Amit Rawal^1,2 (), Michael J. Wolff¹, Rosanne L. Rademaker¹; ¹Ernst Strüngmann Institute (ESI) for Neuroscience in Cooperation with the Max Planck Society, Frankfurt, Germany, ²Faculty of Behavioural and Human Movement Sciences, Vrije Universiteit Amsterdam, The Netherlands

Visual working memory (VWM) allows brief maintenance of information to serve behavioral goals. Recently, Henderson and colleagues (2022) highlighted this goal-oriented nature, demonstrating that a predictable response allows people to flexibly change a visuo-spatial memory representation into a motor-based one. Can such flexibility also be observed between two visual features, color and space? In this eye-tracking study, participants performed two blocks of trials during which they remembered a centrally presented (250ms) color for 3 seconds, subsequently reporting it on a color-wheel via method-of-adjustment. Critically, this color-wheel was either randomly rotated on every trial (random block) or was identical on every trial (fixed block). In a third block, participants remembered a spatial location. When people remember a color in the fixed color-wheel blocks, specific colors are tied to fixed locations. Do people take advantage of this inherent spatial information to aid behavioral performance? And is this reflected in gaze behavior? Confirming that people can utilize color-space associations, we find higher accuracy and quicker responses in fixed compared to random color-wheel blocks. This replicates comparable recent work (Bae & Chen, 2024). Spatial biases in gaze can reflect the location of a prioritized item (van Ede et al., 2019), which we replicate during the delay of the single-item spatial memory block. Importantly, there was also a bias when the color-wheel was fixed, but none when it was random (as color was not predictive of space). During recall, a further gaze bias towards the correct color was observed, which was larger in the random compared to fixed color-wheel blocks. This may reflect a lack of integration of spatial knowledge in random color-wheel blocks. Together, these results show that behavioral performance can be improved by incorporating spatial information to aid color memory, and that systematic changes in gaze can index this flexible utilization of VWM.