Hearing Research, 18 (1985) 159-168 Elsevier

159

HRR 00603

Characteristics of tone-pip response patterns in relationship to spontaneous rate in cat auditory nerve fibers

William S. Rhode and Philip H. Smith Department of Neurophysioiogy, University of Wisconsin, 1300 University Avenue, Madison, WI 53706, U.S.A.

(Received 17 December 1984; accepted 24 April 1985)

The responses of single auditory nerve (AN) fibers in the cat were recorded in response to 25 ms tone pips. Peristimulus time histograms (PSTH) of discharge patterns recorded from fibers with high spontaneous rates (high SRs), show that the discharge rate rapidly adapts to a much lower steady-state level over a 15 ms period with shorter times for units with best frequencies (CFs) greater than S kHz. The PSTHs of auditory nerve fibers with low SRs do not show this pattern of rapid adaptation. Differences between the high and low SR populations indude higher thresholds, better tuning, and longer latency in the low SR population. The peak-to-steady-state discharge ratio is an increasing function of SR and CF; it varies from 1.0 for fibers with SR = 0 to over 8 for fibers with high SRs and CFs near 10 kHz. This ratio increases with increasing stimulus intensity and stimulus recovery time. The high SR population shows a number of responses to transients which are weak or absent in the low SR population. Increasing the recovery time shortened the latency of both high and low SR AN fibers by as much as 1 ms. A number of other response properties of AN fibers are also reported that are important when interpreting the responses of cochlear nucleus neurons to tone pips.

auditory nerve, spontaneous rate, adaptation

Introduction

While recording from units in cat cochlear nuclei, in a previous study [lO,ll], we observed two distinct temporal patterns in response to brief tones (25 ms duration). One pattern corresponded to the ‘classical’ peristimulus time histogram (PSTH) seen in auditory nerve fibers [4]. In this pattern, the neural activity rises quickly to a peak and is followed by a rapid adaptation, with a time constant of l-15 ms, then a subsequent short term adaptation with a time constant which may range from 15 ms to times beyond the duration of our stimulus (30-60 ms [3,15]). This pattern has been called primary-like [8]. A second pattern, which we designate PL,, was characterized by a slower, more gradual rise with little or no initial peaking.

Since no systematic study of the response pat- tern of AN to short tone pips had been reported, we sought to examine such patterns to determine whether the PL, response should be considered a response of a particular cell type of the cochlear nucleus or whether it occurs in the AN.

Liberman [5] demonstrated differences in the characteristics of AN fibers which correlated with the SR of AN fibers, Increases in both threshold and sharpness of tuning with a decrease in SR led to a division of AN fibers into three populations, SR < 0.5, 0.5 < SR > 18 and SR > 18.

In this study, the PL and PL, response patterns are shown to correspond to AN fibers with high and low SRs, respectively.

Methods

Three healthy young cats (2-4 kg) with clean external and middle ears were used. The surgical approach and procedures were similar to those employed in our study of the cochlear nucleus [lO,ll]. Each cat was anesthetized with sodium pentobarbital and tracheotomized. The bulla was intubated with a 20 cm, 0.5 mm polyethylene tube to equalize the static pressure across the drum [17]. A small portion of the cerebellum was aspirated and the eighth nerve was exposed by retracting the cerebellum and packing small pieces of cotton

0378-5955/85/$03.30 Q 1985 Elsevier Science Publishers B.V. (Biomedical Division)

160

immediately anterior and posterior to the internal auditory meatus. A recording chamber was at- tached to the skull and filled with warm mineral oil. The impedance of our glass recording micro-

pipettes, filled with 3 M KCl, was between 20 and

40 M&Z. Stimuli were 25 ms tone pips, with trapezoidally

shaped onsets and offsets of 2 ms duration. Rise/ fall times of 4 and 8 ms were occasionally used.

The search stimulus was a continuously swept tone. Response area (RA) data was collected using 5 or 10 repetitions of a 25 ms tone pip repeated

every 105 ms. Intensity was usually varied from 0 to 90 dB, in 10 dB steps. Frequency was varied linearly, with roughly 20 steps per octave for units with characteristic frequencies (CFs) greater than

2 kHz and 10 steps per octave for units with CFs below 2 kHz. Threshold, Q,,) (= CF/BW where

BW = bandwidth at 10 dB above threshold), spontaneous rate (SR), maximum discharge rate.

dynamic range and CF were all determined from the RA. SR was determined by averaging the spontaneous rates measured with attenuators set to 132 dB prior to the presentation of each

415 iir 9D d6

0.2 sp/5

I I 84003-12

60 dH

36 sp/s

84003-24

frequency in the response area. Series of 250 short tones at CF (STCF), each consisting of 25 ms pips

repeated every 105 ms, were used to determine latency and PSTH patterns and to compute inter- spike interval patterns. Latency was determined by

the average of the first spike arrival times during the stimulus period. A result of this approach is that units with low discharge rates will show longer

latencies especially if they have no onset transient. Using a 10% rise in the PSTH envelope as a criterion will result in shorter latencies but does not accurately reflect the stochastic nature of the individual unit. The first four central moments of the interspike interval distribution, the mean, SD., skew and kurtosis, were also determined.

Results

PL and PL,$ responses

The peristimulus time histogram shapes, PL and PL,, which we observed for cochlear nucleus (CN) units were also observed for AN fibers. PL histograms are associated with AN fibers of high SR (SR > I7 sp/s). Low SR fibers (SR < 2 sp/s)

:/

i;oao Yr

80 dB

43 sp/s

11 84032-124

i2SO Viz i 1!500 Hz 90 dB : 80 dB

0.4 5p/s cl. 4 sp/s 84003-13 i 84032-88

--, 0 10 20 30 40 50 10 20 30 40 SC 0 IO 20 30 =Ic 5:

MILLISEC UILLISEC MILLISEi

Fig. 1. PSTHs for 25 ms tone pips at CF (250 repetitions). Top row: 3 units with high and one (first panel) with medium spontaneous

discharge rates. Bottom row: units with low spontaneous discharge rates. The CF, intensity (in dB SPL), spontaneous rate (in

spikes/s) and unit ID are all indicated in each panel. Threshold for 83003-14 = 30 dB; 84003-12 = 40 dB; 84003-24 =15 dB;

84003-13 = 45 dB; 84032-124 = 10 dB; 84032-88 = 1.5 dB; 83032-99 = 10 dB; 84032-96 = 30 dB. The same vertical scale is used in all

PSTHs to facilitate comparison. Bin width = 100 ps.

161

generate PST histograms similar to PL, units. In- termediate SR units (2 < SR < 16 sp/s) can show PL,-like histograms at low stimulus intensities and PL-like responses at high stimulus intensities. The top row of Fig. 1 illustrates the STCF response pattern of three AN fibers with high SRs and one with a medium SR. Stimuli were presented 30-70 dB above the unit’s discharge rate threshold (see figure legend). The lower row of Fig. 1 illustrates the responses of four AN fibers with low SRs. The CFs of the units in each column of Fig. 1 are comparable and cover a range of frequencies from 475 Hz to 27 kHz. Note that the vertical scales of the PSTHs are the same, to facilitate comparisons between units with different SRs and CFs.

For high SR fibers, the ratio of the tallest bin (100 ,W bin widths) to the average bin height, between 15 and 25 ms, is usually greater than 5. This ratio is usually between 1 and 2 for low SR fibers. The ratio increases with CF. For high SR units with CFs greater than 4 kHz, it is between 4 and 10 and between 1 and 4 for low SR units. For units with CFs less than 1 kHz, the ratio is be- tween 2 and 4 for high SR units, and between 1 and 2 for low SR units. The lower ratio at low frequencies may be due to an interaction between phase-locking and rapid adaptation. The mecha- nism responsible for the initial high discharge rate of high SR fibers at stimulus onset appears to be weak in AN fibers with low SR.

Besides the early rapid onset peak seen in PL units, which may last less than 0.5 ms (see Fig. lA, third and fourth panels), these high SR fibers display a rapid adaptation of firing [14], lasting between 1 and 15 ms. Units with higher CFs have smaller time constants (< 5 ms, see Fig. 1) of rapid adaptation. The short term adaptation time increases with decreasing CF and can easily be greater than 25 ms for units with CFs less than 500 Hz. Since we typically used 25 ms tone pips the time constants of short-term adaptation for these units are inestimable.

Another aspect of AN responses to short tones, which is important when attempting to interpret the responses of the target units in the CN, is the difference seen in the response of PL and PL, units to changes in stimulus intensity and frequency. The STCF responses of representative PL and PL, units over a 60 dB SPL range are

shown in Fig. 2. At low intensities, the latency of the PL unit (lower panel in column A) increases and the peak rate is lower than observed at higher stimulus levels. This somewhat flattened response only occurs at dB levels close to unit threshold and the initial peak in the PSTH is prominent by the time the stimulus level is 30 dB above threshold (40 dB SPL, column A in Fig. 2). At 60 dB SPL, the initial peak is 0.5 ms wide and rapid adapta- tion is completed 15 ms after stimulus onset. The principal change at 80 dB, for this unit, is a period

A

80 d6

401 I 60 dB

40 dB

20 dB

6

80 dB

60 dB

i 40 dB

Fig. 2. The effect of intensity on the PSTH shape. Left cohmn: high SR unit (84032-41, 28.4 sp,‘s), CF= 2500 Hz, threshold = 3 dB. Right column: low SR unit (84032-20, 0.6 sp/s), CF = 8 100 Hz, threshold = 11 dB.

162

of sustained firing that lasts for 10 ms after stimu- lus cessation. This afterdischarge is not typical and will be described in the next section. In contrast, the PSTH shape of the low SR unit (column B. Fig. 2) shows little change as a function of increas- ing intensity over a 60 dB range. Just above threshold the histogram is flat and shows no evi- dence of adaptation. At 80 dB SPL, the PSTH remains nearly flat. The PSTHs of units with SRs between 2 and 16 sp/s often look similar to the PSTHs of low SR units, exhibiting PL, patterns at low stimulus intensities, but are nearly indis- tinguishable from the PSTHs of high SR units at high intensities. The medium SR unit conversion from a PL, to PL pattern does not occur as close lo the unit threshold as the increase in the onset peak of the high SR PSTH, as seen in column A of Fig. 2.

The shape of the PSTH does not change appre- ciably as the stimulus frequency is varied throughout most of the unit’s response area. How- ever, for stimulus frequencies near either edge of the unit’s response region, AN fibers often re- spond to the stimulus onset and offset with a burst of activity. A RA for a high SR unit is shown in the top panel of Fig. 3A along with its response to a frequency at the upper limit of the unit’s re- sponse region (lower panels}. The unit’s response ‘rings’ for 5 ms at both stimulus onset and offset. 4s the rise/fall time is increased from 2 to 8 ms, it becomes apparent that this ‘ringing’ is a response to the two corners of the linear ramp. This ringing response is quite robust. It often extends for several kHz beyond the high frequency edge of the re- sponse area. It is also observed in the low frequency portion of a response area, particularly at frequen- cies below 200 Hz. The ringing occurs even when the stimulus frequency is as low as 20 Hz. Re- sponse areas generated using long duration tones may mask these responses at their edges. This is not due to the absence of the response, but results from the fact that the on and off ringing adds only a fixed number of spikes, regardless of the tone duration. Thus, the proportion of the response attributable to on and off effects must decrease as tone duration is increased, and may be lost in the baseline jitter of the RA. Edge effects are observed

to a lesser degree in the responses of low SR units, often confined primarily to the onset of the stimu- lus. A RA for a low SR unit is shown in Fig. 3B (upper panel). The unit displayed a PL, PSTH at CF (Fig. 3B, middle panel) and an on response at the high frequency edge of the response area (Fig. 3B, lower panel). The low SR unit edge response does not persist over as wide a frequency range as observed for high SR units. The edge effect is strongest for units where CFs are lower than 5 kHz. High SR units do not usually exhibit this effect when the CF is greater than 10 kHz.

Although the majority of AN units cease firing 3-6 ms after stimulus offset (e.g. Figs. 1, 2, 5). some units with CFs below 4 kHz may exhibit up to 20 ms of post-stimulus ringing to tones at CF. In low frequency units, ringing is characterized by phaselocked discharges and is most often found in units with high SR (Fig. 4; see also Fig. 2A, upper panel). These responses are similar to those re- ported by Pfeiffer and Kim [9] who found that a small proportion of AN fibers (about 5%) rang for long periods of time in response to ciicks. This phenomenon does not occur in all cats and is more frequently observed in cats with low unit thresholds. The ringing is most prominent in units with CFs between 1 and 3 kHz, the same region where the on/off transients are most prominent.

Another transient phenomenon is observed in AN fibers with CFs above 7 kHz. There can be a synchronization of the first 2 or 3 spikes at stimu- lus onset (row A, Fig. 1). It appears that the unit is responding maximally and that the synchronous behavior merely reflects the constraints on firing rate imposed by the unit’s refractory properties. This phenomenon has been reported by Smith and Brachman [ 151.

Recovery time A number of the response parameters of AN

fibers are affected by the time allowed for recovery after prior stimulation [3]. Understanding the re- covery of AN fibers as a function of interstimulus interval is important when attempting to explain CN unit responses since they too are affected by recovery time. In our experiments, 25 ms tone pips are typically presented every 105 ms. The brevity of stimulus parameters is dictated by the desire to maximize the amount of data collected during

A 84032-41

163

El 83230-29

450 _

360 I

0 0 900 1800 2700 3600 4500

FREQUENCY WzI.

90

72

72 -

ff54 -

SiI 836 -

R/F 8 ms 80 dB

3900 Hz

320

0 2 0

FREQUENCY (Hz1

32

R/F 2 40

ms 80 d6

3900 Hz 32

E24

5 El6

R/F 2 ms 70 dB

4600 Hz

8

R/F 2 ms 90 d8

6000 Hz

Fig. 3. (A) Response area for a high SR unit (84032-41). Arrow indicates frequency at which PSTHs in lower panels taken. Lower two panels: PSTHs taken at the arrow indicated in the RA above (3900 Hz, 80 dB) where the rise/fall time of the signal was set at 2 and 8 ms as indicated. (B) Response area for a low SR unit (83230-29). Arrow indicates frequency at which PSTH in lower panel taken Middle panel: PL, PSTH taken at CF. Lower panel: transient on PSTH seen at the frequency indicated by the arrow in the upper panel (6000 Hz),

164

Fig. 4. (A) STCF response for a unit where the response rings after stimulus offset. Unit 84032-41, CF= 2500 Hz, Th = 3 dB SPL. SR = 28.4 sp/s. (B) STCF response for a unit where the response rings after stimulus offset. Unit 84032-50, CF = 1800 Hz. Th = 10 dB SPL, SR = 88.5 sp/s.

relatively short unit-holding periods encountered when using ultrafine micropipettes for recording. In order to study AN recovery, the stimulus duty cycle (stimulus duration time/repetition interval time x 100%) was varied between 25% and 5%. As duty cycle is decreased, the time available for AN recovery is increased from 75 ms (25%) to 225 ms (10%) to 475 ms (5%). The initial peak of the PSTH of either the PL or PL, units becomes larger as the duty cycle is decreased. This effect is much less pronounced in low SR (PL,) units where the

84032. YP

.Isjr: 2’.

Fig. 5. (A) PSTHs for STCF responses in which the stimulus duty cycle was decreased from 25% to 5% as indicated, with stimulus duration held constant at 25 ms. Unit 8403244, SR = 6.1 sp/s. (B) PSTHs, for low SR units, where the duty cycle was varied as indicated. Unit 84032-98, CF= 22000 Hz, SR = 1.4 spfs.

initial peak may not increase appreciably even with recovery times of 1 second. Usually there was only a small change in PSTH shape as recovery time increased from 225 to 475 ms, regardless of the unit’s SR. With this peak increase, the latency of the responses decreases (Fig. 5). It was common to record a decrease of 1 ms or more in latency irrespective of the SR of the AN fiber. Some low SR units, however, showed little change as re- covery time was lengthened. Thus, in making latency comparisons between AN and CN units. similar stimulus parameters should be employed.

Dynamic range The m~hanism by which the auditory system

encodes stimulus level is still an open question. We hear over a 120-130 dB range, yet, individual AN fibers vary in response over a 20-40 dB range (e.g. [l]). We have noted some variability in the rate curves for AN fibers. While some units display the ‘classic’ AN rate curve shape, characterized by a 20 dB dynamic range followed by a flat saturation region, there are also a number of AN fibers which continue to show an increase in output rate as intensity increases beyond dB above threshold. This effect is strongest in AN fibers with low SR. The gradual increase with intensity can be due to a change in either the early rapid adaptation phase or the response between 10 ms and stimulus offset. Occasionally, the latter activity reaches a plateau or even decreases as intensity is increased. Others have reported similar results [12].

Other variables A natural division in the distribution of sponta-

neous rates occurs at about 16 sp/s and a large discontinuity at 2 sp/s suggests a second dividing point. Grouping the population into these three categories produces an orderly variation in threshold, spontaneous rate, &, latency and steady-state discharge rate (Table I). To reduce the effect of frequency dependent changes in these parameters, units with CFs above and below 1000 Hz were tabled separately. The thresholds for low SR units are 10 dB higher than the high SR units. We have also noted that, in a single cat, units from a given spontaneous rate population, with similar CFs, show a narrow range of thresholds (20 dB).

Auditory nerve fiber latency is another im-

165

TABLE I

AN RESPONSE CHARACTERISTICS

AN fibers were divided into 6 groups based on CF and spontaneous rate (SR). TH = average threshold in dB SPL. SR = average spontaneous rate. MR = maximum discharge rate. Q,, = CF/BW where BW is the bandwidth measured 10 dB above TH. LAT = latency in ms. PTS = peak to steady-state discharge rate. SK = skew of the interspike-interval distribution (3rd central moment), a measure of the asymmetry: positive - right, negative - left, 0 - symmetric. KU = kurtosis (4th central moment of the interspike interval histogram), a measure of the peakedness of the dis~bution; > 0 more peaked than a normal dist~bution, < 0 less

peaked. n = number of fibers.

CF<lOOO CF>lOOO

SR<2 2<SR<16 SR>16 SR<2 2<SR<16 SR>16

TH 49 40 36 22 21 13

SR 0.5 5.7 48.4 0.5 5.1 46.6

MR 213 280 229 276 290 247

QIO 1.7 1.4 1.4 6.2 4.3 4.9

LAT 7.2 4.7 3.7 3.4 3.1 2.3

PTS 1.4 2.5 4.0 1.9 2.4 5.7

SK 1.1 1.3 1.8 1.5 1.5 1.7

KU 1.4 2.5 3.9 2.9 2.8 3.6

n 6 4 25 58 15 ’ 89

portant parameter to consider when attempting to understand CN unit response characteristics. The latency scatter diagram for the three populations is shown in Fig. 6. As expected, the low SR units have the longest latencies and the highest variabil- ity and spread. The latencies of the high SR units group quite tightly and are the shortest, achieving about 1.45 ms at high CFs. These latencies are

x

x Y

1 I

1000 10000

FREOUENCY (Hz)

Fig. 6. Scatter diagram of the latencies of the three groups of AN fibers plotted vs. log CF. *, units with SR -z 2 sp/s; A, units with 2 < SR e 16 sp/s; q , units with SR > 16 sp/s.

somewhat longer than the value obtained by Kiang et al. [4] using click stimuli. This may, in part, be accounted for by differences in the sound delivery tube length as well as the means of latency calcula- tion (see Methods). The properties of mid-range SR units tend to fall between the other two groups in threshold, latency, and Q,, (Figs. 6, 7 and Table I).

0 / I I

100 1000 1ooon

FREQUENCY (Hz1

Fig. 7. Scatter diagram of the Q,, of the three groups of AN fibers plotted vs. log CF. Symbols same as in Fig. 6.

Discussion

Our goal was to delineate properties of AN fibers in response to short tones for purposes of

comparison with those observed for units of the cochlear nucleus. Because fine tipped, high imped-

ance micropipettes are capable of recording from both axons and from cell bodies, it is important to

differentiate between first and second order re- sponses in the CN. Certain response characteris- tics such as high SR, short latency and monopolar spike waveforms are typically associated with AN

fiber responses in the CN. Our CN data demon- strated the existence of a PL, PSTH pattern, that is characterized by a gradual rise with little or no

peaking at stimulus onset. The PL, response pat- tern is also present in the output of AN fibers. The fact that there are at least two distinct AN popula- tions, distinguished by SR, latency, and PSTH

shapes, makes the task of interpreting coding in the CN more difficult. AN fibers with mid-range SRs have mid-range values for most properties.

Since they represent the smallest group (10%) we will not discuss them further.

A number of transient AN responses are sensi- tive to stimulus frequency, intensity and character- istic frequency. The time constant of rapid adapta-

tion is CF dependent, varying from less than 0.5 to a few milliseconds for low frequency units. A similar result has been reported in the gerbil by

Westerman and Smith [18]. They also reported a short-term time constant on the order of 60 ms. Zwislocki [19] has used Kiang’s data [4] to esti- mate the time constant of short-term adaptation at 30-50 ms without a description of the estimation

technique. Because of our short stimulus times of 25 ms we cannot make definitive estimates of this

short-term adaptation time constant but, based on PSTH shapes of some of the high SR units (see Fig. lA, third and fourth panels), it would appear to us that the values can be lower in some units with high CFs. Other transient effects such as post-stimulus ringing, and on-off edge effects have implications for interpreting CN responses.

The relative numbers of inputs from the differ- ent AN populations and their distribution on a given CN cell would seem to be critical variables if that particular cell type had the assigned function of extracting a particular kind of information about

the auditory environment. An illustration from our intracellular experiments in the AVCN exemplifies the problems of interpretation of CN unit coding [ll]. Intracellular records from cells, subsequently labeled with horseradish peroxidase and shown to

be bushy cells, reveal a high rate of spontaneous input from auditory nerve fibers in the form of large, fast synaptic potentials. These CN cells often respond in an onset fashion with a very precisely

timed spike to stimulus onset. These data imply that these cells may be preferentially innervated by

high SR AN fibers. Their large peak-to-steady-state ratios would suggest a relatively simple integration

of synaptic inputs which is congruent with the existence of the secure synapses (calyces of Held) in this region.

Most of our measures of AN unit behavior

agree with those of Liberman [5]. Despite our different selection of cut points to separate popu-

lations, which we based on the distribution of spontaneous rates, we note that our combined intermediate and low SR populations represent 42% of the total which agrees with the 39% found by Liberman. The peak-to-steady-state ratio is a function of SR and CF; it is lowest (near 1.0) for units with SR = 0 and increases to between 8 and 9 for high SR units with CFs near 10 kHz. Maxi- mal discharge rate, thresholds and tuning (Q,,)

decrease as a function of increasing spontaneous rate. These results are at variance with those from the gerbil, where no relationship between CF, SR and threshold was found [18]. Furthermore, we find that the difference in thresholds between low and high SR units (10 dB) is, on average, only about half as large as reported by Liberman [5]. One possible explanation for this discrepancy is that we find relatively few, low SR units with high

thresholds and sharp tuning curves. Liberman sug- gests that electrical stimulation of the AN per- mitted the location of a larger number of high

threshold units than could be obtained through more conventional search procedures. In one of our cats there were many units with thresholds near 0 dB SPL; the thresholds of the remainder were all within 25 dB of the most sensitive units at any single frequency. There were no units with thresholds between 30 and 70 dB SPL, suggesting that the distribution of thresholds is not continu- ous. This result is similar to that found in cats

reared in a low noise chamber by Liberman (IS], Fig. 3). We did not use search stimuli with SPLs above 70 dB (except at low frequencies where SPLs reached 90 dB) in order to minimize tem- porary threshold shifts 1131. Thus, we cannot rule out the possibility that some units with very high thresholds and no spontaneous firing may have been overlooked. However, because the proportion of low (19%, SR.< 0.5 sp/s) and intermediate (23%) SR fibers isolated is similar to that reported by Liberman (16% and 23%, respectively [5]), we believe it is unlikely that many of these high threshold fibers were missed with our search tech- nique.

The association of PL and PL, responses with AN fibers having different rates of spontaneous activity is of potential relevance to models of the afferent hair cell synapse between the inner hair cell and primary fiber. Smith and Brachman [16] suggest that the source of the rapid adaptation in AN fibers is presynaptic. Their multiple reservoir mode, with its local, intermediate and global neu- rotransmitter stores, is consistent with a large amount of auditory nerve behavior [2]. In their model, Geisler and Schwid [2] attribute rapid adaptation to a depletion of the immediate store. Liberman [6] reconstructed four inner hair cells from serial electron micrographs including the synaptic complex and the synaptic bar which oc- curs at each synapse. He noted that, while every afferent terminates on an inner hair with a single terminal, there are differences in the size and location of the terminations. By labeling individ- ual fibers with horseradish peroxidase ]7], he de- termined that low SR fibers were smaller, had fewer mitochondria and were usually located me- dially on the inner hair cell. High SR fibers oc- curred most often on the pillar side of the inner hair cell but could be found all around the cell. Low SR fibers were never found on the pillar side of the inner hair cell. This result suggests that the site of AN innervation of the inner hair cell, as well as the fiber’s caliber and ~tochondrial con- tent, may influence the adaptive process. Liber- man also describes a ‘complex synapse’ char- acterized by an asymmetrically located, long pre- synaptic bar and more specialized synaptic thick- ening than the other synapses. These complex synapses are associated with the small caliber,

167

~tochond~a-par fibers which have a smaller contact area with the hair cell. Therefore, a pre- synaptic correlate of low SR is not ruled out. Certainly, models of the hair cell synapse will have to pay heed to these details.

Acknowledgements

Special thanks are due to D. Greenwood and S. Greenberg for their superlative effort in reviewing this manuscript. The authors also thank M. Rhode for typing the manuscript and C. Dizack and T. Stewart for artwork and photography, respectively. This research was supported by NIH grant NS 17590.

References

1

2

3

4

5

6

7

8

9

10

11

Evans, E.F. (1981): The dynamic range problem: Place and time coding at the level of cochlear nerve and nucleus. In: Neuronal Mechanisms of Hearing, pp. 69-85. Editors: J. Szka and L. Atkin. Plenum Press, New York. Geisler, CD. and Schwid, H. (1982): Multiple reservoir mode of neurotr~s~tter release by a cochlear inner hair cell. J. Acoust. Sot. Am. 72, 1435-1440. Harris, D.M. and Dallas, P. (1979): Forward masking of auditory nerve fiber responses. J. Neurophysiol. 42, 1083-1107. Kiang, N.Y.S., Watanabe, T., Thomas, E.C. and Clark, L.F. (1965): Discharge Patterns of Single Fibers in the Cat’s Auditory Nerve. Research Monograph. No. 35. MIT Press, Cambridge, MA. Liberman, M.C. (1978): Auditory-nerve response from cats raised in a low-noise chamber. J. Acoust. Sot. Am. 63, 442-455. Liberman, M.C. (1980): Morphological differences among radial afferent fibers in the cat cochlea: an electron-micro- scopic study of serial sections. Hearing Res. 3, 45-63. Liberman, M.C. (1982): Single-neuron labeling in the cat auditory nerve. Science 216, 1239-1241. Pfeiffer, R.R. (1966): Classification of response patterns of spike discharges for units in the cochlear nucleus: Tone burst stimulation. Exp. Brain Res. 1, 220-235. Pfeiffer, R.R. and Rim, DO. (1972): Response patterns of single co&ear nerve fibers to click stimuli: descriptions for cat. J. Acoust. Sot. Am. 52,1669-1677. Rhode, W.S., Smith, P.H. and Oertel, D. (1983) Physiologi- cal response properties of cells labeled intracellularly with horseradish peroxidase in cat dorsal cochlear nucleus. J. Comp. Neurol. 213,426-447. Rhode, W.S., Oertel, D. and Smith, P.H. (1983): Physiologi- cal response properties of cells labeled intracellularly with horseradish peroxidase in the cat ventral cochlear nucleus. J. Comp. Neurol. 213, 448-463.

168

12 Sachs, M.B. and Abbas, P.J. (1974): Rate versus level functions for auditory-nerve fibers in cats: tone-burst stimuli. J. Acoust. Sot. Am. 56, 1835-1847.

13 Siegel, J.H., Rim, D.O. and Molnar, C.E. (1982): Effects of altering organ of Corti on cochlear distortion. J. Neuro- physiol. 47, 303-328.

14 Smith, R.L. (1977): Short-term adaptation in single audi- tory nerve fibers; some poststimulato~ effects. J. Neuro- physiol. 40, 2098-1112.

15 Smith, R.L. and Bra&man, M.L. (1980): Operating range and maximum response of single auditory nerve fibers. Brain Res. 184, 499-505.

16 Smith, R.L. and Bra&man, M.L. (1982): Adaptation in auditory nerve fibers: a revised modei. Biol. Cybernetics 44, 107-120.

17 Tonndorf, J. and Khanna, SM. (1968): Submicroscopic displacement amplitudes of the tympanic membrane (cat) measured by a laser interferometer. J. Acoust. Sot. Am. 44, 15461554.

18 Westerman, L.A. and Smith, R.L. (1984): Rapid and short- term adaptation in auditory nerve responses. Hearing Res. 15. 249-260.

19 Zwislocki, J. (1969): Temporal summation of loudness: an analysis. J. Acoust. Sot. Am. 46. 431-441.