Abstract

The authors review how attention helps track and process dynamic events, selecting and integrating information across time and space to produce a continuing identity for a moving, changing target. Rather than a fixed ‘spotlight’ that helps identify a static target, attention needs a mobile window or ‘pointer’ to track a moving target, picking up pieces of evidence along the way to determine not just what the target is, but what it is doing. Behavioural studies show that this dynamic version of attention is model-based, using familiar trajectories to help identify a target and to guide encoding of continuing input from its path. Attention has very coarse temporal resolution for both static and moving targets. However, when the focus of selection is on the move, a given location on a moving target’s path can be selected for extremely brief instants, as little as 50 ms, compared to the typical ‘dwell time’ or minimum duration of attention selection at a fixed location, of 200 ms or more. To determine the path of a moving object, attention must accurately process and sort the onsets and offsets in order to match an offset to the subsequent onset. This aspect of dynamic attention has been called the ‘when’ pathway and patient studies show that it is a qualitatively different system from spatial attention, being completely based in the right parietal lobe for events in both hemifields. Finally, like the salience map of spatial attention, temporal attention may have its own map that guides allocation to upcoming, current, and recent moments to select information at the appropriate time, changing the experience of time as it does so.