Abstract
How does the visual system keep track of a moving object that temporarily disappears and re-emerges again? We investigated the nature of the tracking mechanism with two tasks. In the first, observers maintained central fixation and covertly tracked a target that translated back and forth on a constant linear trajectory at three velocities, fast (3.9 deg/s), medium (2.6 deg/s) or slow (1.7 deg/s). “Unoccluded” trials consisted of a fully visible trajectory and “occluded” trials were identical except for an invisible occluder that concealed the center of target's trajectory. Observers judged the target's re-emergence from behind the occluder with a button press. These judgments were highly accurate for fast and medium trajectories in both occluded and unoccluded trials, but there was increased variability in these anticipatory judgments of re-emergence during occluded slow trajectory trials, suggesting an unstable representation of the target in space across longer delays.
The second task was identical to the first except observers were not required to press a button, but only to watch the stimulus as we recorded cortical activity with fMRI. Extrastriate regions showed increased activation during unoccluded relative to occluded trials presumably related to the representation of the visible target. Additionally, we identified cortical regions that showed increased activity during periods of temporary dynamic occlusion. The right angular gyrus and posterior portions of superior temporal sulcus showed increased activation during occluded relative to unoccluded trials. Activation during occlusion may be related to maintaining an internal representation of the spatiotemporal properties of the invisible target in the visual system during attentive tracking. These findings may shed light on the cortical mechanisms involved in the phenomenon of perceptual constancy.
NSF grant BCS-0418103 and NIH grant R01-HD40432