Blue-yellow signals are enhanced by spatiotemporal luminance contrast in macaque V1. J Neurophysiol 93: 2263-2278, 2005; doi:10.1152/jn.00743.2004. We measured the color tuning of a population of S-cone-driven V1 neurons in awake, fixating monkeys. Analysis of randomly chosen color stimuli that were effective in evoking action potentials showed that these neurons received opposite sign input from the S cones and a combination of L and M cones. Surprisingly, these cells also responded to LM cone contrast irrespective of polarity, a nonlinear sensitivity that was masked by conventional linear analysis methods. Taken together, these observations can be summarized in a nonlinear model that combines nonopponent and opponent signals such that luminance contrast enhances color processing. These findings indicate that important aspects of the cortical representation of color cannot be described by classical linear analysis, and reveal a possible neural correlate of perceptual color-luminance interactions.
I N T R O D U C T I O NColor perception results from a complex neuronal computation. The early steps of this computation are well understood: the sensitivity of the cone photoreceptors to lights of various wavelength has been characterized thoroughly (Baylor et al. 1987;Wandell 1995) as has the subsequent synergistic and antagonistic combination of cone signals in the retina and lateral geniculate nucleus (Derrington et al. 1984;DeValois 1965; Gouras 1968). How color information is processed in cortex, however, is less clear. The goal of the current experiments was to extend our understanding of how neurons in the primary visual cortex (V1) combine signals that originate in the cones.One possibility is that V1 neurons combine cone signals linearly. This assumption has been made, either implicitly or explicitly, in studies that employ cone-isolating stimuli to estimate the weights with which cone inputs are integrated (Conway 2001;Johnson et al. 2001;Landisman and Ts'o 2002). While the linearity assumption is justified for many V1 neurons, it is clearly inappropriate for others (Conway 2001;Hanazawa et al. 2000;Hubel and Wiesel 1968;Lennie et al. 1990;Vautin and Dow 1985). For example, a V1 neuron that responds to S cone stimulation, but only when L-and M-cone excitations are appropriately balanced, does not integrate cone inputs linearly and thus cannot be characterized with coneisolating stimuli (Hanazawa et al. 2000).Recently developed data-analysis tools have provided new ways to characterize neurons that combine inputs nonlinearly (de Ruyter van Steveninck and Bialek 1988;Paninski 2003;Rust et al. 2004; Schwartz et al. 2001;Sharpee et al. 2004;Simoncelli et al. 2004;Touryan et al. 2002). Here we use one of these techniques to reveal a surprising nonlinear computation performed by blue-yellow neurons in V1.We excited V1 neurons in awake monkeys with a dynamic, randomly colored stimulus and analyzed the stimulus sequences that preceded spikes. Analysis proceeded in two steps. First, we computed the average stimulus ...