Abstract
Multiple object tracking (MOT) is a task wherein observers attempt to track multiple targets among moving distractors. Contour interpolation is an early process that represents non-visible edges on the basis of how surrounding edges (inducers) are spatiotemporally related. Although MOT is usually employed to study attention, here we show that one version of the paradigm—what we call multiple vertex tracking—also reveals interpolation effects. In a typical trial, a target disk and distractor disk orbited within each screen quadrant. During a portion of the movement phase, all disks transformed into pac-men that could either interpolate with one another or not. In Experiment 1, when targets and distractors formed illusory contours with each other, tracking performance was worse than in a control condition having the same edge relations but no interpolation. Experiment 2 showed qualitatively the same outcomes when interpolated contours were occluded. In Experiments 3 and 4, when interpolation strength was decreased by either decreasing support ratio (disk diameter) or increasing inducer rotation angle, interpolation effects on tracking decreased. Across all experiments, performance was best when all four targets interpolated with one another, although the strongest interpolation effects occurred between targets and distractors. These data strongly suggest that contour interpolation automatically directs attention. They further suggest that—contrary to recent theories of tracking—the specific shape of tracked objects can be relevant to MOT, at least when those shapes contribute to unit formation.
This work was supported by a UCLA Graduate Research Fellowship awarded to BPK.