Abstract
We are often bombarded with salient stimuli that can capture our attention and distract us from our current goals. Decades of research has shown the robust detrimental impacts of salient distractors on search performance and, of late, in leading to feature errors. However, recent work has shown that salient distractors in search tasks can be suppressed if they are either biased to one location or indirectly probabilistically cued, thereby reducing the slowing of reaction times on distractor-present trials. While this previous work has shown the beneficial effects of distractor suppression on search tasks, how suppression influences the encoding of features is not well understood. When attempting to encode a target's features, can learned spatial suppression protect against feature errors driven by a salient distractor? To investigate this question, participants briefly viewed four differently colored squares on every trial, with the target square surrounded by a thick, white frame, which participants subsequently reported on a color wheel. On two-thirds of trials, a salient distractor (four white dots) appeared briefly around a nontarget square. We created a predictable distractor location by presenting the salient distractor in one location on 62.5% of distractor-present trials. Participants' responses were fit to a probabilistic mixture model estimating performance parameters we could compare across conditions. Our results showed that performance (including guessing rates, response precision, and probability of misreports) was significantly improved when the salient distractor appeared in a likely location relative to if it appeared elsewhere. Additionally, evidence from our distractor-absent trials suggested that the suppression of this predictable distractor was likely reactive, as there was little difference in performance within these trials depending on whether the target was in the likely distractor location or not. These results suggest that reactive suppression of a salient distractor can help protect against the perceptual disruptions of nontarget features.