Sensory or input factors can influence the strength of interference in the classic Stroop color-word task. Specifically, in a single-trial computerized version of the Stroop task, when color-word pairs were incongruent, opponent color pairs (e.g., the word BLUE in yellow) showed reduced Stroop interference compared with nonopponent color pairs (e.g., BLUE in red). In addition, participants' color discrimination ability was measured by standard color vision tests (i.e., Farnsworth-Munsell 100-Hue Test and Ishihara plates). Error rates in the Farnsworth-Munsell test correlated positively with the amount of Stroop interference. Neural network simulations (variants of J. D. Cohen, K. Dunbar, & J. L. McClelland's, 1990, model) showed that only a distributed trichromatic input layer was able to simulate these findings. Thus, sensory input from the color system needs to be incorporated into current accounts of the Stroop effect.