| Under natural conditions, we are able to track with our eyes an object moving through the environment and can accurately perceive its direction of motion. This research addresses the issue: how accurate are eye movements and the perception of target motion when the trajectory of the moving target contains an illusory component?;Using a variant of the Duncker Illusion in which a vertically moving target appeared to move diagonally, we studied saccadic and smooth tracking in both head-fixed and head-free conditions. We found that while smooth pursuit gaze tracking of a sinusoidal target was predominantly vertical, predictive saccades to a repetitively stepping target were oblique, with horizontal components in the direction of the illusion. Furthermore, head tracking responses for both sinusoidal and stepping targets were always in the direction of the illusion. Unlike the erroneous oblique saccades that were followed by a corrective horizontal saccade, no such correction occurred for the head movements. Thus, perceived rather than real retinal information is used in the programming of predictive saccades and head movements.;We next studied saccades to the remembered locations of targets that were briefly flashed during tracking of an illusory target trajectory. These saccades were grossly inaccurate in the horizontal plane, the plane in which the illusory motion occurred. The greatest errors occurred in the initial saccade, and were followed by corrective saccades even in the absence of additional visual feedback. A secondary finding was that the brain compensated for eye movement during the memory period, i.e., used "spatial" not "retinal error," producing accurate vertical components.;By controlling for both eye position at the time of the target flash and total illusory motion during the memory period, we demonstrated that errors are introduced at both the time of the target flash and the time of saccade generation. Furthermore, it is perceived eye movement and not efference copy that is used to compensate for eye movements during the memory period.;In a final experiment we artificially increased the time of visual processing by delaying by several hundred milliseconds the visual consequences of any eye movements. This procedure, which caused a discrepancy between the expected and actual visual input, induced ocular oscillations that were correlated with the length of the added delay. |
