Assimilation-type and contrast-type bias of motion induced by the surround in a random-dot display: evidence for center-surround antagonism.

As a mechanism to detect differential motion, we have proposed a model of "a motion contrast detector" that has a center-surround antagonistic receptive field with respect to the direction of motion. Supporting evidence has been obtained in the studies of induced motion, motion capture, and motion aftereffect. In order to obtain further evidence in a more strictly controlled situation, we examined the perceptual bias of motion in a center stimulus induced by another, surrounding motion. By using a stochastic random-dot display configured in a center-surround concentric fashion, we measured the % signal in the center stimulus that made the stimulus perceptually stationary in the presence of a moving surround. Measurements were done for various stimulus sizes and eccentricities. The amount of bias changed as a function of stimulus size and eccentricity. At several eccentricities, smaller stimulus sizes tended to yield assimilation-type biases, whereas larger sizes tended to yield contrast-type biases. However, a spatial scaling procedure revealed that the amount of bias was a simpler function of "scaled" stimulus size that was obtained by dividing the physical size by a scaling factor at each eccentricity. In the scaled profile, assimilation-type bias changed to contrast-type bias with increasing size, reached the peak of contrast-type bias at a certain size, and decreased slightly with further increasing size. Furthermore, a model of a difference of Gaussians, DOG, function well approximated the behavior of the profile. From these results, we concluded that the process specific to perceiving relative motion is mediated by a motion contrast detector, which is possibly located in area MT.