Jin Shi¹, Kirsten Adam¹; ¹Rice University

Long-standing accounts of attentional selection propose that visuospatial attention is captured by salient visual information, and attention must subsequently be disengaged and reoriented. Here, we have developed a novel task to measure the disengagement of covert attention from salient information. We call this task the “anti-singleton” task, because it is a covert attention analogue of the well-known “anti-saccade” task used to study overt shifts of attention (Hallett, 1978). On each trial, participants viewed an array of 8 colored items arranged in a circle, and 1 of the items was a unique color singleton (e.g., 1 red item and 7 green items). Participants (n=18) reported the orientation of a line inside a target item (horizontal or vertical). Under the “pro-singleton” condition, the target was defined as the unique color singleton; under the “anti-singleton” condition, the target was defined as the item spatially opposite from the color singleton (i.e., 180 degrees around the circular array). In the pro-singleton condition, participants could shift attention directly to the singleton target. In the anti-singleton condition, participants had to first locate the singleton, and then rapidly shift attention away from the singleton to locate the target. Response times were substantially slower in the anti-singleton task (658 ms) compared to the pro-singleton task (1,008 ms; p < .001), consistent with an added shift away from the singleton. Pro-singleton trials elicited a robust N2pc (~180-280 ms), reflecting efficient target selection. Anti-singleton trials elicited an initial N2pc toward the color singleton (~180-280 ms), followed by a subsequent N2pc to the opposite hemifield containing the target (~280-400 ms). Together, our behavioral and electrophysiological results confirm that the anti-singleton task reliably measures serial shifts of covert attention toward and then away from salient information.