Contrast dependence of short-range apparent motion.

The apparent motion of band-pass filtered random dot kinematograms was assessed by measurements of two alternative direction discrimination performance. In the case of two-dimensional (isotropically filtered) stimuli, the results were largely independent of contrast: at Michelson contrasts of 5 and 50%, near-perfect direction discrimination was obtainable for a limited range of displacements. However for one-dimensional (grating) stimuli, apparent motion seems to be highly dependent on contrast. At 5% contrast, performance was comparable to that obtained with the two-dimensional stimuli. At 50% contrast the motion percept broke down to a large extent, with consistently poor direction discrimination being obtained. The breakdown of apparent motion is interpreted in terms of a decreased signal-to-noise ratio in the pooled response of motion detectors that are tending to contrast saturation. Here, "noise" refers to the sampling components present in any apparent motion sequence. Evidence relating to the sampling frequency of the motion sequence is presented to support the hypothesis. It is argued that the discrepancy between the results obtained with one- and two-dimensional stimuli support the idea that motion is initially encoded by orientationally tuned mechanisms.