Abstract
Moving items cause the position of flashed items to appear shifted in the direction of motion (Whitney & Cavanagh, 2000). Is this motion-induced position capture due to retinal or perceived motion? In order to examine this issue, we employed a slit-view motion display where retinal motion is minimized but motion perception remains vivid. [1] Two white diamonds (3 deg) translated horizontally in opposite directions on a gray background at 7.2 deg/s, one above and one below the fixation. When the diamonds were in vertical alignment, two vertical lines (1 deg) were flashed for 13 ms, one inside each diamond (full-view condition). Observers (N=3) saw the lines shifted in the direction of the moving diamonds; when the top line appeared shifted to the left, the bottom line appeared shifted to the right, and vice versa (p < .05). [2] In the slit-view condition, observers saw exactly the same display except the diamonds were visible only through a narrow (5.8 min) vertical slit (within which the lines were flashed), so that only small elements of the diamonds were visible at any one time. Even though horizontal retinal motion was greatly reduced in this condition, observers perceived the diamonds moving behind the occluders and the motion-induced position capture occurred undiminished (p < .05). [3] Finally, we allowed observers to see the diamonds through only a 1-pixel slit (1.4 min), eliminating any horizontal motion component. Surprisingly, the moving diamonds were still perceived. Although the motion direction was ambiguous, observers tended to see the diamonds moving in the same direction. Despite that this object motion perception was all in the observer's mind, the motion-induced position capture was still significant (p < .05; both lines appeared shifted in the same direction), though somewhat reduced. Thus the motion-induced position capture must be attributed to high-level motion processing responsible for dynamically integrating object motion and shape.