The representations of the steady-state vowel sound /ε/ in the discharge patterns of cat anteroventral cochlear nucleus neurons

1. We have recorded the responses of neurons in the anteroventral cochlear nucleus (AVCN) of barbiturate-anesthetized cats to the synthetic, steady-state vowel sound /ε/, presented over a range of stimulus intensities. 2. The responses of (putative) spherical bushy cells [primary like (Pri) units] to the vowel resemble those of auditory-nerve fibers (ANFs) in terms of both rate and temporal encoding at low and moderate stimulus levels. It was not possible to study the responses of most Pri units at the highest stimulus level because of the large neurophonic component present in recordings from most primarylike units at higher stimulus levels. 3. The responses of many (putative) globular bushy cells [primarylike with notch (PN) units] to the vowel resemble those of ANFs; however, there appears to be greater heterogeneity in the responses of units in the PN population than in the Pri population in terms of both temporal and rate encoding. 4. Populations of stellate cells (chopper units) have degraded representations of the temporal information in ANF population discharge patterns in response to the vowel; this is consistent with the responses of these units to pure tones. Both regular (ChS) and irregular (ChT) chopper subpopulations, however, maintain better rate-place representations of the vowel spectrum than does the population of ANFs as a whole. The rate-place representations of the vowel spectrum by both chopper populations closely resemble those of low and medium spontaneous rate ANFs at most stimulus levels. 5. The data presented in this paper suggest a functional partition of the AVCN chopper population could yield two distinct rate representations in response to a complex stimulus: one that is graded with stimulus level (over a 30 to 40 dB range) and that, even at rate saturation, maintains a 'low contrast' stimulus representation; and a second that maintains a robust, 'high contrast' stimulus representation at all levels but that confers less information about stimulus level.