Perceived speed of compound stimuli is moderated by component contrast, not overal pattern contrast

Abstract

The perception of speed is susceptible to manipulations of image contrast, both for simple sine wave and more complex stimuli, such that low-contrast patterns generally appear slower than their high-contrast equivalents. It is not known whether the crucial factor is the contrast of the underlying Fourier components or the contrast of the overall complex pattern. Here, two experiments investigate this issue using compound gratings, comprising two vertical sine wave stimuli with equal contrast, but a 3:1 spatial frequency ratio. Component gratings were summed in "peaks add" and in "peaks subtract" phase, creating conditions with either (a) identical component contrasts, despite differences in overall pattern contrast or (b) differences in component contrasts despite identical overall pattern contrast. Experiment 1 demonstrated that the perceived speed is determined by the contrast of the components regardless of relative phase and hence of overall pattern contrast. Experiment 2 replicated this result while eliminating potential explanations based on differences in spatial frequency content. Along with previous compound grating and plaid studies, the data support a two-stage velocity estimation process involving the derivation of separate speed signals for each Fourier component, followed by integration of these signals across spatial scales.