The contribution of coherent and incoherent motion to the detection of contours in a macaque monkey.

Shih-Cheng Yen and Charles M. Gray.

Center for Neuroscience, UC Davis, Davis, CA 95616.

Abstract

The goal of this study was to investigate the interaction between common motion and contour smoothness in a contour detection task. Each stimulus was made up of an array of 200 randomly oriented Gabor patches, with a target defined by a group of 8 Gabor patches oriented to form a closed contour. The salience of the contour was varied by increasing the jitter in orientation of the Gabor patches on the contour from 0° to ±75°. A macaque monkey, with implanted eye coils, was trained to perform a detection task by making a saccade to a peripheral contour 9.5° from a central fixation spot. Three sets of experiments were conducted and the threshold orientation jitter, defined at the 75% performance level, was measured. For the first set of experiments, the Gabor patches were presented statically, yielding a threshold jitter of ±35°. In the next two sets of experiments, each Gabor patch in the array drifted within a stationary window in one of two directions orthogonal to its orientation. The temporal frequency of the motion and the phase of the motion were kept constant across the whole array. When the drift direction of the Gabor patches on the contour was set to the same direction so that the contour appeared to move coherently, the threshold jitter increased to ±45° (p<0.001). When the directions of motion of the Gabor patches on the contour were randomized, the incoherent motion of the contour led to a decrease in the threshold jitter to ±28° (p<0.001). Our results suggest that motion coherence and continuity are tightly coupled Gestalt cues of contour salience.

Supported by: McDonnell-Pew Program in Cognitive Neuroscience and the NEI.