Phenomenological model for the adapatation of shape-selective neurons in area IT

Abstract

Shape-selective neurons in inferotemporal cortex show adaptation if the same shape stimulus is shown repeatedly. Recent electrophysiological experiments have provided critical data that constrain possible underlying neural mechanisms. We propose a neural model that accounts in a unifying manner for a number of these critical observations. The reproduction of the experimental phenomenology seems to require a combination of input fatigue and firing rate fatigue mechanisms, and the adaptive processes need to be largely independent of the duration of the adapting stimulus. The proposed model realizes these constraints by combining a set of physiologically-inspired mechanisms.