Isabel S. Jaen¹, Tzu-Yao Chiu¹, Julie D. Golomb¹; ¹The Ohio State University

People make saccades frequently to sample visual information. Since visual input is naturally coded in retinotopic (eye-centered) coordinates, an updating/remapping process is needed to maintain visual stability across saccades. Previous studies found that spatial attention lingers at the old retinotopic location briefly after making a saccade. This retinotopic attentional trace was shown to disrupt visual feature perception, including one’s ability to bind an object’s location to its features. In this study, we investigated whether the predictability of a saccade influences post-saccadic remapping and feature binding. In an eye-tracking experiment, participants were cued at a target location and asked to report the color of the item subsequently appearing at this location, after making a saccade. At variable post-saccadic timepoints, they were presented with an array of four colored squares, appearing at the cue’s spatiotopic (target) location, its retinotopic (non-target) location, and two other non-target locations. Critically, we manipulated saccade predictability, where saccades in each block were either predictable (consistent saccade target location with fixed saccade cue onset) or unpredictable (variable spatially and temporally). Results showed improved general performance when saccades were predictable. Moreover, in the unpredictable saccade condition, there were “swap errors” at the early post-saccadic timepoint, a pattern reported previously where participants report the retinotopic non-target color instead of the correct spatiotopic target color. However, the swapping errors disappeared when saccades were made predictable. These results suggest that systematic color misperceptions associated with the retinotopic attentional trace may be malleable to top-down expectations of the upcoming saccade, highlighting the role of predictive coding in maintaining visual stability across saccades.

Acknowledgements: This study was supported by NIH R01-EY025648 (JG).