Motion Transparency in Superimposed Dense Random-Dot Patterns: Psychophysics and Simulation

Abstract

To determine the mechanism underlying motion transparency, representative examples of motion transparency are listed and possible mechanisms are suggested. Those are feature tracking, multiple spatial-frequency channels, luminance-based transparency rules, and motion energy. Next, an interesting stimulus for motion transparency is introduced, namely superimposed dense random-dot patterns, which is not explained by feature tracking or multiple spatial-frequency channels. A psychophysical experiment reveals that the occurrence of motion transparency in this stimulus depends on three luminance levels assigned to three possible combinations of component dots: (1) white dots superimposed upon white dots, (2) white dots upon black dots, and (3) black dots upon black dots. However, physical rules of luminance-based transparency fail to explain the results. Finally, a computer simulation reveals that a computational model based on motion energy quantitatively predicts the human psychophysical performance. All the results support the idea that motion-energy detection followed by spatial integration is a likely candidate for the mechanism underlying motion transparency.