According to the theory of efficient coding, sensory processing is optimized for representing the information content of natural scenes. This implies that perceptual systems are adapted to the statistical regularities of the environments they are immersed in. In color vision, relatively low sensitivity for discriminating color along blue and yellow axes has been linked to the dominance of blue-yellow color variation in natural scenes. It has been suggested that higher order visual processes, such as aesthetic preferences, could also be adapted to natural environments. Here, we manipulate the chromatic contrast of natural scenes to test whether low-, medium-, and high-level aspects of color perception can be calibrated to the color statistics of visual environments. In three experiments, we measured color discrimination, chromatic balance perception, and aesthetic judgments of colorful Mondrian patterns after adaptation to scenes with natural colors or scenes with manipulated colors. After viewing naturally colored scenes, color discrimination, perception of chromatic balance, and color preferences are biased along the blue-yellow color axis, along which there is most chromatic variance in natural scenes. Blue-yellow biases were reduced or partially inverted following adaptation to color-manipulated scenes, though the extent of the reversal in bias we observed following short-term adaptation appears to have been limited by long-term adaptation to the color statistics of natural environments. Our findings support the efficient coding theory and provide experimental support for the hypothesis that multiple attributes of human color perception adapt to the color statistics of visual environments. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Wozniak et al. (Mon,) studied this question.