The role of motion parallax in the perception of egocentric direction

Abstract

Prisms optically rotate the visual scene relative to the head, but the error in perceived direction that results is less than the optical deflection of the prism. This is known as the immediate correction effect (Rock, Goldberg, & Mack, 1966). The effect could result from cues provided by knowledge of the surrounding environment, and indeed, a recent study found that the immediate correction effect is reduced in the dark (Pochopien & Fahle, 2015). Here we tested an alternative hypothesis: the immediate correction effect is driven by motion parallax resulting from movements of the head, and motion parallax (the speed of head-centric motion that results from a head rotation) is a function of the direction of the object relative to the head. Accordingly, the light/dark effect would result would from the more precise relative motion signals available in the light. Our hypothesis would also predict two other patterns of behaviour: the immediate correction effect should reduce with the distance of the target (motion parallax reduces with distance), and the immediate correction effect should be reduced if the target is only transiently visible (removal of motion signal). We presented participants with illuminated target objects while they wore horizontally shifting prism glasses. Participants were instructed to orient their heads and bodies to face the target objects, and we recorded positional information via a motion-tracking system. We compared orientation errors for targets placed at varying distances when the room was lit, dark, and when the targets were flashing in the dark. As predicted, the immediate correction varied as a function of target distance. Second, the immediate correction effect was larger in the light than in the dark. Third, the dependence of the magnitude of the immediate correction effect on distance was abolished when the target was only transiently visible.