Neural field dynamics for motion perception

Abstract

In this paper we present a model for the perception of apparent motion based on dynamic neural fields. We show that both, the organization of percepts and their specific dynamical properties can be derived from the same perceptual dynamics. We show that neural fields adequately formalize this perceptual dynamics. This leads to a model that is sufficient in the sense that it reproduces correctly the perceptual organization of motion. We show that the model allows also to identify elements which are necessary for understanding perceptual organization together with the dynamic properties of the percepts (e.g. hysteresis and its link to switching rates). Such elements are an activation dynamics which is coupled to adaptation processes with different time-scales, and different mechanisms for perceptual switching. We show how the model can be linked quantitatively to a large set of psychophysical data for the motion quartet.