Abstract

Reviewing experiments of velocity sensitivity and direction discrimination, Nakayama (Vision Research, 1985) hypothesized the existence of a second stage of motion processing where velocity signals are time-averaged. This implies that sensitivity to rapidly changing velocity should be worse than that predicted from the bandpass nature of earlier processing stages. We used three different experimental procedures to measure an observer’s temporal sensitivity to acceleration: discriminating a sinusoidally changing velocity stimulus from a constant velocity stimulus; measuring an observer's ability to discriminate the faster of two contiguous velocity steps; and measuring the ability of a target which was preceded and followed by a reference velocity. No evidence was found for time-averaging which could not be accounted for by the temporal characteristics of early filters. We conclude that at least two branching output pathways exist for motion having different temporal properties.

© 1986 Optical Society of America