Abstract
Observers can use an attentional template consisting of a task-relevant feature (target template) to efficiently guide eyes to matching objects in an array. One explanation for this behavior is that perceptual processing of the currently fixated object biases observers to fixate objects with similar features. Another possible explanation is that observers continue to strengthen this target template by fixating objects that have the relevant feature. In real-world search tasks, however, observers frequently fixate objects that do not have task-relevant features. How do these task-irrelevant saccade targets influence the integrity of the target template and efficiency of search? To test this, observers performed a gaze-contingent search task (find the circle with a top or bottom gap) and we examined how the history of fixating objects with one feature value (e.g., blue) influenced selection of the next object when presented with a choice. Red and blue circles appeared one at a time, and each required fixation before the next would appear. At varying points during a trial, two circles (one blue, one red) appeared equidistantly from the currently fixated circle (choice point). When observers did not know the color of the target item (no target template), they fixated a circle that matched or did not match the choice point color equally often and independently of the number of blue or red circles fixated sequentially before the choice point. When observers did know the target item color (active target template), they consistently fixated the task-relevant circle, regardless of how many task-relevant or irrelevant circles were recently fixated. These results suggest that the target template representation is buffered from both the potential effects of perceptual priming and interference generated by incidental encoding of fixated objects into visual working memory. Guidance appears to be independent of the incidental perceptual events occurring during visual search.
Meeting abstract presented at VSS 2013