Kanathip Jongmekwamsuk^1,2 (), Chaipat Chunharas², Rosanne L. Rademaker¹; ¹Rademaker Lab, Ernst Strüngmann Institute for Neuroscience in cooperation with Max Planck Society, Frankfurt, Germany, ²Cognitive Clinical & Computational Neuroscience, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand

The human brain combines prior knowledge with sensory information to make inferences about the environment and construct our perception of the world. To find out if people prefer certain responses over others when recalling a direction of motion, and to see if such inhomogeneities in response preference persist as increasingly more sensory evidence is provided, we had human participants report the direction of motion of moving dot stimuli shown for 1s at four possible levels of coherence (0%, 25%, 50%, and 75%). Across trials with coherent motion, we ensured uniform sampling of all possible directions of motion, which means that in the case of accurate recall (plus noise), we expect a uniform distribution over all possible response options. Similarly, on trials without coherent motion, random responses should also lead to a uniform response distribution. Our data revealed striking inhomogeneities in reported direction in the absence of coherent motion, with many more reports at cardinal and oblique directions compared to other directions. And while performance increased with higher coherence, with reports becoming more uniform across motion directions, clear non-uniformities persisted even on high coherence trials. Specifically, people continued to over-report oblique (but not cardinal) directions even for 75% coherent dot motion. This suggests an integration of existing inhomogeneities in response preference (revealed in the 0% coherence condition) and sensory information, where direction estimation is increasingly less affected by a “response prior” as sensory certainty goes up, but the weight of the response prior appears to differ between cardinals and obliques.