The Journal of Neurophysiology Vol. 79 No. 5 May 1998, pp. 2416-2431
Copyright ©1998 by the American Physiological Society

IID Sensitivity Differs Between Two Principal Centers in the Interaural Intensity Difference Pathway: the LSO and the IC

Thomas J. Park

Department of Biological Sciences, Neurobiology Group, University of Illinois at Chicago, Chicago, Illinois 60607

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Park, Thomas J. IID sensitivity differs between two principal centers in the interaural intensity difference pathway: the LOS and the IC. J. Neurophysiol. 79: 2416-2431, 1998. Interaural intensity differences (IIDs) are the chief cues that animals use to localize high-frequency sounds. Neurons that are sensitive to IIDs are excited by sound at one ear and inhibited by sound at the other. Thus a given IID generates a combination of excitation and inhibition that is reflected in a cell's spike count. In mammals, the so-called "IID pathway" begins in the lateral superior olive (LSO), which is dominated by the type of IID-sensitive neurons just described. The LSO then sends a prominent projection to the inferior colliculus (IC), which also contains a substantial population of IID-sensitive cells. Recent pharmacological studies have suggested that the response properties of IID-sensitive neurons in the IC undergo considerable processing and thus should not simply reflect the output of the LSO. However, we have no direct evidence as to whether IID sensitivity, the defining response feature of these cells, differs at these two levels. The present study makes this direct comparison in the Mexican free-tailed bat, a species that relies greatly on high-frequency hearing and thus on IIDs for localizing sounds in space. Extracellular recording techniques were used to obtain IID functions from 50 IC neurons. Comparable data from 50 LSO cells were available from a previous study. The main result was that IID sensitivity significantly differed between cells in the LSO and the IC. Among LSO cells, sensitivity was centered ~0 dB (no intensity difference between the ears) whereas, in the IC, sensitivity was biased toward the inhibitory ear: on average, IC cells required a more intense signal at the inhibitory ear to reach the same degree of suppression as observed in LSO cells. Further analysis showed that the vast majority of IC cells (88%) exhibited a mismatch in the latencies of their inputs: inhibition arrived later when an equally strong excitation and inhibition were elicited; this reduced the effectiveness of the inhibition. Because latency shortens with increasing stimulus intensity, an IID with a more intense signal at the inhibitory ear could equate the latencies of excitation and inhibition, increasing the effectiveness of the inhibition. This result suggests that latency mismatches account, to a great extent, for the difference in sensitivity between the LSO and the IC; and when mismatches were negated by electronically time shifting the signals to the ears, sensitivity was no longer significantly different between the two nuclei.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Interaural intensity differences (IIDs) are the main cues mammals use to localize high-frequency sounds in the horizontal plane (Erulkar 1972; Mills 1972). These cues are processed first in the lateral superior olive (LSO), where the coded intensities from the two ears initially are compared (Boudreau and Tsuchitani 1968; Caird and Klinke 1983; Cant and Casseday 1986; Moore and Caspary 1983; Sanes and Rubel 1988). The comparison is a subtractive process in that LSO neurons are excited by sound at the ipsilateral ear and inhibited by sound at the contralateral ear. Thus a given IID generates a combination of excitation and inhibition that is reflected in an LSO cell's spike count. The encoded information is conveyed from the LSO to the contralateral inferior colliculus (IC) via a prominent excitatory projection (Brunso-Bechtold et al. 1981; Glendenning et al. 1992; Saint Marie and Baker 1990; Zook and Casseday 1982). The IC, in turn, also contains a substantial population of IID-sensitive neurons (Irvine and Gago 1990; Pollak et al. 1986; Semple and Kitzes 1987). Consistent with the crossed projection from the LSO to the contralateral IC, the ears providing excitation and inhibition are reversed for these two nuclei: stimulation of the ipsilateral ear is excitatory for the LSO, stimulation of the contralateral ear is excitatory for the IC. Otherwise, the superficial binaural response properties of IID sensitive neurons in both nuclei are similar (e.g., prominent excitation from one ear and prominent inhibition from the other ear). Hence, it had been assumed previously that the binaural response properties of many IID-sensitive cells in the colliculus reflect, to some degree, those established in the LSO.

A number of recent studies have challenged the assumption that IID sensitive cells in the colliculus simply reflect properties established in the LSO. The procedures these studies employed involved local blockade of inhibitory neurotransmitters to individual IC cells (Faingold et al. 1989; Klug et al. 1995; Park and Pollak 1993; Vater et al. 1992), reversible inactivation of nuclei that project to the IC (Faingold et al. 1993; Li and Kelly 1992), and in vivo whole cell patch-clamp recording (Covey et al. 1996). The results of these studies suggest that although the binaural properties of LSO and IC cells are superficially similar, most IC cells receive excitatory and/or inhibitory inputs, in addition to inputs from the LSO, that modify IID sensitivity. In some cases, collicular cells even derive their IID sensitivity de novo, bypassing the LSO and integrating excitatory and inhibitory inputs within the IC.

Although the evidence cited above suggests that, on a population level, IID sensitivity might differ between the LSO and the IC, no direct comparison has been made between those two nuclei to characterize differences and what their functional significance might be. The purpose of the present study was to carry out such a direct comparison within one species, the Mexican free-tailed bat.

	METHODS

Abstract Introduction Methods Results Discussion References

Surgical and recording procedures

Four Mexican free-tailed bats, Tadarida brasiliensis mexicana, were experimental subjects. This dasypygal mammal was chosen for two reasons. First, like all echo-locating bats, the free-tailed bat has excellent high-frequency hearing and relies heavily on IIDs for sound localization. Second, relevant data from the free-tailed bat's LSO was already available (Park et al. 1996). Before surgery, animals were anesthetized with methoxyflurane inhalation and 15 mg/kg pentobarbital sodium injected subcutaneously. The hair on the bat's head was removed with a depilatory, and the head secured in a head holder with a bite bar. The muscles and skin overlying the skull were reflected and 4% lidocaine hydrochloride was applied topically to all open wounds. The surface of the skull was cleared of tissue, and a ground electrode was placed just beneath the skull over the posterior cerebellum. A layer of small glass beads and dental acrylic was placed on the surface of the skull to secure the ground electrode and to serve as a foundation layer to be used later for securing a metal rod to the bat's head.

The bat was transferred to a heated (27-30°C), sound-attenuated room, where it was placed in a restraining apparatus attached to a custom-made stereotaxic instrument (Schuller et al. 1986). A small metal rod was cemented to the foundation layer on the skull and then attached to a bar mounted on the stereotaxic instrument to ensure uniform positioning of the head. A small hole (~0.5- to 1.0-mm diam) then was cut over the inferior colliculus on one side. Position of the hole and positioning of the electrode followed procedures described by Schuller et al. (1986). Recordings were begun after the bat was awake. If the animal struggled or otherwise appeared in discomfort, the local anesthetic was refreshed and an additionalsubanesthetic injection of pentobarbital sodium (10 mg/kg body wt) was given subcutaneously. This dosage of pentobarbital never induced anesthesia: the bats still were awake in that their eyes were open, they drank water when it was offered, and they responded when their face or ears were touched gently. There were no noticeable, systematic changes in neuronal response properties from the pentobarbital. These additional pentobarbital injections were administered on only several occasions and then only once during a given recording session. Recording sessions generally lasted from 3 to 5 h/day to minimize the animals' discomfort from being restrained.

Action potentials were recorded with a glass pipette filled with buffered 1 M NaCl. Electrode impedance ranged from 5 to 20 M. Electrode penetrations were made vertically through the exposed dorsal surface of the inferior colliculus. Subsequently, the electrode was advanced from outside of the experimental chamber with a piezoelectric microdrive.

Acoustic stimuli and data acquisition

Pure tones with a duration of 60 ms were used as search stimuli. When a unit was encountered, its characteristic frequency and absolute threshold were audiovisually determined to set stimulus parameters subsequently controlled by computer. The characteristic frequency was defined as the frequency that elicited responses at the lowest sound intensity to which the unit was sensitive. Binaural stimuli then were presented to determine whether the unit was monaural or binaural and, if it was binaural, whether it was excitatory/inhibitory (EI) or excitatory/excitatory (EE). Units were operationally classified as EI if sound at the inhibitory (ipsilateral) ear predominantly suppressed the responses evoked by sound at the excitatory (contralateral) ear when the stimuli were presented simultaneously. EI cells were classified further by features such as whether they showed facilitation at certain IIDs or whether they had nonmonotonic rate-level functions. Although these features will be considered in the final section of RESULTS, for the purposes of comparing LSO and IC cells, these subdivisions of EI cells were grouped together.

Stimuli used to investigate IID sensitivity were 2-ms frequency sweeps with a rise fall time of 0.2 ms. The frequency of the sweep stimuli swept down from 5 kHz above to 5 kHz below a unit's characteristic frequency. Sweeps at both ears were coherent in that they had the same frequency range and duration and each began and ended with the same phase. The stimuli were presented via Brüel and Kjaer 1/4-in microphones used as ear phones fitted with probe tubes (5 mm diam) that were placed in the funnel of each pinna. Maximum sound intensity was 80 dB SPL measured 0.5 cm from the opening of the probe tubes. Sound pressure and the frequency response of each earphone was measured with a 1/4-in Brüel and Kjaer microphone. Each earphone showed less than ±3 dB variability for the frequency range usually used (15-80 kHz), and intensities between the earphones did not vary more than ±3 dB at any of those frequencies. Stimuli were presented at a rate of four per second. Acoustic isolation between the ears was better than 40 dB and was determined empirically during the course of the experiments by testing units that were operationally defined as monaural: units that were excited by stimulation of the excitatory, contralateral ear but showed no apparent excitatory or inhibitory effects when the ipsilateral ear was stimulated. The intensity at the contralateral ear was set 20 dB above the cell's threshold at its characteristic frequency. Then the ipsilateral ear was stimulated with intensities ranging from ~40 dB below to 40 dB above that at the contralateral ear. Next, the contralateral stimulus was turned off, and the intensity at the ipsilateral ear was increased until spikes were evoked, presumably via cross-talk to the contralateral ear. For fundamentals between 10 and 50 kHz and 74 dB SPL (the highest intensity used), all harmonics were 35 dB less intense than the fundamental. For higher frequencies, the harmonics were even lower.

Rate-intensity and IID functions were generated for each unit. Rate-intensity functions were generated with intensities from 10 dB below to 40 dB above threshold. The values for these intensities were estimated audiovisually before data acquisition. Each intensity was presented 20 times, and the order of presentation was varied pseudorandomly. IID functions were generated by fixing the sound intensity at the excitatory ear at 20 dB above threshold and pseudorandomly varying the intensity at the inhibitory ear, usually from 30 dB below to 40 dB above the intensity at the excitatory ear in 5- or 10-dB steps.

To evaluate the role of coincidence in shaping IID sensitivity, IID functions were measured when the signals at the two ears were presented simultaneously and when the timing of the signals was manipulated by introducing interaural time differences (ITDs). Hence, for each unit, a family of IID functions was generated at a variety of ITDs. The same data also yielded a family of ITD functions measured at a variety of IIDs. ITDs were computer controlled and usually varied in 100- or 200-µs steps. These ITDs were larger than the ITDs the animal normally would experience and were used here as a tool to evaluate the role of coincidence in shaping IID sensitivity. Each stimulus presentation was delivered 10 or 20 times. Only well-isolated spikes were studied. Spikes were fed to a window discriminator, and the output of the discriminator was fed to the computer. Data were displayed on the computer screen for inspection during the experiments and stored on hard disk for later analysis.

Data from cells in the LSO came from a previously published report (Park et al. 1996) that used the same procedures as those described above.

	RESULTS

Abstract Introduction Methods Results Discussion References

This study reports on 50 IID-sensitive neurons recorded from the IC of the Mexican free-tailed bat and compares their binaural response properties to those of 50 IID-sensitive neurons previously recorded from the LSO of the same species (Park et al. 1996). IID functions were measured for all cells with 2-ms-long frequency sweeps that descended from 5 kHz above to 5 kHz below a neuron's characteristic frequency. This stimulus was selected for three reasons. First, these sweeps simulate many aspects of the echolocation calls used by the free-tailed bat, and hence they are biologically relevant stimuli (of course, in nature echolocation calls would not always be centered on each neuron's characteristic frequency). Second, they generate a very similar response from cells in both the IC and the LSO in that they usually evoke a maximum of only one or a few spikes from cells in either nuclei. This feature is illustrated in Fig. 1, which shows the raster plots and IID functions for a typical IC neuron and a typical LSO neuron. Third, these sweeps generate spikes that occur within a very narrow time window (also illustrated in Fig. 1), a feature that made it possible to explore the relative timing of excitation and inhibition to these cells, an issue that will be described in depth in the following sections.

View larger version (19K): [in this window] [in a new window]   FIG. 1. Interaural intensity difference (IID) functions and corresponding raster plots for a typical IID-sensitive neuron from the inferior colliculus (IC; top) and a typical IID-sensitive neuron from the lateral superior olive (LSO; bottom). Positive IIDs indicate a greater intensity at the excitatory ear. Stimuli were 2-ms long, 10-kHz downward frequency sweeps centered at each unit's characteristic frequency. Intensity to the excitatory ear was fixed at 20 dB above threshold, whereas the intensity to the inhibitory ear was varied. Each IID was presented 20 times in pseudorandom order. Inset: rate-level functions for each cell.

Each of the 50 IC neurons exhibited prominent excitation from stimulation of the contralateral ear and prominent inhibition from stimulation of the ipsilateral ear (EI type of neuron). For convenience, I hereafter refer to the ear providing excitation as the excitatory ear and the ear providing inhibition as the inhibitory ear. IID functions were generated for each cell by driving a neuron with a fixed intensity, 20 dB above threshold, to the excitatory ear and then documenting the suppressive influence of increasing intensities to the inhibitory ear. Because the intensity at the excitatory ear was fixed, each intensity at the inhibitory ear generated a different IID. By convention, positive IIDs indicate that the sound was more intense at the excitatory ear.

During the course of this study, 21 additional cells were encountered but not included in the following analyses and comparisons. Of these 21 cells, 7 were responsive only to contralateral stimulation (monaural cells) and 5 were prominently excited by both ears (EE cells). The remaining nine cells were excluded because, although they were classified as EI cells, they did not respond to the 2-ms, 10-kHz frequency sweeps. These cells will be considered in more detail in a following section.

IID sensitivity varied among the IC cells

Each of the 50 IC cells studied showed a steep decline in spike count with increasing intensities to the inhibitory ear until spike activity was inhibited completely (the IID of complete inhibition). This feature is illustrated by the IID functions of five representative IC cells shown in Fig. 2A. Although the general shape of the functions was similar among cells, responsiveness to specific IIDs varied considerably from cell to cell. For example, some cells were inhibited completely when the intensity at the excitatory ear was greater than the intensity at the inhibitory ear (positive IIDs), whereas other cells were inhibited completely when the intensity at the inhibitory ear was greater than the intensity at the excitatory ear (negative IIDs).

View larger version (22K): [in this window] [in a new window]   FIG. 2. Representative IID functions and distributions of IIDs of complete inhibition for 50 IC neurons and 50 LSO neurons. A: IID functions from 6 IC cells illustrate how IID sensitivity varied among the population of cells tested. IID of complete inhibition is indicated on 1 function. B: distribution of IIDs of complete inhibition for the 50 IC cells tested. C: corresponding distribution of IIDs of complete inhibition for the 50 LSO cells from Park et al. (1996).

The IID of complete inhibitionthe point where a cell's IID function first went to zero spikes was selected to characterize the IID function of each cell. This point was chosen because it unambiguously separates all IIDs that evoke responses from all those that do not. Also, the IID of complete inhibition was used in our recent LSO study to characterize IID sensitivity in that nucleus. For the 50 IC cells studied, the distribution of IIDs of complete inhibition is shown in Fig. 2B. IIDs of complete inhibition ranged from +20 dB (excitatory ear more intense) to 40 dB (inhibitory ear more intense). This range of IID sensitivities corresponds to most of the range of IIDs that this species normally would encounter in the free field (Pollak 1988). The average IID of complete inhibition for the 50 IC cells was 18.0 dB (inhibitory ear more intense), corresponding to a sound location lateralized into the ipsilateral sound field. Because a number of previous investigations have used the half-maximum IID (the point on the IID function corresponding to a 50% decline from the peak spike count) as an index of IID sensitivity, the distribution of half-maximum IIDs from the present study also is displayed (Fig. 3).

View larger version (43K): [in this window] [in a new window]   FIG. 3. Distribution of 50% points on the IID functions from the 50 IC neurons. Average 50% point was at an IID of 4.27 dB.

Sensitivity was significantly more lateralized in the IC compared with the LSO

The distribution of IID sensitivities reported above for the IC cells (Fig. 2B) differed from the distribution previously reported for LSO cells from the same species. Figure 2C shows the distribution of IIDs of complete inhibition for the 50 LSO cells taken from Park et al. (1996) (mean IID of complete inhibition = 6.6 dB). As the distributions in Fig. 2 show, the IC cells had IIDs of complete inhibition that generally occurred at more negative values, corresponding to more lateralized sound locations, compared with the LSO cells. A statistical comparison of the two distributions showed that they were significantly different (t = 3.76, df = 98, P < 0.001).

More IC cells had mismatches in the latencies of excitation and inhibition compared with the LSO cells

The latencies of excitation and inhibition were estimated for each of the 50 IC cells to determine whether or not excitation and inhibition arrived at a given cell at the same time. The reasons for this test were that the latencies of excitation and inhibition are known to influence IID sensitivity (Irvine et al. 1995; Joris and Yin 1995; Park et al. 1996; Pollak 1988; Yin et al. 1985) and latency data were available for the 50 LSO cells previously studied. Hence, it seemed logical to acquire the same data for the IC cells and compare the two nuclei to see if latencies might play a part in the difference in sensitivity observed between the IC and LSO populations. The following paragraphs will first digress to describe how input latencies can shape IID sensitivity. I then will describe how the cells in this study were tested for latency effects and how the latency data for the IC cells differed from that of the LSO cells.

The so-called latency hypothesis of IID sensitivity is based on two key features, 1) the latencies of the inputs from the two ears differ among cells such that in some cells inhibition arrives later than excitation and 2) changes in stimulus intensity can change the latency of an input, potentially compensating for a mismatch in the timing of excitation and inhibition to a cell. This second feature, described as an intensity-dependent latency shift, has been well documented in a variety of auditory centers and appears to be a general phenomenon in the auditory system (as well as in other sensory systems).

The rationale of the latency hypothesis is detailed in Fig. 4. Figure 4A shows the excitatory postsynaptic potentials (EPSPs) and inhibitory postsynaptic potentials (IPSPs) of a hypothetical cell. The PSPs show how increasing the intensity to the inhibitory ear causes the strength of inhibition to increase, the latency of inhibition to shorten, and the duration of inhibition to lengthen. These three effects are consistent with data from auditory neurons tested in both in vivo (Irvine et al. 1995; Park et al. 1996; Yin et al. 1985) and slice preparations (Sanes 1990; Wu and Kelly 1992). For the example shown in Fig. 4A, the strengths of excitation and inhibition are equal at the IID of complete inhibition (Fig. 4A, *). The important feature in this example is that the latencies of excitation and inhibition also are matched at this IID. For convenience, cells like this are referred to as "neurons with matched latencies," meaning that the excitation and the inhibition arrive coincidentally at the target cell when their strengths are equal.

View larger version (23K): [in this window] [in a new window]   FIG. 4. Model showing the rationale of the latency hypothesis. Postsynaptic potentials (PSPs) illustrate how increasing the intensity to the inhibitory ear causes the strength of inhibition to increase, the latency of inhibition to shorten, and the duration of inhibition to lengthen. A: PSPs and IID function of a hypothetical IID-sensitive cell with excitatory and inhibitory strengths and latencies that are matched at the IID of complete inhibition (*). B: PSPs and IID function of a hypothetical cell with a mismatch in the latencies of excitation and inhibition. Compared with the cell in A, a greater intensity is required at the inhibitory ear to overcome the latency mismatch and achieve complete inhibition (*). Hence, the strength of inhibition is greater than the strength of excitation at the IID of complete inhibition. C: same cell as in B except that the signal to the inhibitory ear has been electronically advanced to compensate for the mismatch in latencies, thus allowing the cell to reach complete inhibition with a lower intensity at the inhibitory ear (*) than would be required with simultaneous stimulation.

Figure 4B shows how the same hypothetical cell would behave if the latency of inhibition was slightly longer than in the previous example. In this case, excitation and inhibition are not coincident at the IID that evokes equal strengths from the two ears, a situation that allows the cell to discharge at this IID. As a consequence, the IID that evokes equal strengths from the two ears is no longer the IID of complete inhibition. However, an additional intensity increment to the inhibitory ear shortens the latency of inhibition via an intensity-dependent latency shift, thereby creating coincidence and silencing the cell (Fig. 4B, *). For cells like this, the mismatch in latencies causes the IID of complete inhibition to occur with a more intense sound to the inhibitory ear than would otherwise be required. Thus the latency mismatch plays a part in shaping the IID sensitivity of these cells. For convenience, cells like this are referred to as "neurons with mismatched latencies," meaning that inhibition is delayed relative to excitation when the strengths of excitation and inhibition are equal.

A further prediction of the latency hypothesis for neurons with mismatched latencies is illustrated by the drawings in Fig. 4, B and C. An inhibitory signal less intense than that at the IID of complete inhibition should generate an inhibition that, although not coincident with the excitation, is equal to it in strength (middle pair of PSPs in Fig. 4B). Hence, we should be able to reduce the intensity at the inhibitory ear below that which generates the IID of complete inhibition and then electronically advance the inhibitory signal to establish coincidence of excitation and inhibition, reestablishing complete inhibition. Figure 4C illustrates this procedure. In effect, electronically advancing the inhibitory signal mimics the effects of an intensity-dependent latency shift. The difference between the IID that silences the cell when the inhibitory signal is advanced and the IID of complete inhibition obtained when the signals are presented simultaneously should indicate the extent to which a delayed inhibition shapes a cell's IID function via an intensity-dependent latency shift. For cells with matched latencies, the same manipulations produce a different result. Because those cells have an equally strong excitation and inhibition at the IID of complete inhibition, reducing the intensity at the inhibitory ear generates an inhibition that is weaker than the excitation so that complete inhibition cannot be achieved, even when the two inputs are brought into temporal coincidence.

To test the 50 IC cells for latency mismatches, a number of IID functions (usually 20) were generated for each cell with the signal to the inhibitory ear electronically time shifted. Time shifts were varied from function to function in 100- or 200-µs increments. The rationale was that if inhibition was delayed relative to excitation, then advancing the inhibitory signal by an amount equal to that delay should compensate for the latency mismatch, and we should be able to see an effect in the spike counts and the IID functions of the cells (as for the hypothetical cell in Fig. 4C). ITDs were varied systematically in both directions (inhibitory ear leading and lagging) such that a latency mismatch in either direction could be visualized.

For 44 of the 50 IC cells, the manipulations just described indicated that the latency of inhibition was longer than the latency of excitation and that the mismatch in latencies affected the IID of complete inhibition as predicted by the latency hypothesis. The remaining six cells did not display latency mismatches. Figure 5 shows IID functions for three of the cells that appeared to have mismatched latencies and one of the cells that appeared to have matched latencies. For cell A, a cell with mismatched latencies, the solid curve shows the IID function when the signals were presented simultaneously to the two ears. The IID of complete inhibition occurred at an IID of 20 dB (inhibitory ear more intense). The dashed curve shows the IID function when the inhibitory signal was electronically advanced by 400 µs. With this time shift, the cell's IID function reached complete inhibition at a lower intensity to the inhibitory ear (at an IID of 10 dB).

View larger version (26K): [in this window] [in a new window]   FIG. 5. IID functions for 3 of the cells that appeared to have mismatched excitatory and inhibitory latencies (A-C) and 1 of the cells that appeared to have matched latencies (D). Solid curves shows IID functions measured when the signal was presented simultaneously to the 2 ears. Dashed curves shows IID functions measured when the signal to the inhibitory ear was advanced electronically. Note that for cells with mismatched latencies, the time shift that generated the greatest change in the IID of complete inhibition is shown. Bottom: rate-level functions for each cell.

The data presented above for cell A indicates that when the signals were presented with a time shift of 400 µs, equally strong inputs arrived at the IC neuron at an IID of 10 dB, achieving complete inhibition. When no time shift was used, the same IID of 10 dB evoked spikes. Presumably this was because the inhibition arrived ~400 µs later than the excitation and the mismatch in latencies allowed spikes to occur. To establish equivalent latencies when the signals were presented simultaneously, a higher intensity at the inhibitory ear was required. Thus for this cell, a mismatch in latency of ~400 µs appears to have shifted the IID of complete inhibition by 10 dB.

Two additional examples of cells that appeared to have mismatched latencies are shown in Fig. 5, B and C. For these cells, the IID of complete inhibition shifted by 20 and 30 dB when the signal to the inhibitory ear was advanced by 400 and 1,000 µs, respectively. Note that, for each of the cells with mismatched latencies in Fig. 5, the time shift that generated the greatest change in the IID of complete inhibition is shown. Greater time shifts resulted in IID functions that did not go to zero spikes, presumably because, at some point, time shifts surpass the amount of the latency mismatch such that coincidence of excitation and inhibition cannot be established at any IID.

Six of the 50 IC cells behaved differently from those described previously, appearing to have matched latencies. The IIDs of complete inhibition for these six cells did not change when the signal to the inhibitory ear was advanced. Rather, the IID functions of these cells were unable to achieve complete inhibition when the signal to the inhibitory ear was advanced by even small increments. Data from one such cell is presented in Fig. 5D. For this cell, the IID of complete inhibition was 20 dB when the signals were presented simultaneously. When the signal to the inhibitory ear was advanced by just 100 µs, the cell was no longer completely inhibited. Presumably, excitation and inhibition arrived coincidentally when their strengths were equal (at the IID of complete inhibition) so that advancing the signal to the inhibitory ear disrupted the coincidence of excitation and inhibition, allowing spikes to occur.

Latency mismatches, as determined earlier, were much more prevalent among the 50 IC cells compared with the 50 LSO cells reported on previously. As mentioned above, 44 of the 50 IC cells (88%) appeared to have latency mismatches. When the same manipulations were performed on the 50 LSO cells, only 27 cells (54%) showed latency mismatches. The difference in the proportion of cells with latency mismatches in the IC and LSO was statistically significant(² = 23.27, df = 1, P < 0.0001).

Among those IC and LSO cells with mismatched latencies, latency mismatches shifted IIDs of complete inhibition by greater amounts among the IC cells compared with the LSO cells. As seen in the three examples of cells with mismatched latencies in Fig. 5, latency mismatches caused varying degrees of shift in the IIDs of complete inhibition: 10 dB for cell A, 20 dB for cell B, and 30 dB for cell C. The distribution of shifts in IIDs of complete inhibition for the population of IC cells is shown in Fig. 6, top. Figure 6, bottom, shows the distribution for the 50 LSO cells. For the 44 IC cells with mismatched latencies, shifts in the IID of complete inhibition ranged from 5 to 40 dB with an average shift of 16.1 dB. In contrast, shifts in the IID of complete inhibition among the 27 LSO cells with mismatched latencies ranged from 5 to 20 dB with an average shift of 9.1 dB. An unpaired t-test revealed that the difference inthe two distributions (not including cells with matchedlatencies) was statistically significant (df = 69, t = 3.71,P < 0.001).

View larger version (31K): [in this window] [in a new window]   FIG. 6. Distributions of shifts in IIDs of complete inhibition resulting from advancing the signal to the inhibitory ear for the IC cells (A) and the LSO cells (B). Bars at 0 dB correspond to cells with matched latencies.

The differences between the IC cells and the LSO cells in the number of cells that had mismatched latencies, and the amounts by which those mismatches shifted IIDs of complete inhibition suggest that latency mismatches might account, to some extent, for the difference in IID sensitivity observed between the two nuclei. To further explore this idea, the distributions of IIDs of complete inhibition were recalculated for the IC cells and the LSO cells when latency mismatches were factored out by electronically time shifting the signals. In other words, the distributions were recalculated based on the IIDs of complete inhibition when the signals were time shifted to compensate for mismatches. The rationale was that if latency mismatches account for the difference in distributions between the IC cells and the LSO cells, then one might expect that effectively removing the mismatches would make the distributions more similar. Figure 7 displays the distributions of IIDs of complete inhibition obtained from the IC and LSO cells when latency mismatches were effectively removed via time shifts. In general, both distributions were less lateralized compared with the distributions of IIDs of complete inhibition when the signals were presented simultaneously (Fig. 2, B and C). In this case, the average IID of complete inhibition for the IC cells was 1.7 dB and the average IID of complete inhibition for the LSO cells was 4.0 dB. More importantly, the IIDs of complete inhibition obtained for the IC cells with latency mismatches removed were not significantly different from the IIDs of complete inhibition for the LSO cells with latency mismatches removed (df = 98, t =0.773 P = 0.44), supporting the idea that latency mismatches are, to some extent, responsible for the differences in IIDs of complete inhibition between the IC and the LSO.

View larger version (30K): [in this window] [in a new window]   FIG. 7. Distribution of IIDs of complete inhibition obtained from the IC cells (A) and the LSO cells (B) when latency mismatches were effectively removed via time shifts. See text for details.

To summarize this section of the results, 88% of the IC cells appeared to have mismatches in their excitatory and inhibitory latencies that lateralized their IID sensitivities (IIDs of complete inhibition). Because only 54% of the LSO cells had mismatched latencies and because the mismatches for the LSO cells resulted in smaller shifts in the IIDs of complete inhibition compared with the IC cells, latency mismatches appear to be a factor in creating the difference in sensitivity between LSO and IC cells. The next section will assess several aspects of the observed latency mismatches in more detail.

ITD functions were used to characterize latency mismatches

The previous section described how manipulating the timing of the binaural stimuli revealed latency mismatches for excitation and inhibition in the majority of IC cells. The present section describes how ITD functions were used to infer three measures relevant to the observed latency mismatches. The three measures examined were the magnitude of the latency mismatches, the magnitude of the intensity-dependent latency shifts that brought mismatched latencies into coincidence, and the duration of inhibition at the IID of complete inhibition. As the following paragraphs will describe, these measures support the finding that latency mismatches were greater for the IC cells compared with the LSO cells. Furthermore, the ITD functions presented below clearly illustrate how mismatched latencies and intensity-dependent latency shifts together act to lateralize IID sensitivity.

ITD functions were generated in the following way. First, the intensity at each ear was fixed to establish a given IID, then the relative timing of the signals to the two ears was varied in 100- or 200-µs increments. By convention, positive ITDs indicate that the signal at the inhibitory ear was advanced relative to the signal at the excitatory ear, whereas negative ITDs indicate that the signal at the excitatory ear was advanced. These procedures were repeated at a variety of fixed IIDs, including the IID of complete inhibition. The data used here to generate ITD functions came from the same data set that was used in the previous section to generate IID functions at different ITDs.

The next paragraphs first will describe the shape of the ITD functions and how these functions differed for cells with mismatched latencies compared with cells with matched latencies. After that, a description of the three measures outlined above will be presented, and finally those measures will be compared with the same measures previously obtained from the LSO cells.

The ITD functions presented in Fig. 8 illustrate how neurons with mismatched latencies and neurons with matched latencies responded to varying ITDs. The curves in Fig. 8A were measured from a cell that had mismatched latencies as described in the previous section: its IID of complete inhibition shifted when the signal to the inhibitory ear was advanced relative to the signal to the excitatory ear. For this cell, the IID function at the top (measured when the ears were stimulated simultaneously) shows that the IID of complete inhibition occurred at 5 dB (*). The ITD function drawn with * was measured using the same intensities that generated the IID of complete inhibition. Consistent with the IID function, the cell responded with zero spikes when the stimulus was presented simultaneously to the ears (ITD of 0 µs). However, delaying the signal to the inhibitory ear relative to the signal at the excitatory ear (negative ITDs) appeared to disrupt coincidence of excitation and inhibition, allowing spikes to occur. A different result occurred when the signal at the inhibitory ear was advanced relative to the signal at the excitatory ear (positive ITDs): the cell remained completely inhibited over a range of 1,500 µs. In other words, for these intensities, the effective inhibition had a duration of ~1,500 µs so that the signal to the inhibitory ear had to be advanced by >1,500 µs before excitation could express spikes.

View larger version (51K): [in this window] [in a new window]   FIG. 8. IID functions (top), selected ITD functions (middle), and rate-level functions (bottom) for 4 IC cells. A and B: cells that had mismatched latencies; C and D: cells that had matched latencies. Top: IID function for each cell. IID of complete inhibition is indicated (*). Middle: 3 ITD functions for each cell. Negative values on the x axis indicate that the signal to the inhibitory ear was delayed relative to the signal at the excitatory ear, whereas positive values indicate that the signal to the inhibitory ear was advanced relative to the excitatory ear. Each function was measured using a different IID (i.e., a different combination of intensities at the 2 ears). For each cell, the ITD function taken with the intensities that generated the IID of complete inhibition is drawn (*). Note that these curves are U shaped for the cells with mismatched latencies (A and B), and they are V shaped for the cells with matched latencies (C and D). Also note that each cell was capable of generating both U-shaped and V-shaped functions, depending on the intensities presented to the ears.

The shape of the ITD function for cell A changed dramatically when the intensity at the inhibitory ear was reduced below that which generated the IID of complete inhibition (Fig. 8,  on the IID function). The ITD function drawn with  illustrates this result. The key feature of this function is that it did not go to zero spikes for an ITD of 0 µs, rather it went to zero spikes for a positive ITD when the signal to the inhibitory ear was advanced by 400 µs relative to the signal at the excitatory ear. This function also differed from the one previously described in that only one of the ITDs presented, 400 µs, generated a complete suppression of spikes. This result indicates that, with the intensities used here, an ITD of 400 µs brought an excitation and inhibition with equal strengths and durations into coincidence such that even relatively small shifts away from that ITD disrupted coincidence and allowed the cell to spike.

The third ITD function presented for cell A () shows the effect of presenting an even less intense signal to the inhibitory ear ( on the IID function). The decreased intensity at the inhibitory ear resulted in an inhibition that was not strong enough to silence the cell at any ITD.

My interpretation of the three ITD functions shown for cell A is as follows. At the IID of complete inhibition (function with *), there is a coincidence of inputs from the two ears, but both the strength and duration of the inhibition are greater than those of the excitation (e.g., Fig. 4B). Thus advancing the inhibitory signal continued to generate a complete inhibition of spikes because a later inhibitory component of equal strength to the excitation was still coincident with the excitation. The V-shaped ITD function (function with ) indicates that an inhibitory signal less intense than that at the neuron's IID of complete inhibition generated an inhibition that, although not coincident with the excitation, was equal to it in strength. Hence, electronically advancing the inhibitory signal established coincidence and, thus, complete inhibition (e.g., Fig. 4C). In effect, this manipulation mimicked the effects of an intensity-dependent latency shift. Presumably, the difference between the IID that silenced the cell when the inhibitory signal was advanced (+5 dB for cell A in Fig. 8), and the IID of complete inhibition obtained when the signals were presented simultaneously (5 dB for cell A) indicates the extent to which a mismatch in latencies shaped the cell's IID function via an intensity-dependent latency shift. The same effects were observed for each of the 44 IC cells classified as having latency mismatches, and an additional example is shown in Fig. 8B.

Cells with matched latencies generated the same types of ITD curves as did the cells with mismatched latencies, as is illustrated by the ITD curves for cells C and D, Fig. 8. The difference between cells with matched and mismatched latencies was that setting the intensities to those that generated the IID of complete inhibition when the signals were presented simultaneously resulted in a V-shaped ITD function (*) for cells with matched latencies, as opposed to the U-shaped ITD functions observed for the cells with mismatched latencies. However, it was possible to generateU-shaped ITD functions for cells with matched latencies by raising the intensity at the inhibitory ear (cells C and D, ). Hence, it appears that, at the IID of complete inhibition for cells C and D, inhibition and excitation had the same latencies and thus did not need to rely on intensity-dependent latency shifts to establish coincidence. Each of the six cells classified as cells with matched latencies behaved in this manner.

IC cells were characterized by greater latency mismatches, greater intensity-dependent latency shifts, and longer durations of inhibition compared with the LSO cells

The magnitude of latency mismatch was estimated for each of the 44 IC cells with mismatched latencies. Figure 9A shows how this estimation was derived from the ITD functions. The U-shaped ITD function was taken with the intensities that generated complete inhibition. Hence, this curve goes to zero spikes when the signals to the two ears were presented simultaneously at an ITD of 0 µs. TheV-shaped ITD function was taken with a lower intensity at the inhibitory ear (the lowest intensity that generated a complete inhibition). As described above for Fig. 8, this manipulation generated a curve that went to zero spikes at a single, positive ITD. This ITD, corresponding to the zero point on the V-shaped function, was taken to be the amount by which the inhibition lagged the excitation because this ITD appeared to bring the inhibition and excitation into coincidence. In other words, it was assumed that this time shift compensated for a mismatch in latencies that otherwise prevented coincidence of an equally strong excitation and inhibition. Using this estimation procedure, the latency mismatch for the cell in Fig. 9A was estimated to be 400 µs.

View larger version (22K): [in this window] [in a new window]   FIG. 9. Example showing how interaural time difference (ITD) functions were used to estimate the magnitudes of latency mismatches, the magnitudes of intensity-dependent latency shifts, and the durations of inhibition. See text for more detail.

The magnitude of latency mismatch was calculated in the same way for each of the 44 IC cells with mismatched latencies, and the distribution of latency mismatches is shown in Fig. 10, top. On average, inhibition arrived 1,364 µs later than excitation when the excitation and inhibition had equal strengths. The corresponding distribution for the 27 LSO cells with latency mismatches is presented in Fig. 10, bottom. As the two distributions show, latency mismatches were generally much larger among the IC cells compared with the LSO cells, and the difference was statistically significant (df = 69, t = 4.216, P < 0.0001; cells with matched latencies are indicated by the bars at 0 µs on the x axis but were not included in the statistical comparison).

View larger version (20K): [in this window] [in a new window]   FIG. 10. Distributions of latency mismatches for the IC cells (top) and the LSO cells (bottom).

ITD functions also were used to estimate the magnitude of intensity-dependent latency shifts for each cell. The curves in Fig. 9 can again be used to explain how this evaluation was made. As described above, the ITD that generated zero spikes on the V-shaped function, 400 µs in this example, was assumed to correspond to the amount of mismatch in the latencies of excitation and inhibition. When the intensity at the inhibitory ear was increased by 10 dB SPL (theU-shaped function), the range of ITDs over which complete inhibition occurred expanded in both directions. For the purposes of estimating an intensity-dependent latency shift, the important feature of this function is that zero spikes first occurred at an ITD of 0 µs, indicating that the latency of inhibition was effectively equal to the latency of excitation for the intensities used to generate this U-shaped function. Hence, for the example cell in Fig. 9A, increasing the intensity at the inhibitory ear from 20 dB SPL (V-shaped function) to 30 dB SPL (U-shaped function) shortened the latency of inhibition by 400 µs, indicating that this cell showed an intensity-dependent latency shift of 400 µs/10 dB or 40 µs/dB. For the 50 IC cells, the distribution of intensity-dependent latency shifts is shown in Fig. 11, top. On average, the latency of inhibition was estimated to shorten by 59.5 µs/dB for the intensities used here. The corresponding distribution from the 50 LSO cells is presented in Fig. 11, bottom. As the two distributions show, the range of intensity-dependent latency shifts was generally very similar for the two populations. However, on average, the latency of inhibition was estimated to shorten to a lesser degree (41 µs/dB) for the LSO cells, and the difference between the two populations was statistically significant (df = 98, t = 2.882, P < 0.01). The average latency shift reported here for the IC is very similar to that reported previously for the IC of the same species (Pollak 1988) (mean = 47 µs/dB). However, somewhat greater latency shifts were reported for cat IC cells (Yin et al. 1985) (mean = 85 µs/dB).

View larger version (29K): [in this window] [in a new window]   FIG. 11. Distributions of intensity-dependent latency shifts for the IC cells (top) and the LSO cells (bottom).

The third value that was estimated from the ITD functions was the duration of inhibition at the IID of complete inhibition. The rationale for taking this measure comes from previous studies that indicate the effective duration of inhibition increases with stimulus intensity. Because the IC cells studied here had IIDs of complete inhibition that generally required greater intensities at the inhibitory ear compared with the LSO cells, one might expect that the inhibition to the IC cells had a longer duration than the inhibition to the LSO cells. The procedures used to make this estimate again can be illustrated with Fig. 9. In this case, only one ITD function was used: the one that was measured with intensities that generated the IID of complete inhibition on the IID function (the U-shaped ITD function in Fig. 9B). The assumption here was that the range of ITDs over which the cell was completely silent corresponded to the duration of the inhibition (at least when the inhibition was strong enough to completely silence the cell). To be more precise, what this procedure actually measured was the excess duration of inhibition over excitation. For the cell in Fig. 9B, the excess duration of inhibition was estimated to be 1,400 µs. Because cells with matched latencies had V-shaped ITD functions that reached zero spikes at only one point on the function, the excess duration of inhibition for those cells was 0 µs. The distribution of excess inhibitory durations for the 50 IC cells is shown in Fig. 12, top. On average, the excess duration of inhibition at the IID of complete inhibition was 3,105 µs. The corresponding distribution from the LSO cells is presented in Fig. 12, bottom. The average duration of inhibition for the LSO cells was 406 µs. Hence, inhibitory durations were generally much longer among the IC cells compared with the LSO cells and the difference was statistically significant (df = 98, t = 4.458, P < 0.0001).

View larger version (21K): [in this window] [in a new window]   FIG. 12. Distributions of excess inhibitory durations at the IID of complete inhibition for the IC cells (top) and the LSO cells (bottom).

It appears that long-lasting inhibition may be a common feature of IC cells because it also has been reported in the cat (using clicks) (Carney and Yin 1989) and in the Mexican free-tailed bat (using 20-µs tones) (Klug et al. 1997). Those studies reported even longer durations of excess inhibition over excitation (from 2,000 or 3,000 µs up to >100 ms), and the authors suggested that it may play a role in the precedence phenomenon (Yin 1994).

Other ways in which IC cells differed from LSO cells

There were several features, other than IID sensitivity, that differed between IID-sensitive cells in the IC and the LSO. One difference was that some cells in the IC did not respond to the 2-ms, 10-kHz sweep, whereas all of the cells encountered in the LSO responded well to that stimulus. Seven IC cells that were not included in the preceding analyses responded to sweeps that were either >2 ms or <10 kHz but not to 10-kHz sweeps that were 2-ms long. The average IID of complete inhibition for these 7 cells was 15.7 dB, a value similar to the average IID of complete inhibition of 18.0 dB from the 50 IC cells that responded to the 2-ms, 10-kHz sweeps. Two additional cells did not respond to any sweep stimulus.

The two nuclei differed in how they responded to relatively high intensities presented monaurally to the excitatory ear. Each of the 50 LSO cells had monotonic rate-level functions in that their spike counts progressively increased with the intensity to the excitatory ear, or they saturated and reached a plateau. In contrast, 25 of the 50 IC cells (50%) had nonmonotonic rate-level functions that declined by25-100% of the peak spike count at the highest intensities presented. The decrease in spike count is associated with an inhibitory input driven by the contralateral ear, the ear referred to in this manuscript as the "excitatory" ear (e.g., Pollak and Park 1993). The contralaterally driven inhibition usually is recruited at higher intensities than the contralaterally driven excitation. However, the nonmonotonic response pattern indicates that the contralateral input has both excitatory and inhibitory components. In the present study, excitation and inhibition were defined operationally by a cell's net output. However, the underlying inputs can be somewhat more complex than the net output might indicate. [In fact, mixed excitatory/inhibitory inputs driven by the same ear are found among both IC and LSO cells (e.g., Kuwada et al. 1997; Wu and Kelly 1991)].

To determine how tonicity might have affected the response features measured here, I compared the 25 monotonic and 25 nonmonotonic IC cells for five key response features, and the results are presented in Table 1. The distributions of IIDs of complete inhibition when stimuli were presented simultaneously were not significantly different for monotonic and nonmonotonic IC cells nor were the magnitudes of latency mismatches or intensity-dependent latency shifts. Interestingly, the nonmonotonic cells showed significantly longer durations of inhibition (driven by the ipsilateral, "inhibitory" ear) and significantly larger changes in IID of complete inhibition with time shifts compared with themonotonic cells (Table 1). However, it is unclear what, if any, causal relationship may exist between these two binaural features and the tonicity of the contralateral input. (Comparisons also were made between the LSO cells and the subpopulations of monotonic IC cells and nonmonotonic IC cells. In every case, the values for the LSO cells were significantly different from those of the IC cells, as reported in the previous sections when the monotonic and nonmonotonic IC cells were pooled.)

Another difference between the IC and the LSO was that some cells in the IC showed a facilitated response to particular IIDs. For each of the LSO cells, the IID functions peaked and plateaued at IIDs favoring the excitatory ear with spike counts that closely approximated the spike counts evoked when the excitatory ear was stimulated monaurally. On the other hand, seven of the IC cells (14%) had IID functions that peaked near an IID of 0 dB with spike counts that were >25% greater than those evoked by monaural stimulation of the excitatory ear. This type of facilitated response for IC cells has been documented in a number of previous investigations (e.g., Fuzessery and Pollak 1985; Irvine and Gago 1990; Park and Pollak 1994; Semple and Kitzes 1987). The seven facilitated cells did not differ from the nonfacilitated cells on the response features measured here.

There were two features that differed between the IC cells and the LSO cells that were not readily apparent when sweep stimuli were used but were obvious when tone stimuli were used. First, virtually all of the LSO cells had a sustained response pattern to the 60-ms tones used as search stimuli, whereas the majority of the IC cells had a phasic onset response. Second, with sweep stimuli, the IID function for each IC cell and each LSO cell declined to zero spikes as the intensity to the inhibitory ear was increased. However, when tone stimuli were used this was not the case. Although each of the LSO cells also could reach complete inhibition with tone stimuli, the IID functions for 15 of the 50 IC cells did not reach complete inhibition when tone stimuli were used. This finding raises serious questions about the generalizability of data collected with only one type of stimulus, and a more detailed analysis of these cells is presented in two subsequent reports (Park et al. 1998; J. P. Oswald,A. Klug, and T. J. Park, unpublished data).

	DISCUSSION

Abstract Introduction Methods Results Discussion References

The main findings of this report are IID sensitivity differs between the IC and the LSO, two principal centers in the ascending IID pathway, and latency mismatches of binaural inputs appear to play a substantial role in establishing that difference. This second finding suggests that the majority of IID-sensitive cells in the IC do not derive their IID sensitivity directly from the LSO despite the prominent projection from LSO to IC. The reason is that direct projections from the LSO would preserve the relative excitatory and inhibitory latencies established in that nucleus, but the latency mismatches reported here for collicular cells were much larger than those observed in the LSO.

Support for different LSO/IC sensitivities from other studies

Although the present report represents the first systematic comparison of IID sensitivity between the LSO and the IC, the results are consistent with previously published data from those nuclei. Population data for IID sensitivity in both LSO and IC are available for only one other speciesthe mustache bat. For the mustache bat, the average IID of complete inhibition reported for the LSO was 13 dB (Park et al. 1997), and the average IID of complete inhibition for the IC appears to be more lateralized to approximately 22 dB (Park and Pollak 1993). However, the value for the mustache bat IC is only an estimate: that report did not use the IID of complete inhibition as an index of IID sensitivity, but rather it used the 50% point of the IID function. I estimated IIDs of complete inhibition by adding 15 dB to the reported 50% points because 15 dB is approximately one-half the dynamic range of an IC cell's IID function (Irvine and Gago 1990; figures from Park and Pollak 1993).

Data from the gerbil LSO and the cat IC are also consistent with the present findings. The distribution of IIDs of complete inhibition for the gerbil LSO (Sanes and Rubel 1988, Fig. 15) appears to closely match the distribution reported here for the free-tailed bat's LSO. On the other hand, the distribution estimated for the cat IC (Irvine and Gago 1990) is more lateralized and closely approximates that reported here for the free-tailed bat's IC (again, I estimated IIDs of complete inhibition for the cat data, which were reported as 50% points).

IID sensitivity is shaped to a large degree within the IC

The idea that IID sensitivity is shaped to a large degree within the IC is supported by a number of recent studies. Some of these studies have combined extracellular recording with micro-iontophoresis of inhibitory transmitter antagonists (Faingold et al. 1989; Klug et al. 1995; Park and Pollak 1993; Vater et al. 1992). The antagonists functionally disable inhibitory inputs to individual IC cells by blocking postsynaptic receptors. The results showed that although the IID sensitivity of some cells was unaffected, it changed in many others, and in some cases IID sensitivity was eliminated. Because the excitatory inputs from the LSO presumably are unaffected by the inhibitory transmitter antagonists, the authors of these reports concluded that many IC cells derive their sensitivity in part from inhibitory inputs acting directly on collicular cells. For the cells the IID sensitivity of which was eliminated by the antagonists, IID sensitivity appears to be entirely established in the IC.

A second line of investigation also supports the idea that IID sensitivity is shaped to a large degree within the IC. These studies involve a reversible inactivation of the dorsal nucleus of the lateral lemniscus (Faingold et al. 1993; Li and Kelly 1992), a nucleus that sends a prominent inhibitory projection to the IC. Temporarily inactivating this nucleus by injection of kynurenic acid or lidocaine changed the IID sensitivity of cells in the contralateral IC, reducing the effectiveness of stimulation to the inhibitory ear or, in some cases, eliminating IID sensitivity completely. These results indicate that the inhibition originating from the dorsal nucleus of the lateral lemniscus plays a part in shaping IID sensitivity in collicular cells.

Mismatched latencies appear to be a common feature among many IID-sensitive cells

In addition to showing a difference in IID sensitivity between the LSO and the IC, the present study also yielded data indicating that the underlying mechanism for the difference involves more than simply adding direct inhibitory inputs onto IC cells. The data presented here suggest that latency mismatches also play a crucial role. The idea of a mismatch in latencies between the excitatory and inhibitory inputs to an IID-sensitive cell is well established. Lloyd Jeffress first suggested that the relative timing of the inputs from the two ears might play a role in shaping IID sensitivity in auditory neurons (Jeffress 1948). Several researchers (Pollak 1985; Tsuchitani 1988; Yin et al. 1985) have elaborated on this idea, which has become known commonly as the latency hypothesis. In particular, three recent studies support the present finding that latency mismatches shape IID sensitivity in both the LSO and the IC. Joris and Yin (1996) and Irvine et al. (1998) manipulated both IIDs and ITDs while recording from cat and rat LSO, respectively, and Irvine et al. (1995) used similar procedures in rat IC. Although these researchers did not systematically quantify latency mismatches on a population level, their data, in terms of the IID and ITD functions shown, are extremely similar to the data presented here.

It is interesting that the latency mismatches observed in the present study were always in the same directionthe ipsilaterally driven inhibition arrived later than the contralaterally driven excitation when the inhibition and excitation had equal strengths. This finding differs somewhat from the results of a previous study of low-frequency cells in the cat IC (Carney and Yin 1989). Those researchers reported ipsilaterally driven inhibitory components that arrived later than the contralateral excitation, as reported here. However, for 17% of their contralaterally excited cells, they also observed ipsilaterally driven inhibitory components that preceded the excitation. Their data indicated that, for lowfrequency cells in the cat IC, the preceding inhibitorycomponents contribute to shaping ITD sensitivity.

What happens to the information coming from the LSO?

An interesting question raised by this and related studies is, what happens to the information coming from the LSO to the IC? The most likely explanation appears to be the same one previously suggested by the authors of the iontophoresis and reversible inactivation studies. As described earlier, there appears to be some IC cells that receive and accurately reflect IID sensitivity established in the LSO and some IC cells that appear to receive none of their IID sensitivity from the LSO. However, the majority of IC cells appear to receive IID-sensitive properties from the LSO and modify them with additional inhibitory inputs driven by the inhibitory ear. Although not contradicting this scenario, the data presented here suggest a significant complication with regard to the latter group of cells that incorporate but modify LSO inputs. The reason is as follows: a direct input from the LSO would preserve the relative excitatory and inhibitory latencies established in that nucleus. An additional inhibitory input onto the IC cell, no matter when it arrived, could not affect the relative latencies of the excitation and inhibition already established in the LSO. Hence, simply adding an additional inhibitory input to the input from the LSO could not generate the large latency mismatches reported here. However, I propose that adding not only an additional ipsilaterally driven inhibitory input but also an additional contralaterally driven excitatory input could resolve this apparent anomaly. In this elaborated scenario, the additional excitatory projection driven by the excitatory ear would "override" the LSO projection in that it could continue to evoke spikes even when the LSO cell contributing its input is inhibited. Hence, although the input from the LSO would impact the overall spike count of the IC cell, the important latency mismatch would be between the non-LSO excitatory and inhibitory inputs synapsing on the IC cell, not those established in the LSO. The idea of an additional excitatory input onto IC cells is not inconsistent with our current understanding of the brain stem auditory system in that many IC cells receive numerous contralaterally driven excitatory inputs from a variety of lower centers (e.g., Irvine 1992; Oliver et al. 1997). The appeal of this proposed circuitry is that it can explain both the present finding of larger latency mismatches in the IC than in the LSO and the previous findings that blocking local inhibitory inputs usually reduces but does not eliminate the effectiveness of stimulating the "inhibitory" ear.

Functional significance

The greater proportion of cells with IID sensitivities corresponding to peripheral spatial locations in the IC compared with the LSO is consistent with one proposed role of the IC: involvement with control of head, eye, and pinnae movements for rapid orienting toward objects in space. This functional role for the IC recently was suggested by Casseday and Covey (1996) based on anatomic connections. The IC sends major direct and indirect projections to the deep layers of the superior colliculus (SC), which controls these orientation behaviors (Henkel and Edwards 1978; Jay and Sparks 1987; Masino and Knudson 1993: Sparks and Nelson 1987). The SC contains many bimodal cells that comprise both a visual and an acoustic map of space that are roughly in register. Both maps have a good representation of frontal space (i.e., relative to the fovea) (e.g., Jay and Sparks 1987). However, studies of the visual map in the SC indicate that the visual periphery is also well represented in that the SC has a proportionally greater representation of peripheral space compared with retinal ganglion cells and primary visual cortex. This enhancement of peripheral visual space in the SC arises from a greater ratio of peripheral to foveal inputs from the retinal ganglion cells and larger individual receptive fields for SC cells compared with retinal ganglion cells and cells in the primary visual cortex (reviewed by Stone 1983). The underlying assumption associated with the enhanced representation of peripheral visual space in the SC is that good sensitivity to peripheral space is necessary for orientation toward objects in peripheral space. Because the same assumption applies to the acoustic map in the SC, it may be that one function of the IC is to act as a processing station for increasing the peripheral representation of the acoustic map in the SC. The first step toward testing this hypothesis will be to compare IID sensitivities in the SC with those of the IC and LSO to determine if the sensitivities created in the IC correspond to those present in the SC.

	ACKNOWLEDGEMENTS

  I thank A. Klug, C. Comer, and two anonymous reviewers for very thoughtful and helpful suggestions. C. Resler provided excellent technical support and M. R. Kokomo provided excellent environmental support. Special thanks to Dr. George Pollak, my postdoctoral mentor, for sculpting my development as a neurobiologist.

  This work was supported by the Department of Biological Sciences, University of Illinois at Chicago, and National Institute of Deafness and Other Communication Disorders Grant DC-02850.

	FOOTNOTES

  Received 13 August 1997; accepted in final form 23 January 1998.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• BOUDREAU, J. C., TSUCHITANI, C. Binaural interaction in the cat superior olivary S-segment. J. Neurophysiol. 31: 442-254, 1968.[Free Full Text]
• BRUNSO-BECHTOLD, J. K., THOMPSON, G. C., MASTERTON, R. B. HRP study of the organization of auditory afferents ascending to the central nucleus of the inferior colliculus in the cat. J. Comp. Neurol. 97: 705-722, 1981.
• CAIRD, D., KLINKE, R. Processing of binaural stimuli by cat superior olivary complex neurons. Exp. Brain Res. 52: 385-399, 1983.[Medline]
• CANT, N., CASSEDAY, J. H. Projections from the anteroventral cochlear nucleus to the lateral and medial superior olivary nuclei. J. Comp. Neurol. 247: 457-476, 1986.[Medline]
• CARNEY, L. H., YIN, T.C.T. Responses of low-frequency cells in the inferior colliculus to interaural time differences of clicks: excitatory and inhibitory components. J. Neurophysiol. 62: 144-161, 1989.[Abstract/Free Full Text]
• CASSEDAY, J. H., COVEY, E. A. Neuroethological theory of the operation of the inferior colliculus. Brain Behav. Evol. 47: 311-336, 1996.[Medline]
• COVEY, E., KAUER, J. A., CASSEDAY, J. H. Whole-cell patch-clamp recording reveals subthreshold sound-evoked postsynaptic currents in the inferior colliculus of awake bats. J. Neurosci. 16: 3009-3018, 1996.[Abstract/Free Full Text]
• ERULKAR, S. Comparative aspects of sound localization. Physiol. Rev. 52: 237-360, 1972.[Free Full Text]
• FAINGOLD, C. L., BOERSMA ANDERSON, C. A., RANDALL, M. E. Stimulation or blockade of the dorsal nucleus of the lateral lemniscus alters binaural and tonic inhibition in contralateral inferior colliculus neurons. Hear. Res. 69: 98-106, 1993.[Medline]
• FAINGOLD, C. L., GEHLBACH, G., CASPARY, D. M. On the role of GABA as an inhibitory neurotransmitter in inferior colliculus neurons: Iontophoretic studies. Brain Res. 500: 302-312, 1989.[Medline]
• FUZESSERY, Z. M., POLLAK, G. D. Determinants of sound location selectivity in bat inferior colliculus: a combined dichotic and free-field study. J. Neurophysiol. 54: 757-781, 1985.[Abstract/Free Full Text]
• GLENDENNING, K. K., BAKER, B. N., HUTSON, K. A., MASTERTON, R. B. Acoustic Chiasm V: inhibition and excitation in the ipsilateral and contralateral projections of LSO. J. Comp. Neurol. 319: 100-122, 1992.[Medline]
• HENKEL, C. K., EDWARDS, S. B. The superior colliculus control of pinna movements in the cat: possible anatomical connections. J. Comp. Neurol. 182: 763-776, 1978.[Medline]
• IRVINE, D.R.F. Auditory brainstem processing. In: The Mammalian auditory pathway: Neurophysiology, edited by A. N. Popper, and R. R. Fay New York: Springer, 1992, p. 153-231.
• IRVINE, D.R.F., GAGO, G. Binaural interaction in high-frequency neurons in the inferior colliculus of the cat: effects of variation in sound pressure level on sensitivity to interaural intensity disparities. J. Neurophysiol. 63: 570-591, 1990.[Abstract/Free Full Text]
• IRVINE, D.R.F., PARK, V. N., MATTINGLEY, J. B. Responses of neurons in the inferior colliculus of the rat to interaural time and intensity differences in transient stimuli: implications for the latency hypothesis. Hear. Res. 85: 127-141, 1995.[Medline]
• IRVINE, D.R.F., PARK, V. N., AND MCCORMICK, L. Contributions of changes in the timing and amplitude of synaptic inputs to neural sensitivity to interaural intensity differences. In: Psychophysical and Physiological Advances in Hearing, edited by A. R. Palmer, A. Reese, A. Q. Summerfield, and R. Meddis. London: Whurr Publishers, 1998, p. 359-367.
• JAY, M. F., SPARKS, D. L. Sensorimotor integration in the primate superior colliculus. II. Coordinates of auditory signals. J. Neurophysiol. 57: 35-55, 1987.[Abstract/Free Full Text]
• JEFFRESS, L. A. A place theory of sound localization. J. Comp. Physiol. Psychol. 41: 35-39, 1948.[Medline]
• JORIS, P. X., YIN, T.C.T. Envelope coding in the lateral superior olive. I. Sensitivity to interaural time differences. J. Neurophysiol. 73: 1043-1062, 1995.[Abstract/Free Full Text]
• KLUG, A., BAUER, E. E., POLLAK, G. D. A direct demonstration of multiple long-lasting inhibitions on inferior colliculus neurons. Soc. Neurosci. Abstr. 23: 734, 1997.
• KLUG, A., PARK, T. J., POLLAK, G. D. Glycine and GABA influence binaural processing in the inferior colliculus of the mustache bat. J. Neurophysiol. 74: 1701-1713, 1995.[Abstract/Free Full Text]
• KUWADA, S., BATRA, R., YIN, T.C.T., OLIVER, D. L., HABERLY, L. B., STANFORD, T. R. Intracellular recordings in response to monaural and binaural stimulation of neurons in the inferior colliculus of the cat. J. Neurosci. 17: 7565-7581, 1997.[Abstract/Free Full Text]
• LI, L., KELLY, J. B. Inhibitory influence of the dorsal nucleus of the lateral lemniscus on binaural responses in the rat's inferior colliculus. J. Neurosci. 12: 4530-4539, 1992.[Abstract]
• MASINO, T., KNUDSON, E. I. Orienting head movements resulting from electrical microstimulation of the brainstem tegmentum in the barn owl. J. Neurosci. 13: 351-370, 1993.[Abstract]
• MILLS, A. W. Auditory localization. In: Foundations of Modern Auditory Theory, edited by and J. V. Tobias New York: Academic, 1972, vol. II, p. 303-348.
• MOORE, M. J., CASPARY, D. M. Strychnine blocks binaural inhibition in lateral superior olivary nucleus. J. Neurosci. 3: 237-242, 1983.[Abstract]
• OLIVER, D. L., BECKIUS, G. E., BISHOP, D. C., KUWADA, S. Simultanious anterograde labeling of axonal layers from lateral superior olive and dorsal cochlear nucleus in the inferior colliculus of the cat. J. Comp. Neurol. 382: 215-229, 1997.[Medline]
• PARK, T. J., GROTHE, B., POLLAK, G. D., SCHULLER, G., KOCK, U. Neural delays shape selectivity to interaural intensity differences in the lateral superior olive. J. Neurosci. 16: 6554-6566, 1996.[Abstract/Free Full Text]
• PARK, T. J., KLUG, A., OSWALD, J. P., AND GROTHE, B. A novel circuit in the bat's midbrain recruits neurons into sound localization processing. Naturwissenschaften In press.
• PARK, T. J., MONSIVAIS, P., POLLAK, G. D. Processing of interaural intensity differences in the LSO: the role of interaural threshold differences. J. Neurophysiol. 77: 2863-2878, 1997.[Abstract/Free Full Text]
• PARK, T. J., POLLAK, G. D. GABA shapes sensitivity to interaural intensity disparities in the mustache bat's inferior colliculus: implications for encoding sound location. J. Neurosci. 13: 2050-2067, 1993.[Abstract]
• PARK, T. J., POLLAK, G. D. Receptive fields in the inferior colliculus of the mustache bat are shaped by GABAergic inhibition. J. Neurophysiol. 72: 1080-1102, 1994.[Abstract/Free Full Text]
• POLLAK, G. D. Time is traded for intensity in the bat's auditory system. Hear. Res. 36: 107-124, 1988.[Medline]
• POLLAK, G. D., PARK, T. J. The effects of GABAergic inhibition on monaural response properties of neurons in the mustache bat's inferior colliculus. Hear. Res. 65: 99-117, 1993.[Medline]
• POLLAK, G. D., WENSTRUP, J. J., FUZESSERY, Z. M. Auditory processing in the mustache bat's inferior colliculus. Trends Neurosci. 9: 556-561, 1986.
• SAINT MARIE, R. L., BAKER, R. A. Neurotransmitter specific uptake and retrograde transport of [³H] glycine from the inferior colliculus by ipsilateral projections of the superior olivary complex and nuclei of the lateral lemniscus. Brain Res. 524: 244-253, 1990.[Medline]
• SANES, D. H. An in vitro analysis of sound localization mechanisms in the gerbil lateral superior olive. J. Neurosci. 10: 3494-3506, 1990.[Abstract]
• SANES, D. H., RUBEL, E. W. The ontogeny of inhibition and excitation in the gerbil lateral superior olive. J. Neurosci. 8: 682-700, 1988.[Abstract]
• SCHULLER, G., RADTKE-SCHULLER, S., BETZ, M. A stereotaxic method for small animals using experimentally determined reference profiles. J. Neurosci. Methods 18: 339-350, 1986.[Medline]
• SCHLEGEL, P. Calibrated earphones for the echolocating bat, Rhinolophus ferrumequinum. J. Comp. Physiol. [A] 118: 353-356, 1977.
• SEMPLE, M. N., KITZES, L. M. Binaural processing of sound pressure level in the inferior colliculus. J. Neurophysiol. 57: 1130-1147, 1987.[Abstract/Free Full Text]
• SPARKS, D. L., NELSON, J. S. Sensory and motor maps in the in the mammalian superior colliculus. Trends Neurosci. 10: 312-317, 1987.
• STONE, J. In: Parallel Processing in the Visual System. New York: Plenum Press, 1983.
• VATER, M., HABBICHT, H., KÖSSL, M., GROTHE, B. The functional role of GABA and glycine in monaural and binaural processing in the inferior colliculus of horseshoe bats. J. Comp. Physiol. [A] 171: 541-553, 1992.[Medline]
• WENSTRUP, J. J., FUZESSERY, Z. M., POLLAK, G. D. Binaural neurons in the Mustache bat's inferior colliculus. I. Responses of 60-kHz EI Units to dichotic sound stimulation. J. Neurophysiol. 60: 1369-1383, 1988.[Abstract/Free Full Text]
• WU, S. H., KELLY, J. B. Physiological properties of neurons in the mouse superior olive: membrane characteristics and postsynaptic responses studied in vitro. J. Neurophysiol. 65: 230-246, 1991.[Abstract/Free Full Text]
• WU, S. H., KELLY, J. B. Binaural interaction in the lateral superior olive: time difference sensitivity in the mouse brain slice. J. Neurophysiol. 68: 1151-1159, 1992.[Abstract/Free Full Text]
• YIN, T.C.T. Physiological correlates of the precedence effect and summing localization in the inferior colliculus of the cat. J. Neurosci. 14: 5170-5186.
• YIN, T.C.T., HIRSCH, J. A., CHAN, J.C.K. Response of neurons in the cat's superior colliculus to acoustic stimuli. II. A model of interaural intensity sensitivity. J. Neurophysiol. 53: 746-758, 1985.[Abstract/Free Full Text]
• ZOOK, J. M., CASSEDAY, J. H. Origin of ascending projections to inferior colliculus in the mustache bat, Pteronotus parnellii. J. Comp. Neurol. 207: 14-28, 1982.[Medline]

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society