Intracellular Recordings from Cat Cochlear Nucleus During Tone Stimulation

Author(s): Starr, A; Britt, R | Abstract: ALL OF TIIE vmth NERVE FIBERS originating in the coch lea terminate in the cochlear nucleus, the first central station of the auditory pathway. Extracellular recordings from units in this nucleus have shown a diversity of both spontaneous and tone-evoked activities compared to activity in Vlllth nerve fibers. Tone-evoked inhibition of spontaneous activity occurs with certainty in the cochlear nucleus (8, 9, 12), but has not been securely demonstrated in VIIlth nerve (5, 9, 10, 18, 22), and response patterns to steady tones are quite varied in cochlear nucleus (15), whereas a single response pattern characterizes VIII th nerve fibers (I 0). The synaptic events in cochlear nucleus that accompany acoustic stimulation have only recently begun to be explored (4, 6). The present study utilizes intracellular recordings and stimulating techniques in the cochlear nucleus to further delineate the synaptic basis for central processing of auditory signals.

ALL OF TIIE vmth NERVE FIBERS originating in the coch lea terminate in the cochlear nucleus, the first central station of the a uditory pathway. Extracellular recordings from units in this nucleus have shown a diversity of both spontaneous and tone-evoked activities compared to activity in Vlllth nerve fibers. Tone-evoked inhibition of spontaneous activ ity occurs with certainty in the cochlear nucleus (8,9,12), but has not been securely demonstrated in VIIlth nerve (5,9,10,18,22), and response patterns to steady tones are quite varied in cochlear nucleus (15), whereas a single response pattern characterizes VIII th nerve fibers (I 0).
The synaptic events in cochlear nucleus that accompany acoustic stimulation have only recently begun to be explored (4,6).
The present study utilizes intracellular recordings and stimulating techniques in the cochlear nucleus to further delineate the synaptic basis for central processing of auditory signals.

MATERIALS AND METHODS
Thirteen cats were studied. They were anesthetized with sodium pentobarbital (35 mg/ kg initially and 30 mg every 3 hr thereafter), intubated, paralyzed with gaHamine triethiodide (Flaxedil, I ml / hr), and artificially respired. Rody temperature was maintained between 37 and 39 C. The animal was held by a face clamp, leaving the pinna and external auditory canals exposed. The lateral portion of the cerebellum was aspirated to expose the right dorsal cochlear nucleus, and a recording micropipette lowered to the surface of the nucleus. Brain pulsations were reduced by an agar solution placed over the exposed portions of the brain stem aud by a pneumothorax. The animal was then placed in a sound-attenuating room separated from the experimenters and the recording equipment.

Sound signals
Sinusoidal voltages for tonal signals were generated by a voltage-controlled oscillator, passed through an electronic switch for regulation of rise and decay times (5 msec), attenuated, and presented by an Altec 802 D speaker placed 15 cm in front of the right pinna. A staircase integrator that generated 42 equal voltage levels over a 20-sec period provided the voltages for regulating the frequency oC the oscillator. The resulting acoustic signals were a series of ascending frequency tone bursts (/1, f2, f3 . . . /42) with a constant number oC Hertz separating adjacent tones in the sequence. The electronic switch gated each tone burst to be approximately 150-200 msec in duration. The output of this system measured by a Y2 inch Brue! and Kjaer microphone at the level of the external auditory canal was ± 5 db from I to 14 kHz and ± 9 db from .5 to 22.0 kHz. Intensities up to 90 db SPL were used.

Data aqu isition
The micropipeues were filled with 2 M potassium citrate and had impedanc.es between I 0 and 35 megohms. A negative-capacitance circuit (bandpass dc-30 kHz) amplified the potential difference between the micropipette and an indifferent Ag-AgCl electrode placed on the tongue. A modified bridge circuit (2) was used, in the last seven animals, to pass current through the micropipette. The amplified activ-iLy was displayed on an oscilloswpt: and recorded on magnetic tape along with the oscillator control voltages, oscillator output, current pulses, and event markers for subsequent data processing.

Experimental procedures
The micropipette was slowly advanced by remote control during tone presenlation. Extracellular recordings of single units were characterized by biphasic spikes and little or no change in the d-c level. The electrode was considered intracellular or quasi-intracellular (1 1) if there was a drop in the cl-c level and cell d ischarges were of a positive polarity. The experimental protocol consisted of 1) presenting the sequence of ascending frequency tone bursts at several different intensity levels, beginning at 90 db SPL and then decreasing in IO db steps; 2) presenting 40 trials of just one particular frequency at different intensity levels; and 3) passing intracellular currents during tone stimulation. Most of the units were lost before the entire protocol was completed.

Data analysis
Resting membrane level, spike heights, and effects of sou nd stimulation were determined from kymographic records made of every unit recorded on magnetic tape. The tapes were also analyzed by a digital computer (LINC) as follows.
RESPONSE-AREA HISTOGRAM. The number o[ discharges occurring during ead1 frequency tone burst for the different intensities tested.
POSTSTll\IULUS T IME lilSTOGRAM (PSTH). The number of discharges occurring during sequential 2-msec epochs of a tone burst. The analysis was initiated I 0 msec before tone onset and included 40 trials of one tonal frequency. AVERAGE MEMBRANE POTENTIAL. The shape of the membrane potential during tone stimulation for subsequent comparison with the shape of the PSTH. The analysis was initiated I 0 msec before tone onset and included 40 trials of one tonal frequency; the digitizing rate was 2 msec. A low-pass filter (3 db down at 37.5 Hz, 20 db/ octave slope) was used to attenuate spike discharges in the computation. I NTEGRATED MEMBRANE LEVEL. A sum of the difference between the membrane potential 10 msec before tone onset and the membrane potential every I 0 msec during each frequency tone burst for the different intensities tested. The low-pass filter in (3) was used to attenuate spike discharges in the computation. The imegrated membrane level is a measure of membrane potential change as a function of tonal frequency and intensity and can be compared with the response-area histogram.

RESULTS
Intracellular and quasi-intracellu lar records were obtained from 58 cells in the cochlear nucleus. The electrode was judged in tracellular if 1) a large negative d-c shift (20-70 mv) developed abruptly on encountering the cell, 2) the spike discharges were of a positive polarity, and 3) both resting membrane level and discharges disappeared abruptly with movement of the electrode or in the course of study. Satisfactory intracellular recordings were maintained in 20 cells for periods ranging from 20 sec to IO min. Data derived from the other 38 cells seemed compatible with an electrode in close proximity to the cell, but not yet intracellular, a recording situation termed quasi-intracellular b y Mcilwain and Creutzfeldt (1 1). Membrane potentials and spike amplitudes were of low amplitude when cells were first encountered and both would increase gradually on advancing the electrode. Eventually, injury discharges and loss of membrane poten tial accompanied one of the electrode movements. Our experimental procedure was to stop the electrode and to study responses to tone stimuli when the membrane potential was steady, the synaptic potentials clear, and the spikes of positive polari ty. Recordings of up to 40 min were achieved with this technique from cells that had membrane levels between -IO and -40 mv and spike ampliludes of 1-15 mv.
The potential changes that accompanied tonal stimulation are illustrated in the intracellular records from one of our better units (Fig. I) Frequency of the tone is indicated on alternate tone bursts. Note depolarizing shifts with tones that increase discharge rate, and h yperpolarizing shifts wiLh tones that suppress activity. A large hypcrpolarization occurs at offset of most of the tones. i n discharge ra te. Hyperpolarization during excitatory tone stimulation was occasionally noted when the electrode was in an extracellular location. On advancing the electrode to penetrate the cell, the hyperpolarizing response, which was never greater than l mv in amplitude, inverted to a depolarization of considerably larger m agnitude ( well as wi th the frequency and intensity of the tonal stimuli. In Fig. 1, for example, depolarizing shift was interrupted shortly a fter LOne onset by a brief hyperpolarization durjng som e frequencies (6.0-8.0 kHz) but not others (3.5-4.5 kHz). The relationship of the depolarizing shifts to the pattern of cell firings will be discussed in detail in a later section of the paper.
A sustained hyperpolarization was noted during stimulation with tones that sup- Three types of membrane potential changes were noted on terminating the tone bursts. In 33 cells there was a profound h yperpolariza tion (u p to 20 mv) that was maximal at tone offset and then grauually d ecreased in amplitude over the subsequent 200 msec. Spontaneous activity was suppressed. throughout this period. H yperpolarization at tone offset was prominent following excitatory tones but, as is evident in Fig. I, also a ppeared following tones that either had no effect on spontaneous activity (3.0 and 8.5 kHz) or even suppressed spontaneous firin gs (9.0 kHz). In 8 cells, the hyperpolarization was interru pte<l 20-50 m sec after tone offset by an abrupt depolarization that then persisted for 100-200 msec (see Fig. 5, 13 kHz, 90 db).
T his rel>uuntl uepularization was restrkted to tones in the region of the "best" frequency if their intensity were at least 60 <lb above threshold. Fina ll y, in 5 cells, stimul ation with inten e excitatory tone bursts (> 70 db SPL) resulted in a de polarization shift that persisted for 20-100 m sec beyond the tone's termina tion (Fig. 7, cell 13-1) . The depolarizing shift, in 4 of these cells, was followed by the usual off-hyperpolar-i1a t io n.
The amplitudes of the depolarizing and hyperpolarizing shifts d uring tones and the hyperpolarization at tone offset varied from I to 20 mv and are plotted in Fig. 3 A. STARR AND R. BRITT as a function of the unit's resting membrane level. A plot of spike amplitude as a function of membrane level is also included in this figure. All of these measures tended to be larger in cells with greater resting membrane potenLials.

R elation between membrane potential changes and spike discharges
In 12 unils the number of discharges evoked by the different frequency tone bursts presented over at least a 60-db intensity range (response-area histogram) were compared with the integrated membrane level evoked by these same signals. The measures were displayed as two tuning curves; one based on spike counts and the other on membrane potential change. Results from one of the intracellularly recorded units are graphed in Fig. 4 and show the two tuning c:urves to be in general accord except for the following: First, th e range of tonal frequencies that excitated the cell was more extensive when defined by membrane depolarization than by spike count. Second, tonal frequencies adjacent to the excitatory area could be classified as inhibitory by membrane hyperpolarization, whereas the absence of spontaneous discharges m ade it impossible to classify a ny frequency as inhibitory in spike-count measurements. Third, threshold as defined by membrane depolarization was at least IO db lower at every frequency than threshold defined by the occurrence of cell disch arges. Thus, for this cell, membrane potential measurements seem an accurate reflection of the syn aptic events governing cell discharges. H owever, in 8 of the other 11 cells examined by thi s method, membrane potential changes were often a t va riance with the spike-count data. The results  from one of these cells (cell 20-6, Fig. 5) demonstrate the types of noncorrespondence noted. In the absence of a tone signal, cell discharges were always preceded by an EPSP. However, not every EPSP that depolarized the cell to its firing level was followed by spike initiation. For instance, during 15 sec of spontaneous activity, only 286 of the 350 EPSPs that depolarized the cell to its firing level were followed by a cell discharge. Low-intensity tone signals (up to 20 db SPL) did not alter the relationship between membrane potential and spike initiation, but changes a ppearecl as the tone intensity was r aised. At 90 db SPL, a level of depolarization sufficient to evoke discharges at approximatel y 100/sec during the best frequency signal (13 kHz) was ineffective in evoking even a single discharge with a tone only slightly removed in frequency (15 kHz). In add ition , a small spike (.5-1.5 mv in amplitude, < I msec in duration), distinct from the cell's larger spike (5 mv in amplitude, 1-2 msec in dur ation), appeared in the period after the tone termination. This smaller spike was never noted to h ave a preceding EPSP or to occur simultaneously with the large spike. When the cell was lost both spikes disappeared. Finally, the extent of the inhibitory surround, as defined by suppression of spontaneous discharges, involved a wider range of tonal freq uencies than those eliciting IPSPs. The observations that with intense tonal signals spike firings might not be related to the level of evoked depolarization and that there could be suppression of spontaneous activity without observable IPSPs were also characteristic of the other 7 cells in wh ich membrane potential and spike firings had a variable relation. However, the finding of additional small spikes seeme<l to be a peculiarity of cell 20-6.

Poslstimulus tim<' histog1·ams and average membrane level
In 21 units the pattern of discharges evoked by a tone stimulus (PSTH) was compared with the simultaneous changes in membrane level (averaged membrane level). The two measures were analyzed by visua l inspection, taking into <1ccount the phase shift in the r ise and decay portions of the aver<1ged membrane levels introduced by the low-pass filter. The shapes of the PSTH and averaged membrane level seem ed comparable whenever tones suppressed spontaneous activity and also evoked a sustained h yperpol<1rizing shift (Fig. 6). In contrast, firin g· pallern and membrane potentials had a variable relationship during excitatory tone bursts. In some cells the two measures h ad comparable shapes (left side Fig. 7, Fig. 8), whereas signifi cant d ifferences were apparent in responses obtained from other cells (right side Fig-. 7). In the latter group for instance, the occurrence of cell discharges did not necess<1rily correla te with the level of membrane depolarization. Cell di charges might be initiated afler a delay of 20-100 msec following tone onset and then increase gradu <1 ll y in freq uency to a steady r ate even though the membrane was depolarized at either the same level or in a gradually decrea!iing fashion throughout the tone burst. Features of the acoustic stimu lus proved to be an important vari able affecting the correspondence between firing pattern and averaged membr ane level. In records from the same cell the two measures could be markedly different during one tonal frequency but well correlated if that same tone were reduced in frequency (cf. respomes to 13 kH1. in  increa ed with ano<lal currents and <lecreased with cathoda l cu rrents. Thresholds for affecting discharge rates ranged from 0.2 to 3 n a and are comparable to current strengths used by other investiga tors for altering cell discha rge rates in situa tions where the electrode is known to be in an intracellular position. In the present experiments cell discharges evoked by applied currents could be modified by the simultaneous presentation o( tones. Frequencies that suppressed spontaneous activity also ~u ppressed discharges evoked by the passage of currents (Fig. 9). Furthermore, even in Lhose cells wi thout spontaneous acuv1 ty, disch arges evoked by anodal currents were also r educed in ra te whenever tones adj acenl to the exci tatory frequencies were preented (Fig. 10). Thus, cellular excitability to injected currents is m arkedly reduced during the presentation of tones in the inhibitory surround. In cell 18-15, for instance, the threshold for firing the cell by  anodal currents increased from 1.2 na, in the absen ce of tonal stimulation, to 6 na when a tone in the inhibitory surround was concomitantly presented . Finally, the excitability of cells to anodal currents was also markedly depressed in the period imm ediately followin g tone offset when both Each bar in the histogram represents the total num bcr of spikes evoked by one tone freq ucncy. In B, disd1argc ra tes duri ng two different st rengths of depolari7.i11g currents arc shown in the photographs la beled I na, 8 na and arc to be compared with clisc:h argc rales whe n these same currents were applied during th e tonal signals at 60 db (+60). Note the d ecrca~ed rate of current -evoked disch a rges whe n tones in the inhibitory surround (f 8-11 , 5-6.5 kHz and f 17-26, 9.5-14 kHz) were presented. spontaneous activity was suppressed an cl the membrane hyperpolarized .

DISC USSIO
The presen1 experiments show excita tory and inhibilory syn aptic events in cells in the cochlear nucleus during acoustic stimulation. Depolarizing shifts accompanied tones that increased firing rate, and hyperpolarizing shifts accompan ied tones that su ppressed activi ty. T h e classification of the hyperpolarizing response as an inhibitory postsyna ptic potential is suggested by the marked increase in intracellular depo-·"" B fJG. 10. Effect of stead y depola rizing current (B) o n response-area histogram (A) in a cell with very low rates of spontaneous activity (< I /sec). T hree na of depolarizing currcm increased discharge rate to approximately 50/ sec. Two distinct bands of tonal freq uencies < :an be seen to decrease the ra te of current-evoked disch arges. The first extending from (4-9 , 3-5.5 k Hz e n com passing freque ncies adjacent to t he unit's cxci rnrory response area; the second, extending from /26-32, 14-17 kHz seen best with tone inte nsities between 20 a nd 60 db PL. larizing currents required to fire the cell during-the hyperpolarizing shift com pared to the current levels neecled to fire the cell in the r esting state (7). These results complement Gerstein, Butler, and Erulkar 's (6) recent intracellular stu<ly in cochlear nucleus of cat in which IPSPs were not seen though the inhibitory effects of tones were apparent from discharge patterns and spike-count data. Our results a ttest to the difficulties of recording IPSPs in cochlear nucleus. Of the 29 cells in which discharges were suppressed by tones, IPSPs could be detected in only 16. Furthermore, even jn, these 16 cells, IPSPs were no t seen with ever y tone th at suppressed spontaneous activity though the inhi bitor y postsyna ptic efiects of these tones were apparent from the marked depression in responsiveness to intracellular depolarizing currents that accompanied their presenta tion (7).
The fai lure to deL ect JPSPs suggests that the recording electrodes were remote from the zone of spike generation a nd thereby were unable to accurately sample all of th~ syn a p~ic .even ts governing spike initia uon. Tl11 s inter pretation could also account for the frequent inconsistencies in fi ring levels noted for these cells. An elec-tro~e remote from the zone of spike genera t1o n can provide in form ation about the distri bution o f excitatory and inhibitory syn a pses on the cell surface. The finding th at d epo.larizing shifts almost in var ia bly accompanied tones tha t increased firing rates (43 of 4.9 cells) suggests that excitatory synapses are widely distributed over the cell surface and can be detected by an electrode almost anywhere in the cell. Inhibitor y syna pses, on th~ other h and, appear to be restricted to regions close to the spike generator, as suppression of discharges was freq \1Pn tly norecl in th e absence of a recordable hyperpol arizing shift. Furthermore, depolarizing shifts were not always associated ~~th spike discharges, whereas hyperpolar-1zmg res~onses were, without exception, acc~mpan1~d b~ a ~up~ression of cell firing. A differential d1stnbuuon of exci taLOry and inhibitory input h as been defined for the anterior horn cdl of the cal with inhibitory syn apses located close to the cell som a but excitatory syn apses distribu ted over both the soma and dendrites (19). In the cochlear nu cleus the occurrence of inhi bitor y synap ses d ose to the spike-generating region would be an efficient means for restrictinolhe cell's acouslic r esponse characteristics~ ~oreover, some of th e patterning of spike disch arges demonstrated for cells in this nucleus (15) could be accoun w 1 for by ~em poraJ cl~aracteristics of the inhibitory m pUl. F~r m stance, in the presen t experim ent units that had a b urst of discharges a t ~one on set followed by a brief period of silence and then sustained firi ngs showed depolarizing shifts to tones th at were interru pL e<l shortly after onset by an IPSP of short duration.
The present experiments show that inhibition in the cochlear nucleus is clearly postsyn aptic but provide no explanation as to the origin of the inhibi tor y in put. It is apparent that inhibition has a well-defined rel a tionsl~i p to. th~ cell's acou stic response ar ea. This spec1fic1ty could d erive from inhibitory i nterneurons within the nucleus that a~e r~spo ?si ve to a lim ited range of acoustic stimuli, or from VIIIth nerve fibers themselves that are inhi bitory i n action.

A f terefjects of sound
Cells in t he cochlear nucleus have a m arked su ppression of spontaneous activity foll owing soun d stimulation tha t can persist for many seconds (2 1) or even minutes (20). It is no t known whether the change in spontaneous acti vity is due to al ternations in afferen t input since VIIIth 1~e~ve fibers ~lso show a suppression of ac-tlv1ty followmg tone stimula tion (10), or whe~her lhe cochlear nucleus ch anges are mediated cent rally. T he finding in the presen t experiments of h yper polarization in cochlear n ucleus cells at tone offset when spontaneous activity was suppressed is evidence in support of central neu ronal mechanism for this aftereffect. Furthermore, the observa tion tha t d isch arges evoked by intracellular d epolarizing currents were also su ppressed at tone offset ind icates that the excitability cha nge occurs a t the level of the cochlear nucleus cells themselves. The hyper polarizing sh ift at tone offset could represent a change i n permeability to Clor K • ions .as occurs typically in IPSPs (3) o.r to ) fa+ ions as occurs following activation of the electrogenic sodium pump (I). The la tter m ech ani sm has been implicated for h yperpolarization and inhibition following a period of in tense stimulation that occurs in inver tebrate receptors (13), invertebrate cen tral nervou s system neurons (14, I 6), and mammalian C fi bers (17). The observations, in the presen t experiments, that h yperpolarization at tone offset increased in amplitude as tone intensity was raised and occurred in cells in which IPSPs to inhib itor y tone bursts could n ot be detected raise the possibility that an electrogenic sodi~m pump might participate in its generation. Our data cannot d istinguish between the alterna tive ionic mechanisms for hyper polarization at tone offset as no attem p t was made to measure the equilibri um potential or to cleterm ine the effects of intracellular injections of appropriate ions. In either case, hyperpolarization at tone offset is a powerful means of regula ting spontaneous activity and m ay be the neural equivalent o( some of the perceptual aftereffec:ts of sound stimu]ation.

SUMMA RY
Intracell ular and quasi-intracellular recordings were made from 58 cells in the cochlear nucleus of pentobarbital-anesthetized cats during the presenta tion of tone bursts covering the frequencies from 0.5 to 22.0 kHz and intensities up to 90 db SPL.
In some cells there was a good correspondence between membrane potential changes and spike firings . D e polarizing sh i fLc; accompanied tones that increased discharge rates, and hyperpolarizing shifts accompanied tones that suppressed activity. The configuration of the membrane potential changes closely resembled the tempor al patterning of spike disch arges. In other cells membrane potential changes often failed to correspond to cell discharges. Suppression of activity could occur without a detectable hyperpol ar izing shift, and the shape of the depolarizing sh if ts was of ten at variance with the pattern of cell activity. It is assumed that the electrode, in the latter group of cells, was in a site r emote from the zone of spike generation.
Passage of intracellular currents defined that tones in the inhibitory surround always effected a marked depression in cel-lular excitability even in the absence of a h yperpolarizing response.
These results suggest that excitatory and inhibitory synaptic i nputs are distributed differently in cochlear nucleus cells: inhibitory inputs occurred principally near the zone of spike generation, so their effects may not be detected b y an electrode in a remote portion of the cell, whereas excitatory synapses are widely distributed over the cell surface.
Three types of membrane potential changes were note<l at ton e offset: 1) hyperpolarization, 2) hyperpolarization followed by a rebound depolarization, and 3) a sustained depolarization. H yperpolarization at tone offset occurred most frequen tly and was accompanied b y a suppression of spontaneous activity and a decrease in responsiveness to stimulation with intracellular d epolarizing currents.
ACKNOWLEDGMENTS W e express our appreciation to Birthe Nyholm for her illustrations and Pamela Barrett for secretarial assistance. 14. NICHOLLS, J. G. AND BAYLOR, D . A. Long-