287-297, 1963.
Fletcher, H. Auditory patterns. Review of Modern Physics.
12:47-65, 1940.
Flottorp, G. and Djupesland, G. Diphasic impedance change
and its applicability in clinical work. Acta Oto-
Laryngologica Supplement 263:200-204, 1970.
, Djupesland, G., and Winther, F. The acoustic
stapedius reflex in relation to critical bandwidth.
Journal of the Acoustical Society of America, 49:457-
461, 1971.
Fowler, E. P. A method for the early detection of oto
sclerosis. Archives of Otolaryngology, 24:731-741,
1936.
. The diagnosis of diseases of the neural
mechanism of hearing by the aid of sounds well above
threshold. Transactions of the American Otological
Society, 27:207-219, 1937.
. The recruitment of loudness phenomenon.
Laryngoscope, 60:680-695, 1950.
Fulton, R. T. and Lamb, L. E. Acoustic impedance and
tympanometry with the retarded: A normative study.
Audiology, 11:199-208, 1972.
Garner, W. R. The effect of frequency spectrum on temporal
integration of energy in the ear. Journal of the
Acoustical Society of America, 19:808-815, 1947a.


69
made for each signal duration at each frequency in order to
obtain a stable threshold value. The experimenter started
below the acoustic-reflex threshold and increased the
stimulus level in 2-dB steps until a minimum reflex was
detected. At this point, if the response was a definite
change from the ongoing pattern, the stimulus level was re
corded as the ART. The experimenter started the sequence
over again by lowering the intensity 10 dB. If the response
was in doubt, then another response was evoked at the same
signal level or at the next 2-dB higher intensity increment.
If the reflex response was definite at the second response
or the 2-dB higher response, the initially observed sound
pressure level of the reflex was recorded as the ART. Figure
12 illustrates a tracing of a typical acoustic-reflex thres-
old response. In this instance, a 2-dB intensity increase
caused a definite change from the previously recorded pattern.
The tenth ascending threshold response was stored on
channel two of the oscilloscope. The eliciting signal was
increased 5 dB above the reflex threshold and superimposed
upon the ART on channel two of the oscilloscope. The two
superimposed acoustic reflex images on channel two were then
photographed to retain a permanent record.
During the acousticreflex testing procedure, subjects
were requested to glance quietly through some pictorial
magazines in an attempt to equate subject state of alertness
and focus of attention. Subject state and attention have


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EFXU5FO5W_OXRHH9 INGEST_TIME 2014-11-25T19:40:28Z PACKAGE AA00026428_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


135
Garner, W. R. Auditory thresholds of short tones as a
function of repetition rates. Journal of the Acous
tical Society of America, 19:600-608, 1947b.
and Miller, G. A. The masked threshold of pure
tones as a function of duration. Journal of Experi
mental Psychology, 37:293-303, 1947.
Gengel, R. W. Auditory temporal integration at relatively
high masked threshold levels. Journal of the Acoustical
Society of America, 51:1849-1851, 1972.
and Watson, C. S. Temporal integration: I.
Clinical implications of a laboratory study. II.
Additional data from hearing-impaired subjects.
Journal of Speech and Hearing Disorders, 36:213-224,
1971.
Goldstein, R. and Kramer, J. C. Factors affecting thresh
old for short tones. Journal of Speech and Hearing
Research, 3:249-256, I960.
Graham, A. B. (editor). Alternate loudness balance
techniques. In Sensorineural Hearing Processes and
Disorders. Boston: Little, Brown and Company, 1967.
Grason-Stadler. Otoadmittance Handbook 2: A guide to
users of the Grason-Stadler Model 1720 Otoadmittance
Meter. Concord, Mass: Grason-Stadler Company, 1973.
Green, D. M., Birdsall, T. G., and Tanner, W. P. Jr. Signal
detection as a function of signal intensity and
duration. Journal of the Acoustical Society of
America, 29:523-531, 1957.
Gunn, W. J. Effects of attending to auditory signals on
the magnitude of the acoustic reflex. U. S. Army
Research Laboratory Report No. 751, Fort Knox,
Kentucky, 1967.
Harford, E. and Liden, G. Acoustic impedance, the middle-
ear muscle reflex, tympanometry and extratympanic
manmet.ry. Annual Report of the Auditory Research
Labortories. Northwestern University, 1967-1968.
Harris, J. D. Peak vs. total energy in threshold for very
short tones. Acta Oto-Laryngologica, 47:134-140, 1957.
, Haines, H. L., and Myers, C. K. Brief-tone
audiometry: Temporal Integration in the hypacusic.
Archives of Otolaryngology, 67:699-713, 1958.


APPENDIX A
MEDICAL HISTORY FORM


Facial
Nucleus
SOC
Lateral
Lemniscus
Temporal
Cortex
Trigeminal
Nucleus
Figure 2. The acoustic-reflex arc. The stapedial-reflex arc consists of the end-
organ, the primary nuclei, the SOC, the facial nucleus, and the stapedial muscle. The
tensor tympanireflex arc proceeds from the SOC through the lateral lemniscus to the
trigeminal nucleus, ending in the tensor tympani muscle. (Based on Rasmussen, 1946.)
o


CHAPTER Page
V. CONCLUSIONS AND SUMMARY 100
APPENDICES 104
A. MEDICAL HISTORY FORM 105
B. PURE TONE HEARING LEVEL (ANSI-1969) OF ALL
SUBJECTS 107
C. VOLTAGE DIVIDER 109
D. CALIBRATION DATA TABLES Ill
E. RAW DATA OF THE ACOUSTIC-REFLEX THRESHOLD . 115
F. DISCRIMINANT ANALYSIS 125
G. RAW DATA OF THRESHOLD OF AUDIBILITY 127
H. CORRELATION COEFFICIENTS OF THE TEMPORAL
INTEGRATION MEASURES 129
REFERENCES 131
BIOGRAPHICAL SKETCH 147
v


3
this standard detector of recruitment cannot be administered
to a majority of hearing-impaired persons to gain diagnostic
information.
The Monaural Bi-frequency Loudness Balance (MBLB) test
(Reger, 1936) compares the loudness above threshold at one
frequency with the loudness of another frequency. Again, the
threshold at one of the test frequencies must be normal with
the other elevated. In this way, recruitment can be tested
unilaterally, thereby allowing a larger segment of the
population to be tested diagnostically for cochlear lesions.
It is perceptually much more difficult for the patient to
perform than the ABLB and is, therefore, less reliable. This
is especially true as the difference between the standard
and comparison frequency increases (Graham, 1967) .
The Short Increment Sensitivity Index (SISI) test
(Jerger, Shedd, and Harford, 1959) indicates the ability of
the cochlear mechanism to respond to small changes in signal
amplitude. It is an indicator of cochlear disorders if a
high percentage of 1 dB increments is detected by the sub
ject, but it apparently is not as reliable a detector of
cochlear impairment as the ABLB measure (Owens, 1971; Tillman,
1969). Its lower reliability may be related to the patient
sophistication necessary to complete the procedure, which
is demonstrated by the difficulty in orienting some patients
to the auditory listening task.
The four Bekesy Types, introduced by Jerger (1960),
were based upon the difference in self-tracked threshold


128
Appendix G. Raw Data of Threshold of Audibility
500 Hz 1000 Hz 2000 Hz
Signal Duration in Msec
ject
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
Normal
Ears
CPL
31
25
16
17
12
26
24
14
15
14
29
25
20
19
17
CPR
27
25
13
13
12
31
26
19
13
11
33
30
23
20
20
jgl
42
37
27
27
26
38
33
25
26
25
39
34
28
27
27
jgr
42
38
35
31
31
36
36
26
24
23
38
33
27
26
26
JPL
26
16
9
6
3
20
18
10
10
8
30
26
18
14
13
JPR
24
17
7
6
3
24
16
12
11
9
24
21
14
15
13
fsl
21
13
6
6
5
10
8
0
1
0
14
9
4
3
1
fsr
15
10
3
2
1
26
15
7
6
2
18
12
11
8
6
gpl
32
29
19
18
18
31
24
18
16
16
35
32
23
22
24
gpr
40
36
27
25
21
29
22
12
8
8
22
19
11
11
11
Cochlear
-Good
Ears
ecr
35
27
21
19
17
30
25
20
15
12
32
26
22
21
17
fhl
37
29
23
21
17
36
34
26
25
22
42
34
28
24
23
acr
43
38
32
29
30
39
33
27
25
23
36
28
24
22
22
BCS
37
30
21
19
19
38
35
28
23
18
36
29
25
23
22
LSl Hearing Impaired Fre- 39 34 28 27 26 34 27 20 18 16
quency
Cochlear-Bad Ears
lcl
80
77
72
73
72
72
67
59
57
57
72
70
67
66
66
fiir
56
51
46
45
41
53
48
40
43
42
63
61
59
58
58
acl
85
78
72
72
70
85
81
76
77
75
87
82
77
75
74
bcl !
45
40
34
29
29
37
37
31
30
29
35
31
28
28
24
lsr
68
63
59
58
58
64
61
55
55
53
61
56
53
50
49
lsl
61
57
52
51
50
f


101
acuity. This is not to say that temporal integration of
the acoustic-reflex threshold can detect incipient hearing
loss or minor cochlear impairment. Instead, it reflects
the complexity of auditory processing.
In conclusion, there were some apparent, but subtle,
differences between the normal-hearing and hearing-impaired
groups in temporal integration at the ART.
o The mean integration slopes of the cochlear-impaired group
were depressed.
o The time constant of integration was shorter than 20 msec
for some of the cochlear-impaired groups,
o If the time constant is not unusually short, the integra
tion function changed gradually, then abrubtly after 20
msec in the cochlear-impaired ear.
o A cochlear-impaired ear can have both a flattened slope,
as well as a short time constant,
o One did not have to have a hearing loss in the cochlear-
impaired ear to demonstrate either a depressed slope or
a delayed time constant.
Since it has been stated that there are probably differ
ences between normal-hearing and cochlear-impaired groups,
this type of study is open for further investigation. A
similar experiment would require more signal durations,
expecially below 100 msec. Other parameters can be expanded
upon, such as type of stimuli or measuring ipsilaterally
instead of contralaterally to the test ear. Temporal


7
and outward from the oval window, as well as causing the
tympanic membrane to be pushed in a slightly outward direc
tion. The malleus and tympanic membrane swing medially upon
tensor tympani activation (Jepsen, 1963; Kobrak, 1959).
This is not the case, however, upon the contraction of both
muscles simultaneously. The movement of the tympanic membrane
depends upon the relative strength, latency, and contraction
time of each muscle. Figure 1 demonstrates a possible resul
tant of a stronger and faster-acting stapedius muscle con
traction partially opposed by the later and weaker tensor tym
pani muscle contraction. It is important to remember that it
does not matter acoustically which muscle predominates because,
in either case, the acoustic impedance will be increased.
The acoustic-reflex arc consists of an afferent neuron,
a reflex center, and an efferent neuron (Lorente de No, 1933,
1935; Rasmussen, 1946). The afferent portion is the same
afferent pathway for audition starting from the sensory end-
organ but terminating at the level of the superior olivary
complex (SOC). The SOC consists of at least five cellular
groups, but the accessory, or medial, nucleus is felt to be
the central mediator of both reflex arcs, the stapedial and
the tensor tympani. The accessory superior olive gives off
a few fibers to the ipsilateral motor nucleus of the facial
nerve (n. VII) (Rasmussen, 1946), which constitutes the
central portion of the stapedial-reflex arc. The facial
nerve innervates the stapedial muscle to complete the efferent
portion of this reflex arc.


BIOGRAPHICAL SKETCH
William Lee Parker was born in 1945, earned his first
dollar in 1954, and was married in 1966 to the delightful
Cristine Mary Sisson. He received his academic degrees
of Bachelor and Master of Arts with a major in speech
pathology at California State College at Long Beach. In
1973, he received the degree Doctor of Philosophy with a
major in speech, the field of audiology, from the University
of Florida, Gainesville, Florida. At the time of his gradua
tion from California State College at Long Beach, he was
the father of Andrea, Mark, and Joshuaall of whose help
he could not have done without if he had wanted to get out
of school any slower, but whose presence made every moment
a memorable one.
147


137
Jepsen, 0. Middle ear muscle reflexes in man. In Jerger,
J. (editor), Modern Developments in Audiology.
New York: Academic Press, 1963.
Jerger, J. Influence of stimulus duration on the pure-tone
threshold during recovery from auditory fatigue.
Journal of the Acoustical Society of America, 27:121-
124, 1955.
. Bekesy audiometry in analysis of auditory dis
orders. Journal of Speech and Hearing Research, 3:
275-237, 1960.
. The audiological examination as an aid in
diagnosis. Archives of Otolaryngology, 85: 552-
554, 1967.
. Clinical experience with impedance audiometry.
Archives of Otolaryngology, 92:311-324, 1970.
f Jerger, S., Ainsworth, J., and Caram, P.
Recovery of auditory function after surgical removal
of cerebellar tumors. Journal of Speech and Hearing
Disorders, 31:377-382, 1966.
, Jerger, S., and Mauldin, L. Studies in im
pedance autiometry: I. Normal and sensorineural
ears. Archives of Otolaryngology, 96: 513-523, 1972.
, Shedd, J., and Harford, E. On the detection of
extremely small changes in sound intensity. Archives
of Otolaryngology, 69:200-211, 1959.
I
Johansson, B., Kylin, B., and Langfly, M. Acoustic reflex
as a test of individual susceptibility to noise.
Acta Oto-Laryngologica, 64:256-262, 1967.
Karlovich, R. S., Lane, R. H., Smith, L. L., Tarlow, A. J.,
Thompson, D., and Vivion, M. C. Auditory threshold
at 125 Hz as a function of signal duration and signal
filtering. Journal of the Acoustical Society of
America, 49 :1897-1899, 1971.
Klockhoff, I. Middle ear muscle reflexes in man: A
clinical and experimental study with special reference
to diagnostic problems in hearing impairment. Acta
Oto-Laryngologica, Supplement 164:1-92, 1961
Kobrak, H. G. The present status of objective hearing tests.
Annals of Otology, Rhinology and Laryngology, 57:1018-
1026, 1948.


14
from an experimental method (Metz, 1946; Zwislocki, 1957) into
the clinical armamentarium with the introduction of several
commercially available acoustic-impedance meters (Grason-
Stadler, 1973; Madsen, n.d.; Peters, n.d.; Terkildsen and
Scott Nielsen, 1960; Zwislocki, 1963). With the increasing
interest in the United States concerning this technique as
a diagnostic tool, critical evaluations of methods and the
commercially available equipment have been made. These instru
ments have been found to be reliable (Nixon and Glorig, 1964;
Tillman, Dallos, and Kurvilla, 1963), sensitive (Moller,
1964), and also relatively easy to utilize (Brooks, 1971).
Threshold of the Acoustic Reflex (ART)
The acoustic reflex first occurs at a predictable level
above the normal threshold of audibility. The predictability
of this reflex accounts for its clinical usefulness; there
fore, it is forthwith described in detail along with factors
which can affect its performance and measurement.
For pure tones from 250 Hz to 4000 Hz the range of the
acousticreflex threshold (ART) is 70 dB to 90dB above the nor
mal-hearing individual's threshold (sensation level: SL) with
a mean of approximately 80 dB SL (Deutsch, 1968, 1972; Jepsen,
1951). Jerger, Jerger, and Mauldin (1972) report the range
measurement technique. If the reflex threshold concerns the
left ear, the reflex eliciting tone will be delivered to the
left ear. Because both sides will contract to unilateral
stimulation, the reflex action in the above example will be
recorded in the right ear. This eliminates acoustic inter
ference of the eliciting tone with the probe tone, which may
cause misleading results.


CHAPTER III
METHODS AND PROCEDURES
Temporal summation at both threshold of audibility and
threshold of the acoustic reflex were measured by the follow
ing experimental design. First, a baseline of temporal
integration at threshold of audibility was obtained in order
to compare the results in this study to previously reported
data on normal-hearing and cochlearimpaired subjects. This
information delineated normal integration from reduced inte
gration of cochlear-impaired ears. Second, the results of
temporal summation of the acoustic reflex determined whether
the null hypothesis should be rejected; that is, temporal
summation of the acoustic reflex does not function differ
ently in those subjects who have normal hearing and in those
subjects who have a known end-organ auditory lesion. The
results of this second auditory procedure were obtained with
out the subject's overt cooperation.
Ten cooperative adult subjects were tested. One-half
of the group had normal hearing and were the control subjects.
The other half of the group evidenced unilateral end-organ
hearing impairments and served as the experimental subjects.
These two groups were classified homogeneously and screened
for possible problems that might interfere with normal
acoustic-reflex function.
55


136
Hattler, K. W. and Northern, J. L. Clinical application
of temporal summation. Journal of Auditory Research,
10:72-78, 1970.
Hempstock, T. J., Bryan, M. E., and Tempest, W. A. A
redetermination of quiet thresholds as a function of
stimulus duration. Sound and Vibration, 1:365-380,
1964.
Hirsh, I. J., Palva, T., and Goodman, A. Difference limen
and recruitment. Archives of Otolaryngology, 60:525-
540, 1954.
Holst, H., Ingelstedt, S., and Ortegren, U. Ear drum move
ments following stimulation of the middle ear muscles.
Acta Oto-Laryngologica, Supplement 182:73-83, 1963.
Hood, J. D. Basic audiological requirements in neuro
otology. Journal of Laryngology and Otology, 83:
695-711, 1969.
Hughes, J. R. Auditory sensitization. Journal to the
Acoustical Society of America, 26:1064-1070, 1954.
. Electrophysical evidence or auditory sensiti
zation. Journal of the Acoustical Society of America,
29:275-280, 1957.
Hughes, J. W. The threshold of audition for short periods
of stimulation. Proceedings of the Royal Society of
Medicine, B133 : 486-490, 1946.
Hung, I. J. and Dallos, P. Study of the acoustic reflex
in human beings. I. Dynamic characteristics.
Journal of the Acoustical Society of America, 52:
1168-1180, 1972.
International Organization for Standardization. Standard
reference zero for the calibration of pure tone
audiometers. ISO R 389-1964. Geneva, Switzerland.
Jepsen, O. The threshold of the reflexes of the intra-
tympanic muscles in a normal material examined by
means of the impedance method. Acta Oto-Laryngologica,
39:406-408, 1951.
. Intratympanic muscle reflexes in psychogenic
deafness. Acta Oto-Laryngologica, Supplement 109:
61-69, 1953.


Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of
the Requirements for the Degree of Doctor of Philosophy
TEMPORAL SUMMATION OF THE ACOUSTIC-REFLEX THRESHOLD:
A POSSIBLE INDICATOR OF COCHLEAR ABNORMALITIES
By
William Lee Parker
August, 1973
Chairman: Kenneth C. Pollock, Ph.D.
Major Department: Speech
It has been suggested that temporal summation of the
ART might be a clinical tool distinguishing normal from
cochlear-impaired ears. If the ART measure could yield
differences at least as distinctive as those of temporal
summation at threshold of audibility, the ART could be
used with patients for whom threshold measures are not
appropriate.
In this study, temporal summation of the acoustic-
reflex threshold (ART), as well as of the threshold of
audibility, was measured in five normal-hearing and five
unilaterally hearing-impaired subjects by varying signal
duration of 500 Hz, 1000 Hz, and 2000 Hz tones. The signal
durations employed were 500 msec, 200 msec, 100 msec, 20
msec, and 10 msec.
There were statistically significant differences be
tween normal ears and cochlear-impaired ears at 2 out of
12 comparisons. In addition, there were some descriptive
differences between normal ears and cochlear-impaired ears
IX


60
Repetition rate
1 stim./1200 msec
Acoustic Muscle
reflex relaxation
Figure 8. Acoustic-reflex contraction. The intra-
aural muscles can contract, relax, and return to the pre
contraction baseline before contracting again. The
repetition rate is once every 1200 msec for a 500-msec tone
burst at 5 dB above the ART.


APPENDICES


Appendix E(continued)
Sub
ject
500 Hz
1000 Hz
2000 Hz
Signal Duration in Msec
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
7
119
109
103
99
97
111
102
92
90
92
117
114
108
104
102
8
121
111
105
101
97
111
100
90
94
90
119
114
108
106
102
9
121
109
103
101
99
109
102
92
92
92
121
114
108
104
104
10
119
111
103
103
99
107
102
90
92
88
119
114
108
104
102
1
137
123
109
107
107
132
116
106
102
100
NR
131
112
106
98
2
139
121
109
109
107
130
120
108
104
100
135
108
104
102
3
137
123
111
107
103
134
118
108
102
100
1 :1
108
104
102
4
135
123
109
107
105
136
120
106
102
98
129
108
104
102
5
133
125
111
107
109
130
118
106
102
100
133
106
102
100
6
139
121
111
107
107
134
120
108
100
98
135
110
102
100
7
139
125
111
103
109
134
118
110
102
98
133
108
102
102
8
139
123
111
105
107
134
116
108
102
98
133
110
102
102
9
137
121
111
107
107
136
118
108
100
100
133
108
104
102
10
135
121
111
109
109
136
118
106
102
100
133
108
104
102
1
138
119
107
105
103
130
114
106
104
102
123
115
98
94
94
2
132
119
109
107
105
128
116
104
102
102
125
115
98
94
96
3
138
121
109
107
103
126
116
106
102
102
129
113
100
94
94
4
134
119
109
107
103
126
112
108
102
100
133
115
100
92
94
5
134
123
109
105
103
128
114
106
102
102
129
111
100
96
96
6
134
123
109
107
105
128
114
108
102
100
131
111
100
96
92
7
136
121
109
105
101
130
116
106
102
102
131
113
100
94
94
8
134
123
109
105
103
124
114
106
102
102
135
123
100
96
94
9
136
121
109
105
103
130
114
104
104
102
131
119
98
96
94
10
138
121
109
105
105
132
114
108
102
102
131
119
100
96
96
GP,
GP,
119


42
1967; Sanders, Josey, and Kemer, 1971; Wright, 1968b; Wright
and Cannella, 1969). The integration slope is not affected
by either conductive problems (Miskolczy-Fodor, 1956; Harris
et al. 1958; Wright, 1968b; Wright and Cannella, 1969) or by
retrocochlear problems (Olsen et al., 1973; Sanders et al.,
1971). The critical time at which the temporal integration
begins, or the time constant of integration, also seems to
be shortened to about 100 msec or less depending on the par
ticular frequency (Miskolczy-Fodor, 1953, 1956, 1960; Harris
et al., 1958; Sanders and Honig, 1967; Wright, 1968a). The
abnormal results of cochlear damage to slope and time con
stant can be seen in Figure 5, i.e., the slope is flatter,
the time constant is shorter, or both.
It has also been suggested that brief-tone audiometry
may even be so sensitive as to detect incipient damage to the
cochlea (Campbell and Counter, 1969; Sanders and Honig, 1967;
Wright, 1968a) to detect recruitment (Miskolczy-Fodor, 1956,
1960), and to differentiate between various cochlear
lesions (Harris et al., 1958). Invariably, these attributes
are artifacts of unaccounted frequency effects, energy spread,
and/or inter-subject variability (Gengel and Watson, 1971;
Karlovich et ad., 1971). Olsen et a^. (1973) cast doubt on
the efficacy of temporal integration to differentiate be
tween end-organ and eighth nerve lesions because of the large
overlap between populations.


40
the theory of temporal summation is based on neural decay,
the minimum off-time should be 200 msec (Zwislocki, 1960).
In addition, a "critical off-time" has been calculated where
by successive stimuli having less than a 200 msec inter
stimulus interval take on the threshold value of one contin
uous tone (Jerger, Jerger, Ainsworth, and Caram, 1966) .
Doyle (1970) in a temporal integration study varied the inter
stimulus interval from 100 msec to greater than 500 msec.
There were no differences in temporalintegration results as
long as the dead time was a minimum of 500 msec, but the
slopes flattened as the off-time was shortened to 100 msec.
In practice, therefore, the inter-stimulus interval should
exceed the theoretical minimum, and probably be no less than
500 msec.
The repetition rate of the signal per unit of time,
usually one second, is also based on consideration of decay
of neural excitation. If one desires to keep the repetition
rate constant, then the rate is determined by the length of
the longestduration test signal and the desired inter-stimulus
interval. Wright (1968a) recommends the use of the same
repetition rate for all test stimuli so that sampling per unit
of time will be uniform.
Some attention has been given to attenuation rate for
those test procedures employing semi-automated equipment, for
example, Bekesy tracking. Hempstock et al. (1964) observed
an increase in the standard deviation of the threshold as the


35
interval, attenuation rate, method of threshold calculation,
and psychophysical procedure.
Earlier studies indicate that for a narrow band of
frequencies between 1000 Hz and 4000 Hz a 10 dB/decade slope
is generated (Garner, 1947a; Garner and Miller, 1947; Hughes,
1946; Munson, 1947), while narrow-or wide-band noise bursts
demonstrate about a 8 dB/decade change in threshold inte
gration (Garner, 1947b; Miller, 1948; Small, Brandt, and
Cox, 1962). There are no systematic differences in the slope
between threshold determination in quiet (Garner, 1947b;
Hempstock, Bryan, and Tempest, 1964; and others) and in back
ground noise (Garner and Miller, 1947; Blodgett, Jeffress, and
Taylor, 1958; Hempstock et a_l. 1964; Gengel, 1972; Plomp
and Bouman, 1959; and others).
More recently, a controversy concerning frequency effects
has developed. While 1000 Hz elicits an average slope of 10
dB per decade, lower frequencies have a steeper slope and an
increase in frequency produces a flatter slope, as seen in
Figure 4 (Brahe Pedersen and Elberling, 1972; Elliott, 1963;
Gengel, 1972; Gengel and Watson, 1971; Hattler and Northern,
1970; Hempstock et ad., 1964; Miskolczy-Fodor, 1953; Northern,
1967; Olsen and Carhart, 1966; Sanders, Josey, and Kemer,
1971; Sheeley and Bilger, 1964; Simon, 1963; Tempest and
Bryan, 1971; Watson and Gengel, 1969). Some investigators
feel that there is no frequency dependence (Clack, 1966;
Martin and Wofford, 1970; Wright, 1968a, 1972; Zwicker and


84
would account for the statistically significant findings at
500 Hz. Using discriminant analysis, the more distinct the
groups the less chance of error in classification (Rao, 1952).
It is notable that the point of greatest significant differ
ence (p<.05 at 500 Hz for normal vs. cochlear-bad) yields the
smallest difference between mean values (Figure 14).
Figure 18 compares the temporal integration slope of
the acoustic reflex of normal-hearing persons in three
studies. The individual mean scores from this study and
Djupesland and Zwislocki (1971) are plotted as well as the
median values. The group mean values from McRobert et al.
(1968) are also included. The average change in slope per
decade in time at 1000 Hz for this study is 6 dB less than
the 21 dB/decade from Djupesland and Zwislocki, as well as
2 dB less than the 25 dB/decade from McRobert et a_l. In
addition, it appears that the individual mean values for this
study were more scattered than those of Djupesland and
Zwislocki (1971), especially at the 10-msec signal duration.
This large scatter of individual scores may be repre
sentative of the true population, since no more than nine
normal-hearing subjects were employed in any one study.
These differences may also be a result of the type of
equipment employed, since the two lower values of 19 dB and
21 dB/decade were obtained with a Peters electroacoustic
bridge and a Madsen electroacoustic bridge, respectively.
The steeper slope of 25 dB/decade was obtained with a


51
a difference in reflex eliciting intensity levels for pure
tones versus white noise. The intensity level difference
between the two types of stimuli is much larger in normals
than the range in cochlearimpaired ears.
Norris, Stelmachowicz, and Taylor (1972) have reported
the most significant difference in the acousticreflex action
between normal and pathological cochleas. They measured the
difference in levels necessary to obtain the acoustic-reflex
threshold. The stimulus is pulsed at 2.5 pulses per second,
a 50 percent duty cycle with a 200-msec signal duration.
The resulting modulation of the reflex in response to the
pulsed tone is greater and more regular in normal ears than
in sensorineural hearing losses. This may reflect an in
creased latency of acousticreflex response in cochlear-
impaired ears (Johansson et ad., 1967; Simmons, 1963) or a
distorted spatial summation involving rate of reflex growth
with increased stimulus level (Simmons, 1963).
If effects due to disturbed spatial summation are present
in the acoustic reflex as a result of cochlear pathologies, it
is reasonable to assume that end-organ lesions can also
disturb temporal summation. Cochlear lesions can and do
disturb temporal integration at threshold of audibility, that
is, by flattening the slope of temporal integration and/or
shortening the critical duration. Intuitively, then, tem
poral integration of the acoustic reflex may also be affected
by cochlear lesions, for example, a flattened slope and/or
a shortened critical duration.


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Kenneth C.
Associate
Pollock,
Professor
Pn:D., Chairman
of Speech
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
7
F. Owen Black, M.D.
Associate Professor of Surgery
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Assistant Professor of Speech


87
<
Figure 19. Signal-duration interval in which the time
constant of temporal integration for each test subject occurred
across all frequencies.


Correlation Coefficients of the Measures Between
Temporal Integration of the Acoustic-Reflex Threshold and
Temporal Integration of the Threshold of Audibility.
Frequency
Correlation
Signifi
cance
20 msec-200 msec measure
Normal ears
500
Hz
0.3763
ns
1000
Hz
0.6926
ns
2000
Hz
0,6349
ns
Cochlear-good ears
500
Hz
0.9177
.05
1000
Hz
-0.9768
.01
2000
Hz
0.0186
ns
Cochlear-bad ears
500
Hz
-0.2654
ns
1000
Hz
-0.0995
ns
2000
Hz
0,6880
ns
10 msec-200 msec measure
Normal ears
500
Hz
0,0943
ns
1000
Hz
0.3887
ns
2000
Hz
0.6397
ns
Cochlear-good ears
500
Hz
-0.7857
ns
1000
Hz
0.3493
ns
2000
Hz
-0.9661
.01
Cochlear-bad ears
500
Hz
-0.1461
ns
1000
Hz
0.1460
ns
2000
Hz
0.7719
ns
ns = not statistically significant.
.05 = significant at the 5 percent level.
.01 = significant at the 1 percent level.
130


20
critical bandwidth. The ART is more or less constant as the
bandwidth is increased to a critical value. Beyond this
critical bandwidth the acoustic-reflex threshold will decrease
as the bandwidth increases (Flottorp, Djupesland, and
Winther, 1971; McRobert et al., 1968). Moller (1962b), in
contrast, did not see a consistent change in all subjects as
bandwidth was increased, but it is possible that he did not
extend his bandwidths far enough. Moller did obtain improve
ment in threshold for one subject for which the bandwidth sur
passed the critical width determined by Flottorp et al.
(1971) at 525 Hz fc. The work by McRobert et al. (1968)
established that the lowering of ART with increasing band
width is dependent upon the center frequency, and is more
pronounced at 1000 Hz fc. Apparently, the critical bandwidth
with which the reflex mechanism responds to auditory stimuli
widens with increases in sound pressure level (Hung and
Dallos, 1972).
The acousticreflex threshold does not seem to differ
regardless of ascending or descending stimulus presentation
approach (Beedle, 1970; Harford and Liden, 1967-1968; Peterson
and Liden, 1970, 1972), although Deutsch (1968, 1972) reports
a systematic improvement of threshold over three test trials.
Deutsch attributes the ART improvement of approximately 2 dB
from trial to trial to "auditory sensitization" (Hughes,
1954, 1957). Simmons (1960) labels this response condition
"post-tetanic potentiation" and attributes this sensitivity


40
30
20
10
0
40
30
20
10
0
40
30
20
10
0
96
I 1
l_ H
J 1
hi
J rl
i
i
Normal ears
Cochlear-good ears
Cochlear-bad ears
500 Hz
1000 Hz
i
2000 Hz
i i ¡ 1 r
10 20 100 200 500
SIGNAL DURATION IN MSEC
23. Inter-subject range of thresholds of audi-
function of temporal integration.


Page
Figure
15.
16.
17.
18.
19.
20.
21.
22.
23.
Acoustic-reflex thresholds of normal and
cochlear-impaired (bad) ears as a function
of temporal summation 81
Acoustic-reflex thresholds of normal (good)
and impaired (bad) ears of the Meniere's
group as a function of temporal summation . 82
Inter-subject range of acoustic-reflex thresh
olds as a function of temporal integration 83
A comparison of mean acoustic-reflex thresh
olds as a function of temporal summation in
normal-hearing ears 85
Signal-duration interval in which the time
constant of temporal integration for each
test subject occurred across all frequencies 87
The delayed time constant in some cochlear-
impaired ears contrasted with the normal
time constant of temporal integration at the
acoustic-reflex threshold 89
Thresholds, audibility, and acoustic reflex of
the test subjects at 500-msec signal duration 92
Mean thresholds of audibility as a function of
temporal integration 94
Inter-subject range of thresholds of audibility
as a function of temporal integration .... 96
viii


71
been reported to affect the acoustic reflex (Durrant and
Shallop, 1969? Gunn, 1967). As long as the subject's atten
tion is generally diffuse and not under intense concentra
tion, e-g., distinguishing visual word choices or auditory
intelligibility, the variability is probably slight (Durrant
and Shallop, 1969).
Experimental Safeguards
In performing temporal integration at the threshold of
the acoustic reflex, sound pressure levels normally asso
ciated with thresholds of discomfort, tickle and pain, and
temporary and permanent threshold shifts were required. It
was necessary to safeguard the subjects against any harm or
unnecessary discomfort. Possible response artifacts result
ing from high intensity signals were also investigated.
It was conceivable that all subjects might be subjected
to stimulus levels in excess of 130 dB SPL, as seen in Table
5. This table demonstrates a possible intensity level neces
sary to elicit a reflex by a 10-msec tone burst at 500 Hz.
For this hypothetical person, the maximum earphone output
would be 136.5 dB SPL.
The Committee on Hearing, Bioacoustics and Biomechanics
of the National Academy of Sciences has published upper limits
of acceptable exposure to impulse noise as a function of
sould pressure level and signal duration as seen in Figure 13
(Ward, 1968). The solid line represents the upper acceptable
limit of exposure for probably 95 percent of the population


THRESHOLD dB RE HEARING LEVEL (ANSI-1969)
92
Figure 21. Thresholds, audibility, and acoustic reflex,
of the test subjects at 500-msec signal duration.


47
time constant is about 200 msec and the threshold changes
2 dB to 3 db per doubling of signal duration.
There are relatively few data on temporal integration
of the acoustic-reflex threshold in humans. The systematic
published works are by McRobert, Bryan, and Tempest (1968)
and Djupesland and Zwislocki (1971). In both instances data
are plotted in terms of acoustic-power growth as signal dura
tion decreased while maintaining acoustic-reflex threshold.
Their slopes, reported in power to maintain reflex threshold
per average decade change in signal duration, are in the
range reported by Simmons (1963). The slopes range from 14
dB/decade for 300 Hz, to 21dB/decade for 2500 Hz (McRobert
et al., 1968), and to 25 dB/decade for 1000 Hz (Djupesland
and Zwislocki, 1971). Although there are no reports of a
systematic frequency effect (McRobert et a_l. 1968; Simmons,
1963), the middle frequencies apparently have steeper slopes.
Figure 6 represents the effects of temporal integration
upon the threshold of the acoustic reflex in six normal hear
ing subjects by Djupesland and Zwislocki (1971). The median
of the individual threshold means demonstrates a regular
decrease of 5 dB to 7 dB per halving of duration, or a slope
of about 25 dB/decade. The time constant is seen to be, for
the most part, about 200 msec. One subject appears to begin
integration at 100 msec with no obvious change in slope, as
seen by the dashed line connecting the lowest mean ART data.
McRobert et a^. (1968) also present slopes which break
linearity between 100 msec and 200 msec.


142
Norris, T. W., Stelmachowicz, P., and Taylor, D. Effects
of stimulus variations on acoustic reflex patterns.
Presented at International Audiology Congress, Budapest,
1972.
Northern, J. L. Temporal summation for critical bandwidth
signals. Journal of the Acoustical Society of America,
42:456-461, 1967.
Olsen, W. 0. and Carhart, R. Integration of acoustic power
at threshold by normal listeners. Journal of the
Acoustical Society of America, 40:591-599, 1966.
, Rose, D. E., and Noffsinger, D. Brief tone
audiometry with normal, cochlear, and VIII nerve
tumor patients. A paper accepted for publication
Archives of Otolaryngology, 1973.
Owens, E. Bekesy tracing and site of lesion. Journal of
Speech and Hearing Disorders, 29:456-468, 1964.
. Audiologic evaluation in cochlear versus
retrocochlear lesions. Acta Oto-Laryngologica,
Supplement 283:1-45, 1971.
Perlman, H. B. and Case, T. J. Latent period of the crossed
stapedius reflex in man. Annals of Otology, Rhinology
and Laryngology, 48:663-675, 1939.
Peters Ltd. Peters AP 61 Acoustic Impedance Meter: Oper
ating instructions. New York: Lehr Instrument
Corporation, n.d.
Peterson, J. L. and Liden, G. Dynamics of the stapedial
muscle reflex. Presented at the 10th International
Audiology Congress, Dallas, Texas, 1970.
and Liden, G. Some static characteristics of
the stapedial muscle reflex. Audiology, 11:97-114, 1972.
Plomp, R. and Bouman, M. A. Relation between hearing
threshold and duration for tone pulses. Journal of
the Acoustical Society of America, 31:749-758, 1959.
Price, G. R. Influence of external ear acoustics on im
pulse arriving at the ear drum. Journal of the
Acoustical Society of America, 52:129 (a), 1972.
Rao, C. R. Advanced Statistical Methods in Biometric
Research. New York: John Wiley and Sons, 1952.


Appendix E(continued)
500 iHz
1000 Hz
2000 Hz
Sub
ject
Signal Duration in Msec
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
1
117
104
i 93
93
91
111
100
94
96
94
120
110
104
106
102
2
119
103
93
93
91
112
102
96
98
94
123
111
106
101
101
3
112
104
93
91
91
112
102
96
98
92
124
111
106
101
101
4
113
104
95
93
91
114
103
96
98
90
126
111
106
102
102
5
117
103
95
91
93
112
103
96
98
92
127
111
108
102
102
6
113
108
95
91
95
112
103
94
96
92
124
110
110
102
101
7
113
103
93
91
93
112
105
96
98
94
124
110
108
104
101
8
117
104
93
93
95
112
103
96
96
92
124
111
106
106
99
9
115
103
95
93
91
114
103
96
98
92
125
111
106
106
101
10
113
104
95
93
91
112
103
96
98
96
125
110
106
104
102
1
113
108
99
93
83
105
102
96
92
90
108
102
91
91
85
2
111
106
97
95
83
109
100
96
94
88
106
102
95
91
87
3
111
106
99
99
85
109
100
98
96
90
110
101
95
93
87
4
110
106
101
99
89
109
102
96
90
88
111
101
95
91
85
5
111
108
99
97
87
109
102
94
94
92
111
102
95
91
87
6
113
111
101
99
89
109
102
96
94
88
113
102
93
93
85
7
108
110
101
99
91
109
103
98
94
88
110
99
93
93
89
8
115
110
99
101
91
107
98
98
94
90
111
102
93
91
87
9
113
108
101
97
89
107
102
96
96
90
110
101
95
89
87
10
115
100
101
99
93
107
102
96
94
88
111
102
93
89
91
1
117
111
101
99
99
107
98
94
90
88
119
112
110
104
102
2
125
111
105
101
99
111
100
92
88
88
123
112
106
104
100
3
119
111
105
101
99
111
100
92
90
88
119
115
108
106
104
4
119
111
101
101
97
111
98
92
92
88
119
114
110
104
102
5
117
113
103
103
99
111
98
94
90
88
119
112
110
104
104
6
123
111
103
101
97
112
100
92
92
90
119
115
110
106
104
JP,
FS,
FS
R
\
118


53
systematic fashion in the normal ear as a function of stimu
lus duration, i.e., a tenfold change in time can cause a
median change of 25 dB of acoustic power. The null hypo
thesis is: temporal summation of the acoustic reflex does
not function differently in those subjects who have normal
hearing and in those subjects who have a known end-organ
auditory lesion.
The questions that this study attempted to investigate
are:
1. What is the most sensitive measure of the acoustic-re
flex threshold when using increments of 2 dB: the first
response, the first level to occur 60 percent of the
time, a measure of central tendency of either 5 or 10
trials, or the lowest response of 10 trials?
2. What is the group average and range of acousticreflex
thresholds for a normal-hearing population as a function
of temporal integration?
3. What is the group average and range of acoustic-reflex
thresholds for an unilaterally cochlear impaired popula
tion as a function of temporal integration?
4. Is there a significant difference in temporal integra
tion of the acoustic reflex between normal hearing ears
and cochlear impaired ears?
5. Is there a significant difference in temporal integration
of the acoustic reflex between the ears of normalhearing
subjects and the normalhearing ears of unilaterally
cochlear impaired subjects?


Appendix E(continued)
Sub
ject
500 Hz
1000 Hz
2000 Hz
Signal Duration in Msec
Run
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
6
129
119
113
109
109
116
106
96
94
94
123
104
102
98
94
7
129
117
111
109
107
120
106
98
96
94
129
110
100
98
94
8
129
115
115
113
105
120
110
98
96
94
129
110
102
98
96
9
127
119
113
111
105
120
108
98
96
92
131
108
104
96
94
10
129
119
113
111
107
118
102
98
96
92
129
110
102
100
96
1
105
107
97
103
95
100
90
90
88
90
108
98
106
102
98
2
107
105
99
99
99
100
90
92
90
90
108
98
100
100
100
3
109
107
99
103
97
98
90
92
90
90
108
102
104
104
100
4
109
109
99
99
97
100
90
90
88
90
106
100
100
102
100
5
109
105
95
97
97
100
92
90
86
88
104
102
108
102
96
6
107
105
101
97
95
100
90
90
90
90
110
104
98
102
100
7
109
101
103
103
95
100
90
90
90
90
104
102
100
100
102
8
107
105
101
101
95
102
90
88
90
90
108
102
100
104
104
9
107
107
103
101
97
102
90
90
88
90
110
102
102
102
102
10
107
107
101
99
95
100
88
90
90
90
108
102
100
102
100
1
117
103
93
93
91
135
122
110
106
104
135
128
116
104
104
2
117
103
93
95
91
135
124
110
106
106
135
130
120
104
104
3
119
101
93
95
93
134
124
114
104
106
135
130
112
104
102
4
119
103
93
95
93
135
126
112
106
106
137
130
116
104
102
5
119
101
93
93
91
139
130
110
110
104
134
135
116
102
104
6
119
101
95
93
91
139
126
110
112
104
137
130
116
104
102
7
115
103
95
93
93
139
124
110
112
106
134
130
116
102
102
8
113
101
95
93
93
139
126
108
110
104
137
132
114
104
104
9
119
103
95
95
93
137
122
110
110
104
135
134
116
102
104
10
117
101
95
95
93
139
128
108
110
106
137
134
120
104
102
FH
R
AC,
122


102
integration measures do not have to take place at threshold,
nor by changing the stimulus durations only. This type of
investigation should be pursued because the acoustic reflex
can reflect a cochlear abnormality without the subject's
cooperation.
In summary, five normal-hearing and five unilaterally
hearing-impaired persons were test subjects. Their results
determined the efficacy of temporal summation of the acoustic
reflex threshold as a possible predictor of cochlear abnor
malities. Temporal summation of the acoustic-reflex thresh
old was obtained by maintaining a constant intra-aural muscle
reflex strength as a function of increased acoustic power
while signal duration was decreased.
A measure of temporal integration was made at both
threshold of audibility and threshold of the acoustic reflex,
the first auditory measure was determined by the subject and
the second by the experimenter. Bekesy tracking was employed
to self-record threshold of audibility, while a modified
method of limits, using 10 ascending trials, determined the
acoustic-reflex threshold. The signal durations used were
500 msec, 200 msec, 100 msec, and 20 msec at the test fre
quencies of 500 Hz, 1000 Hz, and 2000 Hz. As much as 139
dB sound pressure level was employed to maintain constant
energy at threshold. The reflex was monitored from an
oscilloscope connected to a Peters 61AP electroacoustic
bridge.


ApRsn<3:j.?c R(pqn^iRVi^d))
50Q Hz 1000 Hz 2000 Hz
Sulp- Test Signal Duration in Msec
ject
Run,
10,
20,
100
200
500
10
20
100
200
500
10
20
100
200
500
5,
117,
108
106
108
130
108
106
102
120
104
102
97
6
| | '1
117
108
106
108
T
130
108
108
102
120
106
100
97
7,
1?4
108
100,
106
r-
130
110
104
100
119
106
102
97
8,
1 1
118,
108
108
108
130
110
106
104
117
106
102
97
9
l l i
117
108
104,
108
128
108
108
102

106
102
98
10
1 1 !
123
108
108
108

130
108
106
100
119
104
104
97
JGr
1
1 1 !
NR,
} i >
127,
115
110
115
NR
130
110
106
102
NR
139
114
112
103
2
i l '
125
113
110
113
12 8
110
106
102
139
114
114
103
3
127
113
111
113
130
112
106
100
137
116
112
101
4
127
113
111
115
130
112
108
102
139
114
112
101
i :
5,
131
113
113
115
130
110
108
104
135
116
112
101
i.
6
1 I l
131
117
111
113
134
112
108
102
137
114
114
103
7
1 1 1
129
115
111
113
! 1 '
130
114
106
102
137
114
112
101
8
||ft
129
115
111
115
1 *
128
110
108
104
137
114
114
103
9
1 l 1
129
115
113
115
j \ 1
130
110
110
102
139
118
112
105
10
127
113
113
113
1 '
132
110
108
100
139
114
112
103
JPL
1
122
110
101
99
99
121
116
96
100
96
138
133
117
115
111
2
122
113
101
101
99
120
116
98
102
96

131
119
115
111
3
126
115
99
101
99
122
116
98
102
96
138
133
115
115
111
4
127
115
101
101
97
120
118
96
100
96
138
131
117
113
111
i :
5,
125
113
101
103
99
123
118
98
102
96
138
133
117
115
109
6
12 3
115
101
103
97
122
119
98
100
94

135
117
115
115
7
126
115;
99
101
99
125
118
94
98
98

135
119
117
111
8
126
117
103
103
97
125
118
96
98
96
138
135
117
119
113
9
127
115
103
101
99
123
120
98
98
96
138
131
117
113
111
10
126
115
103
101
97
122
119
98
98
98
138
129
117
115
111
117


32
occurs when the involuntary acoustic-reflex threshold is
compared with the threshold of audibility. When the dynamic
range between the two thresholds is 55 dB or less, there is
cochlear impairment. By independently varying the signal
duration at threshold of audibility, more diagnostic informa
tion is gained. This manipulation of auditory processing, as
seen in the next section, may be applied to the acoustic re
flex threshold as well.
Temporal Summation
There is a psychophysical assumption that the normal
auditory system can summate, or integrate, acoustic power over
some critical period of time. Since the initial work of Garner
(1947b), Garner and Miller (1947), Hughes (1946), Munson (1947),
and de Vries (1948), considerable interest has been generated
over the concept of "trading" increased acoustic power with
a decreased signal duration to maintain a constant loudness
level. When the on-time of the auditory stimulus is sequen
tially halved, starting from some critical duration, the
signal power is reduced 2 dB to 3 dB with each decrement, i.e.,
from 200 msec to 100 msec, from 100 msec to 50 msec, etc.
The critical long duration is thought to be about 200 msec
and a linear relationship holds to a critical short signal
length of approximately 10 msec. ~
Perfect integration
If one assumes perfect integration (Garner, 1947a,b;


144
Simmons, F. B. and Dixon, R. F. Clinical implications of
loudness balancing. Archives of Otolaryngology,
83:449-454, 1966.
Simon, G. R. The critical bandwidth level in recruiting
ears and its relation to temporal summation. Journal
of Auditory Research, 3:109-119, 1963.
Small, A. M., Jr., Brandt, J. F., and Cox, P. G. Loudness
as a function of signal duration. Journal of the
Acoustical Society of America, 34:513-514, 1962.
Swannie, E. M. Impedance audiometry in clinical practice.
Proceedings of the Royal Society of Medicine, 59:
971-974, 1966.
Tempest, W. and Bryan, M. E. The auditory threshold for
short-duration pulses. Journal of the Acoustical
Society of America, 49:1901-1902, 1971.
Terkildsen, K. Movements of the eardrums following intra-
aural muscle reflexes. Archives of Otolaryngology,
66:484-488, 1957.
. The intra-aural muscle reflexes in normal
persons and in workers exposed in intense industrial
noise. Acta Oto-Laryngologica, 52:384-396, 1960a.
. An evaluation of perceptive hearing losses in
children, based on recruitment determinations. Acta
Oto-Laryngologica, 51:476-484, 1960b.
. Acoustic reflexes of the human muscle tensor
tympani. Acta Oto-Laryngologica, Supplement 158:
230-238, 1960c.
. Clinical application of impedance measurements
with a fixed frequency technique. Journal of Inter
national Audiology, 3:147-155, 1964.
/ Osterhammel, P., and Scott Nielsen, S.
Impedance measurements: Probe-tone intensity and
middle-ear reflexes. Acta Oto-Laryngologica, Supplement
263:205-207, 1970.
and Scott Nielsen, S. Electroacoustic impedance
measuring bridge for clinical use. Archives of
Otolaryngology, 72:339-346, 1960.
Thomsen, K. A. The Metz recruitment test. Acta Oto-
Laryngologica, 45:544-552, 1955a.


15
10
5
0
15
10
5
0
15
10
5
0
94
. a Normal ears
Cochlear-good ears
Cochlear-bad ears
i i 1 r
10 20 100 200
SIGNAL DURATION IN MSEC
22. Mean thresholds of audibility as a function
integration.


6
pathways, a bilateral change in acoustic impedance^ in both
ears to acoustic stimuli. Upon simultaneous contraction,
the two intra-aural muscles, the stapedius and tensor tympani,
act in a physiologically antagonistic manner but produce a
synergistic impedance against sound energy.
The stapedius muscle is the smallest muscle in the human
body. It originates from the pyramidal eminence on the
posterior wall of the tympanic cavity and its tendon inserts
on the head, neck, or posterior crus of the stapes. The
tensor tympani muscle arises from the bony semicanal above
the Eustachian tube. Its tendon traverses the tympanum to
insert on the mallar manubrium (Jepsen, 1963; Kobrak, 1959) .
The direction of pull of these two muscles is at right angles
to the axis of their corresponding ossicles, making the func
tional action of the two muscles almost in direct opposition
to one another (Wever and Lawrence, 1954).
If the movement caused by each of the muscles is con
sidered independently of the other, contraction of the
stapedius muscle causes the stapes to be pulled posteriorly
3 .
It is not within the scope of this paper to differ
entiate between the types of impedance present within the
middle ear, nor between the various contributing factors of
increased impedances (see Hung and Dallos, 1972; Lilly,
1972, 1973; Zwislocki, 1961, 1962). It is important, however,
to know that the contraction of one or both of the intra-
aural muscles will produce an opposition to energy flow
through the middle-ear cavity, thereby reducing the acoustic
energy reaching the sensory end-organ.


test series was 700 msec. Preliminary investigation also
indicated that this rate would allow for the muscle contrac
tion to return to the pre-contraction baseline before re
sponding to the next signal burst. Figure 8 illustrates a
tracing of an oscilloscope recording of the acoustic reflex
evoked by a 500-msec signal at 5 dB above the ART. The
muscle activity returned to the pre-contraction baseling
within the 700-msec inter-pulse interval. It should also be
noted that this recording was made towards the end of an
hour and one-half test session, and middle-ear muscle fatigue
was not observed.
The rise and decay times of the stimulus envelope, as
controlled by a pre-set electronic switch, were approximately
5 msec. The 5-msec ramp was verified on an oscilloscope by
determining the duration between the upper 90 percent and
lower 10 percent of the slope.
Figure 9 delineates the spectral characteristics of two
1000-Hz tone burst stimuli measured at the earphone output.
The bandwidth of the 10-msec duration tone pulse is about
120 Hz wide in Figure 9a. This is about 20 Hz wider than
might be expected theoretically, indicating some incidental
spread of energy. When the tone burst is lengthened to 20-
msec duration in Figure 9b, the spectral energy fits nicely
into the theoretical bandwidth of 50 Hz, The spectrum
of these tone bursts suggests that no audible spread of energy
should be detected; this is supported by the smooth envelope


Appendix E. Raw Data of the Acoustic Reflex Threshold
CTl
Sub
ject
JGL
500 Hz
1000 Hz
2000 Hz
Test
Signal Duration in Msec
nun
10
20
100
200
500
10
20
100
200
500
10
20
100
200
500
Normal
Ears
i
123
108
98
96
94
126
124
100
98
94
127
120
99
97
95
2
119
111
98
96
94
124
116
100
98
96
127
125
97
97
93
3
119
111
101
98
96
126
116
100
96
96
125
122
97
97
97
4
126
113
98
96
96
124
120
100
100
96
127
125
97
95
95
5
121
112
101
94
94
126
114
101
96
96
127
123
97
93
95
6
128
111
100
98
94

116
101
98
96
127
123
97
97
97
7
123
111
100
98
94
126
116
101
96
96
127
127
95
95
97
8
121
111
100
98
96
126
112
101
98
96

120
97
97
95
9
126
113
100
96
96
126
116
103
100
96
127
118
97
97
93
10
126
117
101
96
96

116
101
98
96
125
123
97
95
95
1
126
116
98
90
84
126
105
94
88
84
121
110
97
95
93
2
124
110
98
88
84
124
107
92
88
86
125
113
97
93
91
3
126
112
98
88
86
128
107
92
90
84
127
110
95
93
93
4
126
116
100
92
86
126
105
94
90
86
125
115
97
95
91
5
124
114
98
92
88
124
109
92
84
86
'
115
95
95
91
6
126
116
98
94
84
126
107
92
86
86
125
110
97
95
91
7
124
116
100
92
86
124
107
94
84
84

113
95
95
91
8

114
98
92
84
128
109
92
84
88

115
97
93
93
9
128
114
98
94
86
126
114
92
88
88
125
110
97
93
91
10
128
114
98
96
86
128
109
92
86
86
125
115
97
95
93
1
NR
115
108
106
104
139
128
110
106
102
NR
118
106
102
97
2
117
108
104
106

124
108
106
102
119
106
100
100
3
120
104
104
108
139
126
110
104
100
118
106
102
97
4
118
108
108
108
139
128
110
106
102
123
102
100
97


79
good ear of the cochlear-impaired group and the normal
hearing group. However, neither the mean values, nor the
average slope per decade of stimulus time, were significantly
different among the three groups (Table 7 ) .
Table 6. Discriminant Analysis of the Acoustic-Reflex
Threshold *
500 Hz
1000 Hz
2000 Hz
Normal vs. cochlear-good
ears
10 msec-200 msec
ns
ns
ns
20 msec-200 msec
ns
ns
ns
Normal vs. cochlear-bad ears
10 msec-200 msec
ns
ns
ns
20 msec-200 msec
.05
ns
ns
Cochlear-good vs. cochlear-bad
ears
10 msec-200 msec
ns
ns
ns
20 msec-200 msec
.10
ns
ns
ns = not significant
.05 = significant at
.10 = significant at
the 5
the 10
percent
percent
level.
level.
Table 7. Average Slope Change Per Decade of
Acoustic-Reflex Threshold
Time
of the
500 Hz
1000 Hz
2000 Hz
Cochlear-good ears
23 dB
22
dB
25 dB
Normal ears
17 dB
19
dB
19 dB
Cochlear-bad ears
15 dB
14
dB
14 dB


To my wife, Christine