Abstract
Background: The visual direction of an intended heading point has been shown to play an important role in guiding locomotion toward close targets (Rushton et al Curr. Biol. 8,1191, 1998, Rogers and Dalton, Invest. Ophthal. Vis. Sci. 40, s764, 1999). With distant targets, however, the error signals from eye and head position (whenever the observer strays off a direct heading path) are extremely small, making it more likely that optic flow influences locomotion. Aims: (i) to investigate whether observers' walking paths show an increased variability of direction when optic flow from the ground plane is occluded compared to unoccluded viewing; (ii) to see whether the veridical optic flow information about heading direction might be used to recalibrate eye and head position (and hence visual direction) when prisms are used to alter visual direction. Methods: Observers walked towards a distant visual target (150 m) with small angle displacing prisms (5–7 deg) either present or absent. The prisms were mounted close to the observer's eyes to increase the field of view to at least 60 by 60 deg. The presence or absence of occluders (masking the lower half of the visual field below the distant target) controlled to availability of ground plane optic flow. Observers' walking paths were continuously monitored by video. Results: (i) walking paths were typically straight, both with and without occluders, but the average unsigned deviation of the path directions was significantly greater when ground plane optic flow was unavailable; (ii) initial walking paths with prisms over the eyes were significantly deviated (70–90% of the prism angle) for both occluded and unoccluded viewing of the ground plane; (iii) the deviation of subsequent walking paths from the correct heading direction became significantly smaller over time but only when ground plane optic flow was available during those subsequent trails. These results provide convincing evidence for the role of optic flow in recalibrating visual direction.