Many of the methods used for these experiments have been described in detail elsewhere (Smith and Rhode 1985; Smith et al.1991, 1993b) and are summarized here. Animals were maintainedin an American Association for Accreditation of Laboratory AnimalCare (AAALAC)-approved animal care facility, and all methods havebeen approved by the University of Wisconsin Institutional AnimalCare And Use committee.
Young adult cats were anesthetized and maintained in an areflexive state with pentobarbital sodium (35 mg/kg). Pinnae wereremoved, and the external auditory meati cut transversely forinsertion of metal ear pieces through which acoustic stimuli,calibrated from 60 Hz to 40 kHz, were delivered. The MNTB wassurgically approached from the ventral surface of the brain stem.A small hole was drilled in the basioccipital bone just lateralto the pyramidal tract, and a slit was made in the dura. The smallrootlets of cranial nerve VI served as an external landmark forthe more deeply located MNTB.
Acoustic stimuli and data collection
Calibrated acoustic stimuli generated by a computer-controlled digital stimulus system (Rhode 1976) were delivered from Telex140 earphones or Radioshack supertweeters. Spike-triggered pulseswere sent to a unit event timer and stored for on-line and subsequentanalysis. As the electrode was advanced toward the MNTB, a searchstimulus of short tone bursts with variable frequency was presentedto both ears until an axon or a cell body was encountered.
Glass electrodesrecording and injection
Intracellular glass electrodes were filled either with a buffered (pH 7.6), filtered 5% HRP (Sigma) solution in 0.5 M KClor a 2% Neurobiotin (Vector Labs) solution in 0.5 M KCl. The electrodewas lowered over the hole drilled in the skull and advanced in1-µm steps. Extracellular and intracellular signals were monitoredusing standard techniques for DC monitoring, amplification, filtering,and display. Recordings from MNTB cells using the glass microelectrodescould only be verified by subsequent location and identificationof the labeled cells. However, some features of axonal responseswere used to distinguish presumed MNTB axonal responses from thoseof other axons: 1) because the primary excitatory input to a MNTBcell arises from a globular bushy axon from the contralateralcochlear nucleus, only units driven from the contralateral earwere considered. 2) Because of the interposing synapse betweenglobular bushy axon and MNTB cell, first spike latencies takenfrom peristimulus time histograms (PSTHs) to STCFs should be slightlylonger for a MNTB cell axon than for a bushy cell axon with thesame characteristic frequency.
While tones at the unit's CF were presented and the response monitored, entry into the axon, as signaled by a DC shift offrom 30 to 60 mV, was accomplished with 100-ms current pulses.After physiological characterization of the cell (see further)HRP or neurobiotin was injected for 2-10 min using 100 ms, 1-5nA current pulses while CF tones were presented continuously.Current injection was terminated if the response during pausesin the injection changed, which would indicate that the electrodehad slipped into a different cell. After data collection, theelectrode was withdrawn, the DC shift noted and a 10-mV calibrationpulse recorded on the intracellular channel of the tape monitor.
Commercially available tungsten metal electrodes (Microprobe, 10-20 µM exposed tips, 5 M impedances) were used in separateextracellular experiments. Electrode placement, method of advance,stimulus presentation, and unit response characterization wereas described above for glass electrodes. Metal electrode recordingsfrom MNTB cells were distinguished by their prepotentials (Guinanand Li 1990; Guinan et al. 1972a,b; Li and Guinan 1971) and bysubsequent histological location of the recording site from lesionsmade at several points in each penetration (except for 4 cellsin 1 animal where histology was not available).
For metal electrode experiments, we perfused the animal with formalin shortly after a lethal dose of pentobarbital sodiumwas administered after the last penetration. For the HRP/Neurobiotinexperiments, the animal was maintained in an areflexive statefor 24-36 h after the last penetration using a drip solution ofsodium pentobarbital (3% in 5% dextrose) while heart rate, respiration,and withdrawal reflexes were checked every 20 min by one of theexperimenters. Wound areas were swabbed at 2-h intervals withlidocaine. After a lethal dose of sodium pentobarbital, the catthen was perfused transcardially with saline followed by two concentrationsof phosphate-buffered, calcium-containing glutaraldehyde:paraformaldehydefixative (0.01:0.01 and 0.02:0.01), and the brain stored eitherin sucrose buffer (for frozen sectioning) or 1:1 glutaraldehyde:paraformaldehydefixative (for vibratome sectioning).
For the HRP injections, the following steps were taken. Coronal or horizontal sections were cut at 60- or 70-µm thicknesswith a vibratome or after freezing and then were reacted usingthe 3,3'-diaminobenzidine (DAB)-nickel/cobalt intensificationmethod (Adams 1981). For light microscopy, sections were mountedon glass slides, counterstained with cresyl violet, and coverslipped.For electron microscopy, vibratomed, HRP-reacted sections werefixed in 2% osmium tetroxide, dehydrated, and flat-embedded inplastic resin. After a camera lucida drawing of the injected axonwas made, thin sections were taken, counterstained with uranylacetate and lead citrate, and observed with a JOEL 100CX electronmicroscope.
For the Neurobiotin injections, the following steps were taken. Vibratomed 70-µm section were cut into 0.1 M phosphate buffer,washed, incubated in 0.5% H202 in phosphate buffer, then reactedovernight in ABC reagent (Vector ABC kit, in 0.1 M phosphate buffer,pH 7.4 containing 0.3% TritonX, 2% bovine serum albumin, 2 dropsof A and 2 drops of B reagent of ABC kit/10 ml). The next daythe sections were rinsed in 0.1 M phosphate buffer and reactedusing the Adams (1981) DAB-nickel/cobalt intensification methodthen prepared for light and/or electron microscopy as describedabove.
When a cell or axon was isolated, stimuli were presented to each ear separately to determine which ear(s) drove the unit.For all cells, we determined the characteristic frequency (CF,the frequency at which the threshold intensity was the lowest),spontaneous discharge rate, and Q10 (CF/bandwidth at 10 dB abovethreshold) using an automated threshold tuning curve program.We also measured the STCF at 10- to 20-dB intervals from nearthreshold to 50-70 dB above threshold using 100 or 200 repetitionsof 25-ms tones presented every 100 ms with a 3.9-ms rise/falltime. Amplitude-modulated and click stimuli also were used butthe data will not be presented here.
From the STCF response, we derived the mean and standard deviation of first spike latency (see Smith et al. 1991), mean coefficientof variation over the interval 12-18 ms (Young et al. 1988), andthe sustained discharge rate and synchronization coefficient (Goldbergand Brown 1969) during the last 15 ms of the tone burst. To correctfor spontaneous activity in the computation of the mean and standarddeviation of the first spike latency, we eliminated all spikesin the time window from 0 to 2.4 ms after stimulus onset and alltrials in which a spike occurred in the last half of this window.
Labeled axons were drawn in the coronal or horizontal plane using a camera lucida with ×63 oil objective. Measurements ofthe features of terminals on MNTB principal cell bodies and dendriteswere quantified using a Summagraphics bit pad Two graphics tabletinterfaced to the same microVax computer used to generate auditorystimuli.