Temporal coding of amplitude modulated stimuli: Firing rate modulation and entrainment in cochlear nucleus cells and simple neuron models

The cochlear nucleus (CN) is a primary station in the auditory brainstem where neural response of auditory nerve fibers to acoustic stimuli is processed, encoded and transmitted to the central auditory system (CAS). As reviewed in Langner (1992), the firing rate of a large number of cells in the dorsal cochlear nucleus (DCN) modulates in response to sinusoidally amplitude modulated (SAM) acoustic stimuli. Rhode and Smith (1986b) showed that cells in the DCN also exhibit entrainment. Entrainment is described as the occurrence of at least one spike in every stimulus modulation cycle, in response to a SAM acoustic stimulus.; Firing rate estimates have been developed in this study using spike trains recorded during multiple presentations of a short-duration stimulus. These estimates depend upon the inter-spike interval (ISI), and the distance of nearest neighbor (DNN) at any post-stimulus time. In response to SAM stimuli, the modulation strength of these estimates is directly related to the strength of entrainment. Therefore, the maximum modulation frequency for which the firing rate estimates exhibit strong modulation is treated as a measure of temporal coding through entrainment.; Simple refractory computational models have been developed previously (Ghoshal et al. 1992, Kim et al. 1994, Hewitt et al. 1995) which exhibit entrainment and rate modulation similar to experimental data. In the present study, entrainment and rate modulation of experimental data are analyzed to test these computational models. For simple refractory neuron models, both rate modulation properties and entrainment are an effect of refractoriness. As a result, strong correlation exists between rate modulation measures and entrainment measures, for models with refractoriness. However, such a correlation does not exist for more complex models.; Analysis of experimental data shows that the data can be explained better using models that include integrate-and-fire mechanisms, inhibition, and complex neural circuitry, than with simple refractory neuron models. In addition to the SAM response, the intrinsic oscillation properties of DCN single units are analyzed in order to identify possible neural mechanisms present in the DCN. Such analysis assists in determining the possible contribution, of different neural mechanisms in the DCN, to temporal coding of acoustic stimuli.