Color Opponency Modulates Feature Integration Through Bayesian Priors

Stimulus color, even irrelevant to the task, has been shown to modulate high-level cognitive functions, producing different behavioral outcomes. However, the effect of color on lower-level perceptual processes remains unclear. To address this gap, we investigated whether color affects feature integration, an early-stage visual process, using the flash-jump illusion. Our results demonstrated that color modulates the integration of color and motion features, as red and blue flashes resulted in more veridical estimates of flash location, compared to green and yellow. We provide a novel interpretation of our current results in respect to a Bayesian perceptual framework, where the color of the flash is inherently assigned different Bayesian weights, resulting in different levels of perceptual shifts. Furthermore, in reviewing pertinent literature and empirical evidence, we have proposed a novel theory outlining three putative mechanisms, predictions and underlying neural circuitry using Bayesian frameworks to explain such color-dependent modulations in visual processing.