Abstract
Most attempts to examine distractor suppression have restricted their distractor criteria to basic visual features (e.g. color, orientation). In two experiments, we tested whether distractor suppression extends to top-down categorical information. To do so, we created two categorical stimulus sets (real world images of teddy bears and butterflies) from which we derived silhouettes (preserving form information). To provide a ground truth that the silhouettes would induce traditional distractor suppression effects, in Experiment 1 we replicated the paradigm of Geng and DiQuattro (2010), using our categorical form silhouettes as the objects in place of simple shapes. Thirty-six participants indicated the location of an oriented bar embedded in an object silhouette (teddy bears or butterflies). Importantly, they were also told to actively suppress salient objects regardless of categorical information. We found that despite significantly fewer saccades towards the target on distractor-salient trials (p < .001), RTs were ~154ms faster compared to distractor-neutral trials (p < .001). This replicates prior findings and confirms that distractor suppression effects are broadly observable with our stimulus set. In Experiment 2, 79 participants were told to actively suppress either the teddy bear or butterfly object categories. The non-suppressed object category was considered the neutral object. This time, stimulus salience was held constant across all items and three trial types were presented: isolated (one object from the neutral category), neutral (two neutral objects), or suppress (1 neutral object and 1 object from the category to be suppressed). Critically, on trials where an item from the category to be suppressed was present, RTs were ~119ms faster than neutral trials (p < .001). This was driven by shorter dwell times on the suppressed distractors than neutral distractors (p < .001), providing evidence that top-down knowledge of categorical information can facilitate the rapid rejection of non-targets.