Motion perception: The effects of perceived three-dimensional distance.

Abstract

Contemporary computational models of motion perception assume that in processing continuous or near-continuous motion information, the visual system measures spatial displacement in retinal coordinates over a series of time-varying images. Additional three-dimensional information possessed by the system purportedly does not influence this low-level motion analysis. The present research investigated the influence of static three-dimensional distance information recovered from binocular disparity on the perceived direction of motion. It was assumed that if a stereoscopic display context influenced perceived motion direction, the apparent velocity of a moving element would increase in order to traverse the greater apparent distance. This would be reflected in a predictable pattern of errors when the true angular velocity was the same, slower or faster than that of the standard. The stimuli consisted of random-dot stereograms depicting surfaces at varying distances and orientations. In one stereoscopic display, the disparity information indicated a surface sloping smoothly in depth from crossed to uncrossed disparity. The second display contained two fronto-parallel planes at discrete distances from the observed. Motion stimuli were single element translating horizontally and presented monocularly to the observer's right eye. Experiment 1 compared differential velocity judgments in the contexts of the sloped surface and a control condition at zero disparity. The results indicated an overall increase in the perceived velocity of the element moving in the context of the sloped surface. The pattern of results was replicated in experiment 2, but an additional effect of the relative positions of the two surfaces also was obtained. Experiment 3 explored the case of two discrete fronto-parallel planes, one at crossed disparity and the other at uncrossed disparity. This experiment also produced a position effect, but indicated that the perceived distance of the two planes did not differentially affect observer's velocity judgments. It was concluded that in some cases, the metric of motion analysis could be affected by three-dimensional information recovered from binocular disparity. The particular case discovered in these experiments was a surface that appeared to slope smoothly in depth. Discrete depth planes produced no such effect.