Abstract
We are constantly perceiving our dynamic world and encoding its features. Prior research showed that attention plays a crucial role in feature perception and encoding, with attentional capture eliciting a distinct pattern of feature errors. However, the magnitude of these feature errors has differed across studies, despite largely similar designs. One possibility for this variation could be the presence/absence of overlapping target and distractor cues, i.e., “valid” trials. Could the learned relevance of a distracting cue affect our ability to suppress it and reduce feature errors? To learn more about the importance of learned distractor relevance, a largely similar delayed estimation task was again used, but the experiment was split into two contexts. The contexts differed by the prevalence of target and distracting cue overlap. One context (0% match) had no overlap of the target and distracting cue while the other context (50% match) had a target-distracting cue overlap on half of the distractor-present trials. The two contexts were presented in separate halves of the experiment, with order counterbalanced across participants. We used a probabilistic mixture model to estimate parameters of interest including the swap rate, mean shift, standard deviation, and guess rate for each condition. Our results replicated previous findings of significant swap errors on “invalid” trials, where subjects reported the color at the distracting cue location instead of the target location. Interestingly, we found a difference in the magnitude of the swap effects over time, depending on which context was experienced first. These results suggest that the learned relevance of a distractor cue can affect how likely participants were to be captured by a salient distractor and its resulting impact on target feature perception, and that statistical regularities relating to salient items affect the perception and encoding of stimulus features.