Abstract
Space and time support visual working memory (VWM) by providing reference frames, in which objects are incidentally represented relative to other objects. Whereas this has been shown for stationary items appearing sequentially at different locations, natural scenes also contain moving objects, whose positions change over time, often in a predictable manner. VWM may bridge the gaps when objects temporarily move out of view. Here, we investigated if predictable motion is utilised to facilitate VWM for an object’s surface features – that is, if spatiotemporal reference frames are updated for moving objects. In a dynamic change-detection task, participants memorised the colours of three disks moving at constant speed in different directions before disappearing. After a retention interval, the disks then reappeared (I) at the movement endpoint locations where they had disappeared (II) at positions spatiotemporally congruent with their previous motion direction and speed (i.e., where they would have been had they continued their movement), (III) at positions spatially congruent but with a temporal offset (consistent with a change in speed while the objects were out of view) or (IV) at positions temporally congruent but with a spatial offset (consistent with a change in movement direction). Memory performance decreased with an increasing temporal or spatial offset relative to spatiotemporally congruent positions, indicating that spatiotemporal reference frames incorporated an extrapolation of the memorised objects’ motion to facilitate retrieval. This congruency effect diminished over time, however, in a task context with unreliable motion patterns: When items were likely to reappear at spatiotemporally incongruent positions, performance increased at the movement endpoint locations. Thus, objects’ motion and resulting changes in their configuration can be leveraged to support VWM for surface features, but this updating of spatiotemporal reference frames only occurs if the mental motion extrapolation allows for a reliable prediction of where occluded objects will reappear.