Abstract

Temporal acuity for acoustic transients in rats with bilateral auditory cortex lesions (n = 6) was compared with that of sham-surgery control rats (n = 4), using a standard gap-detection method. A comparison of sensitivity to quiet gaps in noise and dark gaps in light tested for a cross-modal effect of the lesion. The groups were compared also in their sensitivity to noise offset, to noise increments, and to noise pulses presented in quiet. Stimulus detection was assessed with the startle reflex modification procedure, which uses changes in reflex expression caused by stimuli presented immediately before reflex elicitation as the objective evidence for their detection. There were no group differences in sensitivity to noise offset, noise pulses, or dark gaps in light. In contrast, the lesion reduced sensitivity to noise increments and eliminated gap detection. These deficits were maintained for 1 month and only partially recovered 2 months after surgery. The data indicate that the auditory cortex is critically important for temporal acuity in hearing, and suggest that its contribution to gap detection is to enhance the salience of noise increments at the end of the gap.

Introduction

Detecting a brief quiet period in a background noise is a simple psychophysical task that is frequently used to assess temporal acuity in the auditory system. This method yields similar gap thresholds in studies of human listeners and laboratory animals, and these thresholds agree with those obtained at different levels of the auditory system in electrophysiological research. Minimal gap thresholds of ~2–3 ms are found in sensory and behavioral experiments in all of mammalian species so far examined, in mice (Barsz et al., 2002) and rats (Ison, 1982); chinchillas (Giraudi et al., 1980) and gerbils (Wagner et al., 2000); and in humans (Plomp, 1964). Comparable neural gap thresholds have been reported in the auditory nerve (AN) of chinchillas (Zhang et al., 1990), the inferior colliculus (IC) of mice (Walton et al., 1997), and the auditory cortex (AC) of cats (Eggermont, 1999). One influential theory that underpins the study of temporal acuity in the sensory laboratory begins with the hypothesis that the cue for gap detection is the increment in sensation evoked by noise onset at its end (Plomp, 1964). The sensory representation of the pre-gap noise level decays during the gap, but for a brief gap some residual sensory activity persists and serves to reduce the contrast between the quiet period during the gap and the subsequent return of the noise. Thus threshold acuity for gap detection depends in part on the rapidity of the decay of sensation during the gap and in part on a more general sensitivity to noise increments. More recent formal ‘energy detector models’ stress the decrement in sensation during the gap as being sufficient for its detection (Buus and Florentine, 1985; Forrest and Green, 1987), while theories that emphasize the perceptual primacy of acoustic onsets rather than offsets (Oxenham, 1997; Phillips et al., 2002) retain a focus on the return of the noise at the end of the gap as being the critical event.

Conceptual models are not necessarily designed to correspond to identifiable physiological events, but, nonetheless, the observation that sensory thresholds approximate the neural thresholds obtained in the AN suggests that they simply reflect the rostral passage of a dip in neural activity from the cochlear receptors to the auditory nerve and thence to higher auditory levels, this dip corresponding to the momentary loss of acoustic energy during the quiet period of the gap. The AN data thus seem consistent with energy detector models of temporal acuity. However, two observations challenge this simple idea, and argue for a more complex neural interpretation of the apparently simple gap detection experiment. The first is that the physiological correlate for a brief gap changes within the brain, its more rostral neural effect in the IC or AC being not a reduction in firing rate within the quiet period but a burst of activity evoked by noise onset at its end (Walton et al., 1997; Eggermont, 1999). The second is that sensitivity to gaps is very much diminished in humans with temporal lobe lesions (Buchtel and Stewart, 1989) and in laboratory animals with diffuse functional decortication produced by spreading depression (Ison et al., 1991) or by more focused bilateral AC ablation (Kelly et al., 1996). The equivalence of behavioral gap thresholds with neural onset response thresholds in IC and AC is consistent with sensory models that stress the perceptual primacy of the noise marking the end of the gap, and the data obtained in brain-damaged subjects suggest that the perceptual response to noise onset is shaped by cortical activity.

The objective of the experiments reported here was to evaluate the theoretical models for gap detection in the context provided by the sensory deficit that accompanies cortical dys-function. Most simply it can be argued from loudness detector models that the effect of the lesion is to retard the decay of sensation within the quiet period of the gap, while from onset models its effect is to reduce the response to the noise marking its end. Each of these will account for the loss of temporal acuity in the brain-damaged subject, but the hypothesis that cortical activity contributes to the strength of the onset response predicts also that subjects with cortical lesions should be less sensitive to noise increments as well as to gaps in noise. We thus examined behavioral sensitivity to gaps in noise and to brief noise increments, and also to noise offset alone and to brief noise pulses presented in quiet. We also studied effects of dark gaps in a light background to determine if the deficits in temporal acuity were cross-modal or, like those of a brain-damaged human patient (Buchtel and Stewart, 1989), were restricted to the auditory domain.

Our subjects were AC-lesioned and sham-operated control rats, and the behavioral method was based on Reflex Modification Audiometry (Young and Fechter, 1983). Weak stimuli presented just prior to a stimulus that elicits an acoustic startle reflex (ASR) modify reflex expression in laboratory animals and in humans, and this change in expression provides objective evidence that the prestimulus has been detected. Most of the stimuli that we used were presented at lead intervals on the order of 100 ms, intervals at which ASR inhibition depends on the integrity of the IC (Leitner and Cohen, 1985; Li et al., 1998). In other experiments the stimuli were presented within 10 ms prior to reflex elicitation, intervals at which noise increments facilitate and noise decrements inhibit the ASR. The neural sites responsible for short lead time effects are thought to reside in the caudal brainstem, affecting afferent processing within the ASR pathways of the cochlear nucleus. Comparing the effects of cortical lesions on the response to gaps and noise pulses or increments should be informative about whether the lesion affects only the rate of decay of excitation during the gap or the response to its end, or has a more general effect on responses to transient onsets. Comparing these effects at long versus short lead times may be informative about the way in which the gap is processed at different levels of the auditory system.

Materials and Methods

Subjects

The subjects were 10 male albino rats of the Fischer 344 strain that were ~3 months of age at the beginning of the work. Six were assigned to a group for which the auditory cortex was ablated following a pretest, and four to a control group that received a similar surgical procedure save for the final aspiration of the cortex. They were raised in the vivarium at the University of Rochester, and housed in individual cages in controlled temperature and humidity rooms, on a 12:12 light:dark schedule. Food and water were provided ad libitum except during testing. All experiments took place in the light part of the cycle, and were conducted over a period of 2.5 months. The procedures were approved by the University Committee on Animal Resources, and were governed by the NIH Principles of Animal Care.

Apparatus and Stimuli

The subject was confined for testing in a wire mesh cage, 13 cm wide, 18 cm long and 9 cm high, with a rounded roof. The cage was mounted on a suspended acrylic platform over an attached accelerometer. This apparatus was placed in a double-walled sound-attenuating room, ~2.5 m on a side in its internal dimensions. The startle reflex was elicited by a 20 ms noise burst of 115 dB (SPL linear scale) provided by a white noise generator, gated through an electronic switch with <0.5 ms rise/fall times to a programmable attenuator, then amplified and delivered through a high frequency tweeter with its maximum output at 16 kHz and a 5 dB/octave rolloff. (For human listeners a noise burst at this level and duration approximates the loudness of a vigorous hand clap at a distance of ~20 cm from the ear.) The white noise used as a carrier for gaps and noise increments or as a pulse in one experiment was provided by a second noise generator, and was gated through a second electronic switch to a second programmable attenuator. The sound was then amplified and delivered through a wide spectrum speaker that varied by no more than ±2 dB from 1 kHz to 32 kHz when measured in one octave bands. The visual stimulus that was used in one experiment was produced by a fluorescent lamp (Sylvania F4TS-CWX), that delivered a 50 lux background illumination. A custom-built relay was used to momentarily turn off the light, this producing a brief dark period having rise and fall times less than 0.1 ms at its edge. The force exerted on the cage floor by the ASR flinch response was detected by an accelerometer. The force output passed through a bridge amplifier, and was integrated over a 100 ms period beginning with the onset of the startle stimulus. The integrated responses were stored on a computer which controlled stimulus presentation with custom-built programs running on a National Instruments board.

Surgery and Histological Procedures

Surgery was performed after initial behavioral pretests for gap detection, as described below. All rats were first anesthetized with xylazine (5 mg/kg) and ketamine (60 mg/kg) and placed in a stereotaxic frame. Blunt ear bars were used to fix the head in place. A midline incision was made in the scalp and the skin and muscles were gently detached from the edges of the dorsal surface of the skull, which was then trephined along its lateral edge from 3 to 7 mm posterior to Bregma. For the rats in the sham-lesion control group the dura remained intact and no tissue was removed. For rats in the experimental group the dura was opened on both sides so that the underlying cortex could be directly visualized and then aspirated. The auditory areas Te1, Te2 and Te3 (Zilles, 1985) were the target zones. When bleeding stopped the area was packed with gel foam and closed. Antibacterial ointment was applied to the outside of the wound. At the conclusion of testing all rats in the lesioned group and one control rat were overdosed with pentobarbital and perfused transcardially with physiological saline and then 4% formalin solution. The brains were removed and fixed further in a formalin–sucrose solution. Sequential 40 μm coronal sections were mounted on glass slides and stained with cresyl violet. The cortical lesions were reconstructed from serial sections corresponding to plates in a stereotaxic atlas (Paxinos and Watson, 1986). The drawings of the surface areas of the lesion on each side were superimposed, and the effective bilateral damage taken to be the area of overlap. In order to provide a second measure of effective lesion size we also measured the volume of the medial geniculate body (MGB) in lesioned rats relative to the single control subject. The volume of the MGB was determined for matched sections according to anatomical landmarks (Paxinos and Watson, 1986), using a microscope with camera lucida. The drawings were traced onto a digitizing tablet interfaced with a microcomputer that calculated their area. The MGB volume was estimated from sequential sections along its rostral–caudal extent. Differences between animals were assumed to reflect the varying amount of degeneration in the pathways to the AC from the MGB.

General Testing Procedures

All tests began with the rat being placed in the testing cage for 5 min, in background noise for noise gap and noise increment experiments, in quiet for noise pulse and light gap experiments. The acoustic experiments took place in an unlit test room, the visual experiment in the light. The different stimulus conditions were given in randomized blocks of trials in which every condition was presented before any were repeated, and the intertrial intervals between startle stimuli averaged 20 s, range 15–25 s. Descriptive and inferential statistical analyses of the data compared mean relative response levels across conditions and across groups. These values were obtained by taking each subject’s mean response for a particular stimulus condition on a test day and expressing it as a proportion of its mean baseline ASR with no prestimulus. Inferential analyses used within-subject and mixed-design analyses of variance (ANOVA). The P-values associated with the F-values were calculated using the Huynh–Feldt adjustments for d.f. in repeated measurement designs with non-homogeneous correlations across conditions.

Specific Experimental Designs

1. Pretraining for Gap Detection: Gaps Presented at Different Lead Times

The pre-surgery series began with a training day in which the startle stimulus was given alone or following a 10 ms gap in a 70 dB noise with onset-to-onset lead times of 10, 15, 20, 30, 40, 60, 100, 160 and 300 ms, in six blocks of 11 trials. Each trial block consisted of two presentations of the baseline condition (a startle stimulus with no preceding gap) and a single presentation of each of the nine prestimulus-startle conditions described above, these provided in random order. Pretraining was designed to assure optimum performance prior to the experiment as the inhibitory effects of weak stimuli increase with test experience (Ison and Bowen, 2000).

2. Inhibition by a Gap Presented at Long Lead Times before Reflex Elicitation

In this experimental procedure the continuous 70 dB background was interrupted by brief periods of quiet, 1, 2, 3, 4, 5, 6, 8 or 10 ms in duration, and the startle stimulus was presented 70 ms after the end of the quiet period. These stimuli were given in 11 blocks of 11 trials, each block consisting of a single presentation of each the prestimulus conditions listed above plus two presentations of the startle stimulus given alone (0 ms gap). Nine of these tests were completed: one as a pretest prior to surgery, the others at weekly intervals after surgery, continuing for 2 months. These are the typical gap stimuli used to assess temporal acuity, and this procedure yields similar thresholds in rats and humans (Ison, 1982; Ison and Pinckney, 1983).

3. Inhibitory Effect of Noise Offset at Short Lead Times before Reflex Elicitation

In this experiment a 70 dB noise was interrupted by quiet periods of 0, 1, 2, 3, 4, 5, 6, 8 or 10 ms duration, and the startle stimulus was presented at the end of the quiet period. This test was given in 11 blocks of 11 trials, each block consisting of a single presentation of the eight gap conditions listed above, two presentations of the startle stimulus with the gap duration of 0 ms, and a ‘blank stimulus’ condition in which spontaneous activity was recorded for a 100 ms time period. Three of these tests were given, one before surgery and two post tests, 1 and 2 weeks after surgery. The rationale for this experiment is that the growth of inhibition with increasing noise offset lead time depends on the same periods of quiet provided in the previous gap experiment, but here the quiet period lies between noise offset and the startle stimulus. If the growth of inhibition with increased gap duration is the same in the lesioned and control groups in this experiment, then this must indicate that brief silent periods affect neural activity at a relatively caudal level of the auditory brain stem to the same extent in both groups.

4. Inhibitory Effect of a Gap in Background Illumination at Long Lead Times

In this experiment startle stimuli were presented in an illuminated background, with the light interrupted by brief dark periods of 5, 10, 15, 20 or 30 ms, these ending 100 ms prior to startle elicitation. These stimuli were given in 11 blocks of seven trials, each block including two 0 ms gap ASR baseline trials. The intent of this experiment was to determine whether loss of temporal acuity following surgery was modality specific or a sign of a general deficit. The tests were given twice, 1 and 3 weeks following surgery.

5. Inhibition by a Noise Pulse Presented at Different Lead Times

In this experiment the stimuli were 60 dB (SPL) noise bursts presented in quiet. The bursts were 20 ms long and their onset preceded the ASR stimulus by 10, 15, 20, 30, 40, 60, 100 and 160 ms. These conditions were given in five blocks of 10 trials, which included two ASR baseline trials. This test was given once, 3 weeks after surgery. The objective of this test was to determine whether reflex inhibition provided by a noise pulse would be affected the lesion. If the lesion has no effect then any deficits in inhibition noted in the gap detection tests could not be attributed to a direct effect of the AC on the neural mechanisms responsible for reflex inhibition.

6. Facilitatory Effect of a Noise Increment at a Short Lead Time

In this procedure increments in noise level were added to a continuous 58 dB (SPL) background, these ending with the startle stimulus that was presented 10 ms after their onset. The increments were 0, 1.5, 3, 6, 9, 12 and 15 dB, and were implemented as a reduction in the attenuation of the noise by a programmable attenuator with near 0 ms rise and fall times. The stimuli were given in 11 blocks of eight trials, including two ASR alone baseline trials. This test was given once, 4 weeks after surgery. A brief increment that overlaps the startle stimulus facilitates the startle reflex, but it is the same event that at longer lead times inhibits the reflex. If reflex facilitation is systematically affected by increments that just precede the startle in lesioned rats, then the increments in noise level must have been detected at a relatively caudal level of the central auditory system.

7. Inhibitory Effect of a Noise Increment at a Long Lead Time

In this procedure the same increments of 0, 1.5, 3, 6, 9, 12 and 15 dB over a 58 dB noise background were used, but the increments were 20 ms in duration presented 110 ms prior to the startle stimulus. The stimuli were given in 11 blocks of eight trials, including two ASR baseline trials. The importance of this test is that a lesion effect would suggest that the AC contributes to temporal acuity by enhancing the detection of the noise at the end of the gap. This test was given twice: at 4 and 8 weeks after surgery.

Results

A. Anatomy

Figure 1 shows the subdivisions of the cortex in the rat seen from the left side, as described on the basis of cytoarchitectonic evidence (Zilles, 1985). The central auditory area is Te1 and the secondary areas are Te2 and Te3, these being heavily outlined. Overlaid on this figure are drawings of the surface extent of the smallest and the largest bilateral lesions (horizontal vs vertical stripes). Drawings of the lesion in both hemispheres for each rat are presented in Bowen (Bowen, 1996) (see pp. 111–112). Compared to the size of the normal auditory areas, the median lesioned surface area within Te1 was 89% (range, 57–100%), and within Te2 and Te3 combined the median lesioned surface area was 81% (range, 36–90%). The median MGB volume was 51% of the control value (range, 33–58%). There was no relationship between MGB volume and the surface area of the AC lesion (r = 0.11, NS).

B. Behavioral Results

1. An Overview

Table 1 lists each of the six experimental designs described above, and provides the results of the ANOVA for the main comparisons of the two groups. The major finding of theoretical interest is that rats with AC lesions were significantly less sensitive to both auditory gaps and to noise increments. The groups were not different in sensitivity to noise offsets, to visual gaps, or to noise pulses presented in a quiet (the interaction factor resulted because pulse inhibition occurred at shorter lead times in the lesion group, P < 0.05).

2. Gaps in Noise: an Immediate Loss of Temporal Acuity, with Some Recovery

Figure 2 shows the mean relative response (± SEM) for the inhibitory effects of gaps on the ASR for the control group averaged across all tests, and the lesioned group for the pretest and a selection of post tests: the first, 1 week after surgery; the fourth, 1 month after surgery; and the eighth, 2 months after surgery. The pretest scores for the to-be-lesioned group and the scores for all of the tests combined in the control group are typical for the gap detection test in normal rats. The mean amplitude of the ASR was reduced as gap duration increased from 2 to 4 ms, to reach a plateau of ~50% of the baseline value. A conventional threshold for gap detection is the duration at which the ASR is inhibited by one-half of the maximum level of inhibition. This threshold was 2 ms for four rats and 3 ms for the remaining six rats. Sensitivity to gaps was completely disrupted by the surgery but partially recovered towards the end of the test series.

Figure 3 shows the ASR data for each lesioned rat on the pretest and the first and the last post test in this series. The numbers in parentheses give the surface area covered by their lesions as a proportion of the normal area of Te1, Te2 and Te3 combined. One week after surgery, five of the six rats with lesions appeared to respond randomly at near-baseline for these gap durations. There were peaks of facilitated responses for subject no. 4 at 4 ms and no. 5 at 5 ms, but single points in individual subjects could be expected to yield such discrepancies by chance because of the inherent variability in the ASR for small samples. However, subject no. 2 showed a systematic pattern of startle facilitation for brief gaps that is never seen in the normal animal, though, as shown below, noise increments facilitate startle when presented at a very brief lead time. The last post test, given 8 weeks after surgery, found clear evidence of partial if not consistent gap detection for longer gaps in four rats (nos 1, 2, 3 and 5), one of which (no. 5) recovered approximately to its pretest performance. One rat (no. 4) showed no sign of recovered sensitivity, while one rat (no. 6) exhibited the facilitatory pattern shown earlier by no. 2.

The main effect of the lesion in reducing gap detection was highly significant in the ANOVA (see Table 1). Additional analyses indicated that the lesion had an immediate effect on gap sensitivity as shown in a significant difference between the pretest and the first post test, F(1/5) = 49.85, P < 0.01; and also that the effect of the lesion declined over test days, the group being more sensitive to gaps on the last post test compared with the first, F(1/5) = 13.54, P = 0.014. However recovery was still not complete, as their last test was different from the pretest in the rate of growth of inhibition with increased gap duration, F(1/5) = 8.21, P < 0.05. In addition, there was a marginally significant group difference on the last post test, F(1/8) = 5.07, P = 0.054.

3. Reflex Inhibition Associated with Noise Offset, Gaps in Light, and Noise Prepulses Is Not Diminished by Lesions of Auditory Cortex

Figure 4 shows the inhibitory effect of noise offset on the ASR when the startle stimulus was presented after periods of quiet varying between 1 and 10 ms. The mean of the three tests is shown for the control group, each test separately for the lesioned group. Noise offset inhibited the startle reflex, and the ANOVA for the growth of inhibition with increasing lead time provided F(1/8) = 310.82, P < 0.01. The lesion had no effect on this measure (F < 1).

Figure 5 shows reflex inhibition for a brief dark period in an otherwise continuous light. The 2 days for the control subjects were collapsed into a single function, but both tests are depicted for the lesion group. Although the group means for the lesioned subjects suggests an improvement from the first to the second test, overall there was no difference between the groups and no difference between the two tests. The lesion had no significant effect on visual gap sensitivity. The gaps inhibited the ASR (P < 0.01) and durations equal to 15 ms or more were effective in both groups, P < 0.05, Dunnett’s test.

Figure 6 shows the relative ASR after a brief 60 dB noise pulse presented in quiet at intervals of 10–160 ms. The overall shapes of the functions for the two groups were very similar in their reaching a near equal asymptotic level of inhibition at ~20 ms, but the lesion group showed significantly more inhibition at the 10 ms interval (P < 0.05).

4. Reflex Facilitation and Inhibition Associated with Noise Increments Are Affected by Lesions of Auditory Cortex

Figure 7 shows the effect on startle behavior of providing brief 10 ms noise increments in an otherwise continuous 58 dB noise, +1.5 to +15 dB, beginning 10 ms prior to the startle stimulus and ending with its onset. Consistent with previous data (Ison et al., 1997), these increments produced reflex facilitation that first increased and then decreased as the increment increased in level and as the less sensitive inhibitory effect of the increment became more prominent. The lesioned group showed less facilitation overall, as is shown in Table 1, but there was no effect on its threshold: facilitation for the 1.5 dB increment was significant in both groups, P < 0.05.

Figure 8 shows the effect on startle behavior of presenting the same noise increments of 1.5 dB to 15 dB over the 58 dB background, but now the duration of the increment was set at 20 ms and it ended 90 ms prior to the startle stimulus. This test was given twice, four and then 8 weeks after surgery, within 3 days of the gap detection tests shown in Figure 2. This figure depicts the group means for the control subjects, but separate plots for each lesioned subject. At this longer lead time the noise increments inhibited the startle reflex in the control group, and the strength of inhibition was proportional to their intensity. The control group had a threshold for inhibition between 1.5 and 3 dB, whereas the threshold for the lesioned group on this first test was between 12 and 15 dB. Five lesioned subjects showed severe deficits in this test that were evident even at 15 dB. One subject (no. 6) was exceptional in showing normal inhibition for noise increments, but no systematic evidence of gap sensitivity in a test given just 3 days earlier. The lesion in this rat was unusual in that only 12% of Te2 was missing, in contrast to other subjects for which the damage ranged from 78% (no. 1) to 100% (nos 2 and 5). This may be an interesting finding, but from these limited data we cannot exclude the possibility that it is a simple coincidence.

The lesioned group as a whole substantially improved on the second test, given 8 weeks after surgery: for the interaction between Test and Increment Level, F(6/30) = 2.60, P < 0.05. Increments of 6 dB and beyond produced significant inhibition in the lesioned group on the second test, P < 0.05. There was no significant difference between the two groups on this second test day, but this may be attributable to the large variability between the lesioned subjects. The performance of two subjects (nos 1 and 4) was at least as good as the control group mean for increments of 3 dB and above, but overall gap detection had not returned to normal on this final test.

Discussion

The present experiments confirm prior reports that bilateral lesions of AC disrupt temporal acuity as revealed in a diminished sensitivity to brief gaps in noise (Buchtel and Stewart, 1989; Ison et al., 1991; Kelly et al., 1996). The observation that the two groups were equal in their sensitivity to dark gaps in light confirms human data in showing that this cortical effect is modality specific (Buchtel and Stewart, 1989). The groups were also equal in their sensitivity to noise pulses presented in quiet, indicating that the lesion did not affect the neural mechanisms responsible for inhibition or the onset response to a noise burst presented in quiet. Ablation of the AC did disrupt noise increment detection as well as gap detection, thus supporting the hypothesis that its effect on temporal acuity resulted at least in part because it diminished sensitivity to noise onset at the end of the gap.

Reflex inhibition at relatively long lead times depends on neural activity in the IC (Leitner and Cohen, 1985; Li et al., 1998), and IC thresholds for gap detection approximate those found in behavioral experiments using Reflex Modification Audiometry (Walton et al., 1997). This contribution of the IC in these studies suggests the possibility that the lesion effects shown in the present experiments have their basis in the disruption of a downstream influence of the AC on normal IC activity. This conjecture is consistent with data showing that electrical stimulation of the AC sharpens rate-intensity functions in the IC of the bat (Zhou and Jen, 2000), as the increased sensitivity to variation in stimulus level that so results would benefit both the onset response to the end of the gap and the increment threshold. However, it is noteworthy that some of the lesioned subjects on the later tests had normal or near normal increment thresholds but relatively poor sensitivity for gaps. As described earlier, this dissociation of the two phenomena suggests that the rate of decay of afferent activity during the quiet period of the gap is also affected by the AC lesion.

The failure to find lesion effects on the thresholds for gap and increment detection at short intervals indicates that caudal brainstem mechanisms respond appropriately to these types of stimuli in the absence of cortical input. Amplitude modulation in the caudal brainstem is best coded in the octopus cells of the posterior ventral cochlear nucleus. These cells are specialized for onset responding to increments in AN activity by virtue of their intrinsic membrane characteristics (Oertel et al., 2000). The temporal precision of their bursts of activity is observed in slice preparations and thus cannot be the product of complex neural circuitry (Bal and Oertel, 2001). In addition, in the slice preparation their activity is insensitive to inhibitory neuro-transmitters (Golding et al., 1995). Thus, on the assumption that activity in these cells contributes to the types of reflex modification we observe at short lead times, then the lack of effect of AC lesions is to be expected.

Overall our data are best interpreted as indicating that activity in auditory cortex makes two significant contributions to temporal acuity. One is to strengthen the onset response to the noise at the end of the gap, because of its effect on noise-increment thresholds. The second is to increase the rate of decay of activity during the quiet period of the gap, perhaps by maintaining the fidelity of neural coding of gaps in noise in its rostral progression from the cochlear nucleus to higher auditory centers. These conclusions are consistent with models of gap detection that consider the importance of both the rate of sensory adaptation during the quiet period of the gap and the sensory response to the noise increment at its end (Plomp, 1964), with the understanding that both of these sensory effects depend on active neural processing that is at least partially under the control of the auditory cortex.

Notes

This research was supported by USPHS Research Grant AG09524, and by the Rochester International Center for Hearing and Speech Research. These findings formed part of the PhD dissertation work of G. P. Bowen, and a preliminary report was given before the Association for Research in Otolaryngology, St Petersburg Beach, Florida, February 1995. G.P.B. is now at the Wayland Baptist University, Plainview, Texas; D.L. is at Arena Pharmaceuticals, San Diego, California; M.K.T. is at the University of Michigan, Ann Arbor MI.

Address correspondence to James R. Ison, Department of Brain and Cognitive Sciences, University of Rochester, Rochester, NY 14627, USA. Email: ison@bcs.rochester.edu.

Table 1

ANOVA results for AC lesion versus control group effects

Experiment  F-value 
*P < 0.05; **P < 0.01. 
Auditory gap detection 
     Groups F(1/8) = 42.55** 
     Groups × Duration F(8/64) = 12.88** 
Noise offset 
     Groups F(1/8) = 0.1 
Visual gap detection 
     Groups F(1/8) = 0.05 
Noise pulse 
     Groups F(1/8) = 0.61 
     Groups × Lead Time F(8/64) = 3.45* 
Noise increment 
 (1) Facilitation 
     Groups F(1/8) = 31.23** 
 (2) Inhibition 
     Groups F(1/8) = 6.96* 
     Group × Level F(6/48) = 4.65** 
Experiment  F-value 
*P < 0.05; **P < 0.01. 
Auditory gap detection 
     Groups F(1/8) = 42.55** 
     Groups × Duration F(8/64) = 12.88** 
Noise offset 
     Groups F(1/8) = 0.1 
Visual gap detection 
     Groups F(1/8) = 0.05 
Noise pulse 
     Groups F(1/8) = 0.61 
     Groups × Lead Time F(8/64) = 3.45* 
Noise increment 
 (1) Facilitation 
     Groups F(1/8) = 31.23** 
 (2) Inhibition 
     Groups F(1/8) = 6.96* 
     Group × Level F(6/48) = 4.65** 
Figure 1.

A diagram of the left surface of the cortex in the rat, adapted from Figure 46 in Zilles (Zilles, 1985) with the permission of the author and the publisher (©Springer-Verlag GmbH & Co). The tick marks provide a scale in mm. The areas of auditory cortex, Te1, Te2 and Te3, are marked by a thick outline. The surface area of the largest bilateral cortical lesion is marked with vertical stripes, and the area of the smallest bilateral lesion is marked with horizontal stripes. The median extent of the bilateral lesion within Te1 was 89% of its area, and within Te1 and Te2 combined it was 81%.

Figure 1.

A diagram of the left surface of the cortex in the rat, adapted from Figure 46 in Zilles (Zilles, 1985) with the permission of the author and the publisher (©Springer-Verlag GmbH & Co). The tick marks provide a scale in mm. The areas of auditory cortex, Te1, Te2 and Te3, are marked by a thick outline. The surface area of the largest bilateral cortical lesion is marked with vertical stripes, and the area of the smallest bilateral lesion is marked with horizontal stripes. The median extent of the bilateral lesion within Te1 was 89% of its area, and within Te1 and Te2 combined it was 81%.

Figure 2.

The mean amplitude of the ASR (± SEM) expressed as a proportion of the startle baseline, when the startle stimulus was preceded at an interval of 70 ms by a quiet gap in a continuous background noise. The heavy solid line depicts the overall mean data for the control group, for the pretest and for the total of eight post-surgery tests given 1 week apart. The heavy dotted line shows the pre-surgery means for the lesioned group. The inhibitory effect of these gaps was eliminated in the post tests given 1 and 4 weeks after surgery, with some but incomplete recovery 8 weeks after surgery.

Figure 2.

The mean amplitude of the ASR (± SEM) expressed as a proportion of the startle baseline, when the startle stimulus was preceded at an interval of 70 ms by a quiet gap in a continuous background noise. The heavy solid line depicts the overall mean data for the control group, for the pretest and for the total of eight post-surgery tests given 1 week apart. The heavy dotted line shows the pre-surgery means for the lesioned group. The inhibitory effect of these gaps was eliminated in the post tests given 1 and 4 weeks after surgery, with some but incomplete recovery 8 weeks after surgery.

Figure 3.

The mean relative amplitude of the ASR when the startle stimulus was preceded at an interval of 70 ms by a quiet gap in a continuous background noise. The pretest and the first and last post surgery tests for each of the subjects with AC lesions are shown, and also the size of the bilateral AC lesion as a percentage of the total area of Te1, Te2 and Te3. No subject had a normal inhibitory effect of the gaps on the first post test, and only one subject (S#5) showed as much inhibition on the last post test as it had on the pretest. Isolated values above and below the baseline value can be expected on the basis of chance for any one condition in any one subject, but two subjects showed a systematic pattern of gap-produced facilitation, S#2 on the first post test and S#6 on the last post test. This effect is never seen in normal subjects.

Figure 3.

The mean relative amplitude of the ASR when the startle stimulus was preceded at an interval of 70 ms by a quiet gap in a continuous background noise. The pretest and the first and last post surgery tests for each of the subjects with AC lesions are shown, and also the size of the bilateral AC lesion as a percentage of the total area of Te1, Te2 and Te3. No subject had a normal inhibitory effect of the gaps on the first post test, and only one subject (S#5) showed as much inhibition on the last post test as it had on the pretest. Isolated values above and below the baseline value can be expected on the basis of chance for any one condition in any one subject, but two subjects showed a systematic pattern of gap-produced facilitation, S#2 on the first post test and S#6 on the last post test. This effect is never seen in normal subjects.

Figure 4.

The mean amplitude of the ASR (± SEM) expressed as a proportion of the startle baseline, when the startle stimulus was preceded by noise offset at lead times of 1–10 ms. The heavy solid line depicts the overall mean data for the control group, for the pretest and for the two post-surgery tests given 1 week apart. The heavy dotted line shows the pre-surgery means for the AC-lesioned group, the others showing that the inhibitory effect of these gaps was maintained in the post tests given 1 and 2 weeks after surgery. The systematic increase in reflex inhibition with the increase in the intervening duration of quiet between noise offset and startle stimulus onset indicates that the AC-lesioned subjects could detect and respond to the same brief intervals of quiet that had no effect or an abnormal facilitatory effect when presented 70 ms prior to the startle stimulus.

Figure 4.

The mean amplitude of the ASR (± SEM) expressed as a proportion of the startle baseline, when the startle stimulus was preceded by noise offset at lead times of 1–10 ms. The heavy solid line depicts the overall mean data for the control group, for the pretest and for the two post-surgery tests given 1 week apart. The heavy dotted line shows the pre-surgery means for the AC-lesioned group, the others showing that the inhibitory effect of these gaps was maintained in the post tests given 1 and 2 weeks after surgery. The systematic increase in reflex inhibition with the increase in the intervening duration of quiet between noise offset and startle stimulus onset indicates that the AC-lesioned subjects could detect and respond to the same brief intervals of quiet that had no effect or an abnormal facilitatory effect when presented 70 ms prior to the startle stimulus.

Figure 5.

The mean amplitude (± SEM) of the ASR expressed as a proportion of the startle baseline, when the startle stimulus was preceded at an interval of 100 ms by a dark period in otherwise continuous background illumination. The heavy solid line depicts the overall mean data for the control group, for tests given 1 and 2 weeks after surgery. The dotted lines show the means for the AC-lesioned group both tests. There was no significant group difference on the first test, and on the second test both groups showed the normal inhibitory effect of photic gaps with a threshold lying between 10 and 15 ms. These results indicate that the AC lesions did not cause a general cross-modal deficit in temporal acuity.

Figure 5.

The mean amplitude (± SEM) of the ASR expressed as a proportion of the startle baseline, when the startle stimulus was preceded at an interval of 100 ms by a dark period in otherwise continuous background illumination. The heavy solid line depicts the overall mean data for the control group, for tests given 1 and 2 weeks after surgery. The dotted lines show the means for the AC-lesioned group both tests. There was no significant group difference on the first test, and on the second test both groups showed the normal inhibitory effect of photic gaps with a threshold lying between 10 and 15 ms. These results indicate that the AC lesions did not cause a general cross-modal deficit in temporal acuity.

Figure 6.

The mean amplitude (± SEM) of the ASR expressed as a proportion of the startle baseline, when the startle stimulus was preceded by a 10 ms 60 dB noise pulse in a quiet background at lead times of 10–160 ms. This test was given 3 weeks after surgery. The noise pulse inhibited the ASR in the lesioned group at least as well as it did in the control group, indicating that AC lesions do not directly affect the neural mechanisms responsible for reflex inhibition. The small but significant difference showing apparent greater inhibition in the lesioned group at a 10 ms lead time is best attributed to the lesion’s reducing reflex facilitation. Weak noise pulses that are presented in a quiet background usually produce a small degree of reflex facilitation that briefly overrides their inhibitory effect.

Figure 6.

The mean amplitude (± SEM) of the ASR expressed as a proportion of the startle baseline, when the startle stimulus was preceded by a 10 ms 60 dB noise pulse in a quiet background at lead times of 10–160 ms. This test was given 3 weeks after surgery. The noise pulse inhibited the ASR in the lesioned group at least as well as it did in the control group, indicating that AC lesions do not directly affect the neural mechanisms responsible for reflex inhibition. The small but significant difference showing apparent greater inhibition in the lesioned group at a 10 ms lead time is best attributed to the lesion’s reducing reflex facilitation. Weak noise pulses that are presented in a quiet background usually produce a small degree of reflex facilitation that briefly overrides their inhibitory effect.

Figure 7.

The mean amplitude (± SEM) of the ASR expressed as a proportion of the startle baseline, when the startle stimulus was preceded by a 10 ms increase in the level of the background noise, with its onset 10 ms prior to the startle stimulus. This test was given 4 weeks after surgery. The normal facilitatory effect of these noise increments was diminished in the lesioned rats, but both groups had a threshold of facilitation below +1.5 dB.

Figure 7.

The mean amplitude (± SEM) of the ASR expressed as a proportion of the startle baseline, when the startle stimulus was preceded by a 10 ms increase in the level of the background noise, with its onset 10 ms prior to the startle stimulus. This test was given 4 weeks after surgery. The normal facilitatory effect of these noise increments was diminished in the lesioned rats, but both groups had a threshold of facilitation below +1.5 dB.

Figure 8.

For the control group, the mean amplitude (± SEM) of the ASR expressed as a proportion of the startle baseline, when the startle stimulus was preceded by a 20 ms increase in the level of the background noise, with its onset 100 ms prior to the startle stimulus. For the lesion group, these data are provided separately for every subject. This test was given twice: 4 weeks and 8 weeks after surgery. The normal inhibitory effect of the noise increments was disrupted for the lesioned animals on the first test, and their inhibitory effect was not significant save for the 15 dB increment. This effect of the lesion was much reduced on the second test, and both groups showed significant inhibition for the 3 dB increment.

Figure 8.

For the control group, the mean amplitude (± SEM) of the ASR expressed as a proportion of the startle baseline, when the startle stimulus was preceded by a 20 ms increase in the level of the background noise, with its onset 100 ms prior to the startle stimulus. For the lesion group, these data are provided separately for every subject. This test was given twice: 4 weeks and 8 weeks after surgery. The normal inhibitory effect of the noise increments was disrupted for the lesioned animals on the first test, and their inhibitory effect was not significant save for the 15 dB increment. This effect of the lesion was much reduced on the second test, and both groups showed significant inhibition for the 3 dB increment.

References

Bal R, Oertel D (
2001
) Potassium currents in octopus cells of the mammalian cochlear nucleus.
J Neurophysiol
 
86
:
2299
–2311.
Barsz K, Ison, JR, Snell KB, Walton JP (
2002
) Behavioral and neural measures of auditory temporal acuity in aging humans and mice.
Neurobiol Aging
 
23
:
565
–578.
Bowen GP (
1996
) Modification of the acoustic startle reflex by complex sensory pre-stimuli: the roles of the auditory cortex and visual cortical areas. Unpublished doctoral dissertation, University of Rochester, Rochester, New York.
Buchtel HA, Stewart JD (
1989
) Auditory agnosia: apperceptive or associative disorder?
Brain Lang
 
37
:
12
–25.
Buus S, Florentine M (
1985
) Gap detection in normal and impaired listeners: the effect of level and frequency. In: Time resolution in auditory systems (Michelsen A, ed.), pp. 159–179. Berlin: Springer.
Eggermont JJ (
1999
) Neural correlates of gap detection in three auditory cortical fields in the cat.
J Neurophysiol
 
81
:
2570
–2581.
Forrest TG, Green DM (
1987
) Detection of partially filled gaps in noise and the temporal modulation transfer function.
J Acoust Soc Am
 
82
:
1933
–1943.
Giraudi D, Salvi R, Henderson D, Hamernik R (
1980
) Gap detection by the chinchilla.
J Acoust Soc Am
 
68
:
802
–806.
Golding NL, Robertson D, Oertel D (
1995
) Recordings from slices indicate that octopus cells of the cochlear nucleus detect coincident firing of auditory nerve fibers with temporal precision.
J Neurosci
 
15
:
3138
–3153.
Ison JR (
1982
) Temporal acuity in auditory function in the rat: reflex inhibition by brief gaps in noise.
J Comp Physiol Psychol
 
96
:
945
–954.
Ison JR, Bowen GP (
2000
) Scopolamine reduces sensitivity to auditory gaps in the rat, suggesting a cholinergic contribution to temporal acuity.
Hear Res
 
145
:
169
–176.
Ison JR, Pinckney LA (
1983
) Reflex inhibition in humans: sensitivity to brief silent periods in white noise.
Percept Psychophys
 
34
:
84
–88.
Ison JR, O’Connor K, Bowen GP, Bocirnea A (
1991
) Temporal resolution of gaps in noise by the rat is lost with functional decortication.
Behav Neurosci
 
105
:
33
–40.
Ison JR, Taylor MK, Bowen GP, Schwarzkopf SB (
1997
) Facilitation and inhibition of the acoustic startle reflex in the rat after a momentary increase in background noise level.
Behav Neurosci
 
111
:
1335
–1352.
Kelly JB, Rooney BJ, Phillips DP (
1996
) Effects of bilateral auditory cortical lesions on gap-detection thresholds in the ferret (Mustela putorius).
Behav Neurosci
 
110
:
542
–550.
Leitner DS, Cohen ME (
1985
) Role of the inferior colliculus in the inhibition of acoustic startle in the rat.
Physiol Behav
 
34
:
65
–70.
Li L, Priebe RP, Yeomans JS (
1998
) Prepulse inhibition of acoustic or trigeminal startle of rats by unilateral electrical stimulation of the inferior colliculus.
Behav Neurosci
 
112
:
1187
–1198.
Oertel D, Bal R, Gardner SM, Smith PH, Joris PX (
2000
) Detection of synchrony in the activity of auditory nerve fibers by octopus cells of the mammalian cochlear nucleus.
Proc Natl Acad Sci USA
 
97
:
11773
–11779.
Oxenham AJ (
1997
) Increment and decrement detection in sinusoids as a measure of temporal resolution.
J Acoust Soc Am
 
102
:
1779
–1790.
Paxinos G, Watson C (
1986
) The rat brain. New York: Academic Press.
Phillips DP, Hall SE, Boehnke SE (
2002
) Central auditory onset responses, and temporal asymmetries in auditory perception.
Hear Res
 
167
:
192
–205.
Plomp R (
1964
) Rate of decay of auditory sensation.
J Acoust Soc Am
 
36
:
277
–282.
Wagner E, Klump GM, Hamann I (
2000
) Gap detection in Mongolian gerbils.
Abstr Assoc Res Otolaryngol
 
23
:
27
.
Walton JP, Frisina RD, Ison JR, O’Neill WE (
1997
) Neural correlates of behavioral gap detection in the inferior colliculus of the young CBA mouse.
J Comp Physiol A
 
181
:
161
–176.
Young JS, Fechter LD (
1983
) Reflex inhibition procedures for animal audiometry: a technique for assessing ototoxicity.
J Acoust Soc Am
 
73
:
1686
–1693.
Zhang W, Salvi RJ, Saunders SS (
1990
) Neural correlates of gap detection in auditory nerve fibers of the chinchilla.
Hear Res
 
46
:
181
–200.
Zhou X, Jen PH-S (
2000
) Corticofugal inhibition compresses all types of rate-intensity functions of inferior collicular neurons in the big brown bat.
Brain Res
 
881
:
62
–68.
Zilles K (
1985
) The cortex of the rat. a stereotactic atlas. Berlin: Springer Verlag.