It is well established that the spectral sensitivity under photopic conditions varies across the human retina. We investigate the mechanisms underlying these spectral changes. Through the use of color appearance, flicker sensitivity, additivity, discrimination at threshold and modeling, we show that the changes in spectral sensitivity on a photopic white background across parafoveal retina are consistent with shifts in cone weightings to (L-M) and (M-L) chromatic channels. This two channel model, developed to account for foveal spectral sensitivity curves, provides a better description of parafoveal data than both a single color channel upper envelope model (comprised of a single red-green opponent channel and an achromatic mechanism) and a vector model (combining a red-green opponent channel with an achromatic component). Thus while the two channel model ([L-M] and [M-L]) of foveal color vision is generalizable to the parafovea, simple models with a unitary red/green process ar e not. Although the two channel model can accurately fit parafoveal spectral sensitivity curves without it, a small contribution from a luminance mechanism might improve the ability of the two channel model to account for threshold discrimination and additivity data.
Verdon, W., & Haegerstrom-Portnoy, G. (1996). Mechanisms underlying the detection of increments in parafoveal retina. Vision Research, 36(3), 373–390. https://doi.org/10.1016/0042-6989(95)00103-4