| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Center for Neuroscience and 2Section of Neurobiology, Physiology and Behavior, University of California at Davis, Davis, California
Submitted 6 July 2006; accepted in final form 22 November 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
24 ms. These data place the primate primary auditory cortex at an early processing stage of temporal rate discrimination. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Scharlock et al. 1965). In macaque monkeys, lesions of either the left or both auditory cortices result in impairments in the monkey's ability to discriminate conspecific vocalizations (Heffner and Heffner 1986). In humans, lesions in the temporal cortex were previously shown to result in compromised speech comprehension (Schnider et al. 1994) and the inability to order two successive tonal stimuli (Swisher and Hirsh 1972). Recent human psychophysics and imaging studies in learning impaired children, children classified as "poor readers," and dyslexics indicated that they have a reduced ability to resolve the duration and rate of acoustic test stimuli, such as processed speech and tones, compared with control children (Merzenich et al. 1996; Nagarajan 1999; Tallal et al. 1996; Temple 2001; Wright et al. 1997). Those authors suggested that these deficits result from a general impairment in the ability to resolve auditory temporal information.
One manner in which single neurons can represent the temporal features of acoustic stimuli is by phase-locking to the envelope of the signal. Such a mechanism was demonstrated throughout the ascending auditory pathway (for reviews see Edelman et al. 1988; Lagner 1992; Moore 1989; Pickles 1992). Despite these high-fidelity representations of temporal information, it is still not clearly understood how these neural codes translate to a perception of order and distinct patterns (for review see Johnson 2000; Parker and Newsome 1998).
Initial inroads into the underlying basis of the cortical computations of temporally complex stimuli were conducted in the somatosensory system (for review see, Romo and Salinas 2003). These studies indicated that the phase-locking of individual neurons are well correlated with the corresponding percepts of the individuals. In the auditory domain, similar studies pointed toward the same general conclusions (Brown and Maloney 1986; Moody 1994), although an overall rate code for high temporal frequencies was also proposed (Lu et al. 2001). To further investigate how auditory cortical neurons could potentially encode low-frequency temporal information, responses of single neurons to four tone-pip sequences were recorded in alert macaque monkeys. Preliminary results from this study were previously published in abstract form (Phan et al. 1998).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
This report is based on data collected from 167 single neurons recorded from the primary auditory cortex of two adult male rhesus monkeys (Macaca mulatta) weighing 7–12 kg over the course of the study. All protocols and procedures followed the guidelines outlined in the Ethical Treatment of Animals (National Institutes of Health) and were approved by the UC Davis animal care and use committee.
For all recording sessions, the monkey sat in a customized primate chair constructed to minimize acoustic reflections with its head fixed and pointed toward the center of a speaker array. All data were collected within a darkened double-walled sound booth (2.4 x 3.0 m, IAC, New York, NY) lined with sound-attenuating foam.
Before recording monkeys were implanted with a restraining head post and recording cylinder (see Pfingst and O'Connor 1980; Recanzone et al. 2000a). Two recording cylinders were implanted bilaterally in monkey A (the two surgeries were separated by 18 mo). One recording cylinder was implanted over the left auditory cortex of monkey L.
Data collection
At the start of each recording session, a plastic grid was placed into the recording chamber (Crist et al. 1988). Electrodes (FHC) were introduced into the cerebral cortex through a guide tube placed into a selected grid location. Search stimuli consisted of noise, tones, and clicks presented through either a speaker located directly contralateral to the recording cylinder or a frontal location (±30° in azimuth, 0° in elevation). All speakers were located 1 m from the center of the interaural axis. Once a neuronal response was encountered, neural waveforms were amplitude-filtered and stored (50-kHz sampling) for off-line sorting using customized software. The latency, threshold, best frequency at threshold [characteristic frequency (CF)] and bandwidth at 10 or 40 dB above threshold was audiovisually determined for the recording site. This allowed us to rapidly assess the basic physiological response properties used in classifying the cortical area and to set the stimulus frequencies. All neurons had response characteristics consistent with being within the primary auditory cortex (A1) including these response parameters, the progression of CF, and the location within the recording cylinder (Recanzone et al. 2000a). This was verified histologically by Nissl, myelin, and cytochrome oxidase staining of the superior temporal gyrus in both monkeys.
Stimuli
Stimulus generation and data collection were controlled by Tucker Davis Technologies (Gainesville, FL) hardware and software interfaced with a personal computer. Stimuli consisted of 20 different patterns of four tone pips. The pip durations (2-ms linear rise/fall) were the same as the interpip intervals and varied from 6 to 30 ms in 6-ms steps. Each pip intensity was 60 dB SPL (A-weighted) with a ±2-dB variation between pip sequences. Frequencies of the first and fourth pips of the sequence were near the CF of the neuron under study. Four different sequences were presented for each stimulus duration by varying the frequency of the second and third tone pips. The frequency of these pips differed from the CF by either 5 or 20% and were presented in either order, for example, the second pip was [CF + (CF/20)] and the third was [CF – (CF/20)] or the reverse. Because sequential presentation of identical tones was previously shown to result in decreases in firing rate with each presentation (Brosch and Schreiner 1997; Calford and Semple 1995; Evans and Whitfield 1964; Ulanovsky et al. 2004; Werner-Reiss et al. 2006), these stimuli were chosen to prevent rapid habituation to each stimulus element, as well as to investigate the possibility that single neurons could encode the temporal order of these sequences.
Behavioral paradigm
The monkey performed a go/no-go sound localization task (see Recanzone et al. 2000b for details) during the recording of neural activity. Each trial consisted of a series of four to seven presentations of the four-pip sequences (S1) delivered from the speaker 90° contralateral to the recording site, followed by one presentation (S2) from a speaker located directly in front of the monkey (0°). Tone sequences were presented with a minimum of 800 ms between them. The monkey was trained to depress a lever to initiate a trial and then to release the lever when the stimulus changed location to receive a fluid reward. Within the S1 series, each sequence could vary in temporal rate (pip duration) and frequency but not in order. The order of the S2 sequence was different from that of the S1 and was always presented with a pip duration of 30 ms. Each recording session consisted of 15–20 randomly interleaved presentations of each stimulus duration, order, and frequency difference (about 20 min). Only responses to stimuli presented from 90° are reported here.
Data analysis
Poststimulus time histograms (PSTHs) were constructed using 1-ms time bins. The latency measured for each cell was defined as the median latency to the first spike that was >10 ms from stimulus onset pooled across all S1 stimulus presentations. This provided a reliable estimate even for neurons with relatively high spontaneous activity. To quantify the degree of phase-locking or synchronization to each tone pip of the sequence the vector strength (VS) was calculated using 1-ms time bins (Goldberg and Brown 1969; Mountcastle et al. 1969; Talbot et al. 1968). Each spike was assigned a vector with a length of 1.0 and phase assigned as the time of occurrence relative to the (pip + pip-interval) duration (e.g., 12 ms for the 6-ms pip duration and interval). These vectors were then summed and divided by the total number of spikes. Thus these values range from 1 (perfect synchrony) to 0 (no synchrony). To determine the statistical significance of these responses a Rayleigh test of uniformity, 2n(VS)2, was calculated for each neuron, where VS is the vector strength and n is the number of spikes used to calculate the vector strength (Buunen and Rhode 1978; Lu and Wang 2000; Mardia and Jupp 2000). The Rayleigh test is an approximation of a chi-square test constrained with 2 degrees of freedom, appropriate for circular distributions (Mardia and Jupp 2000; Zar 1999). Rayleigh test values >13.8 indicated that the synchronization of neuronal activity reflected a sample of an oriented distribution that is not attributed to chance (P < 0.01; Buunen and Rhode 1978).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
). The solid bars in Fig. 1A show the shortest median first-spike latency for all neurons recorded from that grid location on a given day. Because we often recorded from more than one neuron within a grid location on a given day, the open bars show the latencies of all other neurons recorded from on that experimental day. Different neurons recorded from a single grid location on a given experimental day were usually recorded from different depths, but could also be different neurons isolated from the same electrode location. The distribution of the characteristic frequency of these neurons is shown in Fig. 1B in 0.5-octave bins. The sample was biased toward neurons with higher CFs, which is the result of more extensive investigation of the caudal regions of A1. Some neurons were recorded with CFs near 500 Hz, which is in the border region between A1 and R (e.g., Merzenich and Brugge 1973; Morel et al. 1993; Recanzone et al. 2000a). These data, coupled with the CF gradient in each monkey and the histological verification, indicate that the vast majority, and likely all, of the neurons in this report were well within the primary auditory cortex. The recordings done in these monkeys went well beyond A1 as part of different studies, which verified the medial, lateral, and caudal borders of A1 with the belt fields (Recanzone et al. 2000a; see Kaas and Hackett 2002). Thus the region of A1 was evenly sampled across its extent.
|
|
|
). In contrast, other A1 neurons encoded the temporal structure of the stimuli by the periodicity of the response. Although our sample appeared fairly uniform across neurons, it could be that there were similar subpopulations of neurons that showed a stronger relationship between firing rate and the temporal envelope than others. To address this possibility, we subdivided the response based on the firing rates of the neurons to the 6-ms-duration stimuli. The firing rates for the 25% of neurons that had the lowest (L) and highest (H) firing rates to the 6-ms-duration stimuli were compared with the remaining neurons nearest the median of the population (M) using the same analyses shown in Fig. 3. These results are shown in Fig. 4. Although there is an offset between the sets of lines, the slopes are generally similar. We quantified this impression by calculating the regression line of these data and comparing the slopes with 95% confidence intervals. The slopes for the median and lowest firing rate cells were nearly identical (–0.0105 and –0.0107 spike·s–1·s–1 for lowest and median, respectively). The greatest slope was for the high-activity population (–0.0206 spike·s–1·s–1), although the 95% confidence intervals encompassed the slopes for the other two populations (–0.010 to –0.040). This difference was almost entirely eliminated if the 6-ms data were excluded (slopes of –0.002, –0.003, and –0.006 for low, median, and high, respectively). If there was a subpopulation of neurons that could better encode the tone-pip duration using a rate code, this would be revealed by a significant difference in the slopes of these curves. These results suggest that this is not the case; rather, neurons across the sampled population appear to respond similarly across these stimulus types regardless of their absolute firing rates.
|
Temporal precision in the responses
Whereas the firing rates were fairly uniformly distributed across the sample, a second way in which the tone-pip durations within a sequence could be encoded by the neurons is in the periodicity of their responses. Because the carrier frequency was generally very high (see Fig. 1B), the periodicity would be expected to be related to the envelope of the stimulus sequences. Tuning to the stimulus envelope is evident for tone-pip durations >18 ms in Fig. 2. Examples of such tuning are shown for a different example neuron in Fig. 5 for the stimulus train in which the second tone pip was less than the CF by 5% and the third was greater than the CF by 5%. In these plots, the timescales are adjusted such that the stimuli span the same distance in each raster. Again, there is little if any periodicity in the response corresponding to each tone pip in the sequence for pip durations of 6 and 12 ms (Fig. 5, A and B). Periodicity is first noted for 18-ms pip durations (Fig. 5C) and increases for the 24- and 30-pip-duration sequences (Fig. 5, D and E). A similar trend is shown for a third neuron that had a relatively low firing rate in Fig. 6. Although the firing rate was lower for this neuron, the appearance of the periodicity of the response is similar to that shown for the neurons in Figs. 2 and 5.
|
|
5. However, there was a clear trend for the number of neurons showing statistically significant periodicity to increase with increasing pip duration, as evidenced by the initial peak in the distributions becoming progressively smaller and the far-right tail of the distributions becoming progressively larger from 6-ms (gray) to 30-ms (black) pip durations. The percentage of neurons that showed statistically significant periodicity by this metric was 3.66, 23.45, 49.77, 64.61, and 73.57% for 6-, 12-, 18-, 24-, and 30-ms pip durations, respectively, where 1.0% of neurons are expected to show statistically significant tuning by chance.
|
One problem inherent in this analysis is that the vector strength, and therefore the Rayleigh statistic, can be dominated by the response to the first pip in the sequence. For example, a neuron that has only a phasic response to the first tone pip, and does not respond at all to the subsequent three pips, could have a high Rayleigh statistic even though it clearly did not have a periodic response. Within our sample, there were several instances where this occurred to varying degrees. To quantify the effect of such phasic responses to the first pip of the sequence, we compared the overall activity to the second through fourth tone pips to the activity to the first pip in the sequence. In this analysis, the activity during the latency period plus the first tone pip and interpip interval was calculated as a percentage of the total firing rate to each stimulus. The distribution of these percentages is shown in Fig. 8A, where the neuronal responses from each of the four stimulus-frequency orders are pooled. As expected from the example neurons shown in Figs. 2, 5, and 6, the majority of the responses occurred during the first bin for tone-pip durations of 6 ms (gray line), as indicated by the high tail of the distribution near 100. This tail disappears and the peak of the distribution moves leftward with increasing pip durations, indicating that the response to the first pip of the sequence has a progressively smaller dominance of the total response. To determine how this may have influenced the proportion of neurons showing statistically significant periodicity in their responses, the Rayleigh statistic was calculated using only the period corresponding to the responses to the second through fourth tone pips (Fig. 8B). Although there was a decrease in the proportion of neurons with significant periodicity in their response when only the second through fourth tone pips were used to calculate the metric compared with when all four tone pips were used (Fig. 8C), the general trends remained the same. Across the population, very few neurons could faithfully encode the stimulus envelope for tone pip durations of 6 and 12 ms, with a nearly linear rise in that proportion for tone-pip durations >12 ms.
|
|
|
As with the firing rate, we also compared the temporal response properties of these neurons with the characteristic frequency. Because the shorter-duration stimuli had few neurons that reached statistical significance, we restricted our analysis to those conditions in which the tone-pip durations were 30 ms and the neurons showed statistically significant tuning. This analysis revealed that there was no correlation with the characteristic frequency and the value of the Rayleigh statistic (r2 values ranged from 10–5 to 0.02; all P > 0.01). Thus the characteristic frequency does not account for any of the variance in the degree of synchronization as measured by the Rayleigh statistic.
Temporal order discrimination
The stimuli used in these experiments differed not only in the tone duration, but also in the frequency and order of the second and third pips in the sequence. This was partly to prevent the neurons from habituating to these tone pips, but also to allow the opportunity to determine whether the responses of the neurons could differentiate between the different tone-pip sequences. To address this issue we compared the firing rates between stimulus sequences of different frequencies and orders of presentation at each pip duration for each neuron. For clarity, we will consider sequences in which the second pip was lower than the CF by 5% (and the third pip greater than the CF by 5%) as "downward" sequences with small frequency differences [downward-small (DS)]. The reversed order, in which the second pip was 5% greater than CF (and the third 5% less) will be considered upward-small (US). Similarly, sequences will be identified as downward-large (DL) and upward-large (UL) for sequences where the second pip was less than or greater than the CF by 20%, respectively.
Analysis of the overall firing rate across neurons to each of the six possible comparisons between the sequences revealed that very few neurons showed any significant differences in the overall firing rate (Fig. 3; repeated-measures ANOVA; P < 0.05). At best, roughly 10% of the neurons showed a significant difference in firing rate and only at the longest tone durations tested. It could be, however, that the overall rate is not different but the responses to the different pips in the sequence does vary with the type of sequence presented. Figure 10 shows such a neuron to the US and DS stimuli presented with 30-ms-tone durations. This neuron is representative of many in which the overall firing rate is similar between the two conditions but there is a very clear difference in the temporal features of the response, with a greater response to the lower frequency [CF – (CF/20) or 10.45 kHz] than to the higher frequency, regardless of the position of this stimulus in the sequence. Such neurons could therefore encode the temporal order of the stimulus sequence. To investigate how many cells showed such a response, we conducted the same analysis of firing rate but restricted the analysis times to those corresponding to each of the individual pips in the sequence. There were very few neurons that showed significant differences in the response to the first pip, as expected, because the frequency of the first pip was invariant across sequences (ANOVA; P < 0.05 for 0–4.8% of neurons across comparisons and tone duration; percentage expected by chance = 5%). A similar finding was also noted for the response to the fourth pip in the sequence (range 0–3.6% across comparisons). This implies that the frequency of the preceding tone pips had very little influence on the response to the CF tone presented as the fourth pip in the sequence. In contrast, there were relatively large numbers of neurons that did show a difference in the response to the second or third tone pips in the sequence depending on the frequency and order of presentation of the individual pips (Fig. 11). There were very few neurons that showed significantly different responses at the shortest tone-pip durations (6 and 12 ms). For comparisons that differed in the order of presentation only (left panels) far more neurons showed significant differences in the responses to either the second or third pips for the larger frequency difference (DL vs. UL) than for the smaller frequency difference (DS vs. US). For comparisons where only the frequency differed but the order remained the same (middle panels), there were generally fewer neurons showing significantly different responses, again only for the longer tone durations. Finally, and somewhat surprisingly, changing both order and frequency had little additive effect compared with the large frequency difference where only the order was changed (compare DL vs. UL in the left column to the results in the right column). Regression analysis of the individual neurons that showed this effect did not reveal any suggestion that these differences were correlated with the CF of the neuron under study (r2 = 0.03; P > 0.05).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
found that subjects could discriminate the order of long tone durations (500 ms) with temporal separations as short as 2 ms. A second study found 200-ms durations were necessary to discriminate order using different stimuli (Barsz 1988; Warren 1974a), although with training, discriminating the order of noise, tones, and buzz sequences can occur at 
10 ms (Warren 1974b).
Previous studies in the auditory nervous system using discrete stimuli (as opposed to amplitude-modulated stimuli) focused on attenuation and/or enhancement of a response when a previous stimulus was presented (e.g., Bartlett and Wang 2005; Brosch et al. 1999; McKenna et al. 1989). We certainly did see both response facilitation and attenuation in our sample. Key differences between this and previous studies in the monkey, however, are in the relatively short duration and intervals between pips and the fact that we presented four pips instead of two. This allowed us to investigate the limits of temporal resolution that these neurons could encode rapid sequences. Our thresholds are in reasonably good agreement with those seen for amplitude-modulated stimuli seen in a variety of species and auditory areas (e.g., for review see Langer 1992), although we rarely saw synchronized responses much below 18-ms-pip durations (e.g., 
28 Hz), indicating that A1 neurons in alert macaque monkeys have difficulty in reliably following such fast temporal sequences.
Recent studies is the alert marmoset also indicated that A1 neurons could potentially use two different coding schemes to encode temporal rate information. For low-frequency stimuli, neurons can use a temporal code by firing in synchrony with the stimulus. At higher frequencies, the neurons can then shift to a rate code where the overall numbers of action potentials is related to the stimulus rate (Lu et al. 2001). Our results are in agreement with that study in a few respects. As in the marmoset, we did not observe reliably synchronous responses for intervals <18 ms and the degree of synchrony increased with increasing interval duration. However, we saw little if any evidence that there were two distinct populations of neurons. This may be explained by the limited range of temporal rates that we investigated (16–83 Hz). If this is the case, then a neural population that encodes high temporal rates in macaques must have cutoff rates near 100 Hz.
We used four tone pips with varying frequencies in the middle two pips to see whether we could gain some insights into how auditory cortical neurons could potentially differentiate between these different sequences. One potential difficulty is that we defined the stimuli based on the characteristic frequency that was assessed qualitatively by the experimenter. It is possible that using a stimulus that was not accurately centered at the characteristic frequency, or was centered on a subpeak of the tuning curve, could have changed the firing rate. We doubt that this was the case because most neurons responded reliably to each of the tone pips in the sequence (e.g., Figs. 2, 5, 6, and 10), although it remains a possibility. It is also possible that those neurons that did show differences between the stimuli were the most sharply spectrally tuned neurons; one frequency was inside the tuning curve and the other was not. Although we did not quantitatively assess the spectral tuning of these neurons, we feel that this is unlikely. Because there was no real difference between the DL versus UL comparison (20% greater vs. 20% smaller) and the US versus DL comparison (20% greater vs. 5% smaller), this would indicate that 35–40% of neurons have bandwidths of roughly 20% from the CF at this intensity level, which corresponds to a Q-value of 0.40. Previous studies in which spectral bandwidths in A1 were measured under nearly identical conditions (and from these same two monkeys) indicated that only 11/413 neurons had a Q-value in this range at 40 dB above threshold (see Fig. 12G of Recanzone et al. 2000a).
Anecdotally, it was extremely difficult for human observers to differentiate between the different elements of the sequences at the fastest rates, but they could for the slowest rate sequences. It is unlikely that the different sequences could be differentiated by a simple rate code because we saw very little difference between the overall firing rates between the different sequence orders. However, given recent results from Werner-Reiss et al. (2006), which reported that neuronal responses to the second auditory stimulus were often weaker when compared with the preceding stimuli, it is possible that a rate code might be used if the durations of the interstimulus interval (ISI) are in the range of hundreds of milliseconds instead of tens of milliseconds. Likewise, using stimulus and ISI rates slower than ours (230 vs. 736 ms), Ulanovsky et al. (2004), using an oddball paradigm, reported that neurons responded more strongly to the same frequency when it was the deviant stimulus than when it was the standard. Thus within the context of stimulus history and its potential role in differentiating sequences, it is possible for neurons to distinguish among the different sequences if the rates are sufficiently slow. However, at the level of our analysis a temporal code could indeed account for perceptual differences (Figs. 10 and 11). These results imply that the synchronization observed in the neuronal responses is a good indicator of the ability of these neurons to detect each element of these tone-pip sequences. These results also predict that there would be little perceptual gain by changing both the order and frequency once the frequency differences become large. These data also indicate that any type of "feature extraction" of such stimuli must occur at stages subsequent to the primary auditory cortex in the ascending auditory pathway.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Present address of M. L. Phan: Psychology Department, Rutgers University, 152 Frelinghuysen Rd., Piscataway, NJ 08854.
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: G. Recanzone, Center for Neuroscience, University of California at Davis, 1544 Newton Ct., Davis, CA 95618 (E-mail: ghrecanzone{at}ucdavis.edu)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Bartlett EL, Wang X. Long-lasting modulation by stimulus context in primate auditory cortex. J Neurophysiol 94: 83–104, 2005.
Brosch M, Schreiner CE. Time course of forward masking tuning curves in cat primary auditory cortex. J Neurophysiol 77: 923–943, 1997.
Brosch M, Schultz A, Scheich H. Processing of sound sequences in macaque auditory cortex: response enhancement. J Neurophysiol 82: 1542–1559, 1999.
Brown CH, Maloney CG. Temporal integration in two species of Old World monkeys: blue monkeys (Cercopithecus mitis) and grey-cheeked mangabeys (Cercocebus albigena). J Acoust Soc Am 79: 1058–1064, 1986.[CrossRef][ISI][Medline]
Bunnen TJ, Rhode WS. Responses of fibers in the cat's auditory nerve to the cubic difference tone. J Acoust Soc Am 64: 772–781, 1978.[CrossRef][ISI][Medline]
Calford MB, Semple MN. Monaural inhibition in cat auditory cortex. J Neurophysiol 73: 1876–1891, 1995.
Crist CF, Yamasaki SDG, Komatsu H, Wurtz RH. A grid system and a microsyringe for single cell recording. J Neurosci Methods 26: 117–122, 1988.[CrossRef][ISI][Medline]
Diamond IT, Neff WD. Ablation of temporal cortex and discrimination of auditory patterns. J Neurophysiol 20: 300–315, 1957.
Edelman GM, Gall EW, Cowan WM. Auditory Function: Neurobiological Bases of Hearing. New York: Wiley, 1988.
Evans EF, Whitfield IC. Classification of unit responses in the auditory cortex of the unanaesthetized and unrestrained cat. J Physiol 171: 476–493, 1964.
Goldberg JM, Brown PB. Response of binaural neurons of dog superior olivary complex to dichotic tonal stimuli: some physiological mechanisms of sound localization. J Neurophysiol 32: 613–636, 1969.
Heffner HE, Heffner RS. Effect of unilateral and bilateral auditory cortex lesions on the discrimination of vocalizations by Japanese macaques. J Neurophysiol 56: 683–701, 1986.
Hirsh IJ. Auditory perception of temporal order. J Acoust Soc Am 6: 759–767, 1959.[CrossRef]
Johnson KO. Neural coding. Neuron 26: 563–566, 2000.[CrossRef][ISI][Medline]
Kaas JH, Hackett TA. Subdivisions of auditory cortex and processing streams in primates. Proc Natl Acad Sci USA 97: 11793–11799, 2000.
Langer G. Periodicity coding in the auditory system. Hear Res 50: 115–142, 1992.
Lu T, Liang L, Wang X. Temporal and rate representations of time-varying signals in the auditory cortex of awake primates. Nat Neurosci 4: 1131–1138, 2001.[CrossRef][ISI][Medline]
Lu T, Wang X. Temporal discharge patterns evoked by rapid sequences of wide- and narrowband clicks in the primary auditory cortex of cat. J Neurophysiol 84: 236–246, 2000.
Mardia KV, Jupp PE. Directional Statistics. New York: Wiley, 2000.
McKenna TM, Weinberger NM, Diamond DM. Responses of single auditory cortical neurons to tone sequences. Brain Res 481: 142–153, 1989.[CrossRef][ISI][Medline]
Merzenich MM, Brugge JF. Representation of the cochlear partition on the superior temporal plane of the macaque monkey. Brain Res 50: 275–296, 1973.[CrossRef][ISI][Medline]
Merzenich MM, Jenkins WM, Johnston P, Schreiner C, Miller SL, Tallal P. Temporal processing deficits of language-learning impaired children ameliorated by training. Science 271: 77–81, 1996.[Abstract]
Moody DB. Detection and discrimination of amplitude-modulated signals by macaque monkeys. J Acoust Soc Am 95: 3499–3510, 1994.[CrossRef][ISI][Medline]
Moore BCJ. An Introduction to the Psychology of Hearing. San Diego, CA: Academic Press, 1989.
Morel A, Garraghty PE, Kaas JH. Tonotopic organization, architectonic fields, and connections of auditory cortex in macaque monkeys. J Comp Neurol 335: 437–459, 1993.[CrossRef][ISI][Medline]
Mountcastle VB, Talbot WH, Sakata H, Hyvarinen J. Cortical neuronal mechanisms in flutter-vibration studied in unanesthetized monkeys. Neuronal periodicity and frequency discrimination. J Neurophysiol 32: 452–484, 1969.
Nagarajan S, Mahncke H, Salz T, Tallal P, Roberts T, Merzenich MM. Cortical auditory signal processing in poor readers. Proc Natl Acad Sci USA 96: 6483–6488, 1999.
Parker AJ, Newsome WT. Sense and the single neuron: probing the physiology of perception. Ann Rev Neurosci 21: 227–277, 1998.[CrossRef][ISI][Medline]
Pfingst BE, O'Connor TA. A vertical stereotaxic approach to auditory cortex in the anaesthetized monkey. J Neurores Methods 2: 33–45, 1980.
Phan ML, Guard DC, Geiger PG, Recanzone GH. Processing of temporal sequences in the auditory cortex of the macaque monkey. Soc Neurosci Abstr 24: 401, 1998.
Pickles JO. An Introduction to the Physiology of Hearing. San Diego, CA: Academic Press, 1992.
Recanzone GH, Guard DC, Phan ML. Frequency and intensity response properties of single neurons in the auditory cortex of the behaving macaque monkey. J Neurophysiol 83: 2315–2331, 2000a.
Recanzone GH, Guard DC, Phan ML, Su TK. Correlation between the activity of single auditory cortical neurons and sound localization behavior in the macaque monkey. J Neurophysiol 83: 2723–2739, 2000b.
Romo R, Salinas E. Flutter discrimination: neural codes, perception, memory and decision making. Nat Rev Neurosci 4: 203–218, 2003.[CrossRef][ISI][Medline]
Scharlock DP, Neff WD, Strominger NL. Discrimination of tone duration after bilateral ablation of cortical auditory areas. J Neurophysiol 28: 673–681, 1965.
Schnider ADF, Benson DF, Alexander DN, Schnider-Klaus A. Non-verbal environmental sound recognition after unilateral hemispheric stroke. Brain 117: 281–287, 1994.
Swisher L, Hirsh IJ. Brain damage and the ordering of two temporally successive stimuli. Neuropsychologia 10: 137–152, 1972.[CrossRef][ISI][Medline]
Talbot WH, Darian-Smith I, Komhuber HH, Mountcastle VB. The sense of flutter-vibration: comparison of the human capacity with response patterns of mechanoreceptive afferents from the monkey hand. J Neurophysiol 31: 301–334, 1968.
Tallal P, Miller SL, Bedi G, Byma G, Wang X, Nagarajan SS, Schreiner C, Jenkins WM, Merzenich MM. Language comprehension in language-learning impaired children improved with acoustically modified speech. Science 271: 81–84, 1996.[Abstract]
Temple E, Poldrack RA, Salidis J, Deutsch GK, Tallal P, Merzenich MM, Gabrieli JD. Disrupted neural responses to phonological and orthographic processing in dyslexic children: an fMRI study. Neuroreport 12: 299–307, 2001.[CrossRef][ISI][Medline]
Ulanovsky N, Las L, Farkas D, Nelken I. Multiple time scales of adaptation in auditory cortex neurons. J Neurosci 24: 10440–10453, 2004.
Warren RM. Auditory pattern recognition by untrained listeners. Percept Psychophys 15: 495–500, 1974a.
Warren RM. Auditory temporal discrimination by trained listeners. Cogn Psychol 6: 237–256, 1974b.[CrossRef]
Werner-Reiss U, Porter KK, Underhill AM, Groh JM. Long lasting attenuation by prior sounds in auditory cortex of awake primates. Exp Brain Res 168: 272–276, 2006.[CrossRef][ISI][Medline]
Wright BA, Buonomano DV, Mahncke HW, Merzenich MM. Learning and generalization of auditory temporal-interval discrimination in humans. J Neurosci 17: 3956–3963, 1997.
Zar JH. Biostatistical Analysis. Upper Saddle River, NJ: Prentice Hall, 1999.
| ||||||||||||||||||||||||||||||||
