Spatial biases in visual working memory encoding persist despite controlled gaze position

Abstract

In everyday life, goal-driven selection processes allow us to selectively encode the most critical items into working memory. However, in traditional visual working memory experiments we frequently overload capacity without providing information about which items are relevant. It is often assumed that when all items are equally likely to be tested participants select a random subset, but recent findings have instead shown that individuals tend to store items from a particular area of the visual field. One potential explanation for these findings is that subjects move their eyes in preparation for an upcoming trial, suggesting the spatial bias is driven by gaze position. For example, Chokron & Imbert (1993) showed that initial gaze position determined bias direction in a bisection task. To test whether gaze position causes encoding biases in change detection, we recorded gaze position while participants (N=22) completed 800 change detection trials. On half of these trials, fixation was maintained on the center of the visual field while the to-be-remembered items were displayed; on the other half of trials, no constraints were placed on gaze position. Despite participants maintaining fixation in the center of the visual field, accuracy was still biased in a particular area of the visual field. This result suggests that gaze position alone cannot explain the encoding biases previously observed in change detection.