Item

Abstract

A quantitative, neuronal model is proposed for the computation of interaural time difference (ITD) in the auditory system of the barn owl. The model uses a general, probabilistic approach, and is composed of two stages, the characteristics of which are based on anatomical and physiological evidence. Excitatory inputs from both ears, phase-locked to the waveform of tonal stimuli, together with phase-independent inhibitory inputs are summated linearly. The result is transformed into a probability of spike generation by a sigmoid nonlinearity, constituting a stochastic, ’soft’ threshold with saturation. The model incorporates inhibition as a control parameter on the nonlinearity, and includes the usual crosscorrelation-type models as a special case. It has a minimum number of parameters, the values of which can be estimated from physiological data in a straightforward manner. This simple, general model accounts for the binaural response properties of physiologically recorded neurons. In particular, it explains the experimentally observed ITD-tuning and the increase of phase-locking from input to output neurons. The model predicts that a decrease in inhibition causes a non-monotonic change in sensitivity to ITD.