Fig. 1: Three-stage zone model of color vision (adapted from Stockman and Brainard3) and schematic of M, P, K geniculocortical pathways.
a At the first stage (outer ring), light is transduced by three types of cone photoreceptors, L (long), M (middle), and S (short) -wavelength sensitive. (Colors in the triangles labeled L, M, or S correspond to the appearance of cone-isolating stimuli that would selectively activate each cone type.) In the second stage, cone signals are integrated through four cone-opponent mechanisms (L+M−, M+L−, S+(L+M)−, and (L+M)+S−) instantiated by retinal circuits connecting specific cone types to retinal ganglion cell types. At the hypothesized third stage, cone-opponent signals are predicted to combine in four specific combinations to generate color-opponent mechanisms, as well as neurons whose activities underlie the perception of four “unique colors” (red—RG; green—GR; yellow—YB; blue—BY). b The observed spatial-overlap model of cone-opponent functional domains (COFDs) corresponds to the cone-opponent mixing combinations predicted in the Stage 3 of the Three-stage model. Outer ring illustrates the appearance of colors at different directions in the isoluminant plane of DKL color space. Generation of color preferences in the non-cardinal directions requires mixing of the cone-opponent mechanisms. c Anatomical organization of cone-opponent thalamocortical inputs. Cone-opponent signals are transmitted along Parvocellular (P) and Koniocellular (K) pathways from RGCs to different layers in V1 (See Text). Achromatic inputs from the Magnocellular (M) pathway project to layer 4 Cα. The mixing of cone-opponent signals in Stage 3 (a) requires neural circuits that span layers 4Cβ, 4 A, and blobs in layer 2/3 to allow mixing between the L/M opponent inputs that target layer 4 Cβ and the S/L+M opponent inputs that target layers 4 A and 3B. Note that the pathways and recipient layers are labeled with different colors corresponding to different cone-opponent signals.
