David Tu¹, Shutian Xue¹, Jonathan Winawer¹, Marsia Carrasco¹; ¹NYU

[Introduction] Performance is location-dependent: Typically, it decreases with eccentricity and is higher along the horizontal than vertical meridian–horizontal-vertical anisotropy (HVA)–and along the lower- than upper-vertical meridian–vertical meridian asymmetry (VMA). Both the eccentricity effect and polar angle asymmetries have been related either to cortical surface area or to system-level computations–gain and internal noise. Here we used an equivalent noise protocol to investigate whether and how scaling stimuli size to equate cortical surface area across locations (M-scaling) modulates spatial patterns in performance and system-level computations. [Method] Observers performed an orientation-discrimination task on a Gabor signal embedded in dynamic white noise, presented either at the fovea or 8º to the left, right, above, or below fixation. Stimuli were equated either for retinal size or for cortical size, based on individual retinotopic maps. We estimated contrast thresholds (using a functional adaptive sequential testing method) and fit a perceptual template model to estimate gain and internal noise at each location. [Results] Without external noise, M-scaling eliminated contrast threshold differences across eccentricity, but only reduced polar angle asymmetries. With equal retinal size, internal noise, but not gain, increased with eccentricity. M-scaling eliminated the eccentricity effect in internal noise and gain was higher in the periphery than at the fovea. With equal retinal size, internal noise did not vary around polar angle, and gain was higher along the horizontal than vertical meridian and along the lower- than upper-vertical meridian. With M-scaling, HVA in gain was eliminated, and VMA emerged for internal noise. [Conclusion] Visual performance differences across eccentricity and around polar angle appear to arise from different computations, the former from differences in internal noise and the latter from differences in gain. Critically, despite equated cortical surface area across locations, polar angle asymmetries in performance and in system-level computations are attenuated but persist.

Acknowledgements: NIH NEI R01-EY027401 to MC and JW