Golnaz Forouzandehfar¹ (), Chengrui Li², Aaron T. Buss¹, A. Caglar Tas¹; ¹University of Tennessee, Knoxville, ²School of Computational Science & Engineering, Georgia Institute of Technology

The present study investigated the effects of spatial proximity and memory similarity on saccade curvatures. Participants were asked to learn a feature-space association to guide their saccades (Wifall et al., 2017) (e.g., for red, saccade to the leftmost object). In the learning block, participants were shown four colored circles 8dva above the fixation. The fixation object changed color, and participants were to execute a saccade to the circle that matched that color. After the learning block, they were asked to complete the same task with four black circles. That is, the task required participants to make a memory-based target selection. We manipulated both spatial proximity of the circles and similarity of the colors associated with the circles. In the close location condition (CL), the separation between adjacent squares was ~2.5dva while it was ~7dva in the far condition (FL). In the similar color condition (SC), the colors for adjacent squares were 30° apart while in the dissimilar color condition (DC), they were 90° apart in HSV color space. Participants were randomly assigned to one of the four conditions: SC-CL, SC-FL, DC-CL, DC-FL. Based on spatial proximity effects in previous studies, we expected to find greater curvature in the CL conditions than in the FL conditions. If memory representations can affect saccade curvatures similarly as perceptual representations, then we also predicted larger curvatures in the SC conditions compared to the DC conditions (Mulckhuyse et al., 2009). We found a significant effect of color (p=.006) with similar colors resulting in significantly larger curvatures than dissimilar colors, but this effect was only present when squares were spatially close. The effect of spatial proximity was not significant (p>.05). Importantly, our results show that oculomotor control is not only influenced by similarity in perceptual representations, but also similarity in memory representations.