Color and Spatial Frequency Provide Functional Signatures of Retinotopic Visual Areas

Abstract

Primate vision relies on retinotopically organized cortex defined by representations of hemifield (upper versus lower visual field), eccentricity (fovea versus periphery), and area (V1, V2, V3, V4). To test for the functional signatures of these organizing principles, we used fMRI to measure responses to colored gratings with varying spatial frequency, color, and saturation across rentinotopically defined parcels in macaque monkeys. Using an approach we call Sparse Supervised Embedding (SSE), we identified stimulus features that differed most across the cortical parcels. Constraining the model according to each voxels eccentricity revealed the expected variation of spatial frequency and S-cone modulation with eccentricity. Constraining according to voxel dorsal-ventral location and retinotopic area, revealed two additional components which together provide unexpected functional signatures. First earlier visual areas were more sensitive to contrast and saturation compared to later areas. Second, ventral parcels responded less to luminance contrast and more to color, particularly along the orange-blue intermediate axis in cone-opponent color space. This dorsal-ventral asymmetry resembles differences between candidate dorsal and ventral subdivisions of human V4 but extends to include all retinotopic visual areas, with a magnitude that increased from V1 to V4. Overall, our study demonstrates SSE as a data-driven tool for generating hypotheses about cortical function and behavior, and provides insight into the functional roles of different retinotopic areas. Significance StatementThis study explores the functional roles of visual areas (V1, V2, V3, V4, V3a, MT) parceled by eccentricity preference and visual hemifield. We employ a new approach called Sparse Supervised Embedding (SSE), a promising technique for visualizing and understanding complex neuroimaging datasets. Analyzing fMRI responses to different color grating stimuli, we uncover a functional organization across these parcels, shedding light on their roles in visual computation. This study enhances our knowledge of processing in early visual cortex and how it varies across the retinotopic map. It also aids in resolving apparent differences between prior human and macaque studies.