An electrophysiological study of binaural interaction in the chinchilla auditory cortex

Material Information

An electrophysiological study of binaural interaction in the chinchilla auditory cortex
Benson, Dennis Alan, 1944-
Place of Publication:
Gainesville, FL
University of Florida
Publication Date:
Physical Description:
1 online resource (xvi, 173 leaves.) : ill. ;


Subjects / Keywords:
Audio frequencies ( jstor )
Auditory cortex ( jstor )
Chinchillas ( jstor )
Ears ( jstor )
Electrodes ( jstor )
Histograms ( jstor )
Lesions ( jstor )
Monaural ( jstor )
Sound ( jstor )
Sound localization ( jstor )
Auditory perception ( lcsh )
Chinchillas ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Thesis--University of Florida.
Bibliography: leaves 162-172.
General Note:
Description based on print version record.
General Note:
General Note:
Statement of Responsibility:
by Dennis Alan Benson.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Dennis Alan Benson. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
004850282 ( AlephBibNum )
575365974 ( OCLC )


This item has the following downloads:

Full Text








Mary Katherine


The author wishes to heartily express his gratitude to

the chairman of his supervisory committee, Dr. Donald C. Teas,

for his continual guidance and encouragement throughout all

phases of this study as well as throughout the author's grad-

uate program. Acknowledgment is also due Drs. J. J. Bernstein,

W. W. Dawson, F. A. King, and W. B. Webb, not only for their

efforts as members of the supervisory committee, but also for

their interest and guidance in the formal training of the


A note of sincere thanks is extended to Robert Idzikowski

for his unfailing technical assistance.

Finally, for being a constant source of enthusiasm and

encouragement, a special kind of appreciation is due my wife,

Mary Katherine.



Acknowledgments iii

List of Tables vi

List of Figures vii

Abstract xi

Introduction 1

Methods 19
Surgical Procedures 19
Recording 21
Acoustic Stimulation 24
Data Collection 26
Histology 27
Experimental Protocol 28

Results 34
General Considerations 34
Anatomy 34
Evoked Potentials and Tonotopic
Organization 42
Threshold Frequency Relations 54
Depth of Recorded Units 63
Spontaneous Activity 65
Single-Unit Response Patterns 67
Binaural Representation 77
Effects of Interaural Intensity Differences 89
Effects of Interaural Time Differences 102
Introductory Comments 102
Tonal Stimuli 103
Click Stimuli 121
Units Responsive to Tonal and
Click Stimuli 128

Discussion 134
Introduction 134
Tonotopic Organization 134



Discussion (continued)

Characteristics of Single-Unit Activity 138
Interaural Time and Intensity Effects 145

References 162

Biographical Sketch 173



Table 1 Intensity Function Classification
Based on Responses to Monaural
Stimulation ---- -------------- 85

Table 2 Classification of AI Functions:
Relationship of Type of Al Function
to Interaural Intensity Difference
for Maximum Discharge -------------- 101



Figure 1 A, Photomicrograph of a frontal
section through an area of the
temporal cortex that was unrespon-
sive to acoustic stimuli.
B, Photomicrograph of the auditory
field located 2 mm ventral to the
area in A. Cresyl violet stain.-------- 35

Figure 2 Photomicrograph of frontal section
of auditory field. Protargol
stain. --------------------------------- 37

Figdre 3 A, Photomicrograph of frontal
section of auditory cortex. Golgi-Cox
B, Photomicrograph of upper four
layers of auditory cortex. Golgi-Cox
stain. --------------------------------- 39

Figure 4 A, Detail of pyramidal cells from
layer V. Golgi-Cox stain.
B, Typical glass microelectrode.------- 41

Figure 5 Averaged evoked potentials recorded
from microelectrode at 600-p depth
in animal 3-12. ------------------------ 44

Figure 6 Examples of the correlation between
the averaged evoked response and
the probability of unit discharge.------ 48

Figure 7 A, Relationship of brain to chincilla
B, Relationship of auditory area to
to brain.
C, Suggested schema of tonotopic
D-F, Maps of unit best frequencies
for three animals. --------------------- 52

LIST OF FIGURES (continued)

Figure 8 Unit best frequency versus thres-
hold SPL at best frequency.-------------

Figure 9 A, Histogram of distribution of
unit best frequencies.
B, Histogram of distribution of unit
thresholds with respect to the
inferred threshold SPL.----------------

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Isointensity response areas for
four units.---------------------------

Histogram showing depth distribution
of 95 recorded cells.----------------

Histogram of the distribution
of spontaneous activity rates.---------

PST histograms of tonic and phasic
response patterns with four of the 32
individual spike trains that comprised
the PST histogram.-------------------

Characteristic response patterns to
tonal stimuli.-----------------------

PST histograms of four units showing
effects of stimulus intensity on
response patterns.-------------------

Intensity function and selection of
PST histograms for unit 3.2-16.---------

Classification of intensity functions for
105 units according to monaural response
class and binaural response category.---

Intensity functions and interaural in-
tensity difference (AI) functions for
four units illustrating three principal
types of AI functions and the effects
of intensity level on the shape of the
AI function.-------------------------



LIST OF FIGURES (continued)

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Figure 29

Examples of peaked AI functions
from four units. ----------------

Examples of monotonic AI functions
from four units.--------------------

Graph of discrimination indexes for
72 units as a function of unit best

Interaural phase-difference (A4)
functions for three units.---------------

Effect of overall intensity and inter-
aural intensity differences on A(
functions for four units.---------------

Effect of stimulus frequency on Ag
functions for four units.--------------

A, Plot of optimal A4 as a function
of the frequency of the A4 function.
B, Same data as A replotted as
optimal At versus frequency.------------

Period histograms from three units
showing absence of phase-locked firing.--

PST histograms, number of discharges
for 32 trials, and mean latency to
initial discharge for range of At
intervals with click stimuli for
unit 4.12-14.---- --------------

Examples of At functions for click
stimuli from three different units
showing different degrees of sensi-
tivity to At parameters. ----------------

A, Example of changes in click At func-
tions with increases in overall intensity
B, Click At functions for unit insensi-
tive to At parameters.------------------













LIST OF FIGURES (continued)


Figure 30

Figure 31

Comparison of At functions for
tone and click stimuli for four
different units.----------------------

Interaural intensity differences at
chinchilla tympanum as a function of
frequency for free field sound source
at two azimuths.-- --------------



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




August, 1973

Chairman: Donald C. Teas
Major Department: Psychology

The interaction of acoustic input from the two ears is

essential for accurate localization of sound in space. Al-

though binaural convergence first occurs at lower levels of

the auditory pathway, ablation studies indicate that the

auditory cortex is a necessary structure for accurate sound

localization. The primary aim of this investigation was to

systematically compare for various stimulus conditions the

relative influences of contralateral and ipsilateral acoustic

stimulation on cortical single units in an unanesthetized pre-

paration and to study the effects upon single-unit responses

of the dominant stimulus cues for sound localization--interaural

intensity difference (AI) and interaural time difference (At).

Recordings were obtained with KC1- or NaCl-filled glass

microelectrodes for 133 units from chinchillas immobilized by

gallamine triethiodide. Stimuli were clicks and shaped tone

bursts 200 msec in duration presented through headphones to

allow independent control of AI and At.

Both ears were found to be represented in the responses

of all units studied. An ear was considered "represented" if a

unit responded to stimulation of that ear or if the response to

stimulation of the other ear was modified by stimulation of

that ear. Contralateral stimulation, however, had a much

greater facilitatory effect than ipsilateral stimulation: for

contralateral stimulation unit thresholds were lower and more

discharges were elicited than for ipsilateral stimulation over

a range of intensities from unit threshold intensity to 80 dB

SPL. The most common form of binaural interaction was summation;

when inhibition was seen it was usually contributed by the ipsi-

lateral ear.

A predominance of contralateral influence was also observed

when the number of stimulus-evoked discharges was plotted as a

function of the Al or At. For 62% of the AI functions maximal

responsiveness occurred for binaural stimuli that were more in-

tense at the contralateral ear. Similarly, of the 36 units

that showed sensitivity to At parameters for tone stimuli, 22

(61%) were maximally responsive at the contralateral-leading At

intervals. For click stimuli, At functions did not generally

peak at a particular At interval as they did for tone stimuli,

yet for all 21 click-responsive units that showed sensitivity to

At parameters, maximal responsiveness also occurred for contra-

lateral-leading stimuli.

Certain observations in the study question the generality

of the hypothesis that a- particular cell invariantly encodes a

specific At, i.e., that cells have "characteristic delays."

First, most units tested at more than two frequencies showed

maximal responsiveness at different At intervals depending upon

stimulus frequency. Second, the At intervals for maximal re-

sponsiveness for half of the units tested were greater than the

maximum interaural delays the animal could encounter naturally.

Third, At functions from the same unit for click and tone

stimuli showed poor correspondence. Consequently, the data

were discussed in terms of a population hypothesis, originally

suggested by von Bekesy, in which localization is accomplished

by a comparison of the relative amounts of activity in the two


Incidental to the primary objectives of this study, ob-

servations were made regarding the anatomy and tonotopic


organization of the chinchilla auditory cortex and the response

characteristics of single units in the unanesthetized preparation,

including spontaneous activity, frequency selectivity, temporal

response patterns, and discharge rate-intensity functions.



For the past hundred years investigators have attempted to

determine the specific role the auditory cortex plays in the

overall functioning of the auditory system. As early as 1870,

Ferrier (1876) began extensive stimulation and ablation studies

on the brains of several species. lie noted that electrical

stimulation of the ectosylvian region in cats and the superior

temporal gyrus in monkeys caused the animal to prick up or retract

its ears and turn its head and eyes to the side contralateral

to the stimulation just as if it were orienting to a sound in the

free field. Following the identification of an "auditory area,"

Ferrier ablated it and found that the animals would no longer

orient to the sound.

These pioneering studies of Ferrier established a pattern

of experimentation that was to be followed in nearly all subsequent

work in the area. Research was directed towards two inseparable

goals, first, finding the extent of auditory cortex and, secondly,

determining what loss of auditory function -resulted from its


Rapid progress was made towards the first goal by histo-

logical investigations. The early myelination studies of Vogt

and Vogt (1919) and the cytoarchitectonic studies of Campbell

(1905) both defined in cat an auditory cortex that was only

somewhat larger than what is now commonly considered primary

auditory cortex or AI. Subsequent histological studies have

used both anterograde and retrograde degeneration to subdivide

the auditory cortex of the cat on the basis of the thalamo-

cortical connections (Woollard and Harpman, 1939; Bremer and

Dow, 1939; Rose and Woolsey, 1949). By comparison with the

cat, the number of anatomical studies of the subdivisions of

auditory cortical areas in other species is very limited.

For the dog, Tunturi (1960) has shown anatomically and electro-

physiologically that the middle ectosylvian gyrus (MES area)

represents the homologue of the cat AI area. For the rhesus

monkey, degeneration studies (Poliak, 1932; Clark, 1936;

Walker, 1937) have confirmed Ferrier's earlier findings that

the primary projection area is located on the superior surface

of the superior temporal gyrus.

Once the extent of the auditory cortex was established,

the second goal, its function, became a subject of investigation.

To assess the role of the auditory cortex in hearing, inves-

tigators studied the hearing capacities of animals before and

after cortical lesions. A review of the results of these studies

must be prefaced with the qualification that the following

deficit function studies involve not only the discrimination

of an auditory stimulus but also the ability to correlate

the discrimination with a previously learned response and

then to effect that response. Thus, when a discrimination

cannot be relearned after a lesion it is equally plausible

that either the site of the discrimination has been destroy-

ed, that access to the effector part of the system has been

impaired or that both the discrimination and the response

parts of the system have been affected. Another source of con-

flict in the interpretation of these studies arises from dif-

ferences in the size and location of the lesion or from differ-

ences in the training and testing procedures. Nonetheless, sub-

stantial agreement has evolved as to what auditory functions

may or may not involve the auditory cortex. All the experiments

reviewed were performed on cats unless otherwise noted and all

lesions were confined to the cortex.

It was first assumed that an intact auditory cortex was

essential for normal hearing acuity. Although bilateral

temporal lobe lesions are uncommon in the human clinical

literature, they are considered to cause permanent deafness.

For the monkey, however, bilateral ablation of the primary

auditory area appears to have no effect on post-operative

absolute thresholds (Wendt, 1934; Harris, 1943). Similarly

for the cat, Kryter and Ades (1943) demonstrated that even

after bilateral ablations extending beyond the recognized

auditory areas absolute thresholds did not change. The

animal findings are not surprising in light of Elliot's

demonstration (1967) that up to 90% of first-order fibers

can be destroyed with little effect on absolute threshold.

Since absolute threshold was not affected by cortical lesions,

investigators resorted to differential intensity or frequency

discrimination tasks.

Raab and Ades (1946) and Rosenzweig (1946) both found

that although amnesia occurred after complete bilateral

auditory cortical lesions, i.e., destruction of both primary

and secondary auditory areas, the discrimination of a change

in intensity of tone pulses could be relearned to pre-ablation

levels of accuracy. Oesterreich et al. (1971) found similar

results with even more extensive bilateral ablations than

the above two studies.

Numerous studies of frequency discrimination after bilateral

cortical ablation have revealed a range of effects from permanent

to no impairment in relearning a frequency discrimination task.

Most of the differences between studies can be related to testing

procedures. When the absolute discrimination of two different

frequencies has been tested to go, no-go or two-alternative,

forced-choice procedures after complete bilateral lesions of

auditory cortex, permanent impairments have been seen in cats

(Cornwell, 1967), dogs (Allen, 1945), and monkeys (Massopust

et al., 1965). On the other hand, if the discrimination in-

volves merely the detection of a new frequency, losses are

less severe or non-existent. For example, Butler et al.

(1957) and Goldberg and Neff (1961) found that after extensive

bilateral cortical lesions cats could relearn to detect the

occurrence of a series of high frequency tone bursts in a

continuous background of low frequency tone bursts. The work

of Thompson (1959, 1960) would suggest that the contrasting

results for the two different frequency discrimination tasks

are related to the relative difficulty of initially learning

the two tasks. Since the absolute discrimination task is

extremely difficult to acquire, the impairment would be ex-

pected to be more severe than that seen for the comparatively

simple detection task. Thus, as long as frequency discrimi-

nation is understood as the ability to detect a change in

the frequency parameters of the acoustic signal as opposed

to the recognition of the differences between two frequen-

cies, one can say frequency discrimination is unimpaired

by extensive auditory cortex ablations.

A task, however, that does result in an unequivocal permanent

impairment is tonal pattern discrimination (Diamond and Neff,

1957; Diamond et al., 1962). Tone bursts of two frequencies

were used, the only difference between the neutral and positive

stimuli was the temporal order of high and low tones: high,

low, high versus low, high, low. Complete bilateral removal

of the auditory cortex still permitted the detection type of

frequency discrimination to be made, but the ability to

discriminate the tonal pattern was permanently lost (Diamond

and Neff, 1957).

Since temporal order is the only difference between stimuli

in the tonal pattern discrimination, it is not unexpected that

other discrimination tasks involving temporal cues are severely

impaired after lesions. Sharlock et al. (1965) found the

ability of cats to discriminate a 1-sec tone burst from a

4-sec tone burst permanently lost after an extensive bilateral

lesion. Similarly, Symmes (1966) found that bilateral ablation

of the primary auditory area in the rhesus monkey resulted

in a permanent inability to discriminate continuous noise

from noise that was interrupted for 20 sec at the rate of ten

times per sec. In both the studies above, animals retained

frequency discrimination ability. For dogs, Gershuni et al.

(1968) and Khananashvili (1966) reported disruption of

frequency and intensity discrimination when signals were

short (less than 100 msec in one study, less than 20 msec in

another). Longer signals were not affected. Khananashvili

also found that discrimination of rise-time was severely

affected by bilateral removal of the ectosylvian and

suprasylvian gyri. After the lesion, rise-times of 130 msec

could not be discriminated from rise-times of 30 msec.

Signals, however, with 250-msec and longer rise-times were


The evidence just cited which suggests the importance

of temporal cues in cortical function is further supported

by the results of sound localization studies. Neff et al.

(1956) placed a buzzer behind each of two food boxes and

trained cats to go to the box from which the buzzer sounded.

Although normal cats could discriminate boxes 50 apart, cats

with complete bilateral ablations had difficulty learning

to discriminate boxes 1300 apart. With lesions that included

the auditory areas but also extended into the suprasylvian

gyri, a practically complete loss of localization occurred.

Strominger (1969) used a similar testing situation to

determine the effects of bilateral ablations of various

subdivisions of the auditory cortex. He found that only

ablations of primary auditory cortex (AI) produced deficits

in postoperative learning. Similar results were shown for

rhesus monkey (Wegener, 1964).

Since localization of sound in the free field involves

the interaction of both time of arrival and intensity differ-

ence cues, sound must be presented through headphones in

order to study the two cues'separately. When this is done,

human subjects report a sound image that is either inside

or in close proximity to the head and that can vary from

left to right depending on the time and intensity parameters

of the stimulus. For example, clicks presented simultaneously

and of equal intensity to the two ears will appear as one

click located along the midline; but if one ear receives

the click slightly earlier than the other, the apparent source

moves towards the ear receiving the leading click. The

minimum interaural time difference required to shift the

source from midline is 30 to 40 Usec for man (Wallach et al.,

1949) and approximately 50 psec for cat (Masterton and Diamond,

1964). When the time difference is increased to 500 usec

one click is still heard but it is strongly lateralized to

the leading ear. In the Masterton and Diamond (1964) study

cats readily learned to lateralize dichotic clicks with a

500 usec time difference once they had learned to make an

appropriate response for single clicks in either the left

or right ears. Thus it appears that for the cat as for

humans, the dichotic click pair is perceptually equivalent

to a monaural click in the ear that receives the leading

click. After complete bilateral ablation there was no

transfer from learning the monaural click discrimination

to learning the lateralization of the dichotic clicks,

indicating that the perceptual equivalence of the dichotic

clicks and monaural clicks was lost. Even though the lesioned

cats eventually learned to lateralize the dichotic clicks

after extensive training, the authors concluded that relearning

was probably based on a set of cues, such as intensity or

quality differences between the clicks, that did not involve

the percept of location. Elliott and Trahiotis (1972) have

pointed out that the task which was relearned in the above

study merely involved the detection of a shift from a left-

leading to a right-leading click pair or vice versa. So it

was not surprising that in a study (Axelrod and Diamond,

1965) which involved an absolute discrimination, namely, the

discrimination of the absolute location of a dichotic-click

pair image, bilateral ablation resulted in an inability to

relearn the discrimination.

In all the tasks employed in the above studies it

should be noted that unilateral lesions were ineffective

in producing any significant loss of discrimination even

in those tasks where bilateral lesions were most effective,

i.e., in tonal pattern discrimination, localization, or

lateralization. There have been reports, however, that

indicate unilateral lesions can cause distinct impairments.

As early as 1884 Luciani noted that unilateral lesions

caused a temporary impairment in localizing sound only when

the sound was coming from the side contralateral to the

lesion. Recently, Whitfield et al. (1972) trained cats

to go to food boxes on the side from which tone pulses sounded.

Then the tones originated from both sides but with one side

leading by 5 msec. The animals appeared to treat these com-

pound tones as humans do, i.e., as one sound originating on

the side leading in time and thus immediately transferred to

the new task. After complete unilateral lesions of auditory

cortex, the perception of the fused pair of tones was perma-

nently disrupted when the leading member of the pair was

contralateral to the lesion. When the leading member of the

pair was ipsilateral, responses were normal.

Other instances of auditory deficits with unilateral

lesions can be found in the human clinical literature in cases

where patients with unilateral cortical damage have been

tested for their localization ability. Walsh (1957) tested

the ability of 20 neurosurgical patients with various kinds

of unilateral lesions and found that all were able to use

interaural time differences as localization cues. He did

not report that there was any difference in performance

depending on which ear received the leading click. The

interaural time differences used in the study, however, were

much larger than threshold values for normal subjects.

Sanchez-Longo and Forster (1958) in their study of 50 patients

with unilateral temporal lobe damage found accuracy of local-

ization impaired on both sides but a greater deficit on the

contralateral side. Teuber and Diamond (1956) compared two

groups of subjects, normal subjects and those with unilateral

cortical lesions, for their judgements of the apparent location

of dichotic clicks. Subjects with lesions required more

intensity in the ear contralateral to the lesion than in the

ipsilateral ear in order to judge a sound in the median plane.

Finally, Strominger and Neff (1967) tested localization of a

broad-band noise source by a man with a right hemispherectomy.

Although his overall performance was not severely impaired as

compared to normal subjects, his largest errors occurred in

the auditory field contralateral to the hemispherectomy.

The data from studies of unilateral lesions are very

interesting in that they suggest that for some discrimination

tasks the hemispheres are not equivalent, i.e., the intact

hemisphere cannot take over certain functions of the lesioned

hemisphere. Moreover, in animal studies, there is no indication

of any hemispheric dominance since the difference between

hemispheres appears symmetrical with each hemisphere demonstrating

a predominance of contralateral input.

If there is, indeed, a difference in the representation of

the contralateral and ipsilateral ears at each hemisphere, that

difference should be reflected in electrophysiological studies.

Recall that Ferrier used electrical stimulation on monkeys and

found orienting responses to the contralateral side. Electrical

stimulation was used also by Penfield and Rasmussen (1950) on

conscious surgical patients who usually reported hearing sounds

on the side contralateral to the stimulation; less often, sounds

were said to be in both ears; sounds were never heard ipsilaterally.

Studies of gross-electrode recording of evoked potentials

in the primary auditory area have generally agreed that the

responses to acoustic stimulation of the contralateral ear are

greater than responses to ipsilateral stimulation (Bremer and

Dow, 1939; Tunturi, 1946). Rosenzweig (1951) measured the

amplitudes of evoked responses at 49 electrode sites in five

cats and showed statistically that the contralateral response

was significantly larger than the ipsilateral at 29 locations

and that there were no significant differences at 20 locations.

At only one location was the ipsilateral response significantly

larger. Gross et al. (1967) also noted the generally larger

amplitudes of contralaterally-evoked responses in AI, All, and

Ep (posterior ectosylvian area). They showed that the area

over which responses could be recorded to contralateral and

ipsilateral stimulation overlapped but that the area from

which responses to contralateral stimulation could be evoked

was more extensive.

Evoked potential studies of the auditory cortex which

have used interaural differences in the acoustic stimulus have

also shown a differential representation of the two ears.

Rosenzweig (1954) and Keidel et al. (1960) demonstrated that

the response to dichotic clicks with small interaural time

differences (0 to 2.0 msec) was larger when the leading click

was presented to the contralateral ear. Hirsch (1968) employed

interaural time delays with low frequency pure tones and found

that evoked response amplitude was a periodic function of the

interaural time difference. The period was equal to that of

the pure tone stimulus and the maximum response occurred when

the contralateral stimulus led (seven experiments) or when

both contralateral and ipsilateral stimuli were simultaneous

(two experiments).

Brugge et al. (1969) have shown that there are single

units in the auditory cortex responsive to interaural time

and intensity differences. Such units have been demonstrated

in lower levels of the system: superior olivary complex

(Galambos et al., 1959;.Rupert et al., 1966; Goldberg and

Brown, 1969), lateral lemniscus (Brugge et al., 1970),

inferior colliculus (Erulkar, 1959; Rose et al., 1966;

Geisler et al., 1969; Benevento et al., 1970; Aitkin et al.,

1972), and medial geniculate (Adrian et al., 1966; Aitkin

and Dunlop, 1968; Starr and Don, 1972). At the cortex Brugge

et al. (1969) found that spike counts were periodic functions

of interaural time differences for tones between 200 Hz and

2.4 kHz. It was not reported whether an ipsilateral or

contralateral lead was more effective, although their figures

show examples of both. Units were also found that were

responsive to interaural intensity differences, but again,

the relative proportion of ipsilaterally versus contralaterally

excited units was not reported. Hall and Goldstein (1968)

found that single-unit activity in unanesthetized cats was

influenced by both ears with the contralateral ear predom-

inating. Interaural time or intensity differences, however,

were not varied.

To summarize, first, we have seen that the auditory

cortex cannot be considered an essential structure for all

auditory discrimination. For discrimination which merely

require the detection of a change in stimulus characteristics,

namely, threshold, frequency, or intensity discrimination,

no impairments result from auditory cortical lesions. On

the other hand, when recognition and identification are

required, permanent deficits are seen, e.g., in absolute

frequency discrimination or tonal pattern discrimination.

Second, tasks requiring the localization of sound in space

or lateralization of sounds presented through headphones are

particularly susceptible to permanent impairments. Finally,

evoked potential studies and ablation studies involving

unilateral cortical lesions indicate that the two hemispheres

cannot be regarded as equivalent; rather, there appears to

be a stronger representation of the contralateral ear in

each hemisphere.

The purpose of this dissertation is to pursue at the

single-unit level a search for the possible mechanisms which

underly some of the deficits revealed in the behavioral

studies, specifically those deficits related to a differential

representation of the two ears in each hemisphere and deficits

in localization behavior. By examining in detail the time-

dependent properties of a large sample of neurons from different

experiments one hopes to gain some understanding of the behavior

of a neural population under a given set of stimulus conditions.

The preparation selected to carry out these objectives

is the unanesthetized, muscle-relaxed chinchilla, an animal

whose peripheral system has been widely studied anatomically

(Boord and Rasmussen, 1958; Smith and Rasmussen, 1963) and

electrophysiologically (Dallos, 1966; Strother, 1967).

Single-unit studies have been performed on the cochlear

nucleus (Mast, 1970) and inferior colliculus (Teas and Adams,

1973). Behavioral testing has shown the chinchilla to have

an audibility curve similar to that of man (Miller, 1970)

and an ability to readily acquire various auditory discrimina-

tions (Luz, 1969).

The unanesthetized preparation is desirable from the

standpoint of approximating the natural state of the animal

as closely as possible. Neural activity recorded from the

barbiturate-anesthetized preparation shows a severe reduction

in spontaneous activity (Li and Jasper, 1953) and a difference

in temporal response pattern when compared to the neural

activity from the unanesthetized preparation. With barbiturate

and chloralose anesthesia, Brugge et al. (1969) found that

for cats nearly all units respondedto shaped tone bursts with

either a single spike or a burst of two to five spikes. On

the other hand, using similar stimuli but with the unanesthe-

tized cat, Evans and Whitfield (1964) and de Ribaupierre

et al. (1972) reported a wide variety of response patterns.

The phasic on-burst described by Brugge et al. constituted

only about 12% of the response patterns observed by Evans

and Whitfield and 40% of those seen by de Ribaupierre et al.

Briefly, the aims of the dissertation are as follows:

first, to determine the extent and tonotopicity of

the chinchilla auditory cortex and to gain a general

understanding of single-unit responses in the unanesthe-

tized preparation by studying the response patterns,

frequency selectivity, and intensity functions of

single units;

second, to examine the relative influences of

contralateral and ipsilateral input on single units and

how this representation is affected by overall frequency

and intensity; and,


third, to examine in detail the responsiveness

of single units to the interaural parameters of the

acoustic stimulus that are essential for localization:

intensity and time.


Surgical Procedures

The experiments were carried out on 29 chinchillas that

weighed from 350 to 600 g. All animals were judged to be in good

health; examination of the external ear canal and tympanic mem-

brane through a Zeiss operating microscope failed to reveal any

signs of external ear infection. All animals demonstrated nor-

mal sound reflexes to the presentation of a loud click.

A tracheal cannula was inserted during inhalation of ethyl

chloride vapor. As soon as the cannula was inserted and the

animal placed in a rigid head-holder, the administration of

ethyl chloride was terminated and ventilation by a respirator

pump was begun. The animal's total time of exposure to the

ethyl chloride was less than 5 min; 3 hr usually lapsed before

recording began. Since muscular activity returned almost im-

mediately after cessation of ethyl chloride, 2.4 mg of galla-

mine triethiodide (Flaxedil) was administered intramuscularly

to immobilize the animal. Additional 1.6 mg doses of galla-

mine triethiodide were given throughout the course of the

experiment whenever withdrawal reflexes could be elicited by

pin pricks of the hind limbs. The interval between injections

tended to increase throughout the experiment but the average

interval was 2 hr. To prevent undue discomfort to the animal

care was taken that all incisions and pressure points were kept

heavily infiltrated with lidocaine (Xylocaine, 2% solution).

Pupil size was observed to be small but never completely con-

stricted; rapid dilation occurred when testing for withdrawal

reflexes. Rectal temperature was monitored with a telethermo-

meter and maintained at approximately 37 C by manually adjust-

ing a hot-water heating pad. The electrocardiogram (EKG) as

measured between needle electrodes beneath the scalp and in

a hind limb was monitored continuously.

The scalp was incised along the midline and enough skin

and soft tissue was removed or reflected over the dorsal and

lateral aspects of the cranium to allow clear observation of

the orientation of the cranial sutures. A 0.5-mm-diameter

dental drill was used to bore a conical hole that was approxi-

mately 1.2 mm in diameter at the cranial surface and decreased

to 0.5 mm in diameter at the cranium-periosteum margin. Bleed-

ing during the drilling was controlled by packing the hole with

bone wax and redrilling. By restricting the exposure of the

cortex to a very small opening no problems with drying or pul-

sation were encountered. The dura and arachnoid membrane were

excised with a specially tapered scalpel taking care to avoid

blood vessels. Usually two or three holes were drilled at

the same time to allow other penetrations in different loca-


The headholder was designed to provide sufficient sta-

bility for single unit recording without depending on ear

bars for support. A bite-bar, mandible supports, and a clamp

over the muzzle provided rigid support while still allowing

rotation of the head on its rostro-caudal axis. Each spec-

ulum could be freely inserted into the external ear canal and

adjusted visually to be in the same position with respect to

the tympanic membrane for all animals.


Microelectrodes were made from 0.9-mm-diameter glass

pipettes using a commercial electrode puller (Kopf 700C).

Tips were broken under a microscope to produce a tip diameter

of approximately 2 p. The shaft of the electrode that act-

ually entered the cortex was made as thin as possible to avoid

damage due to tissue displacement; the diameter of the elec-

trode at a distance 2 mm from the tip was commonly 30 V. The

microelectrodes were filled with a solution of 3 M KC1 or

4 M NaC1 by syringe and placed under a vacuum for 24 hr to com-

plete the filling of the tip. The KCl-filled electrodes were

used only in the initial experiments; the majority of experi-

ments used the NaCI electrodes. Since intracellular recording

was not planned, the NaC1 electrodes were more satisfactory

with regard to the possibility of electrolyte leakage. When

tips were successfully filled, impedances at 100 Hz typically

ranged from 5-25 megohms when measured in 0.9% saline.

The electrode was fitted to a plastic carrier that in

turn was clamped to a hydraulic microdrive (Kopf 1207B).

Since the electrode assembly and microdrive were mounted on

a mechanical micromanipulator and since the headholder was

adjustable, there was enough freedom of adjustment to allow

the electrode to be oriented normal to the cortical surface

and to permit repeated penetrations at measured distances

from the original penetration in the same cranial opening.

The electrode was lowered under visual control until it

touched the cortical surface; further lowering was accom-

plished remotely via the microdrive system from outside the

chamber in which the animal was located. The reference elec-

trode was a 24 gauge needle inserted in the neck muscles.

The electrical activity recorded by the microelectrode

was led to the nearby cathode follower input of a Grass P511

amplifier with a passband setting of 0.1 Hz to 3.0 kHz. The

output of the amplifier was directly displayed on one trace

of a dual-beam oscilloscope (Tektronix 565) that had its sweep

synchronized to the presentation of the acoustic stimulus. The

output was also passed through a high-pass filter (Krohn-Hite

3100) with a low-frequency cutoff of 100 Hz to filter out the

evoked potential and to stabilize baseline fluctuations. The

filtered output was led to an audio monitor and to a Schmitt-

trigger circuit which detected each action potential by pro-

ducing a standard pulse if and only if the action potentials

exceeded a certain preset level. The Schmitt-trigger pulses

together with the filtered signal which served as the input

to the Schmitt-trigger circuit were displayed on separate

traces of the oscilloscope and continuously monitored to in-

sure that pulses were not being accidentally triggered by back-

ground noise. Accidental triggering, however, was not a pro-

blem since the amplitude of action potentials recorded with

the fine-tipped glass microelectrodes was considerably greater

than the background noise level. Multi-unit activity was seen

only rarely and in such cases advancing the microdrive a

slight distance permitted the isolation of a single unit.

The Schmitt-trigger pulses and synchronization pulses that

corresponded to the time of stimulus onset were led to a LAB-8e

computer for on-line calculation of poststimulus time (PST)

histograms. The PST histogram represents the distribution of

the number of unit discharges as a function of time from the

onset of the acoustic stimulus; each PST histogram was based on

a standard 32 presentations of a given stimulus condition.

Acoustic Stimulation

Throughout the experiment the animal was within a single-

wall sound-attenuated chamber (IAC). For experiments subse-

quent to December 1, 1972, further sound attenuation was af-

forded by a concrete-block walled enclosure of the IAC cham-


Acoustic signals were normally tone bursts or clicks;

on occasion, noise bursts were used as search stimuli. Noise

was furnished by a noise generator (Grason-Stadler 455C) fol-

lowed by a bandpass filter (Krohn-Hite 3100). Pure tones were

provided by a function generator (Hewlett-Packard 3300A) the

frequency of which was monitored by an electronic counter (Monsanto

101A). The output of the function generator was led to a phase

shifter (General Radio E3520B) which provided both an unshifted

output and an output that could be continuously shifted from

0 to 3600. Thus, interaural time differences for tones were

introduced by generating phase differences. Stimulus presentation

rate was controlled by a pulse generator which delivered gating

pulses at desired intervals to a laboratory-built switching circuit

that could vary the duration and envelope of the tone and

noise bursts. All tone bursts were triggered on positive-going

zero-crossings of the sine wave. Tone bursts were usually of

200-msec duration with an 8-msec rise/decay time. Stim-

ulus presentation rate was typically 1/1.5 sec with each

stimulus condition consisting of 32 stimulus presentations.

Exceptions are noted in the figures and text.

Clicks were produced by'50-vsec pulses from a Grass S8

stimulator. Delayed pulses over the range of 25 psec to 1

msec obtained from the stimulator allowed presentation of

dichotic clicks with various interaural time differences.

Click presentation rate was usually 1/1.5 sec.

Clicks or burst stimuli were led to a two-channel am-

plifier (JBL SE400S) and then into separate attenuators

(Hewlett-Packard 350D). The stimuli were transduced by

0.5-in. condenser microphones (Bruel and Kjaer 4133) fitted

into specula that were fixed tightly in the external ear

canal. Within each speculum were two small pieces of pipe

cleaner for damping and a 0.125-in.-diameter probe tube that

permitted sound-pressure measurements with 2-3 mm of the

tympanic membrane. Acoustic calibrations were obtained with

the specula fitted into the external ear canals and the probe

tube connected to a 0.125-in. condenser microphone (Bruel

and Kjaer 4138). All intensities are expressed in decibels

sound-pressure level (dB SPL), i.e., decibels re 0.0002 dyne/

cm2. Click intensities are expressed as peak equivalent

pressures, i.e., the dB SPL required for a 1 kHz pure tone

to match the peak-to-peak amplitude of the acoustic waveform

of the click.

Acoustic crossover was measured for two animals by exam-

ining cochlear microphonic potentials for ipsilateral and con-

tralateral stimulation. These measurements indicated that a

minimum of 40 dB separation existed for frequencies from 300

Hz to 15 kHz with the acoustic system described above.

Data Collection

The preliminary analysis of data from the unit recordings

was performed by on-line calculation of PST histograms. Most

analyses, however, were performed off-line from FM tape

recordings of the data. The FM channel bandwidth of 0-1.6

kHz was found to cause only slight distortion in the action

potential waveform which had already been filtered at 0.1 Hz

to 3.0 kHz. With respect to the principal measure of data

analysis, discharge count, the distortion was without effect.

For data processing purposes the following signals were recorded

on separate channels of an FM tape recorder (Ampex FR-1300) at

3.75 in./sec:

1. Amplified signal from electrode (0.1 Hz
to 3.0 kHz amplifier bandpass);

2. Output pulses from Schmitt-trigger circuit;

3. Synchronization pulse at 10 msec prior to stimulus
for pre-stimulus triggering;

4. Synchronization pulse for onset of tone

5. Stimulus to contralateral ear (practical upper
frequency limit for recording stimulus was
2.5 kHz due to bandwidth of FM channel); and,

6. Voice commentary.

For off-line data analysis, the LAB-8e computer was used

to produce PST histograms, latency histograms, or trial-by-

trial analyses of responses for single-unit activity in addition

to averages for the evoked potential recordings. Programs were

written to compute descriptive statistics for latency measure-

ments and for the number of discharges in a specified time

interval after the acoustic stimulus.


Routine histological analyses were not performed after

every experiment since multiple penetrations were frequently

made through the same portion of cortex that lay exposed be-

neath the 1-mm-diameter cranial opening. The depth of units

was derived from readings of the microdrive unit. Calibra-

tion of the microdrive showed its accuracy was within 1%,

but actual readings were probably accurate to only 50 p

due to the dimpling of the cortex upon electrode insertion.

Histology was performed on four animals that had been

used for recording. The extent of cortex responsive to acous-

tic stimuli was marked with insect pins and the animal per-

fused with 10% formalin. A large block of tissue containing

the pins was embedded in paraffin and sectioned in the frontal

plane at 15 v. Every fifth section was mounted and stained

in cresyl violet (two animals) or protargol (two animals).

A complete set of serial sections of the chinchilla brain

stained with cresyl violet and by the Weil method were also

available for study and comparison.

In order to examine the types of cells and the dendritic

organization in the chinchilla auditory cortex, Golgi-stained

serial sections were prepared from four animals. Animals were

perfused with 0.9% saline and the tissue stained according to

the Golgi-Cox method (Ramon-Moliner, 1958). The celloidin-

embedded tissue was sectioned at 150 p in the frontal and

sagittal planes and every section mounted.

Experimental Protocol

To obtain units responsive to acoustic stimuli, the elec-

trode was manually advanced slowly into the cortex by the

microdrive from outside the experimental chamber. A wide

range of frequencies was swept to determine if an evoked

response could be elicited from tone bursts. If an evoked

response was seen for a narrow range (1/2 octave approximately)

of tonal stimuli, then tone bursts within that range of fre-

quencies were employed as search stimuli; otherwise, wide-band

(100 Hz to 20 kHz) noise bursts were used. Often a unit with

spontaneous activity would be encountered that was unresponsive

to the search stimuli. In that case, various combinations

of frequency and intensity parameters were presented either

monaurally or binaurally in order to find the stimuli effective

for driving the unit. Hence, units were selected for respon-

siveness to tonal stimuli; units solely responsive to noise

bursts or clicks were excluded from further analyses. Preliminary

investigations and data from Goldstein et al. (1968) indicated

that such units constitute a small percentage of auditory cortical

units in the unanesthetized preparation (11% of the total sample

in the Goldstein et al. study).

After satisfactorily isolating a unit that could be driven

by tonal stimulation, the following protocol was performed in

order to obtain standard information from a large sample of


1. Determination of best frequency. Best frequency

was defined as the frequency of a tone burst that excited

the unit at the lowest intensity in either or both

ears. The excitation of the unit was determined

visually by observing the oscilloscope trace

of the unit's activity or by computing a PST histogram.

2. Repeated measures of spontaneous activity.

Spontaneous activity was sampled at least every 15

min. Increases in the rate of spontaneous firing were

indicative of cell injury and further study of such

cells was discontinued. Since on-line PST histograms

sampled 1 sec of activity following the onset of the

200-msec tone burst, the latter half of the histogram

also afforded an opportunity to detect changes in

spontaneous activity.

3. Monaural intensity functions at best fre-

quency. Measures of the number of discharges per

sample period versus stimulus intensity were ob-

tained in 5-dB increments from threshold to within

20 dB of maximum available sound pressure. The

duration of the sample period depended upon the

unit's discharge pattern: only discharges that

were judged stimulus-related were included in the

sample period. Hence, for units that discharged

only at the onset of a tone burst, the sample

period could be as brief as 50 msec, whereas

for units that discharged at a rate higher than

the spontaneous rate throughout the duration of

the tone burst, the sample period could be as

long as 250-300 msec. Sample period duration is

noted in the figure legends when pertinent.

4. Interaural phase functions for units

with best frequencies below 2500 Hz. Number

of discharges per sample period for 32 trials

versus interaural phase angle was examined

at intensity levels approximately 20 dB above

the unit's threshold. Additional intensity

levels were presented if the unit could be held

long enough to complete the basic protocol

5. Binaural intensity functions. These

functions were identical to those described in

step 3 except that both ears were stimulated

simultaneously. For phase-sensitive units

stimuli were presented at the interaural phase

angle that evoked maximal firing.

6. Interaural intensity difference func-

tions. The number of discharges per sample period

over 32 trials was counted as stimulus inten-

sity was held constant at the excitatory ear and

stimulus intensity at the other ear was varied

from -20 dB to +20 dB of the constant stimulus in

5 dB steps.

7. Interaural time difference functions for

units responsive to clicks. The number of dis-

charges per sample period over 32 trials was summed

as the time difference between clicks arriving

at each ear varied from 0 isec to 500 psec in

intervals of 0, 50, 100, 150, 200, 300, 400, and

500 psec. The ear receiving the leading click

alternated at each time-difference interval. The

usual intensity level was 70 dB peak equivalent

pressure. Additional intensities were presented

if time permitted.

The basic protocol required approximately 2 hr for low

frequency units and slightly less time for high frequency

units since the interaural phase functions were omitted for

the latter. Some units could be held for as long as 6 hr,

which allowed the protocol to be repeated for the same unit

with a different set of intensities or frequencies. For


most units, however, the basic protocol could not be com-

pleted in its entirety. In this report the sample includes

only those units for which steps 1 through 4 of the basic

protocol were completed.


General Considerations


Comparisons were made between cortical tissue from

areas of cortex where single units could be acoustically

driven (auditory cortex) and tissue from adjoining, acous-

tically unresponsive areas (non-auditory cortex). An ex-

ample of the cytoarchitectonic differences in the two areas

is presented in Fig. 1. Figure 1A is a frontal section

from non-auditory cortex located on the same frontal plane

and approximately 2 mm superior to the auditory cortex in

Fig. lB. There are readily apparent differences between

the two areas in both cortical depth and lamination. The

depth of non-auditory cortex was approximately 1150 p (un-

corrected for tissue shrinkage), as compared to 1550 V for

auditory cortex. The non-auditory cortex showed a distinct

laminar pattern characterized by the presence of large py-

ramidal cells in the deeper part of layer III. Cell size

was small in layer IV and the upper part of V, but increased

in layer VI. In contrast, the auditory cortex in Fig. 1B




a 4-1


u.4 o (1


0 4-
W 0 04

C -"

O i-l

4-) 4J -
C 0
0 CH
4-1 tf
40 CO

U4 0 *

p rl 4- a

C ( 41 Cc 0
4 0>

0 P r
Sco c
S 0 4* >

* 0 C
w cO

* uu

0O 41 0 V)
4 X
4 l- l CO
0 'l ca
C 4-4 'n

had a uniformpopulation of small cells and a poorly distin-

guished lamination for layers II and III and for layers IV -

VI. The term "koniocortex" has been applied to the type of

cortex typical of the primary receiving areas of sensory

systems. Insofar as koniocortex in cat and monkey is char-

acterized by the general smallness of cells, small variations

of cell size, and marked blurring of lamination (Walker,

1937; Rose, 1949), then the chinchilla auditory cortex could

qualify as koniocortex. Unlike the cat and monkey auditory

cortex, however, the chinchilla's lacked a well-defined,

densely populated layer VI.

It should be noted that even though Fig. 1B encompasses

less than one-half of the auditory cortex, the cytoarchitec-

ture was not uniform throughout. As the ventral border of

the auditory cortex (right side of photomicrograph) was

approached, an increasing pyramidalization of layer III

appeared along with a blurring of the border between layers

III and IV.

Although the Nissl-stained material showed some sug-

gestion of a columnar arrangement of cells, protargol-

stained sections revealed a distinct columnar array of

fibers in layer III (Fig. 2A, C). Fiber bundles could be

Fig. 2. Photomicrograph of frontal section of auditory
field. Protargol stain. A, Layers I through IV. Calibration
bar: 100 p. B D, Details of areas indicated by superimposed
corresponding letters in A. Calibration bar in B (applies to
C and D also): 20 p.

seen throughout layer III arranged in discrete, radially ori-

ented fascicles, whereas the fibers in layer IV were neither

in bundles nor oriented (Fig. 2D). It cannot be stated with

confidence that these fascicles all represented thalamic affer-

ents since the bulk of thalamic input presumably forms dense

unoriented plexuses in layer IV (Lorente de N6, 1949). Those

afferents, however, that do emerge from layer IV and ascend to

layer III tend to be vertically oriented (Chow and Leiman, 1970).

Other possible sources for the fascicles could be the non-

specific thalamic afferents or the descending axons of cells

from the upper layer. In any case, the presence of these

fascicles in protargol sections distinguished layer III

from layer II (lower part of Fig. 2B), whereas in the Nissl-

stained material these two layers appeared undifferentiated.

The Golgi-stained sections in Fig. 3 complemented the

Nissl-stained and the protargol material by demonstrating

the dendritic arborization in the cortex. A predominantly

vertical orientation of dendrites was evident in all cor-

tical layers (Fig. 3A). The high degree of vertical orien-

tation, however, should not be made to overshadow the ex-

tent to which dendritic elements ramified in the horizon-

tal direction. In Fig. 3A and Fig. 3B (lower part of layer

I to upper part of layer IV) the horizontal spread of some


I -



u 0
bo >1

0 0


u o

1 4



0 0

T-I .

( (0
0 0

0 0
a 0




.H co


dendrites ranged far beyond the 200- to 400- p limits of

the functionally defined columns of the somatosensory or

visual systems (Mountcastle, 1957; Hubel and Wiesel, 1962).

Also to be noted in Fig. 3B is the heterogeneity of

cell types in layers that have been traditionally con-

sidered to contain mainly small and medium size pyramidal

cells (Ram6n y Cajal, 1911; Lorente de N6, 1949). Even

when some of the small cells could be classified as pyram-

idal on the basis of the pyramidal shape of their soma,

their dendritic arborizations gave them the appearance of

stellate cells. In contrast, the pyramidal cells of deeper

layers had well-defined pyramidal cell characteristics; an

example of a medium pyramidal cell from layer V with a

soma diameter of 15 v is shown in Fig. 4A. Pyramidal cells

of this size constituted the most commonly seen cell type

in the lower cortical layers. Their apical dendrites,

which often reached layer I, were largely responsible for

the vertically oriented appearance of Golgi-stained cortex

in low-power photomicrographs (Fig. 3A). Another pyramidal

cell of similar soma size can be seen in the background of

Fig. 4A along with only a small part of its basilar den-

dritic field. The horizontal spread of basilar dendrites


I 0
C ,S
) .0

1-4 (3 )

0) .0

> 00 41
0ha 0H
y C(U HCf

A *H *'-I i-
-^,. tn dI+.
c^ r^r
SMT ~ C>0
Kl^^^ '* c
r^| M

f rq







V p

from pyramidal cells in the lower layers was usually even

more extensive than the horizontal arborization of dendrites

in the upper layers: fields of up to 500 p were not uncommon.

In Fig. 4B one of the glass microelectrodes used for

recording unit activity is reproduced at the same magnifi-

cation as the pyramidal cells in Fig. 4A.

Evoked Potentials and Tonotopic Organization

An extensive preliminary investigation was made to de-

termine the location and extent of the auditory cortex in the

barbiturate-anesthetized chinchilla using combinations of

gross evoked potentials and single-unit recordings. The first

series of experiments, using large metal electrodes (0.5-mm

tip diameter) placed on the dura,were not very successful in

localizing the auditory area since, due to volume conduction,

evoked potentials were recordable over almost the entire lat-

eral and part of the superior surface of the cortex. In the

next stage, smaller, insulated metal electrodes (0.1-mm tip-

diameter) were inserted through the dura to a depth of approx-

imately 1 mm. Auditory evoked potentials from these elec-

trodes were found over a much more limited area of cortex

than was found earlier. Moreover, the evoked potentials

were observed to be tuned to specific frequencies. Even

greater frequency specificity of evoked potentials was at-

tained with glass microelectrodes so they were used for all

subsequent mapping experiments. Barbiturate anesthesia

was eliminated as soon as on-line averaging capability was

acquired; the major difference between the evoked poten-

tials in the unanesthetized and the anesthetized prepara-

tions was the greater variability of responses in the former,

necessitating averaging techniques. No marked differences,

however, were found either in the extent or tonotopicity of

the auditory cortex in the two preparations. All data re-

ported below were obtained from unanesthetized preparations.

Figure 5 demonstrates the tuning that could be observed

when recording evoked potentials with a microelectrode at

a depth of 600 U. The amplitude of the evoked potential

can be seen to grow, reach a maximum, and then decline as

frequency was varied from 4.0 to 6.6 kHz. It should also

be noted that the increase and decrease in the size of the

potential was not symmetrical with respect to frequency.

As frequency increased, the increase in amplitude was grad-

ual until a maximum was reached, but then the amplitude

declined very rapidly with further increases in frequency.











I .

100 pV

58dB SPL



Fig. 5. Averaged evoked potentials recorded from micro-
electrode at 600 p depth in animal 3-12. Each average contains
25 responses. STIM represents the envelope of the tone burst
stimuli which varied in frequency but were of constant inten-
sity (58 dB SPL). Repetition rate: 1/1.5 sec; bandpass.
3-1000 Hz; positive polarity: up.

The frequency at which the maximum amplitude of the responses

occurred remained relatively constant over a wide range of in-


As Fig. 5 illustrates, the waveform of the evoked re-

responses recorded from within the cortex was typically nega-

tive-positive diphasic. When the electrode was on the cort-

ical surface, the waveform was positive-negative; the polar-

ity reversal could occur at depths varying from 100 to 1000

p. After the reversal the potential usually increased rap-

idly in amplitude. As the electrode was advanced beyond

1200 p, the size of the potential slowly decreased with po-

larity unchanged. (References below to the evoked potential

waveform denote the waveform observed after polarity rever-

sal.) Evoked potentials could still be detected by aver-

aging at depths of 2500 3000 p. The latency of the evoked

potential was a function of stimulus parameters and was in-

versely related to response amplitude. Consequently, min-

imal latency occurred at the evoked potential's "best fre-

quency" and at high intensities. Typical minimal latencies

(as measured from the onset of a tone burst with an 8-msec

rise time) were: 14 msec to the peak negative deflection

and 24 msec to the peak positive deflection.

In the early stages of this study considerable attention

was directed to the problem of correlating unit discharges

with the waveform of the evoked response recorded with the

same electrode. With anesthetized preparations most of the

units that were recorded discharged only once per trial at

the stimulus onset. These discharges tended to occur on the

negative slope of the first negative deflection of the evoked

response. Units with similar discharge characteristics were

also seen in the unanesthetized preparation and their dis-

charges also tended to occur during the initial negative de-

flection of the evoked response but irrespective of the

negative or positive slope (Fig. 6A).

With unanesthetized preparations,however, a number of

other response patterns were commonly seen as will be de-

cussed in a following section, and the pattern of their dis-

charges did not always bear a simple relationship to the

averaged evoked response. For example, Fig. 6B illustrates

the PST histogram of unit 1.2-13 which discharged through-

out the duration of the tone-burst stimulus and for 100 msec

after its termination. The evoked response was the typical

negative-positive diphasic response, yet there was no firing

during the negative phase of the evoked response.

Fig. 6. Examples of the correlation between the
averaged evoked response and the probability of unit
discharge as represented by PST histograms. The top
half of each figure is the averaged evoked response
(AER) recorded from the same electrode as the unit
discharges. AER bandpass: 5-250 Hz; calibration:
100 yv; positive polarity: up. Lower half of each
figure is the PST histogram for the same time period
as the AER. In D note that the negative polarity
portion of the AER is repeated below the PST histo-
gram. PST histogram bandpass: 200-1600 Hz; cali-
bration: 10 discharges. Stimulus duration and fre-
quency indicated by labeled bar above time scale.
All responses based on 25 stimulus presentations.
A, Unit 2.3-12. Stimulus intensity: C (contralateral)
70 dB SPL. B, Unit 1.2-13. Intensity: C 65 dB SPL.
C, Unit 1.2-19. Intensity: C 68 dB SPL. D, Unit
6.3-5. Intensity: I (ipsilateral) 70 dB SPL.

5.tAU kHz
0 20 40


0 100 200 300 400

60 80

3.57 kHz

0 100 200 300 400

5.80 kHz

0 200 400 600 800

The correlation between evoked potential waveform and

unit discharge rate was also weak even when only the posi-

tive phase of the evoked potential was considered. In the

preliminary experiments the positive phase appeared related

to discharge suppression but, with the low spontaneous rates

observed with anesthetized preparations the precise relation-

ship was difficult to discern. With the unanesthetized pre-

paration, however, units with moderate rates of spontaneous

activity were often seen. Unit 1.2-19 (Fig. 6C) is such a

unit and showed two periods of spontaneous-activity suppres-

sion: immediately at stimulus onset and immediately after

the initial burst of two or three discharges. Although the

post-burst suppression was well correlated with a positive

potential, the initial suppression occurred while the aver-

aged evoked potential was at baseline. Furthermore, unit

1.2-13 (Fig. 6B) demonstrates that the positive phase of

the waveform is not necessarily correlated with suppression

since in this case unit firing began about the same time

as the evoked potential's peak.

When there was a correlation between the evoked res-

ponse and unit discharge it was usually only seen for the

early part of the discharge pattern, approximately the

first 50 msec. Unit 6.3-5 (Fig. 6D) is an exception to this

observation. Its PST histogram shows an oscillatory sequence

of discharges that were correlated with the negative phases

of the evoked potential for as long as 600 msec after the on-

set of the tone burst. It should be emphasized, however,

that this was a very infrequent form of discharge that usually

occurred only at high intensities and for click stimuli.

The conclusion one might draw from the above examples

is that there was a tendency for discharges to occur during

the negative phase of the evoked response but that any strict

correlation between the waveform of the evoked response and

the probability of unit discharge would be an oversimplifi-

cation. While it is true that an electrode placed close to

the neuron soma should be able to monitor the net conduc-

tance changes across the membrane, it is known that elec-

trodes do not have to be in immediate contact with cells in

order to record their action potentials. Mountcastle et al.

(1969) considered the spatial extent of the electrical field

for a pyramidal cell discharge to range from 50 to 200 p.

Therefore, the fact that an electrode is recording action

potentials is no indication it is close enough to that cell

to detect the events underlying cell discharge, i.e, mem-

brane conductance changes.

Both evoked potentials and acoustically driven unit act-

ivity were used to determine the extent and the tonotopic ar-

rangement of the auditory cortex. Figure 7A shows the re-

lationship of the brain to the chinchilla skull, and Fig. 7B

the relationship of the auditory area, which is included with-

in the rectangle, to the rest of the brain and to some cranial

landmarks. Since only a small portion of the cortex was ex-

posed for unit recording and since the cortex is lissence-

phalic, reference to cranial landmarks was essential for iden-

tifying the auditory area. The most consistently helpful cra-

nial feature was the coronal suture which is depicted as a

diagonal, dashed line in Figs. 7B-F. Auditory evoked res-

ponses and auditory units could be found within an area ap-

proximately 5 mm by 4 mm on the lateral surface of the brain

beneath the coronal suture. Although some units were found

caudal to the coronal suture, the vast majority were located

rostral to the suture.

Based on the tuning of evoked responses and the best fre-

quencies of single units, a generalized schema of tonotopic

organization is presented in Fig. 7C. Despite the two sep-

arate frequency representations suggested by the figure, the

experimental data could just as well be accommodated to a

D /
A /
8-25 /
/ 12.0
/ 1.0-14.0 .8
BULLA .-'. / 10.0
/ 18.0
/1.4 6.0



B E /
i-7'~[ ORBIT /
S V 1-25 /
Sy / 1.4
S/ 2.6 3.1
I j-



3-8 /
GH / 8.0
I LOO 1.0/ 9.0
w> /

Fig. 7. A, Relationship of brain (dotted line) to chin-
chilla cranium. B, Relationship of auditory area (rectangle)
to brain. Dashed lines are relative positions of cranial
landmarks. C, Suggested schema of tonotopic organization
within auditory area showing progression of low to high fre-
quencies within the two possible representations. D-F,
Maps of unit best frequencies (kHz) for three different ani-,
mals, 8-25, 1-25, and 3-8. Rectangle represents same area
as C and dashed line represents coronal suture.

schema which has only one field with a complete frequency

representation plus another field with only low frequency

representation, since no physiological differences were

observed between responses in the two high frequency fields.

The schema in Fig. 7C is merely an attempt to account for

a progression of low, high, and low frequencies as the re-

cording site wasmoved caudally.

Figures 7D-F depict individual maps from three differ-

ent animals showing the distribution of best frequencies of

single units for various penetrations. It is apparent from

the individual maps that there was considerable variability

even for a single animal in the representation of best fre-

quencies. Instead of the orderly progression of best fre-

quencies that is seen in lower levels of the auditory sys-

tem, at the chinchilla cortex there was only a tendency for

a grouping of units with high, low, or intermediate best

frequencies. Data based solely on evoked potentials were

no less variable since unit best frequencies were usually

at the same frequency at which evoked potentials were maxi-

mal. If not, best frequencies were within an octave range

of the evoked potential's best frequency. No abrupt changes

in best frequency were observed either for cells encountered

along a given track or for adjacent penetrations in the same

cranial opening.

Briefly, the chinchilla auditory cortex shows at best a

weak tonotopic organization. Although one high frequency area

and two low frequency areas could be demonstrated, the transi-

tion between the areas was very gradual and within areas no

orderly arrangement of frequencies was evident.

Threshold Frequency Relations

Determination of the threshold SPL at a unit's best fre-

quency was made for 133 units. Figure 8 presents these data

in the form of a scatter plot in which each symbol represents

the threshold SPL at the best frequency of a given unit. The

three symbols indicate whether contralateral, ipsilateral, or

binaural stimuli produced the lowest SPL threshold. It can

be seen that contralateral stimulation provided the lowest

threshold SPL for 61% of the units, followed by binaural and

ipsilateral stimulation which occounted for 28% and 11% of

the units, respectively. The dotted line across the bottom

of the graph represents the SPL that is inferred to be at the

chinchilla's eardrum at the animal's behavioral threshold.

This measure is based on data from Table 2 of a study by


* *


o *

o .





* o **.. I-

'. /
* !
o *'










09OHS3dHI A IV idS 8P

I I I r t r



" 4
1 0
cn 0

a) U '

a) a

0 0



0 401

-,( --J ;
Sa a
I po 00
L^: w )o

*- -4
(u O4CO 0



I 0


LJ 4-J y

U r M4

Cl ) U) r.
1 Crl rl-

*v- ,- r

Miller (1970) in which the chinchilla audibility curve was

measured using an avoidance conditioning task and on data

from Fig. 8 (State I) of a study by von Bismark (1967) in

which the sound pressure transformation from free field to

tympanum was measured for the chinchilla. Parts of the

curve were based on linear interpolation since the two stud-

ies did not use indentical frequencies for their measure-

ments: accuracy is estimated to be + 2 dB.

Since noise in the test chamber could conceivably raise

the thresholds for single units, background noise was mea-

sured with a condenser microphone placed in the chamber at

the animal's head position. Measurements of the spectrum

level of the background noise were possible up to 400 Hz.

Above this frequency the noise level in the room was below

the sensitivity of the microphone. As can be seen from

Fig. 8, the thresholds for low-frequency units are well

above the spectrum level of the background noise. Even if

one assumes that a "critical band" mechanism is involved

in raising the low-frequency threshold, a plot of the spec-

trum level of the noise plus one critical ratio (from Fig.

6, Miller, 1970) would still fall below the observed low-

frequency unit thresholds. Although no explanation is

available for the relatively high thresholds for low-fre-

quency units,background noise can apparently be excluded as

a possible cause.

For frequencies above 1 kHz unit thresholds are closer

to the inferred behavioral thresholds: ten units had thresh-

olds within 10 dB of the inferred threshold and two units

had thresholds actually below the inferred behavioral thresh-

old. The overall distribution of the thresholds of units

with respect to the behavioral thresholds at that unit's

best frequency is replotted from data in Fig. 8 and pre-

sented in Fig. 9B. The distribution approximates a normal

distribution; the mode is the class of units 20 25 dB

above inferred threshold SPL.

The range of unit best frequencies recorded was almost

as extensive as the range of frequencies that have been used

to behaviorally test the chinchilla's range of hearing.

The lowest unit best frequency encountered was 200 Hz; the

chinchilla can hear as low as 90 Hz at a threshold SPL of

31 dB. The highest unit best frequency was 22 kHz; behav-

iorally, the highest frequency presented was 22.8 kHz at

a threshold SPL of 23.1 dB (Miller, 1970). The data from

the scatter plot of Fig. 8 is replotted once again in Fig. 9A




20N 133


z O

2 4 6 .8 I 2 4 6 8 10 12 15 20 >20





N =133


2 10 -

-5 0 5 0 15 20 25 30 35 40 45 50 55

Fig. 9. A, Histogram of distribution of unit
best frequencies. Shaded bars indicate units respon-
sive to both clicks and tone bursts. B, Histogram of
distribution of unit thresholds with respect to the
inferred threshold SPL.

to show the distribution of unit best frequencies. The histo-

gram shows that a wide range of frequencies was sampled and

that the mid-frequencies (1 to 10 kHz) were most commonly re-

corded. The shaded bars represent the 43 units that were re-

sponsive to clicks as well as tone bursts. The other 90 units

were not all tested for responsiveness to click stimuli so

the exact percentage of units responsive to clicks cannot be

determined. Of those units tested, however, 80% (43 out of

54 units) were responsive to clicks. The main purpose of

plotting the unit best frequencies of click-responsive units

in Fig. 9A was to demonstrate that their best frequencies

could be found across the frequency spectrum.

According to the experimental protocol, once the best

frequency of a unit was determined, that frequency was used

to complete the basic protocol. Since additional frequencies

were not employed until the basic protocol had been completed,

detailed study was not made of the effects of stimulus fre-

quency on firing rate. Data are available, however, from

23 units in which a range of frequencies was presented for

at least one suprathreshold intensity level.

The isointensity response areas plotted in Fig. 10 for

four units show how changes in intensity affected the frequency

selectivity of a unit. For unit 3.1-22 (Fig. 10A) it can

be seen that the sets of curves for contralateral and ipsi-

lateral stimulation are symmetrical and centered at the

unit's best frequency. Thus, as intensity increased, the

maximum number of discharges tended to consistently appear

at best frequency which, it should be emphasized, was de-

fined as a threshold measure. Similar results were observed

for the remainder of the units studied at various frequen-

cies. Unlike many units, however, unit 3.1-22 produced a

non-monotonic intensity function, i.e., as stimulus inten-

sity increased the number of discharges first increased,

then reached a plateau, and finally decreased. Although

the 26-dB isointensity curve was only 15 dB above the unit's

threshold for ipsilateral stimulation, it can be noted that

the 26-dB intensity level produced more discharges over most

of the frequency range than did the more intense stimuli.

Even though the contralateral function reached a plateau at

a higher intensity (36 dB) than the ipsilateral function,

it too showed non-monotonic behavior as the intensity was

raised beyond 36 dB.

Figure 10B demonstrates the effect of a 10-dB change

in intensity on frequency responsiveness for two different

120 3&-22

o 31
so 56



O . I I ,
L2 1.4 1.6 1.8 1.2

.-oC 50d8
-C 60dB



3 5

120 -

I 2 3

Fig. 10. Isointensity response areas for
four units. Each point in the graphs represents

the total number of discharges during the sample
period for 32 stimulus presentations of a given
frequency. Points of equal intensity are joined.
Best frequency (BF) is marked on the abscissa.
Level of spontaneous activity (SP) corresponds
to distance on ordinate. Sample periods: A,
0-300 msec; B, 0-250 msec; C, 0-59 msec.

5I 3 5 10

units. Although thresholds at best frequency were similar

for both units, 40 dB for unit 4.6-22 and 45 dB for unit

1.10-12, stimuli at 50 dB revealed two markedly different

response areas. For unit 4.6-22 the response range extended

from 5.5 to 14.0 kHz, while for unit 1.10-12 the range ex-

tended only from 5.0 to 5.3 kHz; a 10-dB increase in inten-

sity expanded the former's response area from 3.0 kHz to

well above 15.0 kHz and the latter's only from 4.6 to 5.6


The symmetrical arrangement of isointensity curves with

respect to best frequency was not seen for unit 4.6-22 (Fig.

10B). Instead, at 50 dB there appeared to be two response

ranges, one centered at best frequency, 11.5 kHz, and a

more effective one in terms of number of discharges at 6.0

kHz. A 10-dB intensity increase was more effective in in-

creasing the number of discharges at 6.0 kHz than at 11.5

kHz so that only one response range centered at 6.0 kHz re-

mained. Three other units demonstrated two response ranges

but only one showed a distinct shift of the center of its

isointensity curves from best frequency.

The most common change of isointensity curves with in-

creases in intensity is shown in Fig. 10C for unit 1.2-16.

As intensity increased, the isointensity curves expanded to

encompass a greater effective frequency range but remained

centered at best frequency.

The bandwidths of the response areas were determined

for five units in which the isointensity curves were at in-

tensities 15 dB above the unit's threshold and for nine units

at 20 dB above threshold. The upper and lower frequencies

were defined as those frequencies at which the discharge rate

was 50% above the spontaneous rate. Median bandwidth at 15

dB above threshold was .58 octave (range: .25 1.55 octaves);

at 20 dB above threshold median bandwidth was 1.25 octaves

(range: .35 2.50 octaves). Bandwidth as expressed in oc-

taves was not systematically related to the units' best fre-


Depth of Recorded Units

Figure 11 depicts the distribution of 95 units according

to their depth from the cortical surface as measured from the

microdrive system. The measurements of the other 38 units in

the sample were considered unreliable and not included. For

this group of units an accumulation of cerebral spinal fluid

(CSF) within the cranial opening prevented an accurate visual







0 *"





SIINn -O 10 389nN

determination of the electrode's distance from the cortical

surface at the beginning of a penetration. The median depth

for recorded units was 950 p. Although units were found from

depths of 100 to 2400 p, 63% of the sample was located between

600 and 1400 p, which would approximately correspond to layers

III, IV, and the upper part of V.

Spontaneous Activity

Most units in the present sample exhibited spontaneous

activity, defined as discharges in the absence of a controlled

acoustic stimulus. A histogram of the distribution of rates

of spontaneous activity is shown in Fig. 12. Sixteen units

had no spontaneous activity whatever; on the other hand, seven

units showed rates of discharge in excess of 20 discharges/

sec. The median discharge rate was 2.5 discharges/sec. Al-

though differences in the rates of spontaneous activity be-

tween different layers have been noted in the monkey somato-

sensory cortex (Whitsel et al., 1972), no marked differences

were found for the present sample with one exception: only

two of the 16 units with no spontaneous activity were found

in depths greater than 1000 p. A depth of 1000 p would cor-

respond approximately to the beginning of layer IV.


OA >


OD o

w m

w o
C) )

0 0


o e 0


S 0 0

-o0- to8N
O 0 Q)4

S I-N n j o H 3---- nN o q

Single-Unit Response Patterns

Recordings were obtained from 133 units for which steps

1 through 4 of the basic protocol (best frequency, sponta-

neous activity, monaural intensity functions, and interaural

phase functions for low frequency units) were completed. Of

the 133 units, binaural intensity functions (step 5) were

obtained for 105 units and both binaural intensity functions

and interaural intensity difference functions (steps 5 and

6) were obtained for 76 units. Interaural time differences

(step 7) were studied for 31 of 42 units that were responsive

to clicks.

The unit potentials had diphasic waveforms lasting ap-

proximately 1 msec; both positive and negative initial waves

were seen but the initially positive wave was more common.

The amplitude of the initially positive spikes ranged from

100 to 800 pv. Positive spikes of higher amplitude were

occasionally observed but these units eventually displayed

an injury discharge. "Thin spikes" (Mountcastle et al.,

1969) with an initially negative waveform of approximately

250-usec duration were noted on two occasions but the units

were lost within minutes.

The basic method of data reduction used in this study

was the computation of PST histograms. From the PST histo-

gram displayed on the oscilloscope of the computer the number

of discharges within a selected portion (sample period) of

the histogram along with descriptive statistics could be

printed out by teletype; the sample period chosen depended

on the unit's response pattern. The sample period was es-

tablished only after observing the response patterns for a

number of different stimulus conditions and then was fixed

for all stimulus conditions. For a unit such as unit 1.1-18

(Fig. 13A) which showed a sustained discharge, the sample

period would extend from the onset of the tone burst to a

time after the tone burst that would include all of the

stimulus-locked discharges. The particular sample period

for this unit was from 0 to 300 msec after stimulus onset.

For units such as unit 2.1-29 (Fig. 13B) where the distinctive

feature of the response pattern was a single discharge or burst

of discharges at the onset of the stimulus, the sample period

was brief (0 to 50 msec for unit 2.1-29) in order to count

only the onset burst. (Since the two peaks that followed

the onset peak in the PST histogram for unit 2.1-29 also

0 *
O *rf (

o m0 k .

0 4- 4J

O P. r 4 0
o m~*J a

0 0 4J 4 T
Ai3 4Jm 0

*C r-1
> (a ) r1 rPl

0 E *0 )

0 a4r 0p 0o

o 0
*H H) i o I

O r4 0

0 & -. 0 4
O & aT|

O r r

O c ,U c

s4 4J CO
t" TO 0 H

0 m cc H-
o T P

o p* 0 0
o rIl O ,
0 (M *

-H 1 40 -o

*4 4J W 0 0
00 43. H n. 0 (U
0 4J 040 .0
_ o,;4
0 ~ a r-1 (0 (0il


represented stimulus-related discharges, an additional sample

period was used in the off-line analysis of this unit for

comparison with the on-burst sample period.)

One of the distinctive characteristics of single-unit

activity in the unanesthetized preparation was the variety

of response patterns observed from different units. Response

patterns were divided into seven classes and are presented

in Fig. 14 together with the number of units in each class.

The sample consists of the 105 units for which binaural

intensity functions were available. A unit was assigned

to one of seven response pattern categories depending on that

unit's response pattern to a stimulus of best frequency

at approximately 20 dB above the unit's threshold for the

mode of stimulation (monaural or binaural) that gave the

strongest response.

Phasic response patterns were characterized by a single

discharge or a burst of discharges that most commonly occur-

red at the onset of the tone burst (31/105 units); the la-

tency of the onset discharge ranged from 10 to 20 msec and

decreased with increases of stimulus intensity. When mod-

erate spontaneous activity was present, a period of suppres-

sion which was intensity-related followed the initial dis-

charge. A phasic response only at the offset of the stimulus











Fig. 14. Characteristic response
stimuli for 105 units.

patterns to tonal





was seen less often (6/105 units); the latency of the

response was quite variable with some responses occurring

at late as 50 60 msec after tone offset. Two units

showed phasic responses at both onset and offset of the

tone burst. The late phasic response pattern (5/105 units)

consisted of one or two discharges to a stimulus presenta-

tion but unlike the phasic onset pattern, the discharges

occurred with considerable variability at long latencies

from'stimulus onset. In the phasic-tonic pattern (30/105

units), a highly time-locked initial burst was followed

by a sustained discharge. A short period of suppression

occurred between the onset burst and the sustained discharge

for some units.

The tonic pattern was only rarely seen in anesthetized

preparations, but was frequently observed with the unanesthe-

tized preparation in this study. Its characteristics were:

a gradual onset of the sustained discharge, a slight decrease

in the number of discharges throughout the duration of the

stimulus, and continued discharges up to 100 msec after

stimulus offset.

The suppressed response pattern (3/105 units) appeared

to be duration- and intensity-dependent. Stimuli of at

least a 50-msec duration were necessary to show suppression

and at high intensities the response patterns of two units

changed to phasic-offset patterns.

It should be pointed out that the response pattern

which was characteristic for a given unit was not necessar-

ily invariant over all stimulus conditions. Changes in

response pattern were commonly seen when frequencies other

than best frequency were employed. In such cases, a tonic

or phasic-tonic response pattern would often shift to a

phasic-onset type of response.

Changes were not as pronounced, however, for the stimu-

lus manipulations in the basic protocol: monaural and bin-

aural stimulation, overall intensity, interaural intensity

and time. Examples of response patterns of four different

units over a range of intensities are shown in Fig. 15.

Units 4.3-29 (Fig. 15A), 3.3-15 (Fig. 15B), and 3.10-5 (Fig.

15D) revealed no substantial change in their response pattern

as stimulus intensity increased, only the total number of dis-

charges within the sample period changed. Unit 9.8-25 (Fig.

15C) did show a change in response pattern that is typical of

phasic-tonic units. Within 10-15 dB of threshold these units

showed a sustained discharge typical of a tonic response pattern.

Fig. 15. PST histograms of four units showing ef-
fects of stimulus intensity on response patterns. Stimu-
lus frequency and duration indicated by line above time
scale. Stimulus intensity indicated in dB SPL in left
columns. N, total number of discharges per 32 stimulus
presentations as indicated in right hand columns. Sample
period is denoted below N. SPON, spontaneous activity;
I, ipsilateral: C, contralateral. A, Unit 4.3-29. B,
Unit 3.3-15. C, Unit 9.8-25. D, Unit 3.10-5.



I 17d8





C 27dB







20-240 msec






400 600

20-270 msec









C 12 dB








C 40dB







1.38 kHz

0 400 800

10-140 msec







1.60 kHz

0 200 400
5-30 msec







3.50 kHz

0 40 80


i.f 1 anz

0 200
0 200

At 20 dB above threshold, however, a phasic onset component

was added that remained the same size even with further in-

tensity increases. Since the sum of discharges over a given

sample period is not necessarily sensitive to changes in re-

sponse pattern, careful monitoring of all PST histograms was

maintained and whatever response pattern changes did occur

were noted in the experimental protocol. Repeated records

were obtained for identical stimulus conditions at various

time intervals to check response stability. Recordings were

terminated for those units showing unstable response activity.

Analyses of variance were performed to determine if the

response patterns were related to cortical depth or to spon-

taneous activity. For this purpose, the units showing on,

off, on-off, and late phasic responses were pooled; the units

with sustained suppression were omitted from the analysis,

leaving three categories: phasic (44 units), phasic-tonic

(30 units), and tonic (28 units).

The mean depths for the three groups showed no signifi-

cant differences. For spontaneous activity, however, a Newman-

Keuls test (Kirk, 1968) showed that the mean rate of sponta-

neous activity for phasic units (1.55 discharges/sec) was

significantly different (P<.01) from that of both phasic-

tonic (7.58 discharges/sec) and tonic units (6.69 discharges/


Binaural Representation

The purpose of this section is twofold: first, to

examine the relative influences of contralateral and ipsi-

lateral stimulation upon unit discharge rate, and,second,

to determine how the unit's discharge rate is affected by

binaural stimulation. The results described in this section

show how the two ears are represented in each hemisphere

through an analysis of that hemisphere's single-unit

responses. An ear was said to be represented if a unit

responded to stimulation of that ear, or if the responses

to stimulation of the other ear were increased or decreased

by stimulation of that ear. Response was defined as any

stimulus-related discharge above the level of spontaneous

activity; the strength of a response referred to the amount

of stimulus-evoked activity as reflected in the total number

of discharges per sample period for a given stimulus condition.

(The three units which showed a characteristic response

pattern of suppression were excluded by this definition

and consequently were analyzed separately.)

The principal measure used to summarize a unit's re-

sponse was the intensity function: a plot of the total

number of discharges over the sample period for 32 to 25

stimulus presentations (hereafter referred to simply as total

number of discharges) as a function of the intensity of the

stimulus. After the unit's best frequency was found, the

contralateral ear and then the ipsilateral ear was stimulated

with a series of 32 or 25 tone bursts of best frequency

starting at threshold and increasing in 5-dB steps to an

intensity of 70-80 dB SPL. Binaural stimulation at the

interaural phase angle that'produced the maximal discharge

rate followed monaural stimulation. Figure 16 illustrates

for unit 3.2-16 how intensity functions were derived.

It can be seen from the PST histograms of Fig. 16

that,with an increase in intensity, this unit's response

pattern remained relatively constant while the total number

of discharges (indicated in right hand columms) increased.

For the graph in Fig. 16, the total number of discharges

was divided by the number of stimulus presentations (32) and

plotted along the abscissa as the mean number of discharges

per trial. The bars at selected points of the intensity

function signify one standard deviation above and below

the mean number of discharges.

The response pattern of the unit in Fig. 16, unit 3.2-16,

was classified as phasic-tonic because of its characteristic

single initial burst of discharges, pause in firing, and

Fig. 16. Intensity function and selection of PST
histograms for unit 3.2-16. Stimulus intensity and total
number of discharges per sample period appear to the left
and right, respectively, of corresponding PST histogram.
Sample period: 0 300 msec; stimulus presentations:
32; stimulus frequency: 950 Hz. Duration of stimulus
(200 msec) indicated by bar above time scale. Binaural
stimuli at 00 phase angle. Vertical bars on intensity
function indicate + 1 standard deviation. SP in this
and subsequent figures indicates level of spontaneous



164 26d8

255 41

322 51





75 31

245 41

311 51

291 61

333 7

I I I i I I I
0 400 800 0 400 800
msec msec

0 400 800

20 30 40 50 60 70


sustained period of discharge that lasted up to 100 msec after

stimulus offset. Since stimulus duration was 200 msec, the

sample period extended from 0 to 300 msec after stimulus on-


Although a wide variety of intensity functions were ob-

served for the present sample of units, there were three gen-

eral properties that, aside from statistical variability,

were common to most intensity functions. The intensity func-

tion of unit 3.2-16 in Fig. 16 will be used to illustrate

these properties.

First, considerable variability was evident in numbers

of discharges from trial to trial. For example, in Fig. 16,

even though the mean number of discharges for this cell

grew with intensity for ipsilateral and for contralateral

stimulation, the variability was so great that the mean

number of discharges were within + 1 standard deviation of

each other over the range of intensities from 35 to 65

dB SPL. Despite the steepness of the binarual intensity

function, the mean number of discharges from 45 to 70

dB SPL were also within + 1 standard deviation.

The second feature to be noted is the general shape

of the intensity function for the strongest mode of

stimulation (ipsilateral, contralateral, or binaural).

The intensity function for the strongest mode of stimulation

was usually linear for the first 20 to 30 dB above threshold

and then reached a plateau as intensity was increased. For

approximately half of the units still further increases in

intensity decreased the total number of discharges (nonmono-

tonic intensity functions). (It is possible that the other

units would have also had nonmonotonic intensity functions

if intensity were increased beyond the standard 70-80 dB

SPL upper limit.) Although the unit in Fig. 16 did not

show nonmonotonic behavior it did show a linear rise in

its intensity function for binaural stimulation over the

20-dB range from threshold to 45 dB SPL, followed by a

plateau at higher intensities.

A third general property is that the curves for the

three modes of stimulation did not usually intersect except

in the case where two or three curves actually overlapped.

In other words, across intensity levels, the ordering of

response strength for the three modes of stimulation tended

to remain the same. For the unit illustrated in Fig. 16,

binaural stimulation produced the strongest response across

all tested intensities, while the curves for contralateral

and ipsilateral stimulation overlapped.

Complete monaural and binaural intensity functions

were obtained for 105 units; the intensity functions were

then categorized on the basis of the relative positions

of the curves for monaural and binaural modes of stimula-

tion. This analysis resulted in the twelve groups of

intensity functions depicted in Fig. 17. Four functions

could not be classified including those from the three

units with suppression response patterns.

The next stage of analysis was to identify common

features within the twelve groups of intensity functions.

It was found that the twelve groups could be assigned to

one of the five classes listed in Table 1 according to

their responses to monaural stimulation.

For classes B and D, which responded to stimulation

of only one ear, it should be repeated that only increases

in the number of discharges were defined as responses.

For some of the B and D class units, suppression of

spontaneous activity occurred when the "non-responding"

ear was stimulated.

Fig. 17. Classification of intensity func-
tions for 105 units according to monaural response
class and binaural response category. See text for
classification criteria. Abbreviations B, C, and I
for the intensity functions refer to binaural, con-
tralatera], and ipsilateral stimulation, respec-



Effective Mode
of Stimulation

Monaural only

Contralateral only

Ipsilateral only

Contralateral and Ipsilateral

Contralateral response

Ipsilateral response

Contralateral =
Ipsilateral response


Number of Units