Abstract:

A new model for firings of coincidence neurons in the MSO is presented. The model is similar to one by Colburn et al. [Hearing Res. 49, 335--346 (1990)]: Both are point processes, both have inputs which are impulse-train representations of the firing events of primarylike neurons, both operate on a low-pass filtered sum of those inputs, and both output a sequence of impulses at the event times of the coincidence neuron's firings. In this model, their threshold-crossing neuron model is replaced by a recovery model similar to that modeling the input firings. The coincidence neuron's drive function is a nonlinearly distorted version of the filtered sum of inputs; for simplicity, its recovery function is the same as that of the input neurons. The model's firing rates, vector strengths, period histograms, and interval histograms are compared to data from dog and cat and with similar responses of the Colburn et al. model.