Rania Ezzo^1,2 (), Marisa Carrasco^1,3, Jonathan Winawer^1,3, Bas Rokers^1,2,3; ¹Department of Psychology, New York University, ²Psychology, New York University Abu Dhabi, ³Center for Neural Science, New York University

Background. Humans are more sensitive to certain motion directions than others. We previously reported three asymmetries that combine to predict motion direction discrimination around the visual field: a large advantage for cardinal over oblique directions; a moderate advantage for radial and tangential directions over other directions; and a small advantage for radial over tangential directions. Here, we assessed corresponding neural asymmetries by comparing BOLD amplitude for various directions of motion across the visual hierarchy. Methods. In an fMRI protocol, observers (n=10) viewed large-field gratings within a stationary circular aperture (12.2° radius, centered at fixation). The gratings were either static (horizontal, vertical, oblique orientations) or drifting (1 of 8 motion directions along the cardinal and oblique meridians). Regions-of-interest (ROIs) were defined based on retinotopic mapping procedures and a motion localizer, including: V1, V2, V3, hV4, hMT+. Response amplitudes were computed by a GLM, and motion-selective responses were quantified as the difference in responses for each drift direction and its corresponding static condition. Each neural asymmetry was then quantified for each polar angle bin (45° width) and for each ROI. Results. Areas V1, V2, V3, hV4, and hMT+ had greater motion-related BOLD amplitude for cardinal than oblique directions. This cardinal advantage occurred across all polar angle bins. Evidence for an effect of radial and tangential motion over other directions was less clear. Finally, on average, we found greater responses to radial than tangential motion directions in V2 and V3v, but this effect was less consistent across polar angles and across individuals. Conclusion. The neural asymmetry between cardinal and oblique motion directions is consistent across individuals, several ROIs and polar angles, while differences between radial and tangential motion directions occur in select regions, and is more variable across polar angles. This pattern of results broadly aligns with our prior behavioral results.

Acknowledgements: (1) US NIH National Eye Institute R01-EY027401 to MC and JW, (2) NYUAD Center for Brain and Health, funded by Tamkeen NYU Abu Dhabi Research Institute #CG012 to BR, (3) ASPIRE Precision Medicine Research Institute Abu Dhabi (ASPIREPMRIAD) #VRI-20-10 to BR and (4) Global PhD Fellowship, NYUAD to RE.