Ensemble responses of the auditory nerve to normal and whispered stop consonants.

Whispered syllables lack many of the frequency and voicing cues of normally voiced speech, but these two acoustically distinct forms of speech are placed into the same linguistic categories. To examine how whispered and voiced speech are encoded in the auditory system, the responses to speech sounds were recorded from 132 single auditory nerve fibers in 20 ketamine anesthetized chinchillas. Stimuli were the naturally produced syllables /da/ and /ta/ presented in whispered and normal voicing. The results for each syllable presented at a fixed intensity were analyzed by pooling the responses from individual auditory nerve fibers across animals to create a global average peri-stimulus time (GAPST) histogram. For each word-initial consonant, the pattern of peaks in the GAPST was the same for both normal and whispered speech. For the vowel the GAPSTs for the whispered speech sounds did not display the synchronization observed in the responses to the voiced syllables. The temporal pattern of the peaks was constant over a 40 dB intensity range, although peak sizes varied. Grouping fibers within different frequency ranges created local averages (LAPST) that revealed the significant contribution of high frequency fibers in the response to the whispered consonants. Responses of individual fibers varied with both the syllable and the voicing. These findings suggest that the encoding of either a whispered or a normal stop consonant results in the same temporal pattern in the ensemble response.

Responses of auditory nerve fibers to trains of clicks.

Responses of auditory nerve fibers to trains of clicks were recorded in ketamine anesthetized chinchillas. By varying the number of clicks and the interclick interval, this study examined whether "post-onset adaptation," described in psychoacoustic experiments on localization, occurred in auditory nerve fibers. The results showed that the number of action potentials recorded from a nerve fiber in response to a train of clicks was a power function of the number of clicks. For interclick intervals of 2 ms or greater the exponent of the power function was 0.5, and this exponent did not change over a 20-dB range of intensities. The timing of action potentials relative to the click stimuli was measured using synchronization coefficients. The coefficients increased with interclick interval, decreased with increasing intensity, and were greater for fibers with low rates of spontaneous activity than for high spontaneous fibers. Recovery functions showed that for interclick intervals of 2 ms or more, the responses to the second click were at least 70% of the response to the initial click. The recovery depended upon the number of clicks in the train. These findings indicate that auditory nerve fibers respond to high rates of stimulus presentation and do not display the adaptation observed in localization studies.

Rapid inhibition in the cochlear nuclear complex of the chinchilla.

This study examined responses to pairs of clicks recorded extracellularly from single units in ventral cochlear nuclei (VCN) of ketamine-anesthetized chinchillas. The response to the trailing click was suppressed for interclick intervals of 1 and 2 ms, but little suppression was observed for an interclick interval of 4 ms. To determine whether any suppression originated in the dorsal cochlear nucleus (DCN), responses to click pairs were recorded before and after injecting lidocaine in the deep layer of the DCN. For 19% of the neurons (7/37), the response to the trailing click increased following the injection, which is consistent with lidocaine reducing delayed inhibition from the DCN. Unexpectedly, for 62% of the units (23/37) the response to the initial click decreased after lidocaine administration. Three units (3/37 or 8%) showed a combination of both responses. For 52% of the units with a decreased response (12/23), the reduction occurred only for loud clicks (> or = 30 dB above threshold), while at intensities 20 dB lower, responses to pairs of clicks were unchanged. No changes in spontaneous rate were observed. Following lidocaine injections, the tuning curves of 7/12 neurons tested had increased thresholds, but only around the characteristic frequency. These results indicate the presence of two rapid inhibitory inputs onto VCN neurons.

In vitro modulation of somatic glycine-like immunoreactivity in presumed glycinergic neurons.

Previous studies indicate that tuberculoventral and cartwheel cells in the dorsal cochlear nucleus as well as a group of stellate cells in the ventral cochlear nucleus are likely to be glycinergic. To test whether these neurons contain higher levels of free glycine than cells that are probably not glycinergic, immunocytochemical studies with antibodies against glycine conjugates were undertaken on slices of the murine cochlear nuclear complex. Present results show that the cell bodies of all three groups of neurons are immunolabeled. However, the somatic labeling of the tuberculoventral and cartwheel cells can be modulated by experimental conditions. In slices fixed immediately after cutting, many cell bodies in the deep layer of the dorsal cochlear nucleus (DCN), presumably tuberculoventral neurons, are labeled. As a slice is incubated in vitro, cell bodies in the deep layer of the DCN lose their glycine-like immunoreactivity. After 7 hours in vitro, labeled cells are absent in the deep DCN, but the immunoreactivity can be regained by electrically stimulating the auditory nerve for 20 minutes. The loss of immunoreactivity is prevented by electrical stimulation, by axotomy, and by inclusion of 0.8 microM tetrodotoxin, or 1 microM strychnine, or 50 microM colchicine or 50 microM beta-lumicolchicine in the bathing saline. Cartwheel cells retain their immunoreactivity during incubation in vitro without electrical stimulation, but lose it under two conditions. One is following a cut across the ventral cochlear nucleus (VCN) that severs most of their granule cell input, and the other is the inclusion of tetrodotoxin in the bathing saline. The labeling of cell bodies in the ventral cochlear nucleus and of puncta and processes is not changed by any of these experimental manipulations.

Tuberculoventral neurons project to the multipolar cell area but not to the octopus cell area of the posteroventral cochlear nucleus.

Tuberculoventral neurons in the deep layer of the dorsal cochlear nucleus (DCN) provide frequency-specific inhibition to neurons in the anteroventral cochlear nucleus (AVCN) of the mouse (Wickesberg and Oertel, '88, '90). The present experiments examine the projection from the deep DCN to the posteroventral cochlear nucleus (PVCN). Horseradish peroxidase (HRP) injections into the PVCN reveal that the multipolar cell area, but not the octopus cell area, is innervated by neurons in the deep layer of the DCN. Injections into the multipolar cell area, in the rostral and ventral PVCN, labeled neurons across the entire rostrocaudal extent of the deep DCN. The labeled tuberculoventral neurons generally lay within the band of labeled auditory nerve terminals in the DCN. Injections of HRP into the octopus cell area, in the dorsal caudal PVCN, labeled almost no cells within the band of auditory nerve fiber terminals that were labeled by the same injection. The inhibition from tuberculoventral neurons onto ventral cochlear nucleus (VCN) neurons is likely to be mediated by glycine (Wickesberg and Oertel, '90). Slices of the cochlear nuclear complex were immunolabeled by an antibody against glycine conjugated with glutaraldehyde to bovine serum albumin (Wenthold et al., '87). Glycine-like immunoreactivity was found throughout the DCN, the AVCN and the multipolar cell area, but there was little labeling in the octopus cell area. This finding provides independent evidence that tuberculoventral neurons do not innervate the octopus cell area and indicates that the octopus cell area is anatomically and functionally distinct.

Delayed, frequency-specific inhibition in the cochlear nuclei of mice: a mechanism for monaural echo suppression.

To understand how auditory information is processed in the cochlear nuclei, it is crucial to know what circuitry exists and how it functions. Previous anatomical experiments have shown that neurons in the deep layer of the dorsal cochlear nucleus (DCN) project topographically to the anteroventral cochlear nucleus (AVCN) (Wickesberg and Oertel, 1988). Because interneurons in the DCN and their targets in AVCN are excited by the same group of auditory nerve fibers, the projection is frequency-specific. Here we report that microinjections of glutamate in the DCN evoke trains of IPSPs in individual, impaled AVCN neurons in brain slices of the cochlear nuclear complex. Only injections along a rostrocaudal band in the DCN, matching the anatomical projection of tuberculoventral neurons, evoke IPSPs; elsewhere, there were no responses to the glutamate. The inhibition is blocked by 0.5 microM strychnine. Both bushy and stellate cells are targets of the inhibitory projection. Inhibition in the AVCN is delayed by an additional synaptic delay with respect to the excitation. Delayed, frequency-specific inhibition allows the first wavefront to be transmitted to higher auditory centers by bushy and stellate cells, while following inputs encoding signals of similar frequencies are attenuated at least for the duration of an IPSP. These findings are consistent with results from psychoacoustic experiments and suggest that this circuit provides a source of monaural echo suppression.

Jaundiced Gunn rats have increased synaptic delays in the ventral cochlear nucleus.

Recordings were made in vitro from cochlear nuclei of Gunn rats, a strain with a recessive mutation that predisposes rats to hyperbilirubinemia at birth. Delays between shocks to the auditory nerve and earliest synaptic responses of the cochlear nuclear neurons were on average longer in Gunn rats than in heterozygotes. Injections of sulfonamide further increased average synaptic delays in jaundiced rats. Responses to injected current in rats were like those in mice.

Auditory nerve neurotransmitter acts on a kainate receptor: evidence from intracellular recordings in brain slices from mice.

Intracellular recordings from neurons in brain slice preparations of the mouse ventral cochlear nucleus (VCN) were used to examine the actions of excitatory amino acid agonists and antagonists. Synaptic responses to electrical stimulation of the auditory nerve root were partially blocked by kynurenic acid, an antagonist that is specific for glutamate receptors. The antagonists specific for N-methyl-D-aspartate (NMDA), DL-2-amino-5-phosphonovalerate (APV) and Mg2+, did not affect the response, arguing against a role for NMDA receptors at the VIIIth nerve synapse. To test postsynaptic sensitivity to excitatory amino acid agonists, responses to bath applications were measured in VCN neurons while synaptic transmission was blocked by the removal of Ca2+ from the bath or by the addition of tetrodotoxin. Neurons in the VCN were 500-1000 times more sensitive to kainate than to glutamate or aspartate. In the absence of Mg2+, they were also sensitive to NMDA. The responses to kainate and glutamate were increased by the removal of calcium from the bath. These results imply that VCN neurons have both kainate and NMDA receptors and that synaptic transmission between auditory nerve fibers and neurons in the cochlear nuclear complex could be mediated by a substance related to kainate.

Tonotopic projection from the dorsal to the anteroventral cochlear nucleus of mice.

To understand how auditory information is processed in the cochlear nuclei, it is crucial to know what circuitry exists and how it functions. In slice preparations, horseradish peroxidase (HRP) injections into the anteroventral cochlear nucleus (AVCN) reveal two circuits: a connection between the dorsal cochlear nucleus (DCN) and AVCN and a local circuit confined to the AVCN. Extracellular injection in the AVCN labels a band of cells in the DCN. The labeled cells in the DCN lie within a band of auditory nerve fiber terminals that are labeled by the same injection, showing that the connection from the DCN to the AVCN is frequency specific. The injections into the AVCN also labeled a cluster of neurons in the AVCN dorsal to the injection site. These cells may be interneurons that relay information from areas encoding higher frequencies to areas encoding lower frequencies within the AVCN. In the parasagittal plane, the AVCN is organized along two orthogonal axes that are indicated with HRP labeling of fibers and cell bodies. The tonotopic axis runs approximately dorsoventrally; the isofrequency axis runs approximately rostrocaudally. The axons of labeled DCN neurons and the cluster lie along the tonotopic axis, whereas the labeled auditory nerve fibers define the isofrequency axis. Where they cross is where HRP is taken up by the fibers. The area of uptake is small and lies in the middle of the darkly stained injection site.

Binaural interaction in low-frequency neurons in inferior colliculus of the cat. IV. Comparison of monaural and binaural response properties.

We studied the monaural and binaural response properties of 82 low-frequency inferior colliculus (IC) neurons that display a clear sensitivity to changes in interaural phase. Most cells (60%) are excited by sound delivered to either ear, the remainder being excited only by stimulation of one ear; 70% of the neurons receive their stronger or sole excitatory input from the contralateral ear. A monotonic relation between spike discharge and sound pressure level (SPL) is seen in 65% of the monaural response areas, i.e., the range of stimulus frequencies and intensities effective in eliciting a response, while 30% show a nonmonotonic response pattern. In 33% of the cases there is a significant shift in the most effective frequency as a function of SPL. Most discharge patterns are classified as sustained (69%) and the remainder as onset. However, there is considerable variability within these patterns and often two types of discharges are present at different points in the same response area of a single cell. The sustained responses show a broad range of latencies, while onset patterns show a tighter distribution and shorter first spike latencies. Thus, IC neurons showing sensitivity to changes in interaural phase can differ in laterality preferences, response area characteristics, discharge patterns, and latency parameters. Given the diversity of inputs to the IC from lower brain stem structures, this heterogeneity is not surprising. For most neurons excited by stimulation to either ear, the characteristic frequencies, discharge patterns, and first spike latencies are similar, suggesting that the monaural inputs to a binaural cell are of the same type. A neuron's most effective frequencies at a particular SPL for monaural and binaural stimulation are, in general, the same. In some cases a neuron's monaural and binaural response areas can show remarkable similarities, suggesting that certain monaural features are intimately related to the binaural response. In 18% of the IC cells, phase locking to the monaural stimulating frequency is seen. When both inputs are phase locked, a simple coincidence model can predict the interaural phase or delay at which the maximal binaural discharge occurs.

Wiener kernel analysis of responses from anteroventral cochlear nucleus neurons.

Responses to pseudo-random Gaussian white noise, tones and clics were recorded from neurons in the anteroventral cochlear nucleus (AVCN) of barbiturate anesthetized cats. The responses to white noise were used to calculate estimates of the zero-, first- and second-order Wiener kernels for these neurons. The Wiener kernels did contain useful information on the fundamental, DC and second harmonic components of the responses of AVCN neurons to tones, clicks and noise. However, they generally did not provide predictions of the difference tone distortion products found in the peripheral auditory system. Overall, the addition of the second kernel improved a prediction based on the zero- and first-order kernels, but not by very much. If the estimates of the Wiener kernels were not very good, then a second-order prediction could be worse than a first-order one. To produce good estimates of the Wiener kernels, many repetitions of very long Gaussian white noise stimuli are necessary. Therefore the technique does not permit rapid data collection. Further, exposure to long duration high intensity noise can result in acoustic trauma. This damage effects the mechanism that generates the difference tone distortion products, and it can also affect the tuning of the auditory neurons. Thus Wiener's nonlinear system identification theory has only limited usefulness in the analysis of the peripheral auditory system.

Time- and intensity-dependent low-pass filtering of auditory brain stem responses.

A new method of filtering auditory brain stem responses (ABR), which is time-and intensity-dependent, is proposed. The epoch containing the averaged response is divided up into 128 overlapping segments and each segment is differently filtered. The characteristics of a filter that is applied to a segment of the response are derived from the time-dependent spectral composition of normal ABRs at the appropriate stimulus intensity. After filtering, the segments are added together to yield a response with no significant amplitude and/or phase distortion, and with distinctly pronounced peaks that are excellently suitable for evaluation with an automated system.

Response of cat inferior colliculus neurons to binaural beat stimuli: possible mechanisms for sound localization.

The interaural phase sensitivity of neurons was studied through the use of binaural beat stimuli. The response of most cells was phase-locked to the beat frequency, which provides a possible neural correlate to the human sensation of binaural beats. In addition, this stimulus allowed the direction and rate of interaural phase change to be varied. Some neurons in our sample responded selectively to manipulations of these two variables, which suggests a sensitivity to direction or speed of movement.