Amplitude-modulated tone encoding behavior of cochlear nucleus neurons: modeling study

A recent study of amplitude-modulated (AM) tone encoding behavior of dorsal and posteroventral cochlear nucleus (DCN and PVCN) neurons by Kim et al. [Hear. Res. 45, 95-113, 1990] observed that certain neurons (e.g., pause/build type-III neurons and chop-S neurons) tended to exhibit band-pass modulation transfer functions (MTFs) and intrinsic oscillations (IOs) whereas certain other neurons (e.g., chop-T neurons) tended to exhibit low-pass MTFs and no IOs. The goal of the present study was to develop models of these response characteristics in an attempt to understand the underlying neuronal mechanisms. We hypothesized that chopper neurons corresponded to stellate cells and pause/build neurons corresponded to fusiform cells. We also hypothesized that, with right input combination, appropriate models of a single stellate and fusiform cell could account for band-pass and low-pass MTFs as well as the associated IOs. The neuron models developed by Arle and Kim [Biol. Cybern. 64, 273-283, 1991] for the stellate and fusiform cells were used in this study. The models are modified versions of MacGregor type neuron model incorporating cell-specific nonlinear voltage-dependent conductances. The AM tone excitation via the auditory nerve fibers was represented by a current at the soma of the neuron model, which consisted of dc, ac and a zero-mean Gaussian noise. The dc, ac and noise represent a high-frequency carrier beyond the neuron's phase-locking limit, an envelope, and randomness of the system, respectively. With systematic variation of dc, ac and noise amplitudes, we observed the following: the band-pass MTF behaviors of pause/build and chop-S neurons were reproduced by the fusiform cell model and the stellate cell model with a strong dc/noise ratio, respectively. The low-pass MTF behavior of a chop-T neuron was reproduced by the stellate cell model with a weak dc/noise ratio. It was observed that the stellate cell model was more susceptible to the noise, in the sense that an increase in noise tended to abolish the IO and change the MTF of the model from band-pass to low-pass more readily in the stellate cell model than in the fusiform cell model. Kim et al. (1990) observed a close correlation between the IO frequency and the best envelope frequency (BEF). In the models, a similar correlation was observed between the two measures for both the stellate and fusiform cell models.(ABSTRACT TRUNCATED AT 400 WORDS)