Abstract
Purpose: Global form is easily perceived in visual patterns where structure is defined purely by spatial or motion cues, but how do spatial and motion cues interact in the computation of global form when presented concurrently? While the influence of spatial cues on motion-cue integration is well documented, the influence of motion cues on spatial-cue integration remains largely unexplored.
Method: In a 2AFC task, observers discriminated between spatial tests embedded in motion masks. Stimuli consisted of 36 colinear Gabor elements positioned along a virtual radial-frequency (RF) contour — a circle deformed by a sinusoidal modulation of radius as a function of polar angle. Sinusoidal amplitudes for discriminating between circular and non-circular patterns (75%-correct) defined thresholds for spatial-cue integration. Motion masks were applied by drifting each Gabor's carrier at a speed determined by a separate variable-amplitude RF contour that coded for speed. Tests and masks were constructive or destructive depending on their relative phases.
Results: Thresholds for static patterns were low and consistent with global spatial-cue integration. Constructive masks produced little facilitation at low mask amplitudes but interfered moderately with discrimination at higher mask amplitudes. Destructive masks interfered considerably with discrimination, and interference increased with mask amplitude. Psychometric functions revealed that motion-defined masks are perceptually confused with spatially-defined tests, and control masks defined by expansion, contraction, flicker, or random speeds, had little effect on spatial-cue integration.
Conclusions: Motion exerts a profound influence on the assembly of spatial cues into global form, but interactions are selective for motion and spatial cues with compatible global properties. Results suggest a model that computes global form separately within cue-specific pathways but allows for competitive cross-pathway interactions.