Why two eyes are better than one for judgements of heading

Abstract

ARE two eyes needed for judging direction of self-motion? Traditional analyses stress that the pattern of optic flow in one eye is sufficient^1–5. The main difficulty is how to deal with the eye or head rotation. Extraretinal signals help^6–8, but humans can also discount the effect of rotation purely on the basis of monocular flow^6,7,9–12 provided the scene contains depth^6,9,10. Depth differences give rise to changing binocular disparities when the observer moves. These disparities are ignored in monocular theories of judgements of heading. Using computer generated displays, we investigated whether stereoscopic presentation improves heading judgements for conditions that pose problems to the monocular observer. We found that adding disparities to simulated ego-motion through a cloud of dots made heading judgements up to four times more tolerant to motion noise. The same improvement was found when the disparities specify the initial distances throughout the motion sequence. We conclude that binocular disparities improve judgements of heading by imposing a depth order on the elements of the scene, not because they provide additional information on the elements' motion in depth.