Isabela Tellez¹, Ramon Bartolo¹, Bevil R. Conway^1,2, Felix Bartsch^1,2,3, Daniel A. Butts²; ¹Laboratory of Sensorimotor Research, Section on Perception, Cognition, and Action, National Eye Institute, National Institutes of Health, ²Department of Biology, University of Maryland, ³Neuroscience of Attention and Perception, Neuroscience Institute, Princeton University

Color responses of fovea V1 cells are largely unknown. Because color appearance and cone distributions vary with eccentricity, the known color-tuning properties of parafoveal V1 cells may not be instructive. Here, we measured foveal V1 receptive fields (RFs, 0 to 1.5°) using chronic Utah arrays and linear v-probes in fixating alert macaques. We obtained: 1) color-tuning responses to briefly flashed color discs (1° diameter, larger than classical RFs) of varying hue and constant saturation and luminance; 2) high-resolution spatiotemporal RFs using a dynamic chromatic stimulus (“clouds”, colors defined by cone-opponent mechanisms) with power across the frequency spectrum. Stimulus locations in retinal coordinates were obtained using neurophysiological and dual-Purkinje eye tracking, providing near photoreceptor resolution. We used Fourier analysis to determine if the color tuning curves for each neuron were captured by the first or second harmonic of a 360° cycle over the color wheel, corresponding to unimodal or bimodal tuning, respectively. Of 595 neurons, 110 were color tuned (107 unimodal, 3 bimodal). The distribution of preferred hues showed a strong bias towards the blue-orange axis and green quadrant of color space, consistent with documented color tuning of parafoveal neurons. Color-tuning curves were related to the RFs by fitting linear models, with spike-triggered averages in three chromatic channels (L, M, S) as linear filters. Projecting the disc stimuli through these filters and regressing the output against spike counts explained, on average, 61±24% of the variance (R2 ± SD) of disc responses. For 21 neurons, the model captured >80% variance, for 11 neurons, >90%. These results show that a substantial fraction of foveal V1 neurons are color tuned and that a linear transfer of the RF can account for some of the color tuning, but implicate nonlinear operations to fully explain foveal color tuning.

Acknowledgements: Supported by NIH, 1ZIAEY000558; NSF, IIS-2113197; NIH, R01 EY037347-01. NIH author contributions are Works of the U.S. Government. Findings and conclusions are those of the authors and do not necessarily reflect the views of NIH or the U.S. Department of Health and Human Services.