What are the Mechanisms of Rivalrous First-Order and Second-Order Motions?

Abstract

There have been many reports in the visual motion literature describing how patterns of contrast reversal in bar stimuli may yield rivalrous motions in opposing directions. One of these, usually termed first-order motion, is generally explained in terms of ‘short range’ matching of same polarity edges or of standard motion analysis of the distribution of energy in the Fourier transform of the stimulus. The opposite, second-order, motion is usually explained in terms either of ‘long range’ matching of whole forms/features or of luminance signal rectification followed by standard motion analysis. It is known that the mechanisms underlying the two motions show differential sensitivity to manipulations of spatial frequency and eccentricity (Mather, Cavanagh, and Anstis, 1985 Perception 14 163 – 166; Chubb and Sperling, 1989 Proceedings of the National Academy of Sciences of the USA 86 2985 – 2989), and this paper reports a number of studies intended to further specify them in terms of their sensitivities and the interactions between them. The experiments employ mainly stimulus sequences in which luminance changes are unaccompanied by displacement of forms/features, and are designed to investigate the effects of manipulations involving the spacing of luminance events, modulations of directional power, camouflage, and attention. The results obtained raise questions about how the underlying mechanisms are best to be characterised. We are greatly indebted to Professor Ted Evans for technical guidance and assistance.