Abstract
Smooth pursuit eye-movements generate large magnitude spurious motion signals that generate a motion after-effect (MAE). White bars of various orientations away from vertical (0,15,30,45,60,75,85,90 degrees) remained present on the screen on a black background, one orientation visible per trial. The fixation point swept from left to right smoothly for a variable duration and at a variable speed. Subjects smoothly pursued this moving point, after which the fixation spot stopped in the center. Subjects indicated when the MAE stopped. The duration of the MAE increased with increasing duration and speed of smooth pursuit. The strength of the MAE was effectively zero for vertical bars, even though this stimulus had the greatest image motion magnitude in the horizontal direction (the direction of smooth pursuit). This was presumably because there was no component of motion in the vertical direction for vertical bars. It was also zero for horizontally aligned bars, which had no motion component in the vertical direction. The MAE was strongest for bars oriented slightly away from horizontal at about 80 degrees from vertical. We hypothesize that the motion processing system is able to discount component motion with less success for bars oriented at angles very close to the direction of smooth pursuit. We suggest that filters tuned to low spatial frequency components of motion determine a global direction of motion. To inhibit spurious motion signals due to the aperture problem, motion signals derived from contours oriented at an angle close to this global direction of motion receive inhibitory feedback such that perpendicular component motions from these contours are suppressed. However, the limits of angular acuity are reached at small angles (i.e. the angle between a contour and the global direction of motion across the retina), and inhibition of perpendicular component motion of these contours is correspondingly weaker.