analysis of the role of inhibition in shaping responses to

Analysis of the Role of Inhibition in Shaping Responses toSinusoidally Amplitude-Modulated Signals in the Inferior Colliculus

R. MICHAEL BURGER AND GEORGE D. POLLAKDepartment of Zoology, University of Texas at Austin, Austin, Texas 78712

Burger, R. Michael and George D. Pollak. Analysis of the role to the ICc are from the cochlear nucleus, from the lateralof inhibition in shaping responses to sinusoidally amplitude-modu- and medial superior olives, and from the nuclei of the laterallated signals in the inferior colliculus. J. Neurophysiol. 80: 1686– lemniscus (Brunso-Bechtold et al. 1981; Oliver and Huerta1701, 1998. Neurons in the central nucleus of the inferior colliculus 1992; Ross et al. 1988; Vater et al. 1995; Zook and Casseday(ICc) typically respond with phase-locked discharges to low rates 1982b). Many of these projections provide excitatory in-of sinusoidal amplitude-modulated (SAM) signals and fail to

nervation to the ICc while others provide inhibitory innerva-phase-lock to higher SAM rates. Previous studies have shown thattion that can be either GABAergic or glycinergic (Glenden-comparable phase-locking to SAM occurs in the dorsal nucleus ofning and Baker 1988; Gonzalez-Hernandez et al. 1996; Fu-the lateral lemniscus (DNLL) and medial superior olive (MSO)bara et al. 1996; Saint Marie and Baker 1990; Semple andof the mustache bat. The studies of MSO and DNLL also showed

that the restricted phase-locking to low SAM rates is created by Aitkin 1980; Shneiderman et al. 1988; Vater et al. 1992b;the coincidence of phase-locked excitatory and inhibitory inputs Winer et al. 1995).that have slightly different latencies. Here we tested the hypothesis In this report, we turn our attention to the roles of excita-that responses to SAM in the mustache bat IC are shaped by the tion and inhibition in shaping temporal discharge patternssame mechanism that shapes responses to SAM in the two lower evoked by ‘‘complex’’ signals in ICc neurons. Specifically,nuclei. We recorded responses from ICc neurons evoked by SAM the report is concerned with two principal issues: the degreesignals before and during the iontophoretic application of several

to which the temporal fluctuations in acoustic signals arepharmacological agents: bicuculline, a competitive antagonist forpreserved in the temporal discharge patterns of ICc neuronsg-aminobutyric acid-A (GABAA) receptors; strychnine, a competi-and what roles the inhibitory innervation of the ICc play intive antagonist for glycine receptors; the GABAB receptor blocker,shaping the temporal discharge patterns of ICc neurons. Thephaclofen, and the N-methyl-D-aspartate (NMDA) receptor

blocker, (0)-2-amino-5-phosphonopentanoic acid (AP5). The hy- temporal fluctuations we focus on are those created by AM,pothesis that inhibition shapes responses to SAM signals in the fluctuations that occur naturally in many acoustic signalsICc was not confirmed. In ú90% of the ICc neurons tested, the (Bregman 1990; Gerhardt 1988; Muller-Preuss et al. 1994;range of SAM rates to which they phase-locked was unchanged Rose and Capranica 1985; Schnitzler 1987; Schuller 1984).after blocking inhibition with bicuculline, strychnine or phaclofen, In the laboratory, the natural AM are typically simulatedapplied either individually or in combination. We also considered with sinusoidal AM (SAM). These are simple signals com-the possibility that faster a-amino-3-hydroxy-5-methylisoxazole-

posed of only three frequencies, a carrier or center frequency4-propionic acid (AMPA) receptors follow high temporal rates ofand two side bands. The frequency of the AM or modulationincoming excitation but that the slower NMDA receptors couldrate is determined by the frequency separation between thefollow only lower rates. Thus at higher SAM rates, NMDA recep-center frequency and side bands.tors might generate a sustained excitation that ‘‘smears’’ the phase-

locked excitation generated by the AMPA receptors. The NMDA Previous studies have shown that there is a pronouncedhypothesis, like the inhibition hypothesis, was also not confirmed. difference in the ways that neurons in most lower nucleiIn none of the cells that we tested did the application of AP5 by respond to SAM signals compared with the ways that neu-itself, or in combination with bicuculline, cause an increase in rons in the ICc respond to the same signals. Beginning atthe range of SAM rates that evoked phase-locking. These results the auditory nerve, and continuing in most nuclei below theillustrate that the same response property, phase-locking restricted ICc, the majority of neurons respond to SAM signals withto low SAM rates, is formed in more than one way in the auditory

discharges that are phase-locked to the cycles of the modula-brain stem. In the MSO and DNLL, the mechanism is coincidencetion waveform, and the phase-locking is evoked by a wideof phase-locked excitation and inhibition, whereas in ICc the samerange of modulation rates (Frisina et al. 1990a,b; Huffmanresponse feature is formed by a different but unknown mechanism.et al. 1998; Javel 1980; Joris and Yin 1992, 1998; Rhodeand Greenberg 1994; Yang and Pollak 1997). In marked

I N T R O D U C T I O N contrast, the vast majority of ICc neurons phase-lock onlyto low SAM rates, typically õ300 Hz (Gooler and FengThe central nucleus of the inferior colliculus (ICc) is an1992; Langner 1992; Langner and Schreiner 1988; Muller-auditory region that receives converging inputs from a largePreuss et al. 1994; Rees and Møller 1983; Reimer 1987;number of lower auditory nuclei. The principal projectionsSchreiner and Langner 1988; Schuller 1984), and respondto higher modulation rates either with a response only to the

The costs of publication of this article were defrayed in part by the onset of the SAM signal or with sustained discharges that arepayment of page charges. The article must therefore be hereby markeduncorrelated with the cycles of the modulating waveform.‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734 solely to

indicate this fact. Given the difference in phase-locking to ‘‘high’’ SAM

1686 0022-3077/98 $5.00 Copyright q 1998 The American Physiological Society

J310-8/ 9k2d$$oc32 09-10-98 07:29:22 neupa LP-Neurophys

by 10.220.33.3 on October 26, 2016

http://jn.physiology.org/D

ownloaded from

http://jn.physiology.org/

RESPONSES TO AM SIGNALS 1687

rates between ICc neurons and neurons in most lower nuclei,the question arises as to what mechanisms in the ICc transforminputs that largely phase-lock to high SAM rates into outputsthat phase-lock only to low SAM rates? Insights into a mecha-nism that can produce such a transformation are given bystudies of the medial superior olive (MSO) and dorsal nucleusof the lateral lemniscus (DNLL) of the mustache bat (Grothe1994; Yang and Pollak 1997). The MSO and the anteriorportion of the DNLL differ from most other lower nuclei inthat their neurons phase-lock only to low SAM rates and inthis regard are strikingly similar to ICc neurons. Of importanceis that the excitatory and inhibitory inputs to both the MSOand anterior DNLL phase-lock to a wide range of SAM rates,as is the case for the inputs to ICc. When blockers of inhibitionare iontophoretically applied to MSO or anterior DNLL cells,their discharge patterns change dramatically; when inhibitionis intact, they phase-lock only to SAM rates õ300 Hz, butwhen inhibitory inputs are blocked the same cells phase-lockto SAM rates of 500–800 Hz.

These findings lead to the proposal that the phase-lockingtransformations are created by the coincidence of phase-lockedexcitatory and inhibitory inputs that have slightly differentlatencies (Grothe 1994; Yang and Pollak 1997). Specifically,the inhibitory inputs have a slightly longer latency than theexcitatory inputs. Thus when a SAM signal with a low-modu-lation frequency is presented, the phase-locked excitatory in-puts and phase-locked inhibitory inputs are interleaved,allowing the excitatory inputs to drive the target cell on acycle-by-cycle basis (Fig. 1). At higher modulation rates,however, the initial excitation leads the inhibition, thereby

FIG. 1. Arrangement of excitatory and inhibitory inputs that could ac-allowing the cell to respond to the onset of the signal, but thecount for phase-locking only to low modulation rates in the mustache batresponses to subsequent cycles of the SAM signal are inhibitedmedial superior olive (MSO) and dorsal nucleus of the lateral lemniscusbecause the phase-locked excitatory and inhibitory inputs are (DNLL). This arrangement is the same as that originally proposed by

now temporally coincident. Blocking inhibition unmasks Grothe (1994). Target neuron receives excitatory and inhibitory inputs thatphase-locked excitatory inputs, which the neurons then follow are phase-locked. Critical feature of the model is that inhibitory inputs have

a constant delay relative to the excitatory inputs. For low modulation rates,on a cycle-by-cycle basis, even at high SAM rates.100 Hz in this case, inhibition is interleaved between excitatory inputs, andIn short, phase-locking restricted to low SAM rates in thethe target cell responds with phase-locked discharges. For the same delay,

MSO and anterior DNLL is similar to the restricted range of higher modulation frequencies, in this case 200 Hz, result in an overlap ofSAM rates that evoke phase-locking in ICc neurons. Moreover, excitatory and inhibitory inputs for all but the initial cycle. Target neuron

can respond to the onset of the signal but not to subsequent cycles. Withall three nuclei receive both excitatory and inhibitory inputs thathigher modulation frequencies, i.e., 500 Hz, the 1st few cycles of excitationphase-lock to high SAM rates. It therefore seems reasonable toreach the target cell before inhibition, thereby allowing the cell to fire tohypothesize that phase-locking limited to low SAM rates in the onset of the signal. For later cycles, however, the higher modulation

the ICc might be a consequence of the integration of phase- frequencies create a phase shift between the 2 inputs. Now the cycles arelocked excitatory and inhibitory inputs with slightly offset la- so short and the repetition so high that the period between inhibition and

the following excitation is too short to allow the inhibition to decay beforetencies as was found in the MSO and anterior DNLL.the next excitation arrives. Reproduced from Yang and Pollak (1997).Here we report on experiments conducted in the mustache

bat ICc that tested the above hypothesis. We chose the mus-operative in the ICc, then blocking inhibitory inputs shouldtache bat as an experimental subject for two principal rea-transform ICc neurons from cells that normally phase-locksons. The first is that the previous experiments on SAMonly to low SAM rates into cells that phase-lock to a muchcoding in the MSO and anterior DNLL were conducted inwider range of temporal fluctuations as was found in themustache bats. The second is that the mustache bat brainprevious studies of the MSO and anterior DNLL.stem auditory system is fundamentally mammalian in struc-

ture, connectivity, and organization (Pollak and Casseday1989; Ross et al. 1988; Winer et al. 1995; Zook and Casseday M E T H O D S1982a,b) . Consequently, features and mechanisms expressed

Surgical proceduresin the mustache bat’s ICc are likely to be present in the ICcof other, less specialized mammals. We tested the hypothesis Seventeen mustache bats Pteronotus parnellii parnellii , wereby recording the discharges from single units in the ICc used in this study. Surgical procedures were the same as thoseevoked by SAM signals both before and after blocking used in previous studies (Park and Pollak 1993a,b) . Briefly, ani-GABAergic and glycinergic inhibition with iontophoreti- mals were anesthetized with methoxyflurane inhalation (Metofane,

Mallinckrodt Veterinary) and 0.02 mg/g neuroleptic, Innovar-Vetcally applied drugs. If the hypothesized mechanism is indeed


by 10.220.33.3 on October 26, 2016


ownloaded from


R. M. BURGER AND G. D. POLLAK1688

(Pitman-Moore) , injected intraperitoneally. The hair on the head a frequency counter. Tone bursts also could be generated by thecomputer as could SAM stimuli having any desired modulationwas removed with a depilatory, and the head was placed in a holder

with a bite bar. The skin and muscle overlying the skull was depth and modulation rate. A 24-bit digital interface NuBus card(National Instruments DIO-24) and a digital distributor (Restekreflected after the topical application of 2% lidocaine (Elkins-Sinn)

to the initial incision. The surface of the skull was scraped free of model 99) communicated a Power Macintosh 7100/66 computerto a two-channel digital attenuator (Wilsonics, model PATT). Thetissue, and a foundation layer of cynoacrylate and a layer of glass

beads was applied to the skull overlying the forebrain and cerebel- outputs of each independently controlled channel of the attenuatorwere sent to two 1/4-in Bruel and Kjaer microphones biased withlum. A small hole then was drilled over the center of the inferior

colliculus, visible through the skull. 200 V DC and driven as speakers. At the start of each experiment,speakers, with protection grids attached, were inserted into theThe bat was transferred to a heated, sound-attenuated recording

chamber, where it was placed in a restraining cushion constructed funnels formed by the bat’s pinnae and positioned adjacent to theexternal auditory meatus. The pinnae were folded onto the housingof foam molded to the animal’s body. The restraining cushion was

attached to a platform mounted on a custom made stereotaxic of the microphones and wrapped with cellophane tape. The acousticisolation with this arrangement was ¢40 dB (Wenstrup et al.instrument (Schuller et al. 1986). A small metal rod was cemented

to the foundation layer on the skull and then attached to a bar 1986).Spikes were fed to a window discriminator and then to a Macin-mounted on the stereotaxic instrument to ensure a uniform position-

ing of the head. A ground electrode was placed between the re- tosh 7100 computer controlled by a Restek Model 45 real timeclock. Peristimulus time histograms (PST) and rate-level functionsflected muscle and the skin. Recordings were begun after the bats

recovered from the anesthetic. The bats typically lay quietly in therestraining cushion and showed no signs of pain or discomfort.Water was provided with a dropper every 1–2 h. Supplementarydoses of the neuroleptic were given if the bat struggled or otherwiseappeared in discomfort.

Electrodes

In a few cases, recordings were taken with a single glass pipettefilled with buffered 1 M NaCl and 2% Fast Green (pH 7.4). Thefast green allowed the electrodes to be easily visualized duringplacement over the inferior colliculus. Pipettes were pulled to a tipdiameter õ1 mm with resistances between 10–20 megohms. Mostrecordings, however, were made with ‘‘piggy back’’ multibarrelelectrodes (Havey and Caspary 1980). A five barrel H configura-tion blank (A-M Systems, No. 6120) was pulled and the tip bluntedto 10–15 mm. Recordings were made with a single barrel micropi-pette, pulled previously, that was glued to the multibarrel array sothat the tip of the recording electrode was 10–20 mm from theblunted end of the multibarrel electrode. The recording electrodeand the central barrel of the multibarrel electrode were filled withbuffered 1 M NaCl and 2% fast green (pH 7.4). The remainingbarrels were each filled with one of the following solutions: 0.5M g-aminobutyric acid (GABA) (in dH2O, pH 3.5–4.0, Sigma);0.1 M glycine (in dH2O, pH 3.5–4.0, Sigma); 10 mM bicucullinemethiodide (in 0.165 M NaCl, pH 3.0, Sigma); 10 mM strychninehydrochloride (in 0.165 M NaCl, pH 3.0, Sigma); 75 mM phaclo-fen (in 0.165 M NaCl, pH 3.0, Research Biochemicals Interna-tional) ; 70 mM (0)-2-amino-5-phosphonopentanoic acid (AP5)(in dH2O, pH 7.4, Research Biochemicals International) . A reten-tion current of 15 nA was applied to drug-containing barrels withopposite polarity for the drug in solution. The barrels of themultibarrel electrode were connected via silver-silver chloridewires to a six-channel microiontophoresis constant-current genera-tor (Medical Systems Neurophore, BH-2). The central barrel wasconnected to the sum channel to balance current in the drug barrelsand to reduce current effects on the recorded neuron. The recordingelectrode was connected via a silver-silver chloride wire to a DaganAC amplifier (model 2400).

Acoustic stimulation, processing of spike trains andFIG . 2. Peristimulus time (PST) histograms of phase-locked re-iontophoresis sponses evoked by tone bursts and sinusoidal amplitude-modulated

(SAM) signals with rates ranging from 50 to 500 Hz. Spike counts (SC)Search stimuli were tone bursts presented to the contralateralevoked by 20 repetitions of each signal together with vector strengths(excitatory) ear at a rate of four per second. The search tones were (VS) are shown (right ) . This was an unusual neuron because the vector

generated by a sine wave function generator (Wavetek, Model strengths were ú0.300 at all SAM rates tested. SAM waveforms ( below136) and shaped by a custom made analog switch (Restek Model each histogram) schematically represent the signals actually presented.15). Tone-bursts shaped by the switch were 20 ms in duration and Excitatory / inhibitory (EI) neuron, best frequency (BF) was 45.1 kHz,

intensity was 30 dB SPL.had 0.5-ms rise-fall times. Tone-burst frequency was monitored by


by 10.220.33.3 on October 26, 2016


ownloaded from



VS Å√[∑ sin (ai ) ] 2 / [∑ cos (ai ) ] 2 /Nwere generated and graphically displayed. The computer also cal-

culated vector strengths, as described by Goldberg and Brown(1969), as well as spike counts. The vector strength (VS) gives where ai is the phase position of a single spike [i.e., (spike time) 1

SAM rate 1 3607] and N is the total number of spikes. A value ofa quantitative value of how well the unit phase-locked to the enve-lope of an SAM signal. The VS was calculated as follows 1.0 indicates perfect phase-locking, whereas a value of 0.0 indicates

FIG. 3. PST histograms of 3 neurons that phase-locked only to SAM ratesõ300 Hz. A : unit that responded with sustaineddischarges to a BF tone burst and with phase-locked discharges to SAM rates õ300 Hz. Response pattern to SAM ratesú300 Hz were similar to pattern evoked by tone burst. EI neuron, BF was 57.7 kHz, intensity was 40 dB SPL. B : unit thatresponded phasically to a BF tone burst and phase-locked to SAM rates õ250 Hz. At higher SAM rates, it responded onlyat signal onset. EI neuron, BF was 61.0 kHz, intensity was 60 dB SPL. C : unit that phase-locked to SAM rates õ250 Hzbut discharged weakly or not at all at higher SAM rates. Monaural neuron, BF was 59.5 kHz, intensity was 40 dB SPL. SCsand VSs are shown (right) .


by 10.220.33.3 on October 26, 2016


ownloaded from



firing is uncorrelated with the waveform. To exclude the contribu- applied were bicuculline methiodide, a competitive antagonist forGABAA receptors (Borman 1988); strychnine, a competitive an-tion of the onset response to vector strength, the first 5–10 ms of

the response was excluded from the calculation. tagonist for glycine receptors (Curtis et al. 1971); phaclofen, acompetitive antagonist for GABAB receptors (Borman 1988; KerrOnce a unit was isolated, the frequency to which it was mostet al. 1987); and AP5, a N-methyl-D-aspartate (NMDA)-receptorsensitive (its best frequency or BF), threshold at BF, and binauralblocker (Davis et al. 1992).type were determined audiovisually. Binaural properties were de-

termined by presenting a BF tone burst, 20 dB above threshold, During drug application, rate-level functions were monitoredwhile ejection currents of from 10 to 60 nA were applied. For eachfirst to the contralateral ear, then to the ipsilateral ear, and then

to both ears simultaneously. When presented simultaneously, the ejection current, rate-level functions were obtained until the rate-level function, and hence response magnitude, had stabilized. Onceintensity at the ipsilateral ear was increased to determine whether

ipsilateral stimulation suppressed or enhanced the response evoked responses were stable, the complement of tone bursts and SAMsignals was presented again, and the same response features wereby the signal at the contralateral ear.obtained for comparison with those obtained before the applicationAfter this initial evaluation of neuronal properties, computer-of drugs.generated tone bursts at the unit’s BF and SAM signals were pre-

sented to the contralateral ear. The center frequency of the SAM After acquiring data while drugs were applied iontophoretically,the ejection current was switched off and the cell was allowed tosignal always was set at the unit’s BF. The computer generated

tone bursts and SAM signals were 70 ms in duration and presented recover. Recovery was complete when both the shape of the rate-level function returned to its predrug form and the maximum spike-at 10–30 dB above the BF threshold. All SAM signals were 100%

modulated and had modulation rates that ranged from 50 to 500 count returned to the predrug value. In many cases, however, con-tact with the unit was lost before recovery was complete. BecauseHz. After recording the discharges evoked by tones and SAM

signals, pharmacological agents were iontophoretically applied and recovery times were usually 10–20 min, in these cases we allowed¢30 min before searching for another neuron.the responses to the same signals were recorded again. The agents

FIG. 4. Four units that either failed to phase-lock to any SAM rate or phase-locked weakly to low SAM rates. A–C :responses of 3 units that were primarily onset to tone bursts and to all SAM stimuli. D : neuron that responded with asustained discharge to tone bursts and did not phase-lock to SAMs. Each unit displayed selectivity for SAM rates bydischarging with higher SCs to some SAM rates than to others. SCs are shown at right; VS calculations for D are right ofthe SCs. A : monaural neuron, BF was 60.8 kHz, intensity was 40 dB SPL. B : EI neuron, BF was 63.6 kHz, intensity was40 dB SPL. C : monaural neuron, BF was 43.0 kHz, intensity was 40 dB SPL. D : monaural neuron, BF was 63.2 kHz,intensity was 50 dB SPL.


by 10.220.33.3 on October 26, 2016


ownloaded from



R E S U L T S Responses of ICc neurons to SAM signals

After this initial evaluation of neuronal properties, we nextHere we report on the response properties of 195 unitsrecorded from the ICc of the mustache bat. BF, the frequency evaluated the ability of ICc neurons to phase-lock to SAM

signals presented to the contralateral ear. A vector strengthto which each neuron was most sensitive, ranged from 40to 65 kHz, although 138 units (70.7%) were tuned to fre- value of 0.300 was arbitrarily chosen as a cutoff criterion

for classifying a response to SAM as being phase-locked.quencies at or around 60 kHz. The reason for the dispropor-tionate number of 60-kHz neurons was that we targeted the Of the 195 neurons in our sample, 73% (142) phase-

locked only to low SAM rates, °300 Hz, while 9%(18)enlarged 60-kHz region of the mustache bat’s ICc (Ross etal. 1988). For each unit, we first determined its BF and phase-locked to SAM rates ú300 Hz. Of the 18 cells that

phase-locked to higher SAM rates, only 9 (4.5%) locked tothen recorded its discharge pattern with BF tone bursts atintensities from threshold to 40 dB above threshold. Of the rates as high as 500 Hz, the highest SAM rate that we tested.

An example of a unit that locked to rates as high as 500 Hz195 units, 68% (nÅ 132) responded with only one or severalspikes to the onset of the tone burst (onset pattern) and 32% with a vector strength of ¢0.300 is shown in Fig. 2.

As noted above, the majority of ICc neurons phase-locked(n Å 63) discharged throughout the duration of the toneburst (sustained pattern) . In 135 cells, we also presented BF only to SAM rates °300 Hz; at higher modulation rates, the

neurons either responded to the SAM signals as they did totone bursts to each ear separately and then simultaneously todetermine whether the cell was monaurally or binaurally a tone burst or did not respond at all, even to the onset of

the SAM signal. Two examples of neurons that phase-lockedinnervated. Of the 135 cells tested with binaural signals, 53cells (39%) were monaural, 78 cells (58%) were excitatory/ only to low modulation rates and continued to discharge to

higher SAM rates with tone burst like responses are showninhibitory (EI) because the discharges evoked by stimula-tion of the contralateral ear were suppressed by stimulation in Fig. 3, A and B. The onset unit shown in Fig. 3C phase-

locked to low rates of modulation but responded weakly orof the ipsilateral ear, while 4 cells (3%) were excitatory/excitatory and were driven by stimulation of either ear. not at all at higher modulation rates.

FIG. 5. Neuron illustrating that the range of SAM rates that evoked phase-locked discharges was unchanged by theapplication of bicuculline and strychnine. Left : responses evoked by SAM signals before application of bicuculline orstrychnine. Middle : responses during bicuculline application. Right : responses while strychnine and bicuculline were appliedsimultaneously. In each case, phase-locked responses were only observed at SAM rates õ200 Hz. SCs evoked by 20repetitions of each signal together with vector strengths are shown (right of each histogram). BF was 60.5 kHz, intensitywas 30 dB SPL. Ejection currents for each drug are shown at top.


by 10.220.33.3 on October 26, 2016


ownloaded from



mask an underlying excitation that is phase-locked even toNot all ICc neurons phase-locked to SAM signals. Thirty-high modulation rates, as seen in previous studies of thefive (18%) of the neurons either failed to phase-lock to anyMSO (Grothe 1994) and DNLL (Yang and Pollak 1997).SAM signal or responded weakly with an occasional phase-

ICc neurons that displayed changes in phase-locking afterlocked discharge at low modulations rates. Four examplesinhibition was blocked were rare. Of the 43 units that phase-are shown in Fig. 4. Depending on their discharge patternlocked to low SAM rates and to which we applied bicucul-to tone-bursts, these neurons responded with either onset orline, strychnine, or both drugs, only 3 (7%) increased thesustained responses to SAM signals.range of SAM rates to which they phase-locked by ¢150Although the cells in Fig. 4 displayed little or no phase-Hz. In the other 40 units (93%), neither the application oflocking, their spike counts changed systematically with SAMbicuculline nor of strychnine substantially changed the rangerate. The spike counts of some cells were highest at lowof SAM rates to which the neurons phase-locked. We illus-SAM rates (Fig. 4A) , some at high SAM rates (Fig. 4B) ,trate the ineffectiveness of blocking inhibition on phase-whereas the spike counts of others were highest at intermedi-locking with the units shown in Figs. 5–7. For the unit inate rates (Fig. 4C) .Fig. 5, clear phase-locked discharges were evoked in thepredrug condition by SAM rates of 50 and 100 Hz, whereasEffects of blocking inhibition on neurons that phase-150 Hz evoked only occasional phase-locked spikes. Higherlocked to low SAM ratesSAM rates evoked responses largely at the onset of the sig-

We next tested the hypothesis that a delayed phase-locked nals. The application of bicuculline increased response mag-inhibition interacts with a phase-locked excitation to limit nitude, an effect of bicuculline commonly seen in this andthe phase-locking of the majority of ICc neurons to low SAM previous studies (Faingold et al. 1989; Le Beau et al. 1996;rates. If the hypothesis is correct, then blocking inhibition Pollak and Park 1993; Vater et al. 1992a), but phase-lockedby the iontophoretic application of antagonists for GABAA discharges still were evoked only by SAM rates °150 Hz.

When strychnine was applied together with bicuculline, no(bicuculline) and glycine (strychnine) receptors should un-

FIG. 6. Another neuron in which the range of SAM rates that evoked phase-locked discharges was not changed wheninhibition was blocked by application of strychnine or bicuculline. Left : predrug condition where phase-locked responseswere evoked only by SAM rates õ200 Hz. Middle : responses during application of strychnine. Right : responses duringapplication of strychnine and bicuculline. SCs and VSs are shown to the right of each histogram. EI neuron, BF was 60.3kHz, intensity was 40 dB SPL. Ejection currents for each drug are shown at top.


by 10.220.33.3 on October 26, 2016


ownloaded from



additional effect was obtained either on response magnitude Although response magnitude increased in a dose-dependentmanner, there was little change in the range of SAM ratesor on range of SAM rates that elicited phase-locking.

The phase-locking of the unit in Fig. 6 was also unaffected that evoked phase-locking from that observed in the predrugcondition.when inhibition was blocked. In this case, the effects of

strychnine alone initially were evaluated. The principal ef- In summary, blocking inhibition in the vast majority ofcells that phase-locked to low SAM rates did not unmaskfect of strychnine was to increase the response magnitude,

but it had little or no effect on the range of SAM rates that an underlying excitatory drive that was phase-locked to allSAM rates tested. The results provide little support for theevoked phase-locking. Bicuculline then was applied with the

strychnine and did not appreciably change the range of SAM hypothesis that phase-locking limited to low SAM rates isdue to inhibitory innervation.rates that evoked phase-locked discharges.

We also considered the possibility that the failure of theantagonists to allow phase-locking to higher SAM rates may Effects of blocking inhibition on neurons that did nothave been due to the application of an insufficient dosage phase-lock to SAMof drugs. We therefore evaluated phase-locking in 10 unitsby applying drugs with increasingly higher ejection currents. The results of blocking inhibition in 11 neurons that dis-

played no phase-locking to SAM signals were only slightlyIncreased dosages, however, did not substantially change therange of SAM rates that evoked phase-locking in any of the different from those obtained with neurons that phase-locked

to low SAM rates described above. In 4 of the 11 units, SAMunits tested. An example is shown in Fig. 7. In this unit, wefirst documented responses to SAM signals when bicuculline signals evoked no phase-locked responses either before or

during the blockage of inhibition. In only 2 of the 11 neuronswas applied with an ejection current of 30 nA and then withhigher ejection currents of 40 nA (not shown) and 60 nA. was there a pronounced change in phase-locking. One is

FIG. 7. Range of SAM rates that evoke phase-locking was not changed by increasing dosages of bicuculline. Left :responses evoked by SAM rates before application of bicuculline. Also shown are responses to same signals evoked whilebicuculline was applied with an ejection current of 30 nA (middle) and then with a current of 60 nA (right) . Regardless tocurrent level, bicuculline caused an increase in response magnitude but did not cause neuron to phase-lock to higher SAMrates than it did before bicuculline was applied. SCs and VSs are shown to the right of each histogram. EI neuron, BF was60.1 kHz, intensity was 40 dB SPL.


by 10.220.33.3 on October 26, 2016


ownloaded from



FIG. 8. Effects of blocking GABAergic inhibitionwith bicuculline on a unit that responded primarily withonset discharges to SAM stimuli. When bicuculline wasapplied, strong phase-locked discharges at SAM rates°150 Hz were evoked, whereas weaker phase-lockedresponses were evoked at 200 Hz. SCs evoked by 20repetitions of each signal together with VSs in the bicu-culline condition are shown to the right of each histo-gram. EI neuron, BF- 63.6 kHz, intensity was 40 dBSPL. Ejection current is shown at top.

shown in Fig. 8. Strychnine was not applied to this cell so the phase of each cycle. Surprisingly, SAM rates of 100 and150 Hz also evoked sustained discharges, but the phase-lockingwe are uncertain as to whether additional changes wouldwas poor (the vector strength wasõ0.300) and the peaks werehave occurred had glycinergic inhibition been blocked asof lower amplitude than at 50 Hz. At SAM rates of ¢200 Hz,well.only weak sustained discharges were evoked that were notFive other units had less substantial changes than the unitphase-locked.in Fig. 8. In at least some of those neurons, the small changes

To summarize, blocking inhibition in neurons that didoccurred when both bicuculline and strychnine were applied.not phase-lock to SAM signals resulted in enhanced phase-Small changes in phase-locking are illustrated by the neuronlocking to low modulation rates (i.e., °250 Hz) in only ain Fig. 9. In the predrug condition, this cell responded almostfew neurons. In most of these cells, blocking inhibitionexclusively with an onset response to SAM signals and had acaused only a small enhancement in phase-locking to modu-much greater spike count to low SAM rates than to higherlation rates of 50 or 100 Hz, although response magnituderates. Occasionally phase-locked discharges were evoked byoften increased markedly. In none of these cells did the50- and 100-Hz rates, but these were rare. Bicuculline caused anapplication of bicuculline or strychnine allow the cell tooverall increase in response magnitude. In addition, it seemed tophase-lock to SAM rates of 500 Hz, as was found for neuronsblock a weak sustained inhibition previously evoked by SAMthat did not phase-lock to SAM in a previous study of therates of 50 and 100 Hz because the size of the phase-lockedmustache bat DNLL (Yang and Pollak 1997).discharge peaks appeared to increase to a somewhat greater

degree than did the onset responses. However, onset dischargeswere evoked almost exclusively by SAM rates of ¢150 Hz. Results of blocking GABAB receptorsThe addition of strychnine transformed the response patternevoked by tone bursts from an onset to a sustained pattern. The results presented above indicate that the ability ofNow 50-Hz SAM signals evoked a strong response. However, ICc neurons to phase-lock only to low SAM rates is not athe phase-locking was not very precise (the vector strength feature shaped by either GABAA or glycine receptors. Al-was 0.323) because the unit responded throughout the duration though these are the principal inhibitory receptors in the ICc

(Fubara et al. 1996), evidence from both receptor bindingof each modulation cycle rather than to a specific portion of


by 10.220.33.3 on October 26, 2016


ownloaded from



FIG. 9. PST histograms showing slight enhancement of phase-locking while inhibition was blocked for a neuron that, inthe predrug condition, did not phase-lock to SAM. SCs and VSs are shown to the right of each histogram. EI neuron, BFwas 60.8 kHz, intensity was 40 dB SPL. Ejection currents for each drug are shown at top.

(Fubara et al. 1996) studies and pharmacological studies Results of blocking NMDA receptors(Faingold et al. 1989) indicate that at least some ICc neurons We next considered the possibility that slower excitatoryalso possess GABAB receptors. In contrast to GABAA and receptors might be mediating the response features evokedglycine receptors, which are directly gated ligand receptors, by SAM signals. If both a-amino-3-hydroxy-5-methyl-4-GABAB receptors act through G proteins (Nicholl 1988) isoxazolepropionic acid (AMPA) and NMDA receptors con-and thus are slower than the GABAA receptors. It seems tribute to the excitatory response of ICc cells, one explana-possible, therefore, that the release of GABA from inputs tion for some units might be that although the faster AMPAthat follow even high SAM rates with phase-locked dis- receptors follow high temporal rates of incoming excitation,charges could activate GABAB receptors on ICc neurons the slower NMDA receptors follow only lower rates. Spe-that, in turn, would generate a more sustained inhibition than cifically, in units that responded with more sustained butthat generated by the directly gated receptors. If this were nonphase-locked responses to higher modulation frequen-the case, then blocking GABAA and glycine receptors with cies, it seems possible that each cycle of a high-frequencybicuculline and strychnine would leave the putative slower SAM signal could generate a phase-locked AMPA receptorGABAB-mediated inhibition unaffected. In this hypothesis, discharge that was smeared by a ‘‘longer lasting’’ dischargeGABAB receptors would be the main agents responsible for due to NMDA currents. If this hypothesis is correct, thenpreventing phase-locking to high SAM rates. blocking inhibition should increase response magnitude but

To test this hypothesis, we blocked GABAB receptors in should have little effect on the range of SAM rates to whicheight neurons that phase-locked only to low SAM rates. In the neurons phase-lock because the excitatory response ofeach neuron, we first applied the GABAB receptor blocker, the cell is itself smeared by the NMDA component of excita-phaclofen, and then added both bicuculline and strychnine tion. On the other hand, blocking both inhibition and NMDAto block both the GABAB as well as the GABAA and glycine receptors should unmask a phase-locked AMPA responsereceptors. The hypothesis was not confirmed in any of the that is neither ‘‘contaminated’’ by NMDA currents nor sup-eight cells tested. A representative example is shown in Fig. pressed by GABAergic inhibition.10. The application of phaclofen caused a small increase in To test the NMDA hypothesis, we recorded from sevenresponse magnitude in this cell, as was typical, whereas neurons that phase-locked only to low SAM rates andbicuculline and strychnine caused a much larger increase in blocked NMDA receptors by the iontophoretic applicationresponse magnitude. However, the range of SAM rates that of AP5 together with bicuculline. The hypothesis predictsevoked phase-locking was unchanged by the application of that after blocking NMDA and GABAA receptors, an unin-phaclofen or by the combined application of phaclofen, bicu- hibited underlying AMPA response should be revealed that

phase-locks even to high SAM rates.culline, and strychnine.


by 10.220.33.3 on October 26, 2016


ownloaded from



FIG. 10. Neuron showing that blocking of g-aminobutyric acid-B (GABAB) receptors with phaclofen as well as blockingGABAA and glycine receptors with bicuculline and strychnine fails to increase range of SAM rates that evoke phase-locking.Left : strong phase-locked responses were elicited only by SAM ratesõ200 Hz before drugs were applied. Middle: when GABAB

receptors were blocked by phaclofen, response magnitude increased but phase-locked discharges were still evoked only by SAMrates õ200 Hz. Right : when bicuculline and strychnine were applied together with phaclofen, there was a further increase inresponse magnitude but there was no increase in the SAM rates that evoked phase-locking. SCs and VSs are shown to the rightof each histogram. EI neuron, BF was 61.2 kHz, intensity was 30 dB SPL. Ejection currents for each drug are shown at top.

The NMDA hypothesis, like the hypothesis we proposed the underlying mechanism that limits phase-locking to lowrates of SAM stimuli in ICc neurons. Coincidence is em-for GABAB receptors, was not confirmed. In none of the

seven cells did the application of AP5 by itself, or in combi- ployed widely by the auditory system for determining a vari-ety of monaural and binaural response features. Among thesenation with bicuculline, cause a significant increase, or any

increase at all, in the range of SAM rates that evoked phase- are the coding of interaural disparities in the superior olivarylocking. An example is shown in Fig. 11. In this unit bicucul- complex (Joris and Yin 1995; Park et al. 1996), the shapingline was first applied and caused an increase in discharge of phase-locked discharges in the mustache bat’s MSO andmagnitude but did not cause the cell to phase-lock to higher DNLL (Grothe 1994; Yang and Pollak 1997), the tuningSAM rates than it did before bicuculline was applied. Adding for signal duration in the ICc (Casseday et al. 1994; EhrlichAP5 to the bicuculline caused a depression of response mag- et al. 1997), and the delay tuning in the ICc, medial genicu-nitude that presumably was due to the blockage of NMDA late body, and cortex of bats (Mittmann and Wenstrup 1995;currents. A depression of evoked responses by AP5 was Olsen and Suga 1991; O’Neill and Suga 1981). The acoustictypically seen when AP5 was applied alone or in combina- stimuli that generate the coincidences are different in eachtion with bicuculline. The significant feature, however, was of these cases, but the process is the same. Thus we reasonedthat the range of SAM rates to which the neuron discharged that if a common mechanism limits phase-locking to lowwith phase-locked responses was unchanged by either bicu- SAM rates in three successively higher regions of the audi-culline or by AP5 in combination with bicuculline. tory pathway, the MSO, DNLL and ICc, it would provide

additional evidence that coincidence is a common strategyD I S C U S S I O N employed by the mammalian auditory system for shaping a

wide variety of response features, including temporal dis-The principal goal of this study was to test the hypothesisthat coincidence of phase-locked excitation and inhibition is charge properties evoked by ‘‘complex’’ signals. It is for


by 10.220.33.3 on October 26, 2016


ownloaded from



FIG. 11. PST histograms from a neuron in which the range of SAM rates that elicited phase-locked responses wasunchanged by simultaneously blocking NMDA receptors with AP5 and GABAA receptors with bicuculline. Left : responsesevoked by SAM rates before AP5 or bicuculline was applied. Middle : when bicuculline was applied by itself, it caused anincrease in response magnitude with little change in the range of SAM rates that evoked phase-locked discharges. Right :responses change little in range of SAM rates that evoked phase-locking during application of AP5 and bicuculline. SCs andVSs are shown on the right of each histogram. BF was 60.1 kHz, intensity was 40 dB SPL. Ejection currents for each drugare shown at top.

these reasons that we were surprised to find that the blockage was applied that caused a dose-dependent increase in re-sponse magnitude without changing the SAM rates thatof inhibitory receptors, whether GABAA, glycine, or GABAB

receptors, had virtually no effect on the range of SAM rates evoked phase-locking. Because there are no high-affinityuptake mechanisms for bicuculline or strychnine, the effectsthat evoked phase-locked discharges in the vast majority of

ICc neurons. Given the difference in the results that we of antagonists last minutes or tens of minutes after applica-tion. These periods of effectiveness allow the antagonists toexpected from those we observed, we consider, in the follow-

ing sections, some potential explanations for this disparity. diffuse up to several hundred micrometers from the ejectionelectrode. We know this because, in some cases, contactwith a unit was lost during application of the antagonist, andDid the antagonists fail to reach appropriate receptors?search for another unit was initiated immediately thereafter.When a second neuron was recorded several hundred mi-The first potential explanation is that coincidence ofcrometers from the previous unit, it was found that the sec-phase-locked excitation and inhibition is, in fact, the mecha-ond unit’s activity was super normal and that its activitynism shaping responses to SAM in the ICc but that thewould recover progressively within 5–20 min, suggestingapplication of antagonists failed to reach the appropriatethat the antagonists applied at the site of the first neuron hadreceptors and thus did not block the critical inhibitory syn-diffused several hundred micrometers to block inhibition atapses. The possibility that drugs failed to reach the appro-the second neuron. These observations suggest that thepriate receptors is an inherent problem with in vivo iontopho-spread of the antagonists used in this study was sufficientretic studies, but the weight of evidence makes this explana-to allow the agents to block receptors on most of the dendritiction unlikely. One piece of the evidence arguing against thistrees of the cells to which they were applied. When theexplanation is that the application of bicuculline or strych-features above are considered together with the paucity ofnine typically caused a substantial increase in response mag-units whose phase-locking to SAM changed after applicationnitude, showing that the antagonists were effective inof antagonists, it seems to us that a failure to block criticalblocking at least some inhibitory receptors. Moreover, thesereceptors is an unlikely explanation for the results we ob-antagonists were applied continuously for periods of 10–30

min, and in several neurons, more than one injection current tained.


by 10.220.33.3 on October 26, 2016


ownloaded from



Do ICc cells simply follow the temporal synchrony of provide phase-locked inhibitory innervation to the ICc. Ante-rior DNLL cells are different. They discharge with onsettheir inputs?patterns to tone bursts and phase-lock only to low SAM

A second potential explanation is that most excitatory rates, features that are virtually indistinguishable from mostinputs to ICc cells phase-lock only to low rates of SAM, onset ICc cells. Unlike ICc cells, however, blocking inhibi-and the assumption that the predominance of excitatory input tion unmasks an underlying phase-locked excitatory inputto ICc cells phase-lock to high SAM rates is incorrect. In that allows anterior DNLL cells to phase-lock to SAM ratesthis explanation, the phase-locking of ICc cells simply fol- of 500–600 Hz and display response properties closely re-lows the phase-locking of their inputs. sembling posterior DNLL cells (Yang and Pollak 1997).

This may well be the explanation for some cells but almost Because the DNLL and EI region of the ICc receive similarsurely is not the explanation for most. The reason is that, with inputs, it seems unreasonable to argue that DNLL receivesone exception, the majority of nuclei that provide excitatory excitatory innervation that phase-locks to high SAM ratesprojections to the ICc have neurons that phase-lock to SAM while the same innervation provides EI cells in the ICc withrates well ú400 Hz, and most phase-lock to rates substan- excitation that phase-locks only to low SAM rates.tially higher. This has been reported for the types of cochlear In summary, although there are some excitatory inputs tonucleus cells that project to the ICc, multipolar, and fusiform ICc that phase-lock only to low SAM rates, the majority ofcells (Frisina et al. 1990a,b; Rhode and Greenberg 1994), excitatory inputs and at least some inhibitory inputs phase-almost all cells in the LSO (Joris and Yin 1998), and for lock to a wide range of SAM rates. Moreover, two regions,most cells in the intermediate nucleus of the lateral lemniscus the DNLL and EI region of the ICc, receive a similar comple-(Huffman et al. 1998). The exception is the mustache bat ment of inputs yet display pronounced differences in theirMSO. The mustache bat MSO is unusual in that it receives responses to SAM and blocking inhibition. Some observa-only sparse projections from the ipsilateral cochlear nucleus tions made by Covey et al. (1996) are also relevant in thisand is innervated largely by the contralateral cochlear nu- regard. They showed, with whole cell recordings, that atcleus and medial nucleus of the trapezoid body (Covey et least some IC neurons receive subthreshold excitatory inputsal. 1991; Grothe et al. 1992). As mentioned previously, that follow high stimulus modulation rates although thoseMSO cells phase-lock only to low SAM rates due to the rates did not evoke phase-locked discharges. These features,interactions of phase-locked excitation from the cochlear then, provide little support for the hypothesis that the phase-nucleus and phase-locked inhibition from the medial nucleus locking of ICc neurons is simply a reflection of inputs thatof the trapezoid body (Grothe 1994; Grothe et al. 1992). lock only to low SAM rates but rather suggest that many,MSO projects heavily onto the ICc in the mustache bat (Ross if not most, excitatory inputs to ICc cells phase-lock to highand Pollak 1989; Ross et al. 1988) and thus could be a SAM rates.source of excitatory inputs to some ICc cells that phase-lockonly to low SAM rates. However, given the large number Are ‘‘other,’’ slower receptors involved?of excitatory inputs that phase-lock to high rates, it seemsunlikely that MSO inputs alone can explain the phase-lock- Another potential explanation that we considered pre-

viously is the involvement of slower receptors, such asing to low SAM rates that predominate in the ICc.GABAB or NMDA receptors. These receptors might not beThe dominance of inputs that phase-lock to high SAMable to follow the higher phase-locked inputs and thus wouldrates is reflected in the responses of neurons in the DNLL,produce ‘‘smeared’’ responses at the higher SAM rates.the auditory nucleus immediately below the ICc. BecauseHowever, phase-locking to higher SAM rates did not occurnumerous features of the DNLL are similar to portions ofwhen these receptors were blocked or when they werethe ICc, it is instructive to compare the phase-locking ofblocked together with GABAA or glycine receptors. Theseneurons in the mustache bat’s DNLL with those in the ven-results, then, do not support the hypothesis that ‘‘slower’’tromedial region of the mustache bat’s ICc. Both regionsreceptors are the principal agents that prevent ICc neuronsare dominated by excitatory/ inhibitory neurons (Ross andfrom phase-locking to higher SAM rates.Pollak 1989; Wenstrup et al. 1986; Yang et al. 1996) and

both receive a similar complement of projections: from theipsilateral MSO, from the LSO bilaterally, from the contra- Are intrinsic membrane properties different in the ICclateral DNLL through the commissure of Probst, and from than they are in lower nuclei?the ventral and intermediate nuclei of the lateral lemniscus(Ross and Pollak 1989; Vater et al. 1995; Yang et al. 1996). If the explanations discussed above are unlikely, then what

explanation might be advanced to account for the resultsAlthough the binaural properties and complement of afferentprojections are similar in the two regions, their phase-locking presented here? One possibility is that the complement of

voltage-sensitive channels in ICc neurons acts to restrict theto SAM and effects of blocking inhibition on phase-lockingare vastly different. The mustache bat DNLL is divided into phase-locking to low SAM rates. There is now a wealth of

studies showing that the particular complement of voltage-a larger posterior division and a smaller anterior division(Yang and Pollak 1997; Yang et al. 1996). Posterior DNLL sensitive channels plays a critical role in determining the

response properties of neurons in a variety of auditory nucleicells discharge with sustained patterns to tone bursts androutinely phase-lock to SAM rates of 600–800 Hz or higher. (Fu et al. 1996; Manis 1990; Oertel et al. 1988; Perney and

Kaczmarek 1997; Reyes and Spain 1994; Reyes et al. 1996;Because DNLL cells are predominately GABAergic (Adamsand Mugniani 1984; Vater et al. 1992b; Winer et al. 1995) Rhode and Smith 1986). However, there is no direct evi-

dence that confirms the hypothesis that voltage-sensitiveand project bilaterally to the ICc, these cells almost certainly


by 10.220.33.3 on October 26, 2016


ownloaded from



This work was supported by National Institute of Deafness and Otherchannels limit phase-locking to low SAM rates in ICc neu-Communications Disorders Grant DC-00268.rons, although the results from a study by Fortune and Rose

Address reprint requests to G. D. Pollak.(1997) are consistent with this hypothesis. They obtained

Received 20 April 1998; accepted in final form 23 June 1998.‘‘whole cell’’ recordings from midbrain neurons in electricfish evoked by amplitude- and frequency-modulated signals.They found that in some neurons, active conductances were REFERENCESprominent and contributed significantly to the low-pass fil-

ADAMS, J. C. AND MUGNIANI, E. Dorsal nucleus of the lateral lemniscus: atering of those cells. It should be noted, however, that thenucleus of GABAergic projection neurons. Brain Res. Bull. 13: 585–modulation frequencies they employed ranged only from 2590, 1984.

to 30 Hz and were much lower than those used in this and BORMAN, J. Electrophysiology of GABAA and GABAB receptor subtypes.other studies of AM coding in the mammal auditory system. Trends Neurosci. 11: 112–116, 1988.

BREGMAN, A. S. Auditory Scene Analysis. Cambridge, MA: MIT Press,An unpublished intracellular study of slices from rat IC neu-1990.rons by Peruzzi and Oliver (personal communication) is

BRUNSO-BECHTOLD, J. K., THOMPSON, G. C., AND MASTERTON, R. B. HRPmore germane. They found that cells that had particular study of the organization of auditory afferents ascending to the centralvoltage-sensitive potassium channels discharged on a one- nucleus of the inferior colliculus in cat. J. Comp. Neurol. 197: 705–722,

1981.to-one basis to brief, depolarizing current pulses with fre-CASSEDAY, J. H., EHRLICH, D., AND COVEY, E. Neural tuning for soundquencies °110 Hz, whereas cells that lacked those channels

duration: role of inhibitory mechanisms in the inferior colliculus. Sciencecould only follow frequencies õ50 Hz.264: 847–850, 1994.

It should be noted that neither the results reported by COVEY, E., VATER, M., AND CASSEDAY, J. H. Binaural properties of singleFortune and Rose (1997) nor the results from the Peruzzi units in the superior olivary complex of the mustached bat. J. Neurophys-

iol. 66: 1080–1094, 1991.and Oliver study can account for the features that we found inCOVEY, E., KAUER, J. A., AND CASSEDAY, J. H. Whole-cell patch-clampthe bats’ IC. Rather those studies only suggest that voltage-

recording reveals subthreshold sound-evoked postsynaptic currents in thesensitive channels may contribute to the low-pass filtering inferior colliculus of awake bats. J. Neurosci. 16: 3009–3018, 1996.of IC neurons in mammals. Consequently, the suggestion CURTIS, D. R., DUGGAN, A. W., AND JOHNSTON, G.A.R. The specificity of

strychnine as a glycine antagonist in the mammalian spinal cord. Exp.that inherent membrane properties might explain our resultsBrain Res. 12: 547–565, 1971.is not based on strong supporting evidence from studies of

DAVIS, S., BUTCHER, S. P., AND MORRIS, R.G.M. The NMDA receptorchannel properties in ICc neurons but rather is proposedantagonist D-2-amino-5-phosphonopentanoate (D-AP5) impairs spatial

because channels, together with innervation patterns, shape learning and LTP in vivo at intracerebral concentrations comparable toneuronal discharge patterns and because other explanations those that block LTP in vitro. J. Neurosci. 12: 21–34, 1992.

EHRLICH, D., CASSEDAY, J. H., AND COVEY, E. Neural tuning to soundcan either be ruled out or are unlikely.duration in the inferior colliculus of the big brown bat, Eptesicus fuscus.J. Neurophysiol. 77: 2360–2372, 1997.

FAINGOLD, C. L., BOERSMA ANDERSON, C. A., AND CASPARY, D. M.Concluding commentsInvolvement of GABA in acoustically-evoked inhibition in inferior colli-culus neurons. Hear. Res. 52: 201–216, 1991.Although we can only offer a speculation of the mecha-

FAINGOLD, C. L., GEHLBACH, G., AND CASPARY, D. M. On the role of GABAnism that limits phase-locking in the ICc, there is a signifi-as an inhibitory neurotransmitter in inferior colliculus neurons: iontopho-cant, yet puzzling feature of the auditory system revealed retic studies. Brain Res. 500: 302–312, 1989.

by the results of this study; specifically, the mechanisms FORTUNE, E. S. AND ROSE, G. J. Passive and active membrane propertiescontribute to the temporal filtering properties of midbrain neurons inthat limit phase-locking in the ICc are different from thevivo. J. Neurosci. 17: 3815–3825, 1997.coincidence detection of excitation and inhibition that limits

FRISINA, R. D., SMITH, R. L., AND CHAMBERLAIN, S. C. Encoding of ampli-phase-locking in the MSO and DNLL. These are puzzlingtude modulation in the gerbil cochlear nucleus. II. Possible neural mecha-

features because they make us wonder why the auditory nisms. Hear. Res. 44: 123–142, 1990a.system recreates what is seemingly the same response prop- FRISINA, R. D., SMITH, R. L., AND CHAMBERLAIN, S. C. Encoding of ampli-

tude modulation in the gerbil cochlear nucleus. I. A hierarchy of enhance-erty at least three times in successively higher order nuclei,ment. Hear. Res. 44: 99–122, 1990b.and why neurons in the ICc employ a different mechanism

FU, X. W., WU., S. H., BREZDEN, B. L., AND KELLEY, J. B. Potassium cur-from the mechanism employed in the MSO and DNLL to rents and membrane excitability of neurons in the rat’s dorsal nucleus ofrecreate the same response property? Our interpretation of the lateral lemniscus. J. Neurophysiol. 76: 1121–1132, 1996.

FUBARA, B. M., CASSEDAY, J. H., COVEY, E., AND SCHWARTZ-BLOOM, R. D.these features is that the mechanisms underlying restrictedDistribution of GABAA, GABAB, and glycine receptors in the centralphase-locking in ICc are not designed for processing AMauditory system of the big brown bat. Eptesicus fuscus. J. Comp. Neurol.per se. Rather they are reflective of a broader processing369: 83–92, 1996.

strategy utilized by the ICc for the integration of temporally GERHARDT, H. C. Acoustic properties used in call recognition by frogs andpatterned excitation and inhibition and are more complex or toads. In: The Evolution of the Amphibian Auditory System, edited by B.

Fritzsch, M. J. Ryan, W. Wilcznski, T. E. Hetherington, and W. Walkow-at least are partially different from the processes employediak. New York: Wiley, 1988, p. 455–483.by lower auditory nuclei. What the full ramifications of these

GLENDENNING, K. K. AND BAKER, B. N. Neuroanatomical distribution ofmechanisms are for information processing of naturally oc- receptors for three potential inhibitory neurotransmitters in the brainstemcurring complex signals at the ICc is, at the moment, unclear, auditory nuclei of the cat. J. Comp. Neurol. 275: 261–285, 1988.

GOLDBERG, J. M. AND BROWN, P. B. Response of binaural neurons of dogbut their clarification represents one of the principal chal-superior olivary complex to dichotic tonal stimuli: some physiologicallenges for auditory neuroscience.mechanisms of sound localization. J. Neurophysiol. 32: 613–636, 1969.

GOOLER, D. M. AND FENG, A. S. Temporal coding in the frog auditorymidbrain: the influence of duration and rise-fall time on the processingWe thank C. Resler for designing and implementing the electronics for

data acquisition. We also thank B. May, E. Bauer, A. Klug, T. Park, and of complex amplitude-modulated stimuli. J. Neurophysiol. 67: 1–22,1992.L. Hurley for critical comments.


by 10.220.33.3 on October 26, 2016


ownloaded from



GONZALEZ-HERNANDEZ, T., MANTOLAN-SARMIENTO, B., GONZALEZ-GON- Neural delays shape selectivity to interaural intensity differences in thelateral superior olive. J. Neurosci. 16: 6554–6566, 1996.ZALEZ, B., AND PEREZ-GONZALEZ, H. Sources of GABAergic input to the

inferior colliculus of the rat. J. Comp. Neurol. 372: 309–326, 1996. PERNEY, T. M. AND KACZMAREK, L. K. Localization of a high thresholdpotassium channel in the rat cochlear nucleus. J. Comp. Neurol. 386:GROTHE, B. Interaction of excitation and inhibition in processing of pure

tone and amplitude-modulated stimuli in the medial superior olive of the 178–202, 1997.mustached bat. J. Neurophysiol. 71: 706–721, 1994. POLLAK, G. D. AND CASSEDAY, J. H. The Neural Basis of Echolocation in

Bats . New York: Springer-Verlag, 1989.GROTHE, B., VATER, M., CASSEDAY, J. H., AND COVEY, E. Monaural interac-tion of excitation and inhibition in the medial superior olive of the mus- POLLAK, G. D. AND PARK, T. J. The effects of GABAergic inhibition ontached bat: an adaptation for biosonar. Proc. Natl. Acad. Sci. USA 89: monaural response properties of neurons in the mustache bat’s inferior5108–5112, 1992. colliculus. Hear. Res. 65: 99–117, 1993.

HAVEY, D. C. AND CASPARY, D. M. A simple technique for constructing REES, A. AND MØLLER, A. R. Responses of neurons in the inferior colliculus‘‘piggy back’’ multibarrel microelectrodes. Electroencephalogr. Clin. of the rat to AM and FM tones. Hear. Res. 10: 301–330, 1983.Neurophysiol. 48: 249–251, 1980. REIMER, K. Coding of sinusoidally amplitude modulated acoustic stimuli

HUFFMAN, R. F., ARGELES, P. C., AND COVEY, E. Processing of sinusoidally in the inferior colliculus of the rufous horseshoe bat. Rhinolophus rouxi.amplitude modulated sound in the nuclei of the lateral lemniscus of the J. Comp. Physiol. [A] 161: 305–313, 1987.big brown bat, Eptesicus fuscus. Hear. Res. In press. REYES, A. D., RUBEL, E. W., AND SPAIN, W. J. In vitro analysis of optimal

JAVEL, E. Coding of AM tones in the chinchilla auditory nerve: implications stimuli for phase-locking and time-delayed modulation of firing in avianfor the pitch of complex tones. J. Acoust. Soc. Am. 68: 133–146, 1980. nucleus Laminaris neurons. J. Neurosci. 16: 993–1007, 1996.

JORIS, P. X. AND YIN, T.C.T. Responses to amplitude-modulated tones in REYES, A. D. AND SPAIN, W. J. Membrane properties underlying the firingthe auditory nerve of the cat. J. Acoust. Soc. Am. 91: 215–232, 1992. of neurons in the avian cochlear nucleus. J. Neurosci. 14: 5352–5364,

1994.JORIS, P. X, AND YIN, T.C.T. Envelope coding in the lateral superior olive.I. Sensitivity to interaural time differences. J. Neurophysiol. 73: 1043– RHODE, W. S. AND GREENBERG, S. Encoding of amplitude modulation in1062, 1995. the cochlear nucleus of the cat. J. Neurophysiol. 71: 1797–1825, 1994.

JORIS, P. X. AND YIN, T.C.T. Envelope coding in the lateral superior olive. RHODE, W. S. AND SMITH, P. H. Physiological studies on neurons in theIII. Comparison with afferent pathways. J. Neurophysiol. 79: 253–269, dorsal cochlear nucleus of the cat. J. Neurophysiol. 56: 287–307, 1986.1998. ROSE, G. J. AND CAPRANICA, R. R. Sensitivity to amplitude modulated

KERR, D.I.B., ONG, J., PRAGER, R. H., GYNTHER, B. D., AND CURTIS, D. R. sounds in the anuran auditory nervous system. J. Neurophysiol. 53: 446–PHACLOFEN. A peripheral and central baclophen antagonist. Brain Res. 465, 1985.405: 150–154, 1987. ROSS, L. S., POLLAK, G. D., AND ZOOK, J. M. Origin of ascending projec-

tions to an isofrequency region of the mustache bat’s inferior colliculus.LANGNER, G. Periodicity coding in the auditory system. Hear. Res. 60:115–142, 1992. J. Comp. Neurol. 270: 488–505, 1988.

ROSS, L. S. AND POLLAK, G. D. Differential ascending projections to auralLANGNER, G. AND SCHREINER, C. E. Periodicity coding in the inferior colli-culus of the cat. I. Neuronal mechanisms. J. Neurophysiol. 60: 1799– regions in the 60 kHz contour of the mustache bat inferior colliculus. J.

Neurosci. 9: 2819–2834, 1989.1822, 1988.LEBEAU, F.E.N., REES, A., AND MALMIERCA, M. S. Contribution of GABA SAINT MARIE, R. L. AND BAKER, R. A. Neurotransmitter-specific uptake

and retrograde transport of [3H]glycine from the inferior colliculus byand glycine mediated inhibition to the monaural temporal response prop-erties of neurons in the inferior colliculus. J. Neurophysiol. 75: 902– ipsilateral projections of the superior olivary complex and nuclei of the

lateral lemniscus. Brain Res. 524: 244–253, 1990.919, 1996.MANIS, P. B. Membrane properties and discharge characteristics of guinea SEMPLE, M. N. AND AITKIN, L. M. Physiology of pathway from dorsal co-

chlear nucleus to inferior colliculus revealed by electrical and auditorypig dorsal cochlear nucleus neurons studied in vitro. J. Neurosci. 10:2338–2351, 1990. stimulation. Exp. Brain Res. 41: 19–28, 1980.

SHNEIDERMAN, A., OLIVER, D. L., AND HENKEL, C. K. The connections ofMITTMANN, D. H. AND WENSTRUP, J. J. Combination-sensitive neurons inthe inferior colliculus. Hear. Res. 90: 185–190, 1995. the dorsal nucleus of the lateral lemniscus. An inhibitory parallel pathway

in the ascending auditory system? J. Comp. Neurol. 276: 188–208, 1988.MULLER-PREUSS, P., FLACHSKAMM, C., AND BEISER, A. Neural encodingwithin the auditory midbrain of squirrel monkeys. Hear. Res. 80: 197– SCHNITZLER, H.-U. Echoes of fluttering insects: information for echolocat-

ing bats. In: Recent Advances in the Study of Bats, edited by M. B.208, 1994.Fenton, P. Racey, and J.M.V. Rayner. Cambridge, UK: Cambridge Univ.NICHOLL, R. A. The coupling of neurotransmitter receptors to ion channelsPress, 1987, p. 226–243.in the brain. Science 241: 545–551, 1988.

SCHREINER, C. E. AND LANGNER, G. Periodicity coding in the inferior colli-OLSEN, J. F. AND SUGA, N. Combination-sensitive neurons in the medialculus of the cat. II. Topographical organization. J. Neurophysiol. 60:geniculate body of the mustached bat: encoding of target range informa-1823–1840, 1988.tion. J. Neurophysiol. 65: 1275–1296, 1991.

SCHULLER, G. Natural ultrasonic echoes from wing beating insects areO’NEILL, W. E. AND SUGA, N. Encoding of target range and its representa-encoded by collicular neurons in the CF-FM bat Rhinolophus ferrumequi-tion in the auditory cortex of the mustache bat. J. Neurosci. 2: 17–31,num. J. Comp. Physiol. [A] 155: 121–128, 1984.1982.

SCHULLER, G., RADTKE-SCHULLER, S., AND BETZ, M. A stereotaxic methodOERTEL, D., WU, S. H., AND HIRSH, J. A. Electrical characteristics of cellsfor small animals using experimentally determined reference profiles. J.and neuronal circuitry in the cochlear nuclei studied with intracellularNeurosci. Methods 18: 339–350, 1986.recordings from brain slices. In: Auditory Function , edited by G. M.

Edelman, W. E. Gall, and W. M. Cowan. New York: John Wiley, 1988, VATER, M., CASSEDAY, J. H., AND COVEY, E. Convergence and divergenceof ascending binaural and monaural pathways from the superior olivesp. 313–336.of the mustache bat. J. Comp. Neurol. 351: 632–646, 1995.OLIVER, D. L. AND HUERTA, M. F. Inferior and superior colliculi. In: The

Mammalian Auditory Pathway: Neuroanatomy, edited by D. B. Webster, VATER, M., HABBICHT, H., KOSSL, M., AND GROTHE, B. The functional roleof GABA and glycine in monaural and binaural processing in the inferiorA. N. Popper and R. R. Fay. New York: Springer-Verlag, 1992, vol. 1,

p. 168–222. colliculus of horseshoe bats. J. Comp. Physiol. [A] 171: 541–553, 1992a.VATER, M, KOSSL, M., AND HORN, A.K.E. GAD- and GABA-immunoreac-PARK, T. J., AND POLLAK, G. D. GABA shapes a topographic organization

of response latency in the mustache bat’s inferior colliculus. J. Neurosci. tivity in the ascending auditory pathway of horseshoe and mustache bats.J. Comp. Neurol. 32: 183–206, 1992b.13: 5172–5187, 1993a.

PARK, T. J. AND POLLAK, G. D. GABA shapes sensitivity to interaural inten- WENSTRUP, J. J., ROSS, L. S., AND POLLAK, G. D. Binaural response organi-zation within a frequency-band representation of the inferior colliculus:sity disparities in the mustache bat’s inferior colliculus: implications for

encoding sound location. J Neurosci 13: 2050–2067, 1993b. implications for sound localization. J. Neurosci. 6: 921–973, 1986.WINER, J. A., LARUE, D. T., AND POLLAK, G. D. GABA and glycine in thePARK, T. J. AND POLLAK, G. D. Azimuthal receptive fields are shaped by

GABAergic inhibition in the inferior colliculus of the mustache bat. J. central auditory system of the mustache bat: structural substrates forinhibitory neuronal organization. J. Comp. Neurol. 355: 317–353, 1995.Neurophysiol. 72: 1080–1102, 1994.

PARK, T. J., GROTHE, B., POLLAK, G. D., SCHULLER, G., AND KOCH, U. WU, S. H. AND OERTEL, D. Intracellular injection with horseradish peroxi-


by 10.220.33.3 on October 26, 2016


ownloaded from



dase of physiologically characterized stellate and bushy cells in slices of modulated signals in the dorsal nucleus of the lateral lemniscus of themustache bat and the roles of GABAergic inhibition. J. Neurophysiol.mouse anteroventral cochlear nuclei. J. Neurosci. 4: 1577–1588, 1984.

YANG, L., LIU, Q., AND POLLAK, G. D. Afferent connections to the dorsal 77: 324–340, 1997.ZOOK, J. M. AND CASSEDAY, J. H. Cytoarchitecture of auditory system innucleus of the lateral lemniscus of the mustache bat: evidence for two

functional subdivisions. J. Comp. Neurol. 373: 575–592, 1996. lower brainstem of the mustache bat, Pteronotus parnellii. J. Comp.Neurol. 207: 1–13, 1982a.YANG, L. AND POLLAK, G. D. GABA and glycine have different effects on

monaural response properties in the dorsal nucleus of the lateral lemniscus ZOOK, J. M. AND CASSEDAY, J. H. Origin of ascending projections to inferiorcolliculus in the mustache bat. Pteronotus parnellii. J. Comp. Neurol.of the mustache bat. J. Neurophysiol. 71: 2014–2024, 1994.

YANG, L. AND POLLAK, G. D. Differential response properties to amplitude 207: 14–28, 1982b.


by 10.220.33.3 on October 26, 2016


ownloaded from


analysis of the role of inhibition in shaping responses to

Documents