Functional modelling of interaural time difference discrimination in acoustical and electrical hearing.

Abstract

Interaural time differences (ITDs) are important for sound source localisation. We present a model to predict the just noticeable differences (JNDs) in ITD discrimination for normal hearing and electric stimulation through a cochlear implant. We combined periphery models of acoustic and electric stimulation with a novel JND in the ITD estimation stage, which consists of a shuffled cross correlogram and a binary classifier characterisation method. Furthermore, an evaluation framework is presented based on a large behavioural dataset. The model correctly predicts behavioural observations for unmodulated stimuli (such as pure tones and electric pulse trains) and modulated stimuli for modulation frequencies below 30 Hz. For higher modulation frequencies, the model predicts the observed behavioural trends, but tends to estimate higher ITD sensitivity. The presented model can be used to investigate the implications of modifying the stimulus waveform on ITD sensitivity, and as such be applied in investigating sound encoding strategies.