| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Section of Neurobiology, Institute for Neuroscience and Center for Perceptual Systems, The University of Texas at Austin, Austin, Texas
Submitted 1 July 2005; accepted in final form 26 August 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
97% of the calls that were presented. Second, most INLL neurons are characterized by wide tuning curves and are unselective for species-specific calls. The third finding is that the IC population is strikingly different from the neuronal populations in the INLL and DNLL. Where DNLL and INLL neurons are unselective and respond to most or all of the calls in the suite we presented, most IC cells are selective for calls and, on average, responded to 50% of the calls we presented. Additionally, the selectivity for calls in the majority of IC cells, as well as their tuning and other response properties, are strongly shaped by inhibitory innervation. Thus we show that inhibition plays only limited roles in the DNLL and INLL but dominates in the IC, where the various patterns of inhibition sculpt a wide variety of emergent response properties from the backdrop of more expansive and far less specific excitatory innervation. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
), songbirds (Boettiger and Doupe 1998; Theunissen and Doupe 1998), and on the cortices of primates (Rauschecker and Tian 2000; Wang 2000; Winter and Funkenstein 1973) and bats (Esser et al. 1997; Kanwal 1999), because these species have a rich vocal repertoire. Only recently have studies begun to evaluate how species-specific communication calls are processed and represented in subcortical nuclei (Bauer et al. 2002; Klug et al. 2002; Pollak et al. 2003b; Portfors 2004; Suta et al. 2003). Such studies are significant because they not only reveal how lower nuclei process this information, but also because they provide insights into which response features might be created in lower nuclei and which are emergent properties of the cortex.
A subcortical nucleus of particular importance is the inferior colliculus (IC). The IC is a nexus in the auditory system because it is the common target of the projections from the majority of lower auditory nuclei (Casseday 2002; Pollak and Casseday 1986; Roth et al. 1978), from the opposite IC through its commissure (Aitkin and Phillips 1984; Malmierca et al. 1995, 2003; Moore et al. 1998), and from descending projections of the auditory cortex (Huffman and Henson 1990; Winer et al. 1998; Zhou and Jen 2000). The IC also provides the principal source of innervation to the medial geniculate body and thus indirectly to the auditory cortex (Clarey 1992; Wenstrup et al. 1994; Winer 1992).
Here we report on some response transformations that occur in three brain stem nuclei of Mexican free-tailed bats. The three nuclei are the IC, the dorsal nucleus of the lateral lemniscus (DNLL), and the intermediate nucleus of the lateral lemniscus (INLL). Each is a successively higher auditory nucleus, where the IC is the highest region, the DNLL is situated just below the IC, and the INLL is located just ventral to the DNLL. Both the DNLL and INLL send inhibitory projections to the IC. Particular attention was given to evaluating tuning curves of neurons in each nucleus, the degree to which lateral inhibition sharpens the tuning curves in each nucleus, and the role that inhibition, especially lateral inhibition, plays in shaping the responses of neurons in each nucleus to species-specific calls.
The impetus for this study follows from two previous studies that investigated the responsiveness of IC (Klug et al. 2002) and DNLL neurons (Bauer et al. 2002) in bats to species-specific calls. Most IC neurons are selective for calls in that they respond to only a subset of a repertoire of natural calls (Klug et al. 2002) and fail to respond to other calls, although the calls to which they fail to respond have suprathreshold energy in their excitatory tuning curves. This selectivity is largely caused by inhibition, because pharmacologically blocking inhibitory receptors results in a dramatic decrease in selectivity due to an increase in the cell's responsiveness to calls to which the neurons were unresponsive before inhibition was blocked. DNLL neurons, in contrast, are unselective for calls and respond to any call so long as the call has energy that stimulates its excitatory tuning curve. Moreover, the responses of DNLL cells to any complex signal, including species-specific calls, can be accurately predicted simply by convolving the responses evoked by tone bursts with the spectrogram of each signal. Thus processing in the DNLL is apparently accomplished largely by the linear integration of excitation, which suggests that inhibition, especially lateral inhibition, plays little or no role in shaping responses of DNLL neurons.
This report evaluates several issues concerned with the hierarchical processing of acoustic information in the INLL, DNLL, and IC that either were not addressed by previous studies or were raised but not answered by them. In this study, we document the response properties of DNLL neurons evoked by both simple (tonal) and complex stimuli (species-specific signals) and evaluate the same features when inhibition was blocked by iontophoretic application of antagonists of GABAA receptors and/or glycine receptors. As mentioned above, previous studies suggest that inhibition plays little or no role in shaping the responses of DNLL neurons to monaural stimulation, a suggestion inconsistent with the excitatory and inhibitory innervation of the DNLL, and one that we resolve in this study. We evaluated the same features in INLL and IC neurons, both before and while inhibition was blocked. Thus we evaluate 1) the ways in which DNLL and INLL neurons respond to both simple and complex signals; 2) the roles that inhibition plays in shaping responses to simple and complex stimuli in the DNLL compared with the roles that inhibition plays in shaping responses of INLL neurons; and 3) how the processing in the DNLL and INLL compares to the more complex processing in the IC, with particular attention given to comparing the roles of inhibition in the three nuclei. We show that inhibition plays only limited roles in the DNLL and INLL but dominates in the IC, where the various patterns of inhibition sculpt a wide variety of emergent response properties from the backdrop of more expansive and far less specific excitatory innervation.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Surgical and pharmacological procedures, electronic equipment, sound generation, and criteria for isolating single neurons are the same as those described in previous publications (Bauer et al. 2002; Klug et al. 2002). In brief, each Mexican free-tailed bat, Tadarida brasiliensis mexicana, was anesthetized with isoflurane inhalation (IsoFlo, Abbott Laboratories), and surgically prepared by reflecting the muscles and skin overlying the skull. Lidocaine (Abbott Laboratories) was applied topically to all open wounds. The surface of the skull was cleared of tissue, and a foundation layer of cyanoacrylate and small glass beads was placed on the surface. A small hole was made in the skull over the IC using landmarks visible through the skull. The bat was transferred to a heated recording chamber, where it was placed in a restraining cushion constructed of foam molded to the animal's body. The restraining cushion was attached to a platform mounted on a custom made stereotaxic instrument (Schuller et al. 1986). A small metal rod was cemented to the foundation layer on the skull and attached to a bar mounted on the stereotaxic instrument to ensure a uniform positioning of the head. A ground electrode was placed between the reflected muscle and the skin.
After the animal was fixed in the stereotaxic instrument, the electrode was positioned over the IC while viewed with an operating microscope. The electrode was advanced to a depth of 300 µm to ensure that recordings were obtained from neurons in the central nucleus of the IC. The electrode was subsequently advanced from outside of the experimental chamber with a piezoelectric microdrive (Burleigh 7121W). Several criteria were used to determine when the electrode was in the IC and when it exited the IC and entered the DNLL and subsequently the INLL. As the electrode was advanced through the IC, there was an abrupt change in the best frequency (BF; the frequency to which the unit or cluster of units was most sensitive) of the background activity at a depth of 1,500–1,600 µm. This change signaled that the electrode had left the IC and entered the DNLL. For the next 300–400 µm, both the multiunit activity and the single units encountered displayed prominent sustained activity in response to contralateral tone bursts, and this activity was strongly suppressed when tone bursts were presented simultaneously to the ipsilateral ear. A second abrupt change in BF and a change from binaural to monaural activity that was only influenced by sound to the contralateral ear signaled that the electrode left the DNLL and entered the INLL. To ensure that these changes in BF, as well as discharge patterns and binaural properties, indicated DNLL and INLL locations, we verified the electrode location from histological sections in several experiments after making small lesions by passing 5 µA of current for 10–15 min. If killed, the bat was removed from the stereotaxic apparatus and from the cushion. It was overdosed with the inhaled anesthetic IsoFlo and perfused immediately with 4% formaldehyde. After removal from the skull, the brain was placed in a 30% sucrose solution for 24–48 h for the purpose of cryoprotection. The brain was sectioned at 50 µm thickness on a freezing microtome, and the brain slices were counterstained with cresyl violet. Electrode positions within the DNLL or INLL were confirmed through visualization of the small lesion. The histological sections confirmed that the response features were reliable indicators of electrode location, as they were in previous studies of the DNLL in this bat and in mustache bats.
Recordings were begun after the bats recovered from the anesthetic, and thus all data were obtained from awake animals. The bats typically lay quietly during the remainder of the experiments. If they showed signs of discomfort, doses of the neuroleptic ketamine hydrochloride (1/40 dilution, 0.01 ml injection; Vetamine, Mallinckrodt Veterinary), were administered. All experimental procedures were in accordance with a protocol approved by the University of Texas Institutional Animal Care Committee.
Electrodes
"Piggyback" multibarrel micropipettes were used for recordings and iontophoresis of drugs (Havey and Caspary 1980). Multibarrel electrodes were pulled from a five-barrel blank (A-M Systems) and blunted to 15–20 µm. A single barrel pipette was attached to the five-barrel pipette and glued with cyanoacrylate so that the tip of the single barrel pipette protruded 10–15 µm from the blunted tip of the five-barrel pipette. The single-barrel micropipette was used for recording single unit activity and was filled with buffered 1 M NaCl and 2% Fast green (pH 7.4) to enhance the visibility of the electrode. One barrel of the five-barrel pipette was the balancing barrel and was filled with buffered 1 M NaCl and 2% Fast green. The other barrels were filled with solutions of bicuculline methiodide (Sigma, St. Louis, MO), an antagonist of GABAA receptors, and with the glycine receptor antagonist, strychnine HCl (both were 10 mM in 0.165 M NaCl, pH 3.0; Sigma). In some experiments, one barrel was also filled with L-glutamic acid (500 mM in dH2O, pH 9–10; Sigma). Drugs were retained in the electrode with a 15- to 20-nA retention of opposite polarity compared with the ejection current. For bicuculline and strychnine, retention currents were negative, and ejection currents were positive, while for glutamate, retention currents were positive and ejection currents negative. The drug and balancing barrels were connected via silver-silver chloride wires to a six-channel microiontophoresis constant current generator (Medical Systems Neurophore BH-2) that was used to generate and monitor ejection and retention currents. The sum channel was used to balance current in the drug barrels and reduce any influence of passing positive or negative current during iontophoresis. The recording barrel was connected by a silver-silver chloride wire to a Dagan AC amplifier (model 2400).
Acoustic stimuli
Tone bursts were digitally generated by a G4 Macintosh computer with custom built hardware and software. Tone bursts had 0.2-ms rise-fall times and durations ranging from 2.0 to 50.0 ms depending on the particular experiment. All tone bursts were set to the same peak intensity that corresponded to 70 dB SPL when measured between 20–25 kHz. The species-specific calls consisted of a suite of 10 Mexican free-tailed bat vocalizations (Fig. 5) that were previously recorded. These were the same suite of calls used in previous studies (Bauer et al. 2002; Klug et al. 2002). Eight of the 10 vocalizations were communication signals of various behavioral contexts [Social Calls (SC 1–8)], and the remaining two vocalizations were echolocation calls (EC 9–10). The calls were all at the same peak intensity that corresponded to the peak intensity of tone bursts and stored as AIFF files. Thus a library of 10 species-specific sounds was created and stored for later playback. All stimuli were uploaded from the Macintosh G4 into a custom made Downloadable Arbitrary Waveform Generator through a 24-bit digital interface (National Instruments DIO-24) and a digital distributor just before that particular sound's presentation. The acoustic signals were sent to custom-made electronic attenuators. The outputs of the attenuators were fed to earphones biased with 200 V DC. The design of the earphones was originally described by Schuller (1997), and each was tested for frequency-intensity response. These earphones are flat ±5 dB from about 10 kHz to 80 kHz. At the start of each experiment, the earphones were inserted into the funnel formed by the bat's pinnae and positioned adjacent to the external auditory meatus. The pinnae were folded onto the housing of the microphones and wrapped with Scotch tape. The acoustic isolation with this arrangement is 
40 dB.
|
After a cell was isolated, its BF and threshold at BF were obtained, followed by rate-level functions and tuning curve. Quality values (Q values) were calculated by dividing the neuron's BF by the bandwidth of the tuning curve at a specified level above threshold. The suite of 10 species-specific calls was presented at 30–50 dB above the neuron's threshold at BF to ensure that each call had suprathreshold energy that encroached on the neuron's tuning curve (Bauer et al. 2002; Klug et al. 2002). Peristimulus time histograms (PSTHs) or raster displays were generated from 20 presentations of each stimulus (bin width was 1.0 ms) presented pseudo-randomly at a rate of 4/s. Thresholds to tones were determined audiovisually. The criterion for a threshold response evoked by each species-specific call was three or more spikes evoked by 20 presentations the call.
Cells were typically held for periods of 45–60 min. After recording a cell's responses to tones and the 10 calls, pharmacological agents were iontophoretically applied, and the responses to the same signals were recorded again. Before evaluating responses while inhibition was blocked by bicuculline or bicuculline and strychnine, we first applied an ejection current that routinely caused a substantial change in responsiveness, typically 20–40 nA. We recorded responses to BF tones until a substantial increase in spike count was obtained and stabilized. Once responses were stable, the complement of tone bursts and communication calls was presented again, and the same response features were obtained for comparison with those obtained before the application of drugs. The ejection current was switched off, and the cell was allowed to recover. Recovery was complete when both the shape and maximum spike count of the rate-level function returned to their predrug values. Because recovery times were usually 30–90 min and most neurons were held for 45–60 min, most neurons were lost before recovery was attained. We allowed 45 min before searching for another neuron in those instances.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Basic features of DNLL neurons
The average BF of the 30 DNLL neurons was 24.6 ± 6.2 kHz (range, 13–43 kHz). The DNLL population is homogeneous because all the DNLL cells we recorded displayed four tone evoked response features. 1) They responded to tone bursts with a sustained discharge train that lasted as long as the duration of the tone burst that evoked it. Moreover, the majority of DNLL neurons displayed an onset-chopping sustained pattern, where the cells initially discharged at a regular interval that was unrelated to the stimulus period followed by a series of irregularly spaced discharges. 2) DNLL cells were binaural, being excited by sound at the contralateral ear and inhibited by sound at the ipsilateral ear, and thus had excitatory–inhibitory (EI) response properties. 3) The rate-level functions were monotonic, where they exhibited increasing spike counts with intensity that typically plateaued at 30–40 dB above threshold. 4) Their tuning curves were all moderately sharp, with an average Q10dB of 6.3 ± 2.0 (n = 25) and an average Q30dB of 2.6 ± 1.0 (n = 26). The above features are in agreement with previous studies of the DNLL in this species and are shown in Fig. 1.
|
). This lack of selectivity for calls was confirmed here. As shown in Fig. 2A, when calls were presented at 30–40 dB above BF threshold, 83% (19/23) of DNLL neurons responded to all 10 calls.
|
100–300%. The main point, however, is that the width of the tuning curve was virtually the same before inhibition was blocked as it was while inhibition was blocked. The effects of blocking GABAergic, glycinergic, or both inhibitory receptors on tuning were evaluated in 13 DNLL neurons. In 9/13 neurons, the tuning was unchanged, whereas in 4/13 neurons, the tuning curve expanded just slightly. These features are shown in Fig. 4A, which shows the Q30dB values of the 13 neurons before and while inhibition was blocked.
|
|
We tested for surround inhibition with iontophoresis of glutamate in 11 other DNLL neurons. In all of the DNLL neurons, tones within the excitatory tuning curve evoked both an initial excitatory discharge followed by a gap produced by the inhibition that followed the excitation. An example is shown in Fig. 5. Presumably, it was this trailing inhibition that was blocked by bicuculline and caused the increase in spike count and response duration shown in Fig. 3. However, the most notable feature was the absence of any gaps evoked by frequencies outside of the excitatory tuning curve, showing that there was no surround inhibition in this DNLL neuron. Comparable results were obtained from the 10 other DNLL cells in which surround inhibition was evaluated by creating background activity with glutamate.
EFFECTS OF BLOCKING INHIBITION IN DNLL NEURONS ON RESPONSES TO SPECIES-SPECIFIC CALLS.
The above results, coupled with the results from the previous studies on DNLL processing of species-specific calls, are consistent with the hypothesis that processing in the DNLL is determined predominantly by a linear integration of excitatory inputs. We evaluated this issue directly by presenting the same suite of 10 species-specific calls that Bauer et al. 2002 presented in the previous report, and recorded the responses from 11 DNLL neurons evoked by those calls before and while inhibition was blocked. Blocking inhibition increased response magnitude but did not change the selectivity for the calls nor did it change the relative responsiveness to each call. These features are shown in Fig. 6A, where the responses evoked by each call are shown before and while inhibition was blocked by bicuculline. Notice that the neuron was relatively unselective because it responded to 9 of the 10 calls both before and while inhibition was blocked. The spike counts evoked by each call before and during blockage of inhibition are plotted graphically in Fig. 6B. The important feature is that the shapes of the two graphs are similar, showing that the relative responsiveness of each call was unchanged. The neuron, for example, responded most vigorously to call SC2, did not respond to call SC5, and responded with the smallest spike count to call SC9, and this result was obtained before and while inhibition was blocked. That inhibition had no influence on the relative responsiveness to each call even more evident when the normalized spike counts in the two conditions were graphed, as shown in Fig. 6C. Similar results were obtained for the 10 other DNLL neurons. The changes or more appropriately the lack of changes in the number of calls to which each neuron responded when inhibition was blocked are shown in Fig. 2A. Taken together, these results show that DNLL neurons have little or no surround inhibition and that processing in the DNLL is determined largely by excitation, where inhibition acts mainly to reduce response magnitude.
|
Response features of INLL neurons were more diverse than those in the DNLL, and this was found for most of the response features that we evaluated. The majority (85%, 41/48) of INLL neurons were monaural, excited by stimulation of the contralateral ear, and unaffected by stimulation of the ipsilateral ear alone or when presented simultaneously with stimulation of the contralateral ear. However, 15% (7/48) of INLL cells were binaural and were EI (excited by contralateral and inhibited by ipsilateral stimulation). Thus in contrast to the DNLL, where every cell was EI and thus binaural, the aural features of INLL population were a mixture of mostly monaural and a smaller number of binaural, EI cells. The binaural INLL cells were surprising because previous studies of the INLL in another species of bat reported that all neurons in this nucleus were monaural (Covey and Casseday 1991; Huffman and Covey 1995).
The average BF of the sample from the INLL was 22.7 ± 9.1 kHz (n = 45), with a range of 8–54 kHz. Most INLL cells (73%, 35/48) responded to tone bursts with a sustained discharge pattern, and a smaller percentage (27%, 13/48) responded only to the onset of the tone (Fig. 7). Two types of sustained response patterns were observed. One type displayed an onset-chopping sustained pattern, similar to the pattern of DNLL neurons. The other was a primary-like pattern, where the tone bursts evoked an initial high firing rate that adapted to a lower rate for the duration of the stimulus. Two types of onset neurons were also found. One type was an onset chopper, where the neurons fired one to five regularly spaced discharges at the onset of the signal and failed to fire to the remainder of the signal duration. The other onset type also fired only at the onset of the signal, but the discharges did not have a temporal regularity. Both types were also seen in previous studies of the INLL (Covey and Casseday 1991) and are shown in Fig. 7.
|
The most distinctive feature, and the one that characterized the majority of INLL cells, was the width of their tuning curves at 30–50 dB above threshold. Of the tuning curves obtained from 36 INLL cells, 30 (83%) were broadly or very broadly tuned, and only 6 (17%) were narrowly tuned (Fig. 8). In the broadly tuned cells, the tuning at 30–40 dB above threshold encompassed a frequency range that spanned one to three octaves. The average Q30dB of the broadly tuned INLL cells was 1.4 ± 0.60. The tuning of these broadly tuned INLL cells was substantially broader than the tuning curves of DNLL or IC cells. Six cells in our sample were narrowly tuned with V-shaped tuning curves comparable to DNLL and IC cells (Fig. 8). The average Q30dB value of these cells was 5.0 ± 0.8 (n = 6). As we show below, the underlying excitatory innervation of one "narrowly tuned cell" actually was the broadest of any cell we had ever seen and encompassed the entire hearing range of this animal.
|
Most INLL neurons (93%, 39/42), like DNLL cells, responded similarly to time-reversed calls and to the same calls presented in their normal temporal sequence (data not shown). Only 7% (3 of 42) responded differently to forward and reversed calls.
EFFECTS OF BLOCKING INHIBITION IN THE INLL ON TONE-EVOKED RESPONSES AND SPECIES-SPECIFIC CALLS. Inhibition played little or no role in shaping response properties in almost all INLL cells. We evaluated changes in the responses to calls caused by blocking inhibition in 20 neurons, and in 16 of those neurons we also evaluated changes in tuning. In 19 of 20 neurons, there were either no changes in the number of calls to which the neurons responded (n = 16) or there were only small changes, where they responded to 1–3 additional calls when inhibition was blocked (Fig. 2B). The minor roles of inhibition in the INLL are further illustrated by the 16 cells in which changes in both responses to calls and in tuning were recorded before and while inhibition was blocked. When inhibition was blocked in 15 of these neurons, there were either no changes in response magnitude, discharge pattern, in the extent of their tuning curves, and in the responses to the species-specific calls, or there was only an increase in response magnitude with no other changes either in tuning or selectivity for calls. The lack of change in the response to calls when inhibition was blocked is shown in Fig. 2B, and the similarity in tuning before and while inhibition was blocked is shown in Fig. 4B.
This remarkable lack of inhibitory influences is further illustrated by the three neurons in Figs. 9 and 10. The neuron in Fig. 9A was a very broadly tuned INLL cell in which the application of both strychnine and bicuculline failed to have any influence on either response magnitude, tuning, or on the responses to species-specific calls. The complete absence of any changes while inhibition was blocked was seen in 5 of 16 cells in which we obtained tuning curves and responses to calls. This was a most surprising finding because we had never before seen neurons in any nucleus whose properties were virtually unchanged when inhibition was blocked. This absence of effects was not caused by a failure to eject drugs because of blockage of the ejection barrels, because the resistance of the barrels were routinely monitored. It was also not caused by any ineffectiveness of the particular batch of drugs, because we documented substantial changes in IC cells with the same electrodes in the same penetrations and with the same ejection currents. Finally, we used long periods (20–30 min) of high ejection currents in these cells, all without affecting the cells discharge properties.
|
|
Ten other INLL cells (10/16) responded similarly when inhibition was blocked, except in these cells, response magnitude increased, although other features, which include tuning and the number of calls to which they responded, were unchanged (Fig. 10).
Only 1 of 16 INLL cells showed a substantial change in all response features when inhibition was blocked, but those changes were extraordinary. As shown in Fig. 11, in the control condition, the tuning curve was narrow and the cell responded to tones with an onset response. The cell also was selective for the suite of calls, where it responded to only 2 of the 10 calls, and it responded weakly to both of those calls. Blocking glycinergic inhibition caused an expansion of the tuning curve, especially at high intensities, and also caused an increase in the number calls to which the neuron responded, from two calls before inhibition was blocked to nine calls. When both glycinergic and GABAergic inhibition were blocked, there was an additional increase in response magnitude, and the cell now responded vigorously to all 10 of the calls. Moreover, there were additional changes in tuning, where thresholds were substantially lowered for frequencies from 30 to 80 kHz. When both GABA and glycinergic inhibition were blocked, the neuron fired to frequencies from 15 to 99 kHz, which encompassed most, if not all, of the animal's hearing range (Bartsch and Schmidt 1993; Vater and Siefer 1995), whereas it originally responded only to frequencies ranging from 20 to 28 kHz.
|
To summarize, DNLL cells had V-shaped tuning curves with little or no surround inhibition, and they responded to all or most of the species-specific calls, and thus were unselective for calls. Inhibition apparently acted to reduce response magnitude and had little or no influence on their tuning or on their selectivities for calls. Most INLL cells had wider or much wider tuning curves than DNLL cells, and their tuning curves had no surround inhibition, a feature they share with DNLL neurons. INLL cells were also similar to DNLL cells in that blocking inhibition had little or no influence on their tuning, responses to tones, or on their responses to calls. The exception was one sharply tuned INLL cell whose response features were strongly shaped by inhibition.
Features of IC neurons
The BFs of the IC neurons in our sample ranged from 12 to 42 kHz (n = 147), with an average of 23.3 ± 5.1 kHz, and were similar to the BFs of the DNLL and INLL neurons in our sample. However, the neuronal population of the IC is markedly different from those of the INLL and DNLL. The most distinctive feature of the IC population is its diversity, where IC neurons displayed a wide variety of different response features to tonal stimuli and to the species-specific calls. The diversity was also seen in the various changes in IC response features that obtain when inhibition was blocked, showing that inhibition in the IC plays a far more prominent role in shaping response properties than it does in the INLL or DNLL. Indeed, it seems that the wide diversity of response properties expressed by the IC population is due largely to the particular pattern of inhibition that innervates each neuron, which more or less uniquely shapes its response features.
The diverse properties were evident from responses evoked by BF tones. Two main types of temporal discharge patterns were evoked by BF tone bursts: 1) onset patterns in which cells discharged one or a few spikes at the onset of the signal and 2) sustained patterns in which the discharges were evoked throughout the duration of the signal. Onset responses were more common (52%, 93/179) than sustained patterns (29%, 52/179). These, however, were not the only types that we observed. Sixteen cells (9%) had high rates of spontaneous activity and responded to tone bursts with inhibitory gaps in the spontaneous background. Four cells responded to offsets, and two responded with ON-OFF responses. However, 12 cells (7%) were unresponsive to any frequency–intensity combination of tone bursts that we presented, although 10 of these cells responded to downward FM sweeps. Most of these patterns have been described in previous studies of the IC of this bat (Bauer et al. 2000; Bodenhamer and Pollak 1981; Burger and Pollak 2001; Hurley and Pollak 2001; Pollak et al. 1977, 1978.
Several types of rate-level functions were also found and included monotonic functions that reached a plateau and monotonic functions that did not plateau. In addition, many cells had nonmonotonic functions and others had upper-threshold functions, the extreme form of a nonmonotonic rate-level function in which the neuron stopped firing at higher intensities. Nonmonotonic functions were slightly more common than monotonic functions. The number of cells displaying each type of rate-level function is shown in Table 1.
|
).
|
EFFECTS OF BLOCKING INHIBITION ON TONE-EVOKED RESPONSES AND TUNING CURVES. Blocking inhibition had a substantial influence on the response properties in almost all IC neurons. In almost every IC cell, blocking inhibition caused an increase in spike count. In many cells, the increase was accompanied by a change in response pattern from an onset to a sustained response, and the magnitude change in these cells was substantial. In other cells, including both onset and sustained neurons, the temporal discharge pattern was not changed. These cells were still onset or sustained while inhibition was blocked, and the spike count increases were caused by an increase in the discharge probability.
Tuning curves in 47 neurons were measured before and while inhibition was blocked, and representative examples are shown in Fig. 12. Thirty-seven of those neurons originally had V-shaped tuning curves, whereas seven were O type and three failed to respond to tones. The distribution of changes in Q30dB caused by blocking inhibition for the 37 V-type neurons is shown in Fig. 4C and contrasts markedly with the near absence of changes in tuning in DNLL and INLL cells. The average change in Q10dB was –3.1 ± 4.4 and Q30dB was –1.5 ± 1.8 (the minus sign indicates that tuning broadened). When we consider all 47 neurons, the tuning curves in a minority of neurons (23%, 10/47) were unaffected by blocking inhibition (Fig. 12A), but the curves of most neurons (76%, 37/47) broadened when inhibition was blocked, showing that surround inhibition acted to sharpen the tuning curves of most IC cells. Substantial changes occurred in cells that had O-type tuning curves. In seven of seven cells with O-type cells, the upper threshold feature was eliminated, and the tuning curves were changed from an island-like O-type into a traditional V-shaped tuning curve (Fig. 12E). Finally, the most dramatic changes were seen in those IC cells that were unresponsive to tones, and thus had no tuning curves. We blocked inhibition in three of these cells, and in all cases, blocking inhibition converted an unresponsive cell into one that had a classical V-type tuning (Fig. 12F).
While substantial widening of tuning was commonly observed when inhibition was blocked in the IC, the extent of surround inhibition may have been underestimated in at least some of those cells. As described previously, blocking inhibition cannot reveal any pure surround inhibition, whereas the method of creating a carpet of background activity with glutamate can. We therefore evaluated the extent of surround inhibition in 16 IC cells with glutamate and found that, in 7 cells (44%), the range of frequencies that evoked inhibition at 30–50 dB above threshold was far wider than any of the tuning curve expansions that we observed when inhibition was blocked. This is shown in Fig. 13, which shows an IC neuron with a BF of 13 kHz and a fairly narrow, V-shaped, excitatory tuning curve. The remarkable feature of this cell is that inhibitory responses, as revealed by the gaps in the glutamate generated background activity, were evoked by frequencies ranging from as low as 19 kHz to as high as 55 kHz, a range that spanned about three octaves.
|
; Pollak et al. 2003b), inhibition plays a prominent role in shaping the selectivities for calls in most IC neurons, and this was confirmed in this study. Representative examples are shown in Fig. 14. The six neurons shown in this figure are the same neurons whose tuning curves are shown in Fig. 12. These six neurons shown two features: 1) that blocking inhibition increased the number of calls to which IC neurons responded (decreased selectivity) and 2) that the change in the number of calls that evoked responses when inhibition was blocked varied from neuron to neuron. In a minority of neurons, such as the neuron in Fig. 14A, blocking inhibition hardly changed selectivity. The tuning curve of this neuron did not expand when inhibition was blocked (Fig. 12A), suggesting little or no surround inhibition. In contrast, the tuning curves of the four neurons shown in Fig. 14, B and E, expanded markedly, and correspondingly, they responded to many more calls when inhibition was blocked. The most dramatic change was seen for the neuron that was originally unresponsive to tone bursts and to any of the calls (Fig. 14F). Blocking inhibition unmasked a V-shaped tuning curve and allowed the neuron to respond to 9 of the 10 calls. The changes in the number of calls to which the 46 IC cells in our sample responded because of blocking inhibition are shown in Fig. 2C and illustrates the dominant role that inhibition plays in sculpting responsiveness to complex signals in the IC of these animals.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Monaural compared with binaural processing in the DNLL
We have shown that inhibition has little or no influence on sculpting DNLL response properties, a finding consistent with previous studies of the DNLL in this and other species of bats (Burger and Pollak 2001; Yang and Pollak 1994b; Yang et al. 1996). It is important to note that the response features that were uninfluenced by inhibition are responses that were evoked by monaural stimulation of the contralateral ear. However, high-frequency DNLL cells are binaural and are driven by stimulation of the contralateral ear and are inhibited by stimulation of the ipsilateral ear (Brugge et al. 1970; Covey 1993; Kelly et al. 1998; Yang and Pollak 1997, 1998; Yang et al. 1996). The ipsilaterally evoked inhibition is both glycinergic and GABAergic and acts to create unique binaural properties in the DNLL (Yang and Pollak 1994c). Those properties endow DNLL neurons with emergent features for the processing of multiple sound sources in space and/or for directional selectivity for moving sounds (Pollak et al. 2003a; Yang and Pollak 1994a,c). Thus DNLL cells appear to be unselective and therefore are generalists for coding "what" the sounds are, but the same cells are far more complex in their coding for "where" the sounds are located in space or whether they are moving and the direction of movement.
Consequences of wide tuning of INLL cells for processing in the IC
One notable feature of the INLL is the extreme that many INLL cells displayed in terms of the lack of inhibitory influences on their response features and in the widths of their tuning curves. Whereas blocking inhibition in some INLL cells increased response magnitude without affecting other response features, an effect similar to that seen in the DNLL, blocking both glycinergic and GABAergic inhibition in other INLL cells had no effects at all on their response properties. In those cells, even the response magnitude was unchanged. This total lack of any change in responsiveness when inhibition was blocked was never seen in any other neuron in any other nucleus that we had studied in this or in any previous study, and we consider it both unusual and remarkable.
The second extreme feature of INLL cells was the broadness of their tuning. To be sure, the fact that many INLL cells were excited by a range of frequencies that spanned two or even three octaves and encompassed a frequency range of 30–40 kHz is impressive. But even more impressive is the sharply tuned INLL cell in which blocking inhibition caused the excitatory tuning curve to expand and encompass the entire hearing range of the animal, from <12 to 99 kHz. This was, by far, the most remarkable expansion of tuning we have ever seen or that has been reported in any previous study.
The INLL sends predominately inhibitory but also some excitatory projections to the IC (Vater et al. 1992b; Winer et al. 1995), and broadly tuned INLL cells must have a significant impact on their targets in the IC. The impacts of those projections, however, were not apparent with most techniques we used to study IC cells. For example, IC tuning curves expand when inhibition is blocked, but the expansion is typically several kilohertz on either or both sides of the original excitatory tuning curve. Expansion of IC tuning curves by 20–30 kHz when inhibition is blocked has not been observed previously. However, if some IC cells receive inhibitory innervation from the INLL that is broader than their excitatory innervation, blocking inhibition would cause their tuning curves to expand but the surround frequencies that evoked pure inhibition would not be evident. One indication that inhibitory tuning may be wider than excitatory tuning in some IC neurons is seen in the experiments in which a background activity is generated by iontophoresis of glutamate. Our results show that at least some IC neurons have very wide, pure inhibitory surrounds (e.g., Fig. 13). Whether those inhibitory surrounds are a consequence of inhibitory innervation from widely tuned INLL cells is a distinct possibility, although these experiments were not designed test this possibility.
Comparative considerations
Anatomical studies conducted on a wide variety of bats and other mammals have all shown that the structural and connectional features of the mammalian auditory system have a common ground plan, with a similar complement of nuclei, similar connections, similar cell types, similar neurochemistry, and similar synaptic morphologies (Casseday 2002; Grothe 2000; Grothe et al. 1994; Pollak and Casseday 1986; Pollak et al. 1995; Vater et al. 1992b; Winer et al. 1995). Consistent with a basic mammalian ground plan, some of the most prominent results we obtained in the DNLL, INLL, and IC are similar to the findings reported in previous studies of these nuclei in a variety of mammals. However, other results we obtained are not. Below we consider the DNLL, INLL, and IC and point out features in each nucleus that vary among mammals and appear to be species-specific and features in each nucleus that are common to both bats and other mammals.
The DNLL is a conservative nucleus, since the response features that we observed in high-frequency neurons are similar to the features reported in studies of the DNLL in other bats, in cats, rats, and gerbils (Aitkin et al. 1970; Brugge et al. 1970; Covey 1993; Kelly et al. 1998; Markovitz and Pollak 1993, 1994). DNLL neurons tuned to high frequencies in all of these animals are EI, have the same complement of inputs, including GABAergic inputs from the opposite DNLL through the commissure of Probst, and the same projections (Chen et al. 1999; Glendenning et al. 1981; Huffman and Covey 1995; Oliver and Shneiderman 1989; Shneiderman et al. 1988; Yang et al. 1996; Zhang et al. 1998). Of particular significance is that the effects of ipsilaterally evoked inhibition operate in the same way in the DNLL of bats (Yang and Pollak 1994a,c) and in high-frequency neurons in the gerbil DNLL (Zahn 2004). In short, those response features that have been studied are similar among species, or at least, there is no evidence of any substantive differences.
The INLL, in contrast, appears to be species-specific, and markedly so. Although chopper and primary-like discharge patterns that we observed in Mexican free-tailed bats were also seen in the INLL of the big brown bat (Covey and Casseday 1991), the most prominent response feature of INLL reported here, the wide tuning curves, were not seen in the INLL in the big brown bat (Covey and Casseday 1991; Haplea et al. 1994). Moreover, recent studies of the INLL in the mustache bat have shown that the majority of INLL neurons in that animal are combination sensitive (Nataraj 2005; Portfors and Wenstrup 2001a). Those neurons integrate excitation from one frequency band and inhibition from another frequency band in a nonlinear manner to produce responses that depend on both the combination of frequencies in a signal and on the temporal relationships of those frequencies. The results found in this study show that inhibition plays, at best, a minor role in shaping INLL responses, and the lack of virtually any change in response selectivity to the suite of calls we presented when inhibition was blocked show that INLL neurons in Mexican free-tailed bats are not combinatorial, as they are in mustache bats. The INLL has received almost no attention in other mammals and thus we cannot comment on how INLL neurons in bats compare with those of the other mammals. Nevertheless, we point out that the INLL is the largest and most well developed of the nuclei of the lateral lemniscus in all bats, but is far less prominent in other mammals. This disparity in the relative development of the INLL among mammals also suggests that the INLL has prominent species-specific features.
Basic response features of IC neurons in bats
Bats are diverse mammals, both in terms of the number of species and the variety of ecological niches that the various species occupy. Consistent with their diversity, pronounced adaptations have been seen in the IC of the more specialized bats, such as the pallid (Fuzessery 1996, 1997; Fuzessery and Hall 1996), vampire (Schmidt 1991) and mustache bats (Leroy and Wenstrup 2000; Mittmann and Wenstrup 1995; Portfors and Wenstrup 2001b). We have not, however, observed any feature in the IC of Mexican free-tailed bats that could be considered to be a special adaptation for echolocation. Rather, we find that the basic response features in the IC of free-tailed bats are strikingly similar to the response features that have been found in the IC of other mammals. As in the free-tailed bat, IC neurons in a wide variety of bats and other mammals have monotonic and nonmonotonic rate-level functions (Faingold et al. 1991; Pollak and Park 1993; Ramachandran et al. 1999; Sivaramakrishnan et al. 2004; Syka et al. 2000), have more or less sharply tuned neurons (Haplea et al. 1994; LeBeau et al. 2001; Ramachandran et al. 1999), have a large population of binaural EI neurons that display time-intensity trading (Klug et al. 2000; Li and Kelly 1992; Pollak 1988; Yin et al. 1985), exhibit a wide range of latencies (Haplea et al. 1994; Park and Pollak 1993; Syka et al. 2000), and express other features such as duration tuned neurons (Brand et al. 2000; Faure et al. 2003; Fuzessery and Hall 1999), and neurons that respond preferentially to low rates of sinusoidal amplitude modulations (Burger and Pollak 1998; Casseday et al. 1997; Joris et al. 2004) and to the direction of FM sweeps (Fuzessery and Hall 1996; Koch and Grothe 1998).
Not only are these universal features of the IC, but most of these features
