The relationship between property-encoding and object-based attention: Evidence from multiple object tracking

When objects are attended, which of their properties are encoded as a result? We explore this question using a ‘multiple object tracking’ (MOT) task, wherein subjects must track multiple identical and unpredictably moving items in a display filled with identical distractors. In three experiments, subjects had to report the locations, directions of motion, colors, or shapes of items in a MOT task when these properties were suddenly obscured in various ways. To investigate whether items’ featural properties are encoded as a result of being tracked, the colors and shapes of items were occasionally permuted when the items occluded or briefly disappeared, and subjects reported these properties when a single probe item disoccluded or reappeared as a simple placeholder. To investigate whether items’ spatiotemporal properties are encoded as a result of being tracked, an item disappeared or stopped moving during MOT, and subjects reported the missing object’s location, or direction in which the now-stationary object had been moving. For location and direction of motion, performance was more accurate for tracked items than for untracked items. In contrast, no such differences were observed for color or shape information. This pattern of results suggests that in certain situations, featural properties may not be encoded even when the objects which have those properties are attended. This surprising finding may result from the complex attentional load induced by MOT, or may reflect the operation of an early visual ‘indexing’ mechanism.