Abstract
Experiments on the neural mechanisms of feature-based attention suggest that features are selected in parallel across the visual field (Treue & Trujillo,1999). Moreover, feature-based attention not only spreads to the task-relevant feature in an unattended location, but it also spreads to a task-irrelevant feature in the unattended location by virtue of it being bound to the same object as the task-relevant feature (Sohn et al., 2005). In all these experiments, however, the unattended location contained the task-relevant feature. We asked whether a similar mechanism is at work when the unattended location does not contain the task-relevant feature, and instead contains a task-irrelevant feature that matches the task-irrelevant feature in the attended location. Specifically, we asked whether attending to the color of moving dots in one visual field (VF) would influence the neural response to achromatic moving dots in the other VF. Subjects were instructed to attend to red or cyan dots (both present at the attended location and moving in opposite directions) and respond when the target dots dimmed. Critically, the achromatic dots in the opposite VF could either match the direction of the attended dots (same), match the direction of the unattended dots (opposite), or match the direction of neither the attended or unattended dots (neutral). Preliminary analysis revealed that, relative to the same and neutral conditions, BOLD activity in human MT/MST corresponding to the unattended location was suppressed when the direction of motion matched the direction of the unattended dots in the attended location. This result suggests that just as the attended color is selected in parallel across the VF, the task-irrelevant motion of the dots in the unattended color can be inhibited across the VF. This was true even though the dots in the unattended location did not contain either the attended or unattended color.