Changes in acoustic reflex activity following two hours of industrial noise exposure in normal hearing subjects


Material Information

Changes in acoustic reflex activity following two hours of industrial noise exposure in normal hearing subjects
Physical Description:
vi, 169 leaves : ill. ; 28 cm.
Rodriguez, Gary Philip, 1956-
Publication Date:


Subjects / Keywords:
Noise -- Physiological effect   ( lcsh )
Acoustic reflex   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1986.
Includes bibliographical references (leaves 157-168).
Statement of Responsibility:
by Gary Philip Rodriguez.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000893991
notis - AEK2518
oclc - 15300487
System ID:

Full Text








I would like to thank the faculty, staff and students

of the Department of Speech for all their help. Their

support and friendship made my stay at the University of

Florida truly enjoyable. My sincere thanks go to Dr. Kenneth

Gerhardt, chairman of my supervisory committee, whose knowledge,

patience and guidance made this project possible. He embodied

all of the qualities student could possibly hope for,

giving freely of his time and himself whenever called upon.

I will always feel indebted to him, and hold his friendship

in the highest regard.

I also wish to thank my supervisory committee, Drs.

Alan Agresti, Alice Holmes, Joseph Kemker, and Patricia

Kricos. Their expertise, advice, and encouragement were

invaluable through my entire doctoral program.

Special thanks go to my parents, whose love and support

remain a source of strength in my life, and my children,

Casey and Rebecca, who helped keep everything in proper

perspective, even during the roughest of times. My deepest

thanks go to my wife, Diane, whose friendship and love

provide meaning to all my endeavors.




ABSTRACT . . . . . v


Introduction . . . 1
Review of the Literature . 4
Temporary Threshold Shift . 4
Physiology of the Acoustic Reflex 12
Theories on Middle Ear Muscle
Function . . . 16
Acoustic Immittance . . 22
Acoustic Reflex Threshold . 33
Acoustic Reflex Magnitude ... 37
Temporal Characteristics of the
Acoustic Reflex ... . 39
Acoustic Reflex Adaptation . 43
Acoustic Reflex Properties and
Temporary Threshold Shift . 51
Statement of Purpose . . 55


Subjects . . . 59
Instrumentation and Procedures . 59
Exposure . . ... 59
Behavioral Hearing Threshold
Measurements . . 62


Pre-Exposure Measures . . 86
Behavioral Thresholds . 86
Acoustic Reflex Threshold . 87
Acoustic Reflex Magnitude . 90
Acoustic Reflex Latency . 92
Acoustic Reflex Adaptation .. 97



Effects of the Industrial Noise
Exposure . . . 106
Behavioral Threshold . 106
Acoustic Reflex Threshold . 111
Acoustic Reflex Magnitude . 114
Acoustic Reflex Latency . 114
Acoustic Reflex Adaptation . 118
Correlations . . . 124
Acoustic Reflex Threshold and
Reflex Threshold Shift . 124
Acoustic Reflex Magnitude . 126
Acoustic Reflex Latency . 126
Adaptation . . . 129


Interpretation of Results: Pre-
Exposure Measures . . 131
Acoustic Reflex Threshold . 131
Acoustic Reflex Magnitude . 132
Acoustic Reflex Latency . 134
Adaptation . . ... 139
Effects of the Industrial Noise
Exposure . . . .. 143
Behavioral Threshold Shift -. 143
Acoustic Reflex Threshold . 146
Acoustic Reflex Magnitude . 147
Acoustic Reflex Latency . 149
Adaptation . . . 151
Correlations . . . 153
Summary . . . 155

REFERENCES . . . . . 157


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



Gary Philip Rodriguez

August, 1986

Chairman: Kenneth J. Gerhardt, Ph.D.
Major Department: Speech

Research indicates that certain parameters of acoustic

reflex (AR) function change following noise exposure, most

notably, AR threshold. One dimension of AR function that has

not been studied extensively is the relaxation or offset

response of the acoustic reflex system. There is also a

paucity of information regarding AR adaptation in response

to industrial noise stimuli. Although continuous noise

exposures may answer some of the questions regarding AR

behavior, they cannot explain the role of the acoustic

reflex system in an industrial environment.

This study evaluated changes in behavioral thresholds

and AR activity following a two-hour industrial noise

exposure. Results show that significant threshold shift

was obtained, along with changes in AR threshold, AR

offset latency, and certain measures of AR adaptation.

Potential relationships between various measures of

acoustic reflex behavior, and expressions of temporary

threshold shift (TTS) were also investigated. Findings

indicate that AR magnitude and offset latency may be useful

in explaining some of the variability typically encountered

in TTS studies.



The acoustic reflex arc is a unique component of the

auditory system because of it's ability to process energy

in several different forms. The middle ear acts as an

impedance-matching device which facilitates the transforma-

tion of air-conducted disturbances into hydromechanical

events. The cochlea then encodes this hydromechanical

action into neural information which is sent to certain

regions of the caudal brainstem. The brainstem nuclei

through their efferent connections cause contraction of

the middle ear muscles which results in an impedance

change back at the tympanic membrane. As one might imagine

the interaction of these various sub-components is quite

complicated. Because of these complex relationships,

however, the acoustic reflex arc provides an excellent

opportunity to study several regions of the auditory system


Research indicates that abnormalities anywhere along

the acoustic reflex (AR) system can cause a change in AR

properties. If a conductive hearing loss is present,

acoustic reflex thresholds may be elevated or absent.

Sensorineural hearing loss also results in a change in

threshold and certain temporal characteristics of the

acoustic reflex response. Retrocochlear pathology is

known to delay or prevent the occurrence of certain AR

events. Noise exposure has also been reported to cause

changes in parameters of the acoustic reflex, namely

magnitude and threshold. This is most likely due to the

changes created within the cochlear partition, which serves

to encode the acoustic signal for neural transmission.

Only limited information is currently available

regarding the temporal properties of the acoustic reflex

arc. The primary focus of recent research efforts has been

in understanding the onset properties of the acoustic

reflex response in normal and some pathologic populations.

For example, it is well established that the onset properties

of the AR are directly related to certain parameters of the

eliciting stimulus. Intensity, duration, rise-time, and

spectral content of the activating signal all have signifi-

cant effects on the acoustic reflex response. Offset

latency, on the other hand, shows little or no dependence

on stimulus parameters. Because of this difference, some

researchers suggest that the process of AR contraction and

relaxation are inherently different and should be studied

separately. It has also been suggested that AR offset

latency may represent a closer estimate of neural conduction

time and, therefore, should be the focus of future research


Currently, data are available which suggest that AR

temporal characteristics in subjects with sensori-neural

hearing loss may differ from normal hearing populations.

The etiology of these apparent delays is still unknown.

Norris, Stelmachowicz, Bowling, and Taylor (1974) reported

offset latencies of greater duration for a group of subjects

with sensorineural hearing loss when compared to a group

of normal hearing listeners. They theorized that differ-

ences observed between the two groups were caused by changes

in cochlear function of the hearing-impaired subjects. In

contrast, Borg (1976) showed longer AR offset latencies in

animals with brainstem lesions. He suggested that the group

of hearing-impaired subjects used in the previously mentioned

study presented concomitant retrocochlear involvement.

Even less evidence exists regarding the effects of noise-

induced hearing loss on temporal properties of the AR.

Temporary hearing loss in normal hearing subjects offers

a unique model of cochlear stress, while controlling for

possible contaminating effects of retrocochlear involvement.

Such experiments may resolve some of the issues currently

under debate regarding AR temporal characteristics.

In an effort to explore the behavior of the acoustic

reflex in industrial environments, this investigator evaluated

changes in AR parameters following a two-hour industrial

noise exposure. This evaluation consisted of several

measures of AR properties including temporal characteristics,

both onset and offset, AR threshold, magnitude, and adaptation.

Potential relationships between various AR properties and

expressions of temporary threshold shift were also investi-


Review of the Literature

Temporary Threshold Shift

Various responses to noise have been reported in the

literature (Davis, 1958; Dougherty, 1970; Kryter, 1970).

The primary effect of noise on health, however, remains the

production of noise-induced hearing loss. This connection

between loud sound and hearing loss has been recognized

for hundreds of years. The first reference noting the

effect of noise on human hearing appears to be credited

to Pliny the Elder, in about 600 A.D. (Ward, 1980a). He

noted that persons living near the waterfalls of the Nile

were "strucken deaf" from the loud roar of the water.

Noise soon became one of the undesirable by-products of

man's progress. Increasing noise exposure in our highly

industrialized society has made the problem of noise-induced

hearing loss difficult to ignore. The effect of noise

on the auditory system is now the subject of great concern

and involves the efforts of many agencies and individuals.

Noise can result in permanent or temporary changes

in hearing sensitivity. The principal method of investigating

noise induced hearing loss is the measurement of threshold

shift. Threshold shift is documented by measuring changes

in hearing sensitivity prior to and following a specified

noise exposure. Temporary threshold shift is known to

recover within some time interval to pre-exposure levels.

Permanent threshold shift on the other hand, is irreversible,

and does not recover throughout the lifetime of the organism.

As a rule, many of the same factors which influence temporary

threshold shift (TTS) also affect the development of perman-

ent threshold shift (PTS). Such factors as intensity,

duration, and spectrum of the noise along with individual

variability have been correlated to the development of

both TTS and PTS. This is not to suggest that the under-

lying properties of TTS and PTS are identical, however,

certain parallels do exist between the two. The inherent

problem in studying PTS is the ethical responsibility of

the researcher not to induce a permanent change in the

auditory sensitivity of his subjects. The investigation

of PTS, therefore, must be limited to animal models which

only serve as approximations of the human auditory system.

Although previous studies indicate that TTS cannot be used

to predict PTS (Mills, 1984; Ward, 1980a, 1980b), both are

dependent on certain cochlear events which are predicated

by parameters of the fatiguing stimulus.

The time interval between the end of the exposure and

when threshold is measured represents a critical variable

in the quantification of TTS (Melnick, 1978). The pattern

of TTS recovery is complex and is influenced by the process

of adaptation and sensitization of the auditory system.

Hirsh and Ward (1952) first reported a polyphasic pattern

of TTS recovery which occurs within the first few minutes

following the cessation of the stimulus. They termed this

rapidly changing recovery pattern the "bounce effect."

As illustrated in Figure 1, there is a rapid recovery of

threshold in the first minute following the cessation of

the noise. This is then followed by an increase in threshold

shift where it reaches a maximum level at approximately two

minutes into the post-exposure period. Recovery then

continues on a monotonic pattern until hearing threshold

returns to pre-exposure levels. To avoid the complicated

interaction of these changing auditory processes, TTS

is usually measured at about two minutes post-exposure time


The degree of temporary threshold shift (TTS) depends

on many factors including the spectrum, intensity, duration,

and temporal pattern of the fatiguing stimulus (Gerhardt,

Melnick & Ferraro, 1979). Broadband stimuli typically

Figure 1.


1 2 3 4 5 6 7 8 9 (mins)

Polyphasic recovery pattern of temporary
threshold shift (adapted from Hirsh & Ward,

have their greatest effect in the 3.0 to 6.0 kHz region.

This is most likely due to several factors including resonance

characteristics of the outer and middle ear, excessive

mechanical stress on certain areas of the cochlear parti-

tion, and circulatory patterns within the inner ear itself

(Wiener & Ross, 1946; Shaw, 1974; Tonndorf, 1976; Pickles,

1982). When discrete frequencies are utilized as the

fatiguing stimulus maximum TTS is typically measured one-

half to one octave above the exposure frequency (Davis,

Morgan, Hawkins, Galambos, & Smith, 1950; Hirsh & Bilger,

1955). Ward (1973) has also reported that in general,

high frequency noise exposures produce greater TTS than

comparable exposure levels of lower frequency stimuli.

Another important factor influencing the development

of TTS is the interaction between sound intensity and

exposure duration. Melnick (1976) reports that TTS

increases as intensity and duration of the exposure increase.

When evaluating the effects of noise neither element can be

considered alone. In an effort to describe the effects of

exposure intensity level however, we must assume that other

variables, such as duration and frequency spectra, are held

constant. Under these constraints TTS usually increases

linearly with increments of sound pressure level (Melnick,

1978). This linear relationship exists only after the

exposure exceeds a lower limiting intensity level of

about 70 to 75 dB SPL (Ward, Cushing, & Burns, 1976). It

is also known that moderate intensity exposures create TTS

that is proportional to the sound pressure level of the

noise, while extremely high intensity stimulation may

actually decrease the TTS developed (Davis, Morgan,

Hawkins, Galambos & Smith, 1950; Miller, 1958). Ward

(1968) speculates that this response may be due to a change

in the way the stapes vibrates at the oval window after

maximum contraction of the middle ear muscles.

Exposure duration represents another critical variable

in the development of TTS. Experiments involving both

human and animal subjects with exposures greater than eight

hours have shown TTS to increase as duration of the exposure

increases up to a certain level and then to plateau (Mills,

Gengel, Watson & Miller, 1970; Melnick & Maves,*1974).

Carder and Miller (1972) observed this TTS growth pattern

in their study of chinchillas and coined the term "asymptotic"

threshold shift (ATS). ATS has been reported to occur

somewhere between eight and sixteen hours after exposure in

human subjects. Idealized curves of TTS development for

the maximally affected frequency as a function of exposure

duration and intensity are illustrated in Figure 2

(Melnick, 1978). Note there is a lower intensity level,

equivalent quiet, which does not produce TTS regardless

of exposure duration. However, when this level is exceeded,

TTS will grow linearly with logarithmic exposure duration

up to eight to twelve hours and then level off at asymptotic








" EQ + 10

Equivalent Quiet (EQ)

I-------!-, I t- -
1 2 4 8 16 24
Hours 1 3
Day Days

Figure 2. Idealized pattern of TTS as a function of
exposure duration and intensity (adapted
from Melnick, 1978).

threshold shift. It should also be apparent that the growth

rate of TTS depends on the intensity of the noise as

previously mentioned.

The time pattern of the exposure is another parameter

which is known to influence the development of TTS. When

the exposure consists of noise that is fluctuating or

intermittent, a complicated relationship begins to develop

between total duration of the noise exposure and the

degree of TTS it produces. Under such circumstances the

on-time and off-time of the fatiguing stimulus must be

considered when relating the exposure to TTS. It can be

assumed, from limited research, that intermittent sounds,

in general, produce less TTS than continuous exposures of

the same total duration. For particular exposure conditions,

TTS has been found to be proportional to the ratio of time

occupied by the sound to the total time of exposure. This

relationship is known as the "on fraction rule" (Ward,

1970; Ward, Glorig & Sklar, 1958). For example, the

on-fraction rule would predict that when a sound occupies

only one-half of the total exposure time, the magnitude

of TTS developed would be about one-half of that which

would have been produced if the fatiguing stimulus had

been continuous. It should be noted, however, that this

relationship is applicable only for sound bursts of 250

msec to two minutes, and noise with frequency spectra

above 1200 Hz (Selters & Ward, 1962).

Although susceptibility to TTS has been found to be

normally distributed, the variance among individuals is

quite large. Not only do differences exist between individ-

uals for a particular noise exposure, but differences

exist in the same person for different types of noise

exposures ( Ward, 1968). Standard deviations between

subjects are typically in the range of 5-10 dB, while

intrasubject variability reveal standard deviations of

approximately 2 to 4 dB. Intersubject variation also

becomes greater with increases in test frequency (Melnick,


Physiology of the Acoustic Reflex

The acoustic reflex is a bilateral non-voluntary

response that occurs in the presence of intense acoustic

stimulation. Although research indicates that the AR can

be elicited by non-acoustic stimuli (Dallos, 1961), this

section will deal with the auditory response of the acoustic

reflex arc. Contraction of the middle ear muscles results

in a change of impedance at the tympanic membrane and

helps control the transmission of sound through the middle

ear system. This control is thought to occur from an

increase in stiffness and a functional decoupling of the

middle ear structures.

Two middle ear muscles are involved in acoustic

reflex action. Both the tensor tympani and stapedius are

pennate muscles which, by design, are able to exert a

great deal of tension with a minimal amount of displacement.

The tensor tympani muscle is approximately 25 mm in length

and rises from the cartilaginous portion of the eustachian

tube. The muscle itself is enclosed in a bony capsule from

which a tendon emerges. The tendon exits the bony enclosure

and turns around the cochleariform process before inserting

on the manubrium of the malleus. This muscle is innervated

by a motor branch of the trigeminal nerve (CN. V) via the

otic ganglion (McPherson & Thompson, 1977). Upon contraction

the muscle pulls the malleus and tympani membrane inward

(Jepsen, 1963).

In man, the stapedius muscle is chiefly responsible

for the impedance changes observed at the tympanic membrane.

The stapedius muscle is approximately 63 mm in length and

arises within the pyramidal eminence. It is completely

enclosed throughout its entire length in a bony capsule.

A tendon from the muscle emerges from the pyramid, and

inserts at the neck of the stapes. The stapedius muscle

is innervated by a motor branch of the facial nerve (CN. VII)

(McPherson & Thompson, 1977). Upon contraction the muscle

pulls the stapes down and outward. This movement causes

a stiffening of the ossicular chain which increases the

impedance of the middle ear system (Jepsen, 1963).

The acoustic reflex arc is made up of neurons connected

to muscle fiber. Borg (1973) described the reflex arc

as being a very secure, short latency pathway. In addition

to the direct neural tracts, there are probably numerous

indirect, parallel, multisynaptic pathways. These neural

pathways are largely unknown but may involve certain

elements of the extrapyramidal system. The direct neuronal

pathway consists of three to four neurons: the primary

afferent neurons carry impulses from the hair cells to the

cochlear nucleus; the second synapse is located at the

ventral cochlear nucleus; the third neuron is in the

superior olivary complex (SOC). The SOC has connections

with both ipsilateral and contralateral nuclei of the facial

nerve (CN. VII), which innervates the stapedius-muscles

(Borg, 1973).

The reflex arc contains motorneurons of different

sizes, and diameters which dictate their particular conduc-

tion properties. Basic neuronal physiology indicates that

in general, smaller diameter neurons will have slower

conduction velocities when compared to larger diameter

neurons. This premise appears to hold true for the reflex

arc. The thinner motorneurons of the stapedius muscle are

known to have lower thresholds and slower conduction

velocities. They also, as a rule, connect to slower

contracting muscle fibers. As a result of this anatomical

arrangement, a weak input to the reflex system will excite

mainly small, slow conducting neurons which activate slow

contracting muscle fibers (Borg, 1976). The intensity

dependent rise-time of the AR is thought to result from an

orderly recruitment of faster units in response to higher

intensity input. Higher intensity stimuli excite larger

diameter, fast conducting neurons, which in turn connect

to fast contracting muscle fibers. Research indicates that

as the stimulus intensity increases, rise-time of the response

decreases. This finding is compatible with an orderly

recruitment of neural elements, though other possibilities

exist (Borg, 1976).

Borg (1972) has also been able to demonstrate that

changes in middle ear impedance closely relate to EMG

activity of the stapedius muscle. Good correlation was

also demonstrated between the amplitude of the integrated

EMG and magnitude of the AR response. In a related study

Zakrisson, Borg, and Blom (1974) made simultaneous EMG

recordings of the stapedius muscle in one ear while monitor-

ing changes of acoustic impedance activity in the other ear.

Results showed that at low intensity levels EMG activity

directly related to changes of impedance at the tympanic

membrane. Cessation of the eliciting stimulus, however,

revealed an apparent "after-discharge" firing in the EMG

activity for up to 400 msec. This finding may partially

explain why offset latencies of the acoustic reflex are

generally longer than onset responses.

The latency from stimulus onset to the first motor-

unit action potential generally decreases with increments

of stimulus intensity. As previously mentioned this may be

the result of an orderly recruitment of motorunits with

differing thresholds and conduction velocities. However,

offset properties have been demonstrated to be relatively

independent of stimulus parameters, and are generally longer

in latency. As demonstrated by Zakrisson, Borg and Blom

(1974) this may be due to basic physiology of the motor

unit and after-discharge firings observed during EMG

recordings. Therefore, from the evidence available, it

appears that acoustic reflex latency characteristics can

be explained by neural conduction properties and after-

discharge firings of the individual motorunits. The

results of this and other research seem to indicate that

the acoustic reflex arc has properties commonly seen in other

polysynaptic pathways (Borg, 1976).

Theories on Middle Ear Muscle Function

The middle ear of man is a sophisticated biological

system that serves as an impedance-matching device between

the sound-conducting medium of air and inner ear fluids.

The structures involved in this transforming process are

the tympanic membrane, ossicular chain, and middle ear

muscles. Understanding the function of the middle ear

muscles has been the focus of scientific investigation for

many centuries. Observations completed on both animal

and human models have resulted in several hypotheses

regarding the contribution of the intra-aural muscles to

the process of audition. Borg, Counter, and Rosler (1984)

have summarized the findings of numerous experiments into

four major theories: 1) the intensity-control protection

theory, 2) the ossicular chain fixation theory, 3) the

accomodation-frequency selection theory, and 4) the

labyrinthine pressure regulation theory.

1) 'Intensity control-protection theory

In the seventeenth century, Fabricius ab Aquapendente

(1600) speculated that the function of the middle ear

muscles was to prevent the tympanic membrane from rupturing

during intense exposure to sound. von Helmholtz (1868)

believed that the middle ear muscles caused a diminution

in the sound pressure level transmitted to the inner ear

labyrinth. Kato (1913) was one of the first to observe

contraction of the middle ear muscles to moderate sound

pressure levels. Because of his systematic approach to

the study of the intra-aural muscles, the intensity-control

theory emerged as the dominant theory of acoustic reflex

physiology during the early decades of the century.

2) The ossicular chain fixation theory

This theory proposed that the major function of the

middle ear muscles was to maintain ossicular chain position

and continuity. The general view was that the muscles

served no functional role in hearing, but were merely

present to maintain the ossicles in their respective

positions in a state of readiness for transmission of sound

to the inner ear labyrinth (Borg, Counter, & Rosler, 1984).

3) The accommodation-frequency selection theory

The accommodation-frequency selection theory states

that middle ear muscle contraction acts like a filter and

allows selective transmission of certain frequencies through

the middle ear system. Certain investigators believed that

the tension exerted on the tympanic membrane by muscle

contraction involuntarily tuned subject's ears to various

sounds (Du Verney, 1683). Lucae (1874) assigned even more

specific roles to middle ear musculature. He postulated

that lower frequency sounds (below 6.0 kHz) were accommodated

by tensor tympani contraction, and assigned the role of

high frequency accommodation to the stapedius muscle. The

most intriguing version of this theory was presented by

Striker (1880), who theorized that the mere thought of

sound or melody could cause involuntary activation of the

middle ear muscles.

4) The labyrinthine pressure regulation theory

This theory states that contraction of middle ear

muscles (mainly the tensor tympani), causes an increase in

the pressure of labyrinthine fluids, and thereby dampens

the acoustic energy reaching the inner ear. Kato (1913)

discredited this theory after his series of systematic

observations of the middle ear system. He found no

pressure increase at the labyrinth during experimentally

induced tensor tympani contraction.

While the intensity-control hypothesis is generally

the theory of choice, certain arguments have been raised

regarding the protective function of the middle ear muscles.

One major objection is that the latency of muscle contraction

is relatively slow and, therefore, cannot effectively

attenuate many types of acoustic stimuli. Secondly, it is

widely recognized that AR contraction fatigues quite rapidly

to continuous intense sounds. Research also appears to

suggest that only low frequency sounds are attenuated by

middle ear muscle contraction, whereas noise-induced hearing

loss occurs in the higher frequencies (3.0 to 6.0 kHz).

For these reasons, the protective function of the acoustic

reflex has been challenged. The protective theory has also

been criticized from an evolutionary point of view. Many

individuals argue that there are no natural environmental

reasons for the development of a system which protects the

ear from intense twentieth century noise.

Although it is true that acoustic reflex is too slow

for a protective function from single impulse sounds, this

may not be the case for other types of noise exposures.

It should be noted that rapid adaptation of the AR activity

has typically been demonstrated for continuous intense

stimuli (Djupesland, Flottorp & Winther, 1966). However,

industrial exposures typically include a rapid succession

of fluctuating or impulse type sounds. Under these circum-

stances the relaxation properties of the AR may be more

important. Borg (1976) demonstrated the AR has the ability

to reactivate and recover during industrial noise exposures

thereby enhancing and preserving its attentuation properties.

In an experiment utilizing shipyard noise, Borg, Nilsson,

and Liden (1979) demonstrated only minor AR fatigue in

workers following a full day of intense industrial noise

exposure. In another study involving Bell's palsy patients,

Zakrisson, Borg, Liden, and Nilsson (1980) revealed

significantly greater TTS in the affected or paralyzed ear

than in the ear with normal acoustic reflex activity. Greater

hearing loss extending in the low and mid frequencies

was evident in the ear without AR function. We can,

therefore, speculate that auditory trauma would be even

greater under certain conditions following noise exposure

without the presence of the acoustic reflex system.

The evolutionary need of the AR may be demonstrated

if one closely examines the intense vocalizations made by

animals and man. The AR is known to be active during many

self-stimulating activities such as talking, chewing,

eating and screaming. Measurements made at 20 cm from the

head of adults during voluntary loud vocalizations have

been reported to reach 126 dB SPL (Borg, Counter & Rosler,

1984). These one-second vocalizations correspond to

approximately 132 dB at 10 cm from the mouth, and have the

same energy as a 15-minute exposure of noise at 102 dB

Leq (Borg, Counter & Rosler, 1984). Therefore, when

activated by self-stimulation or external stimuli, the

middle ear muscles can help maintain the sensitivity of

our auditory system. Without the action of the AR,

auditory fatigue may occur even more rapidly, resulting in

deterioration of auditory performance. Therefore, it seems

reasonable to assume that the AR acts to increase the

dynamic range of the auditory system and helps to maintain

sensitivity in a variety of listening environments.

The AR may also serve to enhance speech perception

abilities. As previously mentioned the chief effect of the

AR activation is the attenuation of sound below 2.0 kHz.

This frequency selectivity may provide a partial remedy for

the masking effect low frequency vowel sounds have on higher

frequency consonants. Research indicates that this low

frequency attenuation may serve to improve speech perception

in many competing listening situations, and thereby improve

communication ability (Zakrisson, 1974).

In light of the evidence presented, Borg, Counter and

Rosler (1984) speculate that the function of the middle

ear muscles may be quite extensive in many listening

situations. They postulate that, under certain circumstances,

the acoustic reflex can prevent interference, minimize

injury, and improve auditory communication ability. If

this is the case, the middle ear muscles appear to provide

a unique mechanism to control auditory input which allows

the organism to separate relevant from irrelevant sounds.

Acoustic Immittance

Acoustic immittance is the general term used to

describe the transfer of acoustic energy through the middle

ear system. This energy transfer can be expressed in terms

of acoustic impedance or acoustic admittance (Wiley &

Block, 1979). One of the most attractive aspects of

acoustic immittance measurements is that clinical use of

the equipment requires very little understanding of

impedance or admittance concepts. The vast number of

terms and procedural differences which exist from one

manufacturer to the next causes a considerable amount of

confusion to the student of impedance audiometry (Popelka,

1984). Audiologists should understand, at least on a

conceptual basis, the physical principles involved in

these measurements to minimize possible errors produced

in their clinical and research efforts. Knowledge of the

limitations of these measurements must also be considered

in order to make full use of immittance technology.

The term impedance is used for mechanical, electrical

and acoustic systems. Mechanical impedance is the ratio

between an applied force and velocity with which the system

is moved. Electrical impedance is the ratio of the applied

voltage to the current flow. Acoustic impedance relates

the applied sound pressure level to volume velocity. Volume

velocity can be understood by considering a sound pressure

being applied to a unit area of the eardrum. If one assumes

that all the particles on that unit area, and in the air for

a certain depth behind the eardrum, move in a uniform manner,

then a unit volume will be set into motion with a certain

velocity, the volume velocity (Bennett, 1984). Acoustic

impedance then is the ratio between the sound pressure

level present, and the volume velocity which is produced.

Therefore, conceptually, acoustic impedance is a way of

noting the physical response of a system. The unit of

measure of acoustic impedance is the ohm. The acoustic ohm

has been arbitrarily defined as occurring when a sound pres-

sure of one dyne produces a volume velocity of air of one

cubic centimeter per second (ANSI, 1960). In other words, anohm

is a unit of measure that expresses the resistance to

energy flow at the surface of the eardrum (Feldman, 1976).

Three inherent characteristics of impedance interact

in a complex manner to determine the mobility of the middle

ear system: mass, resistance and stiffness. Mass is

related to the density of the elements of the system. In

the middle ear, mass is composed of the weight of the

ossicles, and the tympanic membrane. Resistance occurs

whenever there is a change in the applied energy of a moving

system to another form, usually heat. The resistance

factor in the middle ear is minimal, due to the suspension of

the ossicles by muscles and ligaments. Stiffness is the

tendency of a system to retain its original shape and

position. In the middle ear system, it is generally

attributed to the motion of the stapes and resistance of

the inner ear fluids. The effect of mass, friction, and

stiffness generally determines the total impedance of the

system (Northern & Grimes, 1978). Acoustic reactance is

the imaginary component of impedance resulting from the

stiffness and mass of the system, and is the component

that expresses the storage and return of acoustic energy

(Feldman & Wilber, 1976). The formula for impedance is

presented below:

Z = / R2 + (21ff-M S/21rf)2

where Z = impedance, R = resistance, M = mass, S = stiffness,

f = frequency, andlr = 3.14 (Northern & Grimes, 1978).

Any mathematical description of impedance must also

include the parameter of time. It is convenient to express

time in angular fractions of one wavelength of applied

energy. In Figure 3, we can see that a unit volume being

displaced uniformly for a sinusoid signal actually has

three separate characteristics of motion: displacement,

velocity, and acceleration. The phase relationships of

(a) displacement, (b) velocity, and (c) acceleration

differ considerably over time. At the positive and negative

peak of displacement (a) the motion stops momentarily

before starting in the opposite direction, resulting in

zero volume velocity at that point in time. The maximum

velocity (b) occurs as the displacement passes through the

baseline. The acceleration of the system (c) will be at

its maximum when the rate of change in velocity is

greatest (as the velocity curve passes through baseline).

Examination of the curves show that displacement lags the

velocity by 90 degrees, and acceleration leads it by 90

degrees. The relationships between these components of

motion can now serve as references for impedance quantities.

If velocity is arbitrarily placed at 0 degrees phase

angle, we can now discuss certain relationships of the

middle ear system as they relate to the properties of

motion. Recall, that the resistance of the middle ear



Figure 3. Sinusoidal curves for (a) displacement,
(b) velocity, and (c) acceleration. Note that
displacement lags velocity by 90 and accelera-
tion leads the velocity by 90 (adapted from
Bennett, 1984).

system is provided by the fluids in the cochlea, and is

defined as an opposition to the flow of energy. This is

similar to the opposition a spring offers to mechanical

motion. Lagging the resistance by 90 degrees is stiffness

reactance, Xs. This is analogous to displacement of a

spring. The other reactive component, mass reactance (Xm),

derives its properties from the mass multiplied by acceler-

ation, and is the force required to overcome the inertia

of the system. Figure 3 illustrates that the mass reactance,

represented by (c) acceleration, is 180 degrees out of

phase with the stiffness reactance Xs, represented by

displacement (a). To compute the overall impedance of the

system, the net reactance must simply be linked to the

resistive component.

Vectors lend themselves to graphic representation and

are mathematical descriptors of physical events which

involve both magnitude and phase. Figure 4 demonstrates

that the resulting vector has a length /Z/, which is the

magnitude of the impedance without consideration to phase

angle (0). Note if the net reactance (Xm-Xs) is very

large when compared to the resistive component R, then the

phase angle will be large. This is typically the case for

a low frequency probe tone (220 Hz) where stiffness dominates,

resulting in a phase angle towards -90 degrees (Bennett,

1984). It should be noted, however, that mass reactance

Xm, is frequency dependent, and therefore, is affected by


Figure 4. The vectoral components of impedance (adapted
from Bennett, 1984).

probe tone frequency. Differing probe tones may have a

profound effect on impedance vectors quantities. Also as

frequency increases, the middle ear structures must move

faster resulting in less displacement, therefore, the

stiffness reactance (Xs) is reduced. All of these

variables must be considered for any mathematical trans-

formation involving impedance or admittance values (Block

& Wiley, 1979).

While acoustic impedance provides a measure of a

systems opposition to the flow of energy, admittance, which

is the reciprocal of impedance, represents the ease with

which sound is transmitted.. Like impedance, acoustic

admittance is a vector quantity and has two subcomponents,

susceptance (B) and conductance (G). Acoustic conductance

(G) represents the flow of energy through a resistance.

Acoustic susceptance (B) is an expression of the storage of

energy, and represents the reciprocal feature of reactance.

The relationship between admittance (Y) and impedance (Z)

is presented in Table 1 adapted from Northern and Grimes


Because conductance (G) is in phase with the pressure

(displacement) rather than velocity, the positive susceptance

(B) component is related to stiffness, and the negative

one to mass. This results in positive vector relationships

as shown in Figure 5. It should be noted, however, that

conductance and susceptance are not merely inverses of

resistance and reactance. Measuring both (G) and (B) has

Table 1
Immitance terminology
(adapted from Northern & Grimes, 1978)

Impedance Admittance

Units of Units of
Measure Terms Characteristics Terms Measure

Ohms Resistance Friction Conductance Mhos

(R) (G)

Reactance Susceptance

Positive (Xa) Mass Negative (-Ba)

Negative (-Xa) Stiffness Positive (Ba)


z M


Figure 5.

The vectorial components of admittance
(adapted from Bennett, 1984).

distinct advantages when one is adjusting for ear-canal

contribution to overall measurements made at the probe tip

(Block & Wiley, 1979; Bennett, 1984). Therefore, after

adjusting for ear-canal volume, and considering phase

relationships,the overall admittance of the middle ear

system can be determined with the following formula.

|Y = G2 + B2

Tympanometry is the measurement of compliance change

as air pressure is varied in the external canal (Northern

& Grimes, 1978). This procedure is extremely useful in

determining eustachian tube function, and middle ear pressure

values. In essence, the tympanogram is a measure of the

relative change in sound pressure level within the ear-canal

cavity. When the probe tone is introduced into the ear-

canal sound waves hit the tympanic membrane. Some energy

is reflected back at differing phase and amplitude. The

difference in phase and amplitude between the probe frequency,

and the reflected wave depends on the impedance (or admittance)

characteristics of the middle ear system (Northern & Grimes,


Many clinicians conceptualize the tympanogram as the

way the eardrum moves. In reality, the tympanogram is

simply a recording of changes in sound pressure level in

the external ear cavity between the probe tip and tympanic

membrane. The point of maximum compliance of the system is

simply the point with the lowest sound pressure level as

the clinician varies the pressure. The measurement of

tympanometry is extremely important to the study of acoustic

reflex behavior. Research indicates that conductive middle

ear pathology, and abnormal middle ear pressure can serve

to elevate AR thresholds, or even prevent their recording

(Jerger, 1970; Zwislocki & Feldman, 1970; Jerger, Anthony,

Jerger, & Maudlin, 1974).

Acoustic Reflex Threshold

Acoustic reflex threshold may be defined as 'the lowest

stimulus level that produces a measurable change in acoustic

immittance. It should be recognized, however, that the

acoustic reflex threshold is influenced by many factors

which have little to do with auditory performance or

sensitivity. Such variables as the nature of the elicitor,

equipment sensitivity, and manner with which AR activity is

monitored all affect the final definition of acoustic reflex


Most investigators define AR threshold as the smallest

admittance change from baseline activity that occurs in

association with elicitor presentation (Gelfand, Silman, &

Silverman, 1981; Popelka, Margolis, & Wiley, 1976; Peterson

& Liden, 1972). Several authors have proposed a more

operational definition of AR threshold (Borg, 1972; Block

& Wiley, 1979; Hepler, 1984; Moul, 1985). Borg (1972)

defined AR threshold as the presentation level required to

shift acoustic immittance from baseline to 10% of its

maximum immittance change. Block and Wiley (1979)

utilized a statistical technique for threshold determina-

tion. They choose to define AR threshold as the lowest

elicitor level that causes an admittance change from baseline

activity of more than two standard deviations. This

technique has also been utilized with success by Moul

(1985) and Hepler (1984).

The acoustic reflex is a bilateral response to intense

stimuli which can be elicited by monaural or binaural

stimulation. Ipsilateral or contralateral AR activity can

be monitored in most laboratories; however, studies

show that ipsilateral AR thresholds are approximately

5 to 8 dB lower than contralateral AR thresholds (Zakrisson,

1974). When the elicitor is presented bilaterally, there

is approximately a 3 dB decrease in AR threshold (Moller,

1974). These variables must be taken into consideration

when comparing results of one laboratory to another. The

literature also reports high test reliability for AR

threshold measurements (Chun & Raffin, 1979; Chermak,

Dengerink & Dengerink, 1983). Forquer (1979) tested normal

hearing and sensorineural loss subjects eight times over

a period of three days and found largest AR threshold

difference to be 2.4 dB.

At present, there are no apparent sex differences in

AR threshold data for both normal and hearing impaired

listeners (Jerger, Jerger & Mauldin, 1972; Osterhammel &

Osterhammel, 1979; Silverman, Silman, & Miller, 1983). These

authors also showed that AR thresholds are elicited at

lower intensity levels with increasing age. It should be

noted, however, that differences are only encountered if

one compares extremely wide age groupings.

The acoustic reflex is also extremely dependent on the

parameters of the eliciting signal. Stimulus bandwidth

and density have been reported to be critical factors in

AR threshold determination. Research indicates that AR

threshold remains constant as stimulus bandwidth widens,

up to a point. If the bandwidth of the elicitor is widened

beyond that point, AR threshold is lowered (Moller, 1972;

Djupesland & Zwislocki, 1973; Flottorp, Djupesland, &

Winther, 1971; Margolis, Dubno & Wilson, 1980; Popelka,

Margolis & Wiley, 1976). It is also worth noting the

bandwidth for AR threshold increases as stimulus frequency

increases (Flottorp, Djupesland, & Winther, 1971; Popelka,

Margolis & Wiley, 1976).

Spectral density within a signal bandwidth also

affects the AR threshold. Popelka, Margolis and Wiley

(1976) were able to demonstrate this point by experimentally

increasing and decreasing the number of components in a

particular bandwidth signal. The larger the number of

components in the bandwidth of the signal, the greater

its spectral density. Whenever the spectral density of the

activating stimulus was increased a decrement in AR

threshold was observed. It therefore appears, the

critical band concept in acoustic reflex function is quite

similar to other types of auditory performance.

A relationship also exists between the level, and the

duration needed to elicit the acoustic reflex. The eliciting

stimulus must have a certain duration and intensity level

in order to cause middle ear muscle contraction. Reduction

of elicitor duration below approximately 300 ms must be

offset by increases in sound pressure level in order to

reach AR threshold. This phenomenon by which the auditory

system appears to integrate energy within a certain time

frame is known as temporal integration (Gelfand, 1984).

Johnsen and Terkildsen (1980) utilized a series of

click stimuli to demonstrate the temporal integration

properties of the acoustic reflex system. They reported

that AR threshold for a 128/sec click train was approximately

the same as AR threshold for white noise in a group of

normal listeners. However, as the click rate was varied

from 128/sec to 8/sec a change in AR threshold from 75 dB

to 120 dB was observed. This study seems to suggest that

the auditory system integrated the fast click train as a

continuous signal. When the click rates were significantly

reduced, it prevented the individual clicks from being

processed in this fashion. Similar findings have been

reported for tonal stimuli (Barry & Resnick, 1976; Richards,

1975; Woodford, Henderson, Hamernik, & Feldman, 1975).

These findings appear to indicate that temporal integration

properties of the acoustic reflex system are similar to

those found in other types of auditory behavior.

Acoustic Reflex Magnitude

Acoustic reflex magnitude can be defined as the amount

of immittance change resulting from the contraction of the

middle ear muscles (Silman, 1984). Dynamic properties of

AR magnitude have typically been demonstrated utilizing

acoustic reflex growth functions. These graphs are a plot

of immittance change resulting from increases in stimulus

intensity level. Investigations reveal that AR magnitude

is directly related to stimulus intensity level (Wilson &

McBride, 1978; Silman & Gelfand, 1981; Moller, 1962).

Stimulus parameters such as elicitor frequency, mode of

presentation, and probe tone frequency have also been shown

to influence AR magnitude (Wilson & McBride, 1978; Moller,

1962; Silman, 1984). Another factor which greatly influences

the magnitude of the AR is the static immittance at the

tympanic membrane. In general, as the static immittance

of the system increases, so does the measured AR magnitude

(Wilson, 1979; Block & Wiley, 1979).

Various normalization techniques have been used to

describe AR magnitude functions. Borg (1977) and Moller

(1962), expressed AR magnitude as a percent relative to

maximum impedance change. Silman and Gelfand (1981)

transformed the amount of impedance change to static

acoustic impedance in decibels. The reporting of growth

functions also differs in the literature. Several investi-

gators describe acoustic reflex growth functions in terms

of dB SPL (Moller, 1962; Borg, 1977), while others choose

to relate magnitude re: dB above AR threshold (Djupesland,

1967; Thompson, Sills, Recke, & Bui, 1980; Gerhardt &

Hepler, 1983; Hepler, 1984; Moul, 1985). Since saturation

of the acoustic reflex appears to be governed by SL re:

AR threshold rather than SPL of the elicitor, the latter

method seems to be the most appropriate reference of the

two. Such variables must be considered when comparing

results from one laboratory to the next.

AR magnitude is dependent on the nature of the eliciting

signal. Moller (1962) examined the effect of bilateral

stimulation as well as ipsilateral vs. contralateral

stimulation on AR growth functions. Findings indicate

that AR slope functions were steepest for bilateral stimula-

tion, and shallowest for the contralateral condition. Probe

tone frequency has also been shown to influence AR magnitude

functions. Dallos (1964) along with Wilson and McBride

(1978) reported that for a given intensity level, AR magnitude

decreases with increases of probe tone frequency. These

same authors also reported that AR magnitude was greatest

for broadband noise and 1.0 kHz elicitors.

AR magnitude also varies as a function of age.

Thompson, Sills, Recke, and Bui (1980) reported that

growth of AR magnitude decreases for both tonal and

noise stimuli as age increases. Silman and Gelfand (1981)

and Wilson (1981) also found a decrease in AR magnitude for

older populations.

The dynamic range of the acoustic reflex varies with

the spectral content of the elicitor. AR dynamic range is

approximately 30 dB for tonal stimuli (Dallos, 1964;

Silman, 1984) and 50 dB for broadband noise (Wilson &

McBride, 1978). It should be noted, however, that in many

instances equipment limitations and subject discomfort

prevent complete analysis of acoustic reflex growth functions.

Temporal Characteristics of the Acoustic Reflex

Onset latency is the time (in msec) of the first

stimulus related immittance change resulting from middle

ear muscle contraction. Measurement of AR latency depends

on many factors including,characteristics of the eliciting

stimulus, instrument response time,and mode of presentation.

Latency values are also influenced by the operational

definitions utilized by researchers. Borg (1972) and Moller

(1972) define onset latency as the time between stimulus

onset and the point where the AR reaches 10% of its

maximum amplitude. Colletti (1975) utilizes 5% of maximum

immittance change as his definition of onset latency.

Other investigators suggest that any measurable immittance

change from baseline activity could be used to define

onset latency (Metz, 1951; Sunderland, 1974).

Several methods are currently available for the study

of AR activity. The choice of method also has certain

implications for AR latency measurements. Examples of

recording techniques include electromyography (Perlman &

Chase, 1939), cochlear microphonics (Gerhardt, Melnick &

Ferraro, 1979) and optical detection (Gans, Sweetman, &

Carlson, 1972). While measurements obtained using these

methods may be very precise, they require access to the

middle ear or round window,and are generally impractical

for clinical testing. Immittance measurements by far

represent the easiest, most popular, and least invasive

method available for the study of AR activity. Unfortunately,

the time constant of these instruments often produces a

systematic delay resulting from the time needed by the

circuitry to perform their particular measurements. All

latency values obtained must represent actual biological

AR latency times when reported. This can be accomplished

by subtracting the time constants attributed to the instru-

ment involved in the measurements,from total latency values

(Bostra, Russolo, & Silverman, 1984; Lilly, 1984). Unfor-

tunately the time constants vary depending on make and

model of the instrument, and may even among individual

pieces of equipment. Procedures used to obtain instrument

response times of the equipment utilized in this study are

presented in the methods section.

Several stimulus parameters directly affect onset

latency values. Acoustic reflex latency is reported to be

inversely related to stimulus intensity level (Metz, 1951;

Borg, 1972; Dallos, 1964). The more intense the eliciting

stimulus, the shorter the onset latency of the response.

Rise-time of the activator has also been correlated with

changes in onset latency. Studies show that the

faster the rise-time of the stimulus, the shorter the

latency of the response (McPherson & Thompson, 1977).

Duration of the activating elicitor does not seem to affect

latency values until it is shortened below a critical value.

Whenever the activating stimulus has a duration of less

than one second, the intensity of the stimulus needed to

elicit the AR must be increased, and therefore, indirectly

affects AR latency values (Djupesland & Zwislocki, 1971).

The effects of frequency on latency characteristics are

at present equivocal. According to Moller (1958) elicitor

of 500, 1000 and 2000 Hz are the frequencies of choice for

the study of AR latency properties. Bosatra, Russolo, and

Silverman (1984), however, point out that the use of low

and mid frequency tonal elicitors may be best simply

because of their increased stability over other types of

eliciting signals.

Latency characteristics of the AR are often discussed

in terms of sensation level (re: AR threshold). Onset

latency values for normal hearing young adults (20 to 30

years old) have been reported to be in the range of 150-250

msec at threshold (Metz, 1951; Moller, 1958; Dallos, 1964).

When stimulus intensity levels are raised to 30 or 40 dB

above AR threshold, latencies decrease to approximately 25-40

msec (Moller, 1984; Dallos, 1964). As with other AR

parameters, AR temporal characteristics are extremely variable

across individuals. Intrasubject variability, however,

appear to be quite good. Bostra, Rossolo, and Silverman

(1984) measured AR latency in 20 subjects twice a day for

three consecutive days and found mean intrasubject variability

to be only 9.7 msec, with a maximum variability in the order

of 40 ms. In 40% of the subjects, AR latency remained

constant over the entire three day period.

The dependence of onset latency to stimulus parameters

does not hold true for all AR properties. While onset

responses are inversely proportional to the intensity of

the stimulus, offset responses show little or no intensity

dependence (Borg, 1976). McPherson and Thompson (1977)

suggested this nonlinearity may be due to inherent differ-

ences in the process of contraction vs. relaxation of the

AR response, and therefore, should be considered as two

separate events of AR activity. They went on to postulate

that the AR was actually an "energy related phenomenon."

Since stimulus rise-time and intensity level affect the

total energy available to the system, it seems only logical

these properties control the onset latency of the response.

Offset latency on the other hand,is relatively independent

of stimulus parameters,and is thought to represent the

system's response to cessation of the stimulus. Borg

(1976), therefore, suggests that AR offset properties

represent a closer estimate of neural conduction time of

the AR system and should be investigated further.

Acoustic Reflex Adaptation

Magnitude of the acoustic reflex does not remain

constant over time. Research indicates that when the

activating signal is presented even for several seconds

the middle ear muscles begin to relax,and immittance

values change toward the static values that existed

prior to stimulus onset (Wilson, Shanks, & Lilly, 1984;

Fowler & Wilson, 1984). Several problems are typically

encountered during the recording of acoustic reflex

adaptation, which will be defined as the decrease in AR

magnitude over time for sustained and interrupted signals.

The combination of procedural differences, instrument

sensitivity, and baseline drift all contribute to the

variability reported in the literature.

Adaptation is often quantified as the time (in msec)

until the reflex decreases to 50% of its maximum reflex

amplitude (Anderson, Barr & Wedenberg, 1970), or as a

percentage of maximum amplitude at specified time intervals

(Wilson, Shanks, & Lilly, 1984). Shanks (1979) reported

that regardless of the protocol implemented, magnitude

of the AR adaptation will depend on the unit of measure

utilized to report the immittance change over time

(admittance versus impedance). Antablin, Lilly, and Wilson

(1980) demonstrated the effect of differing measuring units

by plotting mean data functions expressed as percentage

of maximum admittance and impedance. Results indicate

that the impedance functions were slightly steeper than

the admittance functions. This resulted in a shorter

half-life values for the impedance (13.9 seconds) versus

the admittance data (15.4 seconds). Such variation could

have significant consequences under certain research


Baseline drift represents still another problem in

the quantification of AR adaptation. Wilson, Steckler,

Jones, and Margolis (1978) reported that acoustic admittance

of the middle ear is constantly changing. During the measure-

ment of reflex adaptation, middle ear pressure may decrease

slightly, resulting in a concommitant decrease in acoustic

admittance at the tympanic membrane. Separating the contri-

butions of instantaneous changes in middle ear pressure

from the adaptation process represents a difficult obstacle

in the measurement of AR adaptation. Tonndorf and Khanna

(1968) suggest this baseline drift can be explained by an

interaction of the middle ear muscosa and the eustachian

tube. As the eustachian tube opens and closes, the middle

ear pressure is equalized to ambient air pressure. The

middle ear mucosa soon begins to absorb the air, creating

slight negative pressure. The baseline drift demonstrated

during adaptation measurement can, therefore, be explained

as a series of tympanograms over time.

The rate of AR adaptation depends on the frequency

of the activating signal (Djupesland, Flottorp, & Winther,

1967; Johansson, Kylin, & Langfy, 1967; Wilson, Shanks, &

Lilly, 1984). In general, lower frequency signals (500

and 1000 Hz) produce less adaptation than higher frequency

signals (3000 and 4000 Hz). Mid-frequency activators (1500

and 2000 Hz) exhibit varying degress of reflex adaptation

(Wilson, Shanks, & Lilly, 1984). Djupesland, Flottorp and

Winther (1967) were one of the first investigators to

demonstrate that AR adaptation was frequency dependent.

They examined the duration of maintained AR activity for

250, 1000, and 4000 Hz elicitors. Results show that

AR activity was maintained the longest for the 500 Hz

elicitor, followed by the 1000 Hz signal. The 4000 Hz

elicitor had the fastest adaptation of approximately 20

seconds when presented at 10 dB SL (re: AR threshold).

Similar findings were reported by Johansson, Kylin,

and Langfy (1967) utilizing 500 Hz and 3000 Hz elicitors.

These authors reported little adaptation for the 500 Hz

activator over a 10 second period, with substantial adapta-

tion occurring for the 3000 Hz signal. Figure 6 from

Wilson, Shanks, and Lilly (1984) demonstrates this frequency

dependence with admittance data that has been corrected

for ear canal volume (presented at 10 dB SL re: AR

threshold). These data demonstrate that AR magnitude as

well as adaptation properties are dependent on the frequency

characteristics of the signal. As illustrated, the largest

magnitude observed is for the mid-frequency signals. The

2000 Hz activator produced the largest magnitude and the

4000 Hz produced the smallest. Also note that virtually no

AR adaptation is present for the low frequency elicitors

(500 and 1000 Hz).

AR adaptation appears to be more resistant for noise

activators than for tonal stimuli (Ward, 1961). This may

be due to the random nature of noise stimuli which serves

to re-elicit the AR response (Ward, 1973). In a study

utilizing octave band noise elicitors, Djupesland, Flottorp,

and Winther (1966) reported that low frequency noise was

more effective at maintaining AR activity than higher

500 Hz

1000 Hz

2000 Hz

4000 Hz


0 5 10 15
Time (Seconds)

Figure 6. Adaptation functions obtained with four 10.2 s
reflex-activator frequencies presented at 10
dB SL (adapted from Wilson, Shanks, & Lilly,

frequency octave-band noise. This finding may be related

to the temporal encoding properties of the auditory system

(Wilson, Shanks, & Lilly, 1984).

The relationship between AR adaptation and intensity

level remains equivocal. Djupesland, Sunby, and Flottorp

(1967) reported that higher intensity levels sustain AR

activity for greater periods of time. Their results

supported a direct relationship between intensity level

and AR adaptation. Wiley and Karlovich (1975) on the other

hand, reported that for a 500 Hz signal, reflex adaptation

increased with subsequent increases in intensity level.

Results to a 4000 Hz elicitor revealed no significant

difference in AR adaptation as intensity level of the

stimulus was varied. Wilson, Steckler, Jones and Margolis

(1978) observed that there is a direct relationship between

AR adaptation, and activator intensity level for 500 and

1000 Hz elicitors. This relationship became less pronounced,

however, for higher frequency signals (3000 and 4000 Hz).

In summary, most investigators report that AR adaptation

decreases as the intensity level of a low frequency

elicitor (500 and 1000 Hz) increases. Activators of higher

frequencies (2000 Hz or above) reveal no systematic change

in reflex adaptation properties with changes in activator

intensity level (Wilson, Shanks, & Lilly, 1984).

Unfortunately much of the existing literature

describing AR properties involves the use of constant pure

tone or noise elicitors (Dallos, 1964; Johansson, Kylin,

& Langfy, 1967; Anderson, Barr, & Wedenberg, 1969). These

studies demonstrate rapid adaptation to a variety of acoustic

stimuli. Research indicates, however, that AR activity can

be reactivated with short pauses, or changes in spectral

characteristics of the eliciting signal (Borg & Odman,

1979; Hetu & Careau, 1977; Gjaevenes & Sohoel, 1966). It

is also known that typical industrial environments create

noise of wide frequency content, and time varying intensity

levels. In light of this fact, several investigators have

evaluated AR dynamic properties in response to industrial


Borg, Nilsson, and Liden (1979) exposed subjects

monaurally to a 30 minute tape of shipyard noise while

continuously monitoring AR activity. The first and last

minute of the exposure were identical in order to evaluate

the short term fatigability of the AR. Results indicate

that AR properties were remarkably resistant to adaptation

for this type of exposure. In a related study, these same

authors evaluated AR performance following a typical workday

of shipyard noise exposure. Employees were monaurally

exposed at their worksite for approximately seven hours

with periodic breaks. AR activity was monitored prior to,

and immediately following the industrial exposure. Their

findings revealed that the acoustic reflex shows very little

adaptation following long term exposure to industrial noise.

The apparent discrepancy between studies involving

industrial versus continuous noise exposures may be related

to the temporal properties of the AR, and the transient

nature of most industrial environments. While it is well

established that single impulse sounds pass through the

middle ear system unattenuated, they may affect the

succeeding stimuli thereby providing some protective influence

to the inner ear mechanism. As previously mentioned, the

offset properties of the AR are much slower than the contrac-

tion process. Estimates based on impedance measurements

indicate that offset times needed to reach 50% of maximum

amplitude are in the order of 100-500 msec, and may exceed

one second under certain conditions (Borg, 1976; Dallos,

1973). These findings along with the results obtained from

such studies as Borg and Odman (1979) and Hetu and Careau

(1977) suggest that when changes of intensity or frequency

occur, as they do in most industrial situations, the AR

undergoes a series of relaxation responses followed by

reactivation. Due to this action, the AR may be able to

maintain tension on the ossicle chain for longer periods of

times without significant adaptation. Therefore, offset

properties of the AR may represent a critical variable in

the system's protective function against hazardous noise

exposures. Further studies involving industrial noise

exposures are needed, however, to validate these assumptions.

The mechanism of AR adaptation is presently unknown.

Research indicates that neurons in the peripheral auditory

system are characterized by a certain amount of adaptation.

Kiang, Watanabe, Thomas, and Clark (1965) reported that

neurons initially respond to a toneburst with an increase

in firing rate that decreases with time. This process is

evident throughout the auditory system. Recall, however,

that low frequency activators demonstrate less adaptation

than higher frequency stimuli (Wilson, Shanks, & Lilly,

1984). The fact that there is minimal adaptation when the

AR is elicited by low frequency stimuli indicates that

response amplitude may not be related to discharge rate,

but rather, to other properties of neural activity (Moller,

1984). It has been suggested that the reflex response may

be related to a phase-locking phenomenon, which have no

adaptation properties (Kiang, 1980). However, additional

research is needed to support this theory.

Acoustic Reflex Properties and Temporary Threshold Shift

Properties of the acoustic reflex, and their relation-

ship to TTS, have previously been investigated in human

and animal models (Gerhardt & Hepler, 1983; Borg, Nilsson,

& Engstrom, 1983; Karlovich, Luterman, & Abbs, 1972;

Gerhardt, Melnick, & Ferraro, 1979). It is well known

that contraction of the stapedius reduces the transmission

of sound through the middle ear system. This reduction

is greatest for the frequencies below 1 kHz and can be as

much as 20 dB (Dallos, 1973). Due to this action, the

acoustic reflex is thought to provide a protective function

to the delicate structures of the cochlea. This reflexive

response changes with various parameters of the stimulus

including frequency spectrum, intensity, temporal pattern

and duration (Gerhardt, Melnick, & Ferraro, 1979). However,

the protective role of the acoustic reflex remains equivocal.

Johansson, Kylin and Langfy (1967) exposed normal

hearing subjects to octave band noise, and found correlations

between TTS developed, and certain latency characteristics

of the AR. Holmes (1978), utilizing a white noise exposure,

revealed significant correlations between AR magnitude at

2.0 kHz and TTS. Karlovich, Luterman and Abbs (1972)

devised a unique experiment that incorporated a dichotic

listening paradigm, to ensure AR contraction in an experi-

mental versus control group of normal hearing subjects.

The experimental group was exposed to noise in one ear,

while an AR activating stimulus was presented to the other

ear. Similar exposures were conducted for the control

group without the AR activating tones present. The author

reported significantly greater TTS in the control group

when compared to the experimental group after the 1.0 kHz

exposure. Zakrisson (1974) evaluated AR properties in

subjects with unilateral Bells palsy. This condition is

known to cause paralysis of the stapedius muscle on the

affected side. His findings indicated that TTS was

significantly greater in the affected ear when compared

to the non-affected ear following low frequency noise


Contrary to these findings Fletcher and King (1963)

observed no significant difference in TTS developed in a

group of stapedectomized patients when compared to the

normal hearing control group. Turner (1974) also questioned

the protective role of the AR. In his study he measured

absolute impedance values and correlated them to TTS in a

group of normal hearing subjects. -Again, no relationship

was observed between the impedance values measured, and

threshold shift produced by the noise exposure. It should

be noted, however, that these studies involved exposures of

short duration (minutes) at high intensity levels. In an

effort to obtain more a representative sample of AR

activity, experiments utilizing exposures of greater duration

have been completed.

Gerhardt, Melnick and Ferraro (1979) utilized an eight-

hour, 95 dB SPL exposure of 0.5 kHz octave band noise on

chinchillas to evaluate the relationship of the acoustic

reflex to TTS. Round window recordings of the cochlear

microphonic were used to study properties of the middle

ear muscles. Measurements of reflex thresholds were obtained

prior to exposure, during quiet intervals of the exposure,

and after cessation of the noise. Results showed that

acoustic reflex thresholds were significantly altered by

the noise exposure. Following eight-hours of exposure

the average reflex threshold shift (RTS) was approximately

14 dB. The authors also concluded that the development of

TTS and RTS followed the same time course. Specifically,

for every 2 dB of TTS there was a 1 dB reflex threshold


Gerhardt and Hepler (1983), in a study involving human

subjects, measured middle-ear muscle activity with an

electroacoustic bridge. This study investigated reflex

threshold shift, and the subsequent development and recovery

of TTS. The experimental group was exposed to -a four-hour,

1.0 kHz noise in sound field at an intensity of 95 dB SPL.

Results revealed that AR magnitude at suprathreshold levels

decreased and recovered with the growth and recovery of

TTS. This finding seems to suggest, that there is a rela-

tionship between AR magnitude and temporary threshold shift.

Hepler (1984) and Moul (1985) demonstrated that average

AR magnitude is inversely related to certain expressions

of TTS following two-hour exposures of broadband, and octave

band noise, respectively. While certain AR parameters

demonstrated to be significantly correlated to expressions

of TTS, only a slight predictive relationship was demon-


Statement of Purpose

Research indicates that certain parameters of AR

function change following noise exposure, most notably,

AR threshold and magnitude. Changes in onset latency

have not been reported in the literature. One AR parameter

which has not been studied extensively following noise

exposure is the relaxation or offset response of the acoustic

reflex system. There is also very little information

regarding the adaptation process of the AR to industrial

noise exposure. While continuous noise exposures may

answer some of the questions regarding AR physiology, they

can not explain the behavior of the acoustic reflex system

in noise with changing temporal and spectral content.

As previously mentioned, single transient sounds pass

through the middle ear system unattenuated. In industrial

situations, however, impulses often occur as a series of

fluctuations in background noise. Due to its relatively

slow relaxation time, the AR offset response may be particu-

larly important in these environments. If interimpulse

intervals are brief, the contraction elicited by each

impulse or intensity change,may affect succeeding stimuli.

Under these conditions, offset properties may represent

a critical variable in understanding the behavior of the

acoustic reflex during industrial noise exposures. In an

effort to explain some of the uncertainties concerning

AR behavior in industrial environments, the following

questions have been formulated.

1) Do differences exist between pre-exposure and

post-exposure values of the following response

variables in normal hearing subjects after a two

hour industrial noise exposure of 90 dB SPL?

a) Behavioral thresholds at octave and half-

octave intervals

b) Acoustic reflex threshold

c) Acoustic reflex magnitude

d) Acoustic reflex latency characteristics

(onset and offset)

e) Acoustic reflex adaptation

2) Do differences exist in percent adaptation for

the three stimuli evaluated in the pre-exposure


a) Continuous broadband noise

b) Industrial noise

c) Interrupted broadband noise

3) Do any of these measures correlate to changes

observed in audiometric behavioral thresholds

following this same noise exposure?


This study evaluated changes in acoustic reflex

activity following a two-hour exposure to industrial

noise. Several parameters of AR activity were sampled

prior to, and following the noise exposure in an effort

to explain acoustic reflex behavior in industrial environ-

ments. As outlined in Table 2, the initial experimental

session consisted of basic screening for subject selection,

and the establishment of a baseline audiogram. During the

second session, comprehensive analysis of the middle ear

system and AR activity was completed. The third session

included the industrial noise exposure, followed by

measurement of acoustic reflex activity and post-exposure

threshold testing. Recovery thresholds were also conducted

24 hours after the noise exposure. This was done in order

to document complete recovery of behavioral thresholds.

Table 2
Experimental procedures.

Session One:

1) History and consent form
2) Otoscopic screening
3) Tympanometry
4) Acoustic reflex screening (BBN, thresholds
< 85 dB SPL)
5) Subject Training--Bekesy tracking procedure
6) Audiometric Baseline--0.5 through 6.0 kHz

Session Two:

1) Otoscopic screening
2) Tympanometry
3) Acoustic reflex threshold (BBN stimuli)
4) Acoustic reflex magnitude/latency function
(BBN stimuli)
5) Standard Adaptation (BBN stimuli) at 15 dB SL
6) Standard Adaptation (Industrial noise stimuli)
at 15 dB SL
7) Reflex Interruption Test (BBN stimuli) at 15

Session Three:

1) Pre-exposure threshold--0-5 through 6.0 kHz
2) Noise exposure (Industrial Noise--Two hours)
3) TTS at discrete frequencies 0.5-6.0 kHz (2
minutes following the exposure)
4) Acoustic reflex threshold (BBN stimuli)
5) Acoustic reflex magnitude/latency function
(BBN stimuli)
6) Standard Adaptation (BBN stimuli) at 15 dB SL
7) Reflex Interruption Test (RIT) at 15 dB SL
8) Behavioral Threshold--0.5 through 6.0 kHz

Session Four:

1) Recovery Thresholds--0.5 through 6.0 kHz
2) Recovery Acoustic Reflex magnitude/latency
functions (BBN stimuli)


Subjects were between 20 and 32 years of age, and

were paid for their participation in ths study. Thirty

subjects were selected from students and university staff

who met the following criteria: 1) normal hearing (<15 dB

HL re: ANSI, 1969) at octave and half-octave intervals

from 0.5 through 6.0 kHz; 2) a normal tympanogram and

otoscopic screening; 3) no history of chronic ear disease;

4) no employment history which required exposure to

hazardous noise for periods in excess of two years;

5) AR thresholds of 85 dB SPL or less to a broadband

noise elicitor; 6) ability to perform standard Bekesy

procedure (modified method of adjustment) defined by

repeatable thresholds (within 3 dB) at 1.0 kHz and stable

tracings for up to one minute; 7) Bekesy excursion widths

of less than 10 dB; 8) signed consent to the experimental

procedures, plus verbal commitment to return in 24 hours

following the noise exposure to assure complete recovery

from TTS.

Instrumentation and Procedures


The two hour exposure consisted of industrial noise

presented bilaterally at 90 dB SPL. The noise sample was

obtained from a metal workshop in order to approximate

an exposure commonly found in industrial environments.

The noise consisted of background activity produced by

several pieces of equipment, however, the main noise source

was a piece of steel on a stone grinder. A 1/2" microphone

(Bruel & Kjaer, type 4165), and sound level meter (Bruel &

Kjaer, type 2203) was connected to a tape recorder (Teac,

X-10/X7) which stored the noise sample. This signal was

then delivered to an IBM-PC microcomputer with PCLAB real-

time software package. This system facilitated the conver-

sion of analog signals into retrievable data files. One

second of industrial noise activity was sampled at a rate

of 12.0 kHz to create the digitized file used for the

noise exposure. The file was then stored to disk.

Upon keyboard command, the file was accessed, fed

through a D/A converter, shaped with spectrum equalizer

(Realistic 31-2008), amplified (Coulbourn Instruments, model

S82-24 mixer-amplifier), attenuated (Hewlett-Packard,

350D attenuator set), and fed through Beyer DT49 earphones

in an Industrial Acoustics audiometric booth. The subjects

were in a reclining chair for the duration of the two hour

noise exposure.

A fast-fourier transform (FFT) completed on stimuli

used to produce the TTS is included in Figure 7. Also

presented is a sample of broadband noise processed through

the same instrumentation. Spectral analysis was accomplished


1GO 200 > 3CS0 4003 SuS03 e30


0 2U10 4000 610O 8003 LOCC3

Figure 7. Fast-fourier transform (FFT) performed on
industrial noise (top), and broadband
noise (bottom) stimuli.

by coupling the Beyer DT48 earphone to a flatplate,

attached to a standard 6cc cavity. Using a Bruel and

Kjaer 1" microphone (type 4132), and amplifier (Bruel &

Kjaer, model 2604), the signals were recorded on tape

(Teac, X-10/X7), and fed into the IBM-PC. An ILS software

package was used to perform the FFT analysis. The spectrum

of the two stimuli were similar, except for a 5 dB peak

present at 1200 Hz in the industrial noise signal.

Behavioral Hearing Threshold Measurements

All subjects were trained during a trial session to

perform standard Bekesy tracking procedures. Stable

tracings with midline deviations of no more than 3 dB, and

excursions of less than 10 dB were required before continuing.

A computer program controlled all stimulus parameters.

Algorithms provide for acceptance or rejection of

particular trials based on the above criteria.

Thresholds for pulsed tones (250 msec with 50% duty

cycle) were established at discrete frequencies of 0.5,

1.0, 2.0, 3.0, 4.0, and 6.0 kHz prior to, following, and

24 hours after the noise exposure. Tonal stimuli were

generated by a Hewlett-Packard 3311A function generator,

and routed through Colbourn Instruments timing, gating,

and digital logic circuitry. A programmable attenuator

activated by a hand-held switch was controlled by the subject.

Signals were delivered to a single earphone (Telephonics

TDH-39) with a standard (MX-41) cushion. Calibration of

all stimuli were performed prior to, and following the

collection of data.

Due to the rapid recovery of behavioral threshold

following noise exposure, only one ear was tested during

the recovery period. The left ear of all subjects was

designated as the "test ear," and the right ear as the

"non-test" ear for all threshold and AR measurements.

Three measures of behavioral thresholds were completed

during the post-exposure period. Pure tone thresholds were

measured two minutes after cessation of the noise, and

following the post-exposure measurement of AR activity

(approximately 30 minutes after cessation of the noise).

This was done to verify that the auditory system was

still under the influence of TTS when the last AR measure-

ment was completed. Recovery thresholds were again

established 24 hours after the noise exposure.

Acoustic reflex measurements

Acoustic reflex measurements were completed in a sound

treated booth (Industrial Acoustics) while in a sitting

position. Prior to placement of the probe assembly,

subjects were screened otoscopically to ensure no active

pathology or excess cerumen was present that would prevent

the recording of AR activity. Following otoscopic

inspection, tympanometry was performed to verify that

ambient pressure of the middle ear system (point of maximum

compliance) was within normal limits. AR threshold

screening was also completed to ensure the presence of

reflex thresholds of at least 85 dB SPL (broadband noise


General instrumentation

Measurements of the middle ear were made with a

Grason Stadler 1723 Middle Ear Analyzer (220 Hz probe

tone). This device allows for the simultaneous recording

of admittance (Y), and its sub-component, susceptance (B).

Outputs of the instrument were fed into two separate Data

Translation analog-to-digital channels connected to an

IBM-PC. The voltage outputs were then evaluated with the

assistance of computer programs specifically designed for

AR analysis. All programs were written to allow for

accurate measurement of AR parameters while adjusting for

temporal characteristics of the instrumentation. Generally,

the program averaged the first 600 msec of baseline voltage

prior to the introduction of the eliciting signal. Threshold

determination was made when the voltage change,caused by

the elicitation of the AR, exceeded the baseline voltage

by two standard deviations. Elicitor presentation levels

were controlled by the experimenter, or by automated

computer program which systematically increased or decreased

attenuator settings. The lowest intensity level that

elicited an AR response was considered threshold. Other

specific programs will be discussed in the following


Volume estimates

Research indicates that corrections must be made for

ear canal volume in order to obtain valid estimates of

admittance at the lateral surface of the tympanic membrane

(Block & Wiley, 1979; Margolis, 1981; Popelka, 1984).

This measure depends on many factors including size of

the probe tip, depth of insertion, and length of time

in the ear canal (Margolis, 1981). Thus, admittance

values (in mmhos) obtained at -350 daPa were used as

estimates of ear canal volume (Moller, 1965; Shanks &

Lilly, 1981). All acoustic reflex measurements were

made at ambient pressure (point of maximum compliance)

with the same test-ear/non-test ear paradigm previously

described for behavioral threshold testing. The elicita-

tion earphone was placed over the left "test ear," while

the probe assembly monitored AR activity in the right

"non-test" ear.

AR data collection

Measures of AR threshold, latency, and magnitude

were obtained prior to,and following the industrial noise

exposure for all thirty subjects. Measures of acoustic

reflex adaptation, however, were assigned to groups of

subjects,and are described in a later section.

Data acquisition could be completed by the experimenter

with keyboard commands, or by programs allowing automatic

data collection. In the automatic mode, the eliciting

stimulus was initially presented at a level below anticipated

AR threshold. The signal was then increased in 2 dB

steps until a response occurred. If a response was present

during the initial presentation of the eliciting stimulus,

the attenuator was adjusted until no AR activity occurred.

Once the initial response was identified, the attenuator

automatically decreased the signal by 4 dB, and three

successive trials occurred at that level. The intensity

level was increased in 2 dB steps, and three presentations

per intensity level were completed until the criterion for

threshold was met (greater than 2 standard deviations of

baseline activity). The lowest intensity level required

to elicit this criterion-based response was considered

threshold. The automated program provided a visual display

of each trial as it was being collected. This allowed

the experimenter to accept or reject a particular trial.

If the presence of artifacts rendered a particular trial

uninterpretable, the trial was repeated.


Magnitude/intensity functions were obtained after

threshold was established. Starting at threshold, the

computer program automatically increased the elicitor

intensity level in 2 dB steps (two presentations at each

level), up to 12 dB sensation level (SL) with respect

to AR threshold. The intensity function was continued

with one presentation per 2 dB increment up to 104 dB SPL.

Magnitude of each acoustic reflex was computed in

terms of admittance (Y). The phase angle with each

acoustic admittance measurement was preserved for all

computations (Block & Wiley, 1979). Therefore, vector

and phase values of susceptance (B), and conductance (G)

were maintained for calculations of admittance during the

contracted,and uncontracted state of the AR. The corrected

values for (Y) and (B) were stored to disk for later

analysis. Calculation of conductance (G) was made using

the following formula:

G = / (Y)2 (B)2

Magnitude of the AR was obtained by subtracting the

admittance of the middle ear prior to elicitation of the

response, from the admittance during maximum acoustic

reflex contraction. Calculation of AR magnitude was

accomplished by averaging the first 600 msec of baseline

activity, and subtracting from it the average admittance

between 1200 and 1600 msec as shown in Figure 8. Following

termination of the eliciting stimulus, the program then

averaged the corrected admittance values from 2100 through

2485 msec. This was used to represent the time period when

the AR had returned to its quiescent state.


Four latency points and two slope functions were

calculated for AR activity elicited at each stimulus

presentation level. Figure 9 illustrates these six

measurements. Latency 1 (Ll) was defined as the time

from the beginning of AR elicitation to 10% of the measured

steady-state admittance change. Latency 2 (L2) was the

time required to reach 90% of maximum admittance change in

the AR response. Latency 3 and 4 are similar measurements

used to describe the offset characteristics of the acoustic

reflex. Offset latencies (L3 and L4) were defined as the

time from the instantaneous termination of the initial

admittance change to 90% and 10% of the steady-state

admittance change, respectively. The rate of admittance

change was also measured to further describe the temporal

characteristics of the AR. Two linear regression models

1.0199 -
1.8874 -
1.7824 -
1.7699 -
1.7449 4V
1.7324 -
1.7199 --
1.7874 -
1.6949 -
1.6824 -
1.6699 -
1.6574 -
1.6449 -
1.6324 -
1.6199 -
1.6874 -
1.5949 -
1.5824 -
1.5699 ,, --i--, ,
(mS) 0 250 588 758 1880 1250 1580 1758 2888 2258 25088

Figure 8. Example of computer averaging technique used
to calculate onset and offset magnitude.

1.8199 -
1.8874 -
1.7949 -
1.7824 -
1.7699 -
1.7574 -
1.7449 0%
1.7324 .,. -- -10%
1.6949 .. /. --50%
1.6699 90%
1.657?4 -- 100%
1.6449 -
1.6324 -
1.6199 -
1.6874 -
1.5949 -
1.5824 -
1.5699 i i i \I i-
Time (mS) 0 250 588 758 1888 1258 1508 1758 2000 2258 2580
Cursor at (x,N)=(8,1.737861).
LT1=(381.8336, 1.728516), LT2=(667.5652, 1.658936), SLl=-1.98249E-84iaho/mS
LT3=(173.5471, 1.668156), LT4=(484.1685, 1.733398), SL2= 2.357925E-84miho/mS

Figure 9. Example of four point latency (Ll, L2, L3, L4)
and slope functions used to analyze acoustic
reflex activity. Latencies are expressed in
msec and slopes are expressed as mmho change
per msec.

were calculated from the initial and final transition of

each response waveform. Slope 1 (Sl) described the admit-

tance change at the onset of the AR, and is expressed as

the change in millimhos per msec from 10% to 90% of the

steady-state admittance change. Slope 2 (S2) is a similar

measurement at the offset of the acoustic reflex. It was

defined as the change in millimhos per msec from 90% to

10% of the steady-state admittance value. Figure 9 shows

the slope measurements for a particular AR response.

AR adaptation

Several measures of perstimulus adaptation were

evaluated in this study. Treatments were randomized in

an effort to control for any possible order

effect. Adaptation has been defined, as the decrease in

AR magnitude over time for sustained and interrupted

signals. One of the problems encountered when measuring

AR adaptation, is that acoustic admittance constantly

changes (Wilson, Steckler, Jones, & Margolis, 1978). Thus,

the admittance measured at one point in time, may be

different than the admittance measured at another point

in time, even when those measures are completed on the

same ear with no stimulus present (Wilson, Shanks & Lilly,

1984). In some individual cases, baseline drift may exceed

total admittance change caused by acoustic reflex contrac-

tion. For this reason, measures of AR magnitude were

obtained at the beginning, and end of the sampling period

only. Reflex adaptation was then expressed as percent

change in AR magnitude (Y) over time as outlined by Wilson,

Steckler, Jones, and Margolis (1978).

Measures of reflex adaptation for a broadband, and

industrial noise stimuli were obtained by monitoring AR

activity at 15 dB SL (re: AR threshold) for a period of

4.0 minutes. As previously mentioned, estimates of AR

magnitude were obtained at the beginning, and end of this

4.0 minute period. Onset magnitude (Ml), was calculated

by subtracting the average baseline admittance (0-600 msec),

from average admittance during maximum AR contraction

(1200-1600 msec). At the end of the 4.0 minute period,

a similar measure of offset magnitude (M2) was -calculated.

This valuewas determined by subtracting the admittance

value just before termination of the eliciting signal

(1000-1400 msec), from the admittance value after AR

activity returned to its quiescent state (2100-2485 msec).

Percent adaptation was then calculated from the admittance

change observed at onset magnitude (Ml) to offset magnitude

(M2). Figure 10 provides an example of how AR adaptation

was measured. The formula used to calculate percent

adaptation is provided below.

Percent Adaptation = (1 M2/Ml) X 100


30 seconds
into 4.0

1.7008 -
1.5683 -
1.558 1
1. 1,33
Time (CS) 0 258 51
Cursor at (x,1)'C(0,1.749268).

Cursor at


1.8880 -
1.5955 --
C(S) 8 250 5sB 75M 1888 1Z58 1508 175a 2830 2258 2508
(x,y)=(', 1.7Z1191). *

1.783 -
1.6663 -
1.6163 -
Time (.S) 6 25h 508 75 1318 128 1588 1758 288 2258 2503
Cursor at (x,)=(0,1l.723633).

Figure 10.

Standard adaptation for broadband noise
stimuli. Ml represents onset magnitude,
M2 represents offset magnitude.

Three measures of perstimulus adaptation were obtained

during the pre-exposure session. Adaptation estimates were

completed for a continuous broadband noise, continuous

industrial noise, and broadband noise with specified

interruption intervals. Signals used for the broadband

elicitor were externally generated by a Colbourn Instru-

ments noise generator, routed through a Colbourn Instruments

rise/fall gate, mixer/amplifier and Hewlett-Packard 350D

attenuator. Timing of stimulus presentation was controlled

by Colbourn Instruments gating and timing devices inter-

faced with the microcomputer (IBM-PC). The stimulus used

for the industrial adaptation series was the same as the

fatiguing stimulus used in the 2.0 hour noise exposure.

Tape output of a Panasonic 612 stereo cassette -recorder

was fed through the same Colbourn Instruments rise/fall

gate, mixer/amplifier and Hewlett-Packard attenuator.

Timing of the taped signal was accomplished by placing a

pulse on the second (left) channel of the tape recording

which initiated sampling activity of the microcomputer

(IBM-PC). Eliciting signals were presented through a

TDH-39 earphone with MX-41/AR cushion.

Standard adaptation for continuous broadband noise

was completedby calculating percent adaptation in the

fashion previously described (see Figure 10). An estimate

of adaptation for industrial noise was also obtained in a

similar manner (see Figure 11). Following the collection

of these measures for continuous signals, other measures


30 seconds
into 4.0

1.6573 -
1.6198 -
1.59Z3 -
1.54 48
1.4198 -I
(cS) 8 250 580 750
Cx,9)=(8, 1.54174B).


1888 1258 1509 175O 20 2258 2588


Figure 11.

Example of industrial noise adaptation
series. Ml represents onset magnitude,
M2 represents offset magnitude.


of AR function were obtained by briefly interrupting the

sustained broadband signal. The signal was introduced

at 15 dB SL (re: AR threshold), but was switched off, then

on again, for specified interruption intervals of 250, 500,

750 msec (see Figure 12). For convenience, this new

procedure was labeled the reflex interruption test (RIT).

Ten subjects were subjected to one of the three different

RIT treatments during the pre-exposure,and post exposure


One possible utility fo the RIT would be in the evaluation

of AR activity over time without the confounding influence

of baseline drift. In this way, AR offset, and onset

properties to a rapidly changing stimulus can be measured.

Due to the unique nature of this response, AR activity

under RIT conditions was evaluated in two separate ways.

For comparative purposes, percent adaptation was calculated

in a similar fashion as the continuous broadband and

industrial noise signals (see Figure 13). Further analysis

was then completed on the RIT procedure. By introducing

interruptions in the eliciting signal, the RIT provides us

with a relative measure of offset as well as re-elicitation

activity of the acoustic reflex system. Figure 14 demon-

strates this effect. R1 in that figure represents the

change in middle ear admittance following cessation of the

eliciting signal, while R2 results from a re-elicitation

of AR activity due to re-introduction of the broadband

noise. Magnitude of each phase of this response (Rl and

2.44O -


EMS) 0 250 51o 758 1808 15N 1580 175 2808 2258 258e

2.1276 I
(uS) 8 258 5S

Figure 12. Example of 250 RIT (top), 500 RIT (middle),
and 750 RIT (bottom) responses.





30 seconds
into 4.0

Carsor at


Figure 13.

1.7925 -
1.7175 -
I. .6S03 Ml
1.6425 -
1.6050 -" --^
Time (aS) 0 250 588 75B 188 1259 1580 1758 2Z 0 2250 25B8
CuIsor at (x,y)=(0,1.77Z461).

L8101 -
1.7976 -
1.7051 -
7726 -
1 7601 -
1.7476 -
1.7351 -
1.7226 -
1.6226 -
1. 6101
1.5601 i i
(aS) 0 250 583 750 1800
(x.)=(0, .673584).

1258 IS1B 1750

20 8 2250 25b0

1.9?85 -
1.7785 -
1.7668 -


Time (Ws) 0 250 500 758 1880 1250 150o 1750 2088 2250 258B
Cursor at (x,y)=(0,1.737061).

Reflex Interruption Test. Ml represents
onset magnitude, M2 represents offset


2.44800 -
2.4358 -
2.4388 -
2.4250 -
2.4288 -
2.4150 -
2.4188 -
2.4850 -
2.4888 -
2.3958 -
2.3758 -
2.3700 -
2.3658 -
2.3688 -
2.3558 -
2.3588 -
Time (mS) 80








/ ////A


is -


Cursor at (x,y)=(0,2.381592

580 750 1888 1258 1580 1750 200888 2258 2588

Figure 14.

Example of RIT offset (Rl), and re-elicitation
(R2) magnitude.


Stimulus On

Stimulus Off

R2), was then evaluated over time for all stimulus condi-

tions (250, 500, 750 msec).

Following the industrial noise exposure, standard

adaptation, and RIT procedures were again completed. This

was done in order to determine if changes in AR adaptation

properties occur following industrial noise exposure.

Measurement of these procedures were identical to the

ones just described. Industrial noise adaptation was not

completed due the rapid recovery of TTS observed during a

pilot study utilizing an exposure of similar intensity

and duration. It was felt that if significant changes

were to be demonstrated in AR behavior following noise

exposure, those measures would have to be completed while

the auditory still under the influence of TTS.


Calibration of the Telephonics TDH-49 earphone with

standard cushion (MX-41/AR) was conducted according to

ANSI S3.7-1973 using a 1" Bruel and Kjaer condenser

microphone coupled to a 6 cc cavity (type 4152). A Bruel

and Kjaer pistonphone (Model 4220) was used to calibrate

the microphone amplifier.

Different earphones were used for the noise exposure

to provide comfort throughout the two-hour procedure.

Circumaural cushions on theBeyer DT48 earphones, however,

prevented calibration in the usual manner with a 6 cc

coupler. Thus, a flat plate coupler designed by Hepler

(1984), was utilized according to procedures published by

Michael and Bienvenue (1976). The flat-plate coupler is

illustrated in Figure 15. This plate was designed to fit

on top of the NBS 9-A coupler so that a flat surface was

provided for the larger circumaural cushions. A vent made

from catheter tubing was inserted between the Beyer earphone

and flat-plate coupler. This was done to prevent an

airtight seal between the coupler and earphone which

previously created calibration difficulties.

General calibration of the Grason Stadler 1723 Middle

Ear Analyzer was conducted according to manufacturer's

specifications. The admittance output voltage was then

calibrated by connecting the probe assembly to -a variable-

volume syringe. Voltage change per unit physical volume

change was calculated, and included in the computer program

used for AR analysis. The system was calibrated so that

a 100 mv change in output voltage resulted from a 1 mmho

change in measured admittance at 220 Hz. Another confounding

variable which required special attention was response

time of the bridge. Response time of the middle ear

analyzer was calibrated according to procedures described

by Wilson, Shanks, Jones, and Danielson (1982). As

recommended, the diaphram of a TDH-49 earphone served as

the floor of calibration cavity. The face of the earphone

was then covered with a wide speculum and sealed to the




S.. -

Bore: .024" 1

-H .275"

.-- 2.25' .


Figure 15. Flat plate used with the NBS 9-A coupler
for Beyer DT 48 earphone calibration
(after Hepler, 1984).

earphone. The earphone was then connected with tubing to

a cut-off section of syringe. The probe tip of the GS 1723

was hermatically sealed to the syringe. With the 200 Hz

probe tone present, an externally generated 1000 Hz tone

was delivered to the cavity. The response time of the

middle ear analyzer was then calculated (see Figure 16).

As illustrated, the initial latency of the instrument was

approximately 34 msec. Dwell time of the bridge at stimulus

offset was estimated at 41 msec. These values were then

entered into the computer program utilized in waveform


Data analysis

Statistical analysis were conducted at the Florida

State University Computing Center utilizing an IBM 5520

computer. All computations were completed using Statistical

Package for the Social Sciences (SPSS) software. Descrip-

tive statistics for all threshold and AR parameters were

obtained. Analysis of variance for dependent measures

were completed according to Marks (1982) to answer the

following experimental questions.

1. Do differences exist between pre-exposure and

post-exposure values of the following response

variables in normal hearing subjects after a

two hour industrial noise exposure of 90 dB



--^ ^^ 34 msec

1.9438 -
1.9313 -
1.9138 -
1.9063 -
1.8688 -
1.3313 -
1.3128 -
1.793B -
1.7313 -

1.7183 -
(tR) a1

2- 0 A 03 75 1i000 12I3 1508 17FQ 2U'3 22S0 250.

Figure 16.

Analog output of a Grason Stadler 1723 middle
ear analyzer when a test pulse is delivered
to a calibration cavity.


a) Behavioral thresholds at octave and half-

octave intervals

b) Acoustic reflex threshold

c) Acoustic reflex magnitude

d) Acoustic reflex latency characteristics

(onset and offset)

e) Acoustic reflex adaptation

2) Do differences exist in percent adaptation for

the three stimuli evaluted in the pre-exposure


a) Continuous broadband noise

b) Industrial noise

c) Interrupted broadband noise

3) Do any of these measures correlate to changes

observed in audiometric behavioral thresholds

following this same noise exposure?

A Tukey multiple comparison procedure was conducted

whenever significant differences between treatments was

obtained. Pearson Product-Moment correlations were utilized

to investigate whether significant relationships existed

between the various measures of AR activity and expressions

of temporary threshold shift.


Pre-Exposure Measures

Behavioral Thresholds

Behavioral thresholds were established utilizing

Bekesy tracking procedures at octave, and half-octave

intervals. Results show that thresholds were within the

range expected for normal listeners.- Average threshold

values for pre-exposure testing were within 2 dB of ANSI

(1969) standards at all test frequencies. As expected,

greatest hearing sensitivity was demonstrated in the mid-

frequencies (750-2000 Hz). Similar findings have fre-

quently been reported in the literature (Shaw, 1974).

Thresholds were also established at the end of the

experiment to verify complete recovery from the industrial

exposure. Analysis conducted between pre-exposure and post-

exposure thresholds, showed no significant differences

present for the two sets of threshold measures (paired

difference t-test). Mean recovery data were within 1 dB

of pre-exposure thresholds, indicating complete recovery

at all test frequencies. Table 3 provides mean behavioral

thresholds,and standard deviations for pre-exposure,and

recovery sessions.

Acoustic Reflex Threshold

Acoustic reflex threshold was operationally defined

as,the lowest elicitor level required to cause a shift in

baseline admittance by at least two standard deviations

(Block & Wiley, 1979; Hepler, 1984; Moul, 1985). Two

responses out of three presentations at the same intensity

level were required for final threshold determination..

Thresholds were obtained for broadband,and industrial noise

stimuli at the beginning of each experimental session.

Descriptive statistics for the two elicitors are presented

in Table 4. This table demonstrates a large range of

threshold values for the two elicitors across subjects.

Values reported are in general agreement with those of

other investigators (Chun & Raffin, 1979; Gerhardt &

Hepler, 1983; Moul, 1985). Also note, that average AR

thresholds for the two elicitors are essentially the same.

Acoustic reflex thresholds for the broadband,and industrial

noise were 72.4 and 73.7 dB SPL, respectively. To verify

this assumption, threshold data were subjected to a One

Way Analysis of Variance for Repeated Measures (subjects X


Table 3.
Mean behavioral thresholds in dB SPL and
standard deviations for pre-exposure and
recovery data (N = 30).

Pre-exposure Recovery

Frequency Threshold Standard Threshold Standard ANSI
(kHz) dB SPL Deviation db SPL Deviation Standards
(X) (SD) (X) (SD) dB SPL

.5 13.5 5.4 12.8 6.5 11.5

.75 8.0 5.3 7.5 5.6 8.0

1.0 6.9 5.3 5.7 5.0 7.0

1.5 7.8 6.4 6.8 6.3 6.5

2.0 7.8 5.9 7.5 6.1 9.0

3.0 9.5 5.7 9.1 4.8 10.0

4.0 9.8 5.9 8.8 5.7 9.5

6.0 16.5

4.7 14.5

4.8 15.5

Table 4.
Descriptive statistics for acoustic reflex
pre-exposure thresholds (N = 30).

Minimum Maximum Mean Standard
Elicitor Threshold Threshold (X) Deviation
(dB) (dB) (dB) (dB)

Broadband 60 88 72.4 8.2

Industrial 58 88 73.7 8.8

elicitor). Findings indicate that significant differences

did not exist between the two elicitors at the p = 0.05


Acoustic Reflex Magnitude

Magnitude of the acoustic reflex was obtained by

subtracting the admittance of the middle ear prior to

stimulus presentation, from the admittance value during

maximum AR contraction. Research indicates that AR

magnitude varies directly with intensity level of the

activating stimulus (Wilson & McBride, 1978; Wilson, 1979;

Silman, 1984). AR magnitude, along with latency, was

obtained at several stimulus presentation levels (0-16 dB

SL re: AR threshold). The plotting of changes in acoustic

admittance as stimulus intensity increase, is typically

called an acoustic reflex growth function.

AR growth functions were established following threshold

determination for each subject. As expected, increases in

elicitor intensity level resulted in growth of AR magnitude

across subjects. Individual values of AR magnitude ranged

from .001 mmhos at threshold, to .242 mmhos at supra-

threshold levels. Figure 17 provides the average growth

function of AR magnitude for thirty subjects. Also as

previously reported, variability of AR magnitude between


~1 ~I.J



-H -


Elicitation Level (dB SL)

Figure 17.

Average growth in AR magnitude measured in
mmhos, per unit increase in elicitor
presentation level for 30 subjects.

subjects was striking (Hepler, 1984; Moul, 1985). Figure 18

provides an example of AR activity for two subjects with

similar thresholds stimulated by a 90 dB SPL broadband

noise elicitor. As demonstrated, even under similar

stimulus conditions, subjects varied considerably in

magnitude of the acoustic reflex response.

Acoustic Reflex Latency

Temporal properties of the acoustic reflex system

were also evaluated in this study. Four latency points

and two slope functions for each stimulus presentation level

were calculated (see Figure 9). Latency 1 (Ll) was defined

as the time in milliseconds from the beginning of an instan-

taneous admittance change to 10% of the steady state admittance

change. Latency 2 (L2) was the time required to reach 90%

of the maximum admittance change of the AR response.

Latency 3 (L3),and Latency 4 (L4) were measures used to

describe the offset characteristics of the acoustic reflex.

They are defined as the time in msec, from termination of

the eliciting signal to 90% and 10% of the steady-state

admittance value, respectively. In an effort to further

describe the dynamic properties of the AR, two additional

slope functions were calculated. Slope 1 (Sl), was used to

describe the admittance change at the onset of the AR, and

was defined as the change in millimhos per millisecond

1.3 -i

1.3055, -r
1.3B5 ^j v


1. 5z5
US) B 250 580 750 1800 125 1500 1758 2080 225 2560

1.7443 1

1.7316 \ /
1.7189 /
1.6491 '
1.6420 \ 1
1.6174 ..
(US) 0 250 500 750 1003 1250 S150 1750 2080 2250 2500

Figure 18. Example of AR magnitude difference between
two subjects for the same elicitor and
stimulus presentation level.

from 10% to 90% of the steady state admittance change. A

similar measurement was made at the offset of the acoustic

reflex. Slope 2 (S2), was calculated as the millimho

per millisecond change from 90% to 10% of the admittance

change following termination of the eliciting signal.

All measures of AR latency were completed at nine

different stimulus presentation levels (0-16 dB SL re:

AR threshold). This allowed for the evaluation of latency

per unit increase in intensity level. Onset of the acoustic

reflex response, measured by Ll, showed a consistent

decrease in latency per unit increase in intensity. In

general, the greater the intensity level of the stimulus,

the shorter the response time of initial AR activity (see

Figure 19). This finding is consistent with the reports

of other investigators (Borg, 1972; Dallos, 1973). The

second measure of AR onset latency, L2, failed to display

a similar relationship to stimulus presentation level. As

illustrated in Figure 20, L2 values did not change consid-

erably with increases in intensity level. The range of

values for this measure was only about 40 msec. This would

seem to suggest that stimulus intensity level has no affect

on L2 measures.

Offset properties of the acoustic reflex system were

evaluated by obtaining measures of L3, and L4. These

values were defined as the time required by the AR system,

to return to 90% and 10% of their respective steady state

admittance value. Findings show that L3 behaved in a