Abstract
Visual attention appears in many cases to operate on objects. Recent research suggests that the visual system segments objects using an inflexible set of rules, such as common features, minima of curvature, and connectivity. In the present study, we explore another critical cue to objecthood — common fate motion. When an object moves relative to an observer, its edges and internal features move together in a systematic way. Common fate motion is a robust source of information for segmenting objects, and is one of the first object segmentation cues used by young infants. Using a multiple object tracking task, we demonstrate that grouping objects by common motion patterns allows observers to greatly increase their tracking capacity. In Experiment 1, observers tracked 8 target squares among 8 distractor squares. While observers could not track all 8 targets when they moved independently, 4 groups of 2 targets were easily tracked when each target within a pair (horizontally separated by 4 degrees) moved with the same motion pattern. Performance for tracking these 4 groups of 2 was equivalent to performance in a control condition in which observers tracked 4 elongated rectangles, suggesting that the capacity increase for common motion pairs is accomplished by grouping the paired targets into a single object representation. Experiment 2 showed that common fate grouping appears to be inflexible and mandatory in this task. Using the same type of grouping, single targets became very difficult to track when each target's motion was locked to the motion of a distractor item, even though their common motion was irrelevant to the task. Grouping by common fate motion appears to be a mandatory rule for grouping regions in this multiple object tracking task. Strikingly, observers seemed to be forced to treat two regions of a display as a single object even though they were not visibly connected.
NDSEG Graduate Fellowship to S.F.