| There are at least two sources of information about direction that observers might use to control an approach or a reach to a target. The visual direction and optic flow both provide information about target direction, and normally are redundant sources of information. In the reported studies, visual direction and optic flow were dissociated by having seated observers view targets through a head-mounted camera that was rotated without the observers' knowledge. This manipulation allowed us to determine whether observers relied upon the visual direction or the optic flow in guiding a head movement and a reach to a target that could appear randomly at any of five directions. Observers executed three head movements toward the target, and then executed three blind reaches. They were instructed to align the reaches to the previously executed head movements. Reaches were aligned with the head movements in all conditions. When head rotations were unconstrained, observers exhibited large heading errors, as predicted by a visual direction strategy. This finding is consistent with the fact that adding a rotational component to the optic flow pattern complicates matters for using optic flow as information about heading direction. When head rotations were prevented, heading errors dropped significantly, consistent with the use of an optic flow based strategy for approaching the target.;Motion parallax can be useful for recovering optic flow based heading information when a rotational component is added to the flow pattern. In the final study, we added depth structure to the display, and heading errors were small, indicating that accurate heading information was recovered when information from motion parallax was available. Under appropriate conditions, optic flow is used to guide approach to a target, and, in conjunction with somatosensory information about the direction of self-movement, can be used to control a blind reach. |
