Caitlin V. Cunningham¹, Julie D. Golomb¹; ¹Department of Psychology, The Ohio State University

In visual contexts, spatial attention facilitates processing for items appearing at an attended location and is crucial for binding multiple features together within an object. Previous studies reported that manipulating external spatial attention (e.g., shifting or splitting top-down attention) can cause errors (swapping and/or distortion) when reporting a target item’s features. However, less work has investigated errors resulting from manipulations of internal spatial attention. Here, we asked whether spatial shifts of internal attention within working memory (WM) have systematic effects on feature reports, and if so, whether these errors differ from those caused by shifts of external attention. In the current study, participants viewed an array of four colored squares. After a 1s retention period, a spatial postcue oriented internal attention toward one of the items’ locations, followed by a 500ms delay. On some trials (two-cue trials), a second postcue appeared, indicating that participants should shift their spatial attention within WM. After a short (50ms) or long (500ms) delay, participants reported the color of the most recently cued item (target) on a continuous color wheel. Using probabilistic mixture modeling, we found that the short delay resulted in performance decrements compared to the long delay in two-cue trials, suggesting that the short delay provided insufficient time for participants to update their spatial attention within WM. Interestingly, while incomplete shifts of external attention have been found to cause swapping errors (higher likelihood of misreporting the color of the first-cued item), we instead found that after the short delay, participants were more likely to report the color opposite, characteristic of feature avoidance errors. These errors were not observed after the long delay, suggesting that while participants were able to shift their internal spatial attention, the new item’s features may not have been accessible by the time of report after the short delay.

Acknowledgements: NIH R01-EY025648 (JG), NSF 1848939 (JG)