A Functional Point-Neuron Model Simulating Cochlear Nucleus Ideal Onset Responses

Abstract

Cochlear nucleus neurons revealing ideal onset (O_I)-type peri-stimulus time histograms (PSTH) encode temporal features of acoustic stimuli with very high precision. These neurons are therefore assumed to be involved in the recognition of natural sounds with temporally varying envelopes such as speech. A functional point-neuron model is presented here for the simulation of O_I-unit responses found in cochlear nucleus octopus cells. The model assumes a biphasic response of the membrane potential to a current impulse, the membrane impulse response, and a dynamic spike-blocking mechanism. The predicted responses to pure tones at low and high frequencies, injected current steps, and amplitude modulated tones are compared to recordings from the literature. The model accounts for the main response properties in the data using the same small set of parameters for all experimental conditions. The assumed biphasic shape of the membrane impulse response, reflecting a higher sensitivity to stimulus transients and fast changes relative to sustained stimulus portions, allows for a description of O_I-unit responses that cannot be accounted for by a coincidence detector model with an integrate-to-threshold dynamic. The presented functional model may be useful as a processing module in more complex models of auditory signal processing and perception.