Abstract
The visual environment is highly stable, where one object often predicts the presence of another object (e.g., a fork is often next to a knife). Such regularities can be useful in guiding visual search (e.g., seeing a fork may help the search for a knife). Here we examine how regularities bias attention to visual targets. In a visual search paradigm, participants searched for the target (a rotated T) among L-shaped distractors as quickly and accurately as they could. The target and distractors were presented in distinct colors. Unbeknownst to the participants, one distractor color (e.g., green) always predicted the nearby location of the target in the structured condition. In a separate random condition, participants viewed the same array except the colors were randomly shuffled for every trial, so the distractor color did not predict target location. We found that participants were faster to find the target in the structured condition than in the random condition (Experiment 1), suggesting a beneficial effect of regularities on visual search, consistent with previous studies. However, when we increased the number of predictive distractor colors (i.e., two distractor colors always predicted the nearby location of the target), we found the exact opposite effect: participants were slower to find the target in the structured condition than in the random condition (Experiment 2). This suggests that the two co-occurring colors may be distracting, drawing attention away from the target to the colors themselves, despite the fact the colors predict the nearby target location. The results demonstrate that regularities do not always benefit target search, in some cases can hurt search performance if regularities themselves compete for attention. The study raises a possible mechanism where the predictor draws attention to itself first and then directs attention to the target.
Acknowledgement: Natural Sciences and Engineering Research Council of Canada