Testing computational theories of motion discontinuities: A psychophysical study

Abstract

This study reports results from three patients with bilateral brain lesions (A.F., CD., and O.S.) and normal observers on psychophysical tasks, which examined the contribution of motion mechanisms to the extraction of image discontinuities. The data do not support the suggestion that the visual system extracts motion discontinuities by comparing fully encoded velocity signals ([NL]; [Clo]). Moreover, the data do not support the suggestion that the computations underlying discontinuity localization must occur simultaneously with the spatial integration of motion signals ([Kea]). We propose a computational scheme that can account for the data.