Coordinates of Attention

Patrick Cavanagh

Vision Lab, Department of Psychology, Harvard University

Attention allows a desired target to be selected from a cluttered field of distractors. We have characterized the dimensions of the region of selection in both space and time. In space, we derive the minimum size of the selection region – the acuity of attention – from measures of “crowding” where a single target is surrounded by distractors. We use attentive tracking tasks to examine how multiple regions of selection are deployed and demonstrate significant interference between nearby selections due to the suppressive surrounds of each selection region. We find the interference to be hemifield limited suggesting an independent control of attention in each hemisphere. This pattern is also seen with parietal patients who fail the tracking task only in the contralesional hemifield.  Recent evidence with small tracking displays shows that the target-to-target interference can be limited to individual quadrants – it is much reduced if two targets straddle either the vertical or horizontal meridian, suggesting involvement V1 through V4, cortices with separate representations of each quadrant outside the fovea. In the time dimension, we find a minimum selection interval of about 150 msec that is followed by a suppressive surround limiting attention to a subsequent target. The temporal duration of the selection interval is not localized in space. If attention is moving, the access to a single location can be as brief as 50 msec or less, allowing us to probe brief moments in a rapidly changing event sequence that cannot be accessed with attention fixed at that location. This moving attention window also demonstrates accumulation of information across locations that suggest position-independent analyses in non-retinotopic cortices.