Temporally offset motion-induced spatial misalignment

Abstract

We investigated the effect of visual motion on apparent spatial location [1]. A rotating (40 rpm) black bar (2.7°) on a white background was flanked by two smaller horizontal bars (10.8 min arc) that flashed in synchrony with the bar. We measured the apparent vertical misalignment of the flanking bars as a function of the angular position of the rotating bar using a binary choice psychophysical procedure. The vertical offset of the flashed bars was varied systematically over trials and subjects reported which was the higher of the two bars. Surprisingly, the greatest effect (perceived misalignment ∼20 min arc) was found when the bar was around 30° of rotation before the horizontal rather than when the bar was closest to the flashed bars. The effect was replicated in a subsequent experiment using a number of naive subjects. We found a significant misalignment 30° before the horizontal and no misalignment 30° before the vertical. This induced effect demonstrates that spatial position information is maximally affected by motion information present 125 msec before the rotating bar reaches the flash position. This critical angle before the horizontal is consistent with previous flash-lag measurements showing a similar sized lag between the rotating bar and flash[2]. Thus the flanker misalignment is maximal when bar and flash are perceived to be aligned rather than when they are truly aligned. We believe that this spatial misalignment may result from a cortical feedback mechanism [1,3] that updates position information during motion.