Double-Opponent Neurons Drive Perception and Early cVEP Responses to Color Patterns

Abstract

To explore the neural mechanisms of color perception we used saturation scaling as a measure of perceived color and the chromatic visual evoked potential (cVEP) as a measure of neural activity. The color of isoluminant (30 cd/m2) squares or 8×8 checkerboards, both of size 1.875° across, was rectangular-wave modulated from gray to color and back to gray (0.5s on, 1.5s off). The patterns had cone contrasts ranging from 0.04 to 0.4. Outside edges of the stimuli faded gradually into the background. With no clear edges, the color squares stimulated only single-opponent neurons while the checkerboards stimulated both single- and double-opponent cells. The participants whose cVEP responses were recorded rated the same color stimuli for apparent saturation. Saturation ratings and cVEP Fourier power were consistently larger for the checkerboard than the square, i.e., an additional response was due to the checkerboard pattern itself. This implies edge-dependent, double-opponent neurons contributed to both the checkerboard's perceived color saturation and cVEP. For the square, cVEP response dynamics were fairly linear with increasing cone contrast. In contrast, the checkerboard responses sped up substantially as cone contrast increased; evidence of dynamic nonlinearity. Perceived saturation of the checkerboard was roughly proportional to cone contrast but total cVEP power grew sublinearly with cone contrast. However, the power of the 6Hz Fourier component in the cVEP aligned closely with the saturation ratings. The observation that color perception is tracked by specific cVEP Fourier components and not others suggests that the cortical coding of color depends on response dynamics.