Marc M. Himmelberg¹, Yuna Kwak¹, Marisa Carrasco¹, Jonathan Winawer¹; ¹New York University

Introduction. The surface area of human primary visual cortex (V1) is highly variable, with a 2-to-3-fold range among healthy adults, and a coefficient of variation (CoV) of around 0.17, much greater than the surface area of the cerebral cortex, with a CoV around 0.06. This variation likely has implications for visual encoding. Here, we used fMRI to ask how the surface area of V1 is related to its spatial frequency tuning. Methods. In each observer (n=34), we used fMRI to derive retinotopic maps and quantified overall V1 surface area (to 12° eccentricity) and cortical magnification as a function of eccentricity and polar angle. In a separate fMRI session, we measured V1 spatial frequency tuning, estimating preferred frequency as a function of eccentricity and polar angle. We also derived an overall spatial frequency tuning metric for each observer –cycles per V1– by integrating the functions over eccentricity and polar angle. Results. Our data showed substantial variability in V1 surface area (CoV=0.15), and even larger variability in overall spatial frequency tuning (CoV=0.36). Across observers, V1 surface area positively correlated with cycles per V1 (r=0.50); indicating that larger V1s preferred higher spatial frequencies. At the group level, the pattern of V1 cortical magnification and preferred spatial frequency systematically covaried with visual field location. Both were greatest near the fovea and decreased with eccentricity. Moreover, cortical magnification and preferred spatial frequency were greatest at the horizontal, intermediate at the lower vertical, and lowest at the upper vertical meridian. Conclusions. These data reveal a link between the surface area of V1 and its spatial frequency tuning –at the individual level, when summed over visual field location, and at the group level, as a function of visual field location.

Acknowledgements: Funding: NIH R01-EY027401 to MC and JW