Abstract
Ballistic eye movements, known as saccades, help us explore complex visual environments by projecting the most relevant visual information to the retinal region with the highest resolution. However, there is also extensive experimental evidence for impaired visual perception for objects presented shortly before or after saccades. Nevertheless, our perceptual experience of the visual world is not interrupted every time we make a saccadic eye movement. Given that real-life scenes often contain regularities and redundancies, we explored if regularities in visual environments can protect us from impaired perception due to saccadic eye movements. Subjects fixated a fixation cross, and on “saccade” trials made an eye movement to a saccade target. Eight colored squares were briefly presented around the saccade target during the perisaccadic interval. The task was to report the color of a single post-cued target square using a continuous color report. Critically, stimulus regularities were induced by separating stimuli into two groups with different mean colors; individual stimulus colors were selected from distributions around these ensemble means. We fitted the error distribution of reported colors into a mixture model which consisted of a target, non-target, and random guess components using MemToolbox (Suchow et al., 2013). In all conditions, individual color reports were significantly biased toward the mean color of their group, indicating robust reliance on group-level visual properties (BF10>3). Moreover, this attraction bias towards the ensemble mean did not significantly differ for saccade trials compared to no-saccade trials (BF10<0.3). The precision of target response, probability of reporting non-target, and random guesses of trials with saccades also did not differ between saccade trials and no-saccade trials (BF10<0.3). While previous studies have found impaired peri-saccadic perception, the present results suggest that redundancy and regularity in visual environments may protect us from abrupt changes in visual input across eye movements.