Abstract
As attention dynamically shifts from location to location, feature perception can be altered depending on where spatial attention is allocated. When spatial attention is captured by distracting stimuli, both swapping errors (misreporting a distractor color instead of the target color) and repulsion (perceptual distortion away from a distractor color) can occur (Chen, Leber & Golomb, 2019). We hypothesized that the nature and level of these feature-binding errors may vary from trial-to-trial depending on the degree of attentional capture and efficiency of disengagement; e.g., such that swapping errors occur during capture while repulsion occurs during disengagement from the distractor. To test this, in the current study, we sequentially presented the distractor and target cues with variable delays, with the target stimulus appearing either 50ms or 150ms after the spatial distractor cue, to manipulate the time course of attentional capture and disengagement. Participants reported the color of the target item with a continuous color report, and we used the same probabilistic modeling approach as Chen et al. (2019) to characterize the different types of errors made and their relative proportions. We predicted that the salient distractor would interfere with performance more at the 50ms condition, whereas in the 150ms condition participants would be more likely to have disengaged from the distractor. In line with our predictions, we saw significantly more swapping errors in the 50ms condition than in the 150ms condition. The same pattern, however, was not found with repulsion errors. These differences do not appear to be driven by task difficulty, as the random guessing rate and precision were similar between the two conditions. These results suggest that swapping and repulsion feature-binding errors are observed following different time courses when attention is captured by and disengages from a distractor.