Adaptation to variance generalizes across visual domains.

Processing the vast amount of visual information available from the world ought to pose a significant challenge to the brain. One of the ways in which the brain appears to encode the structure inherent in the world is through summary statistical representations (e.g., mean size, color etc.). This study investigates whether variance perception can be adapted for color, and then whether the variance adaptation aftereffects generalize from color to another visual domain. In a series of four experiments, we find aftereffects reflecting adaptation to the variance of hues in an ensemble-such that prolonged exposure to a highly variable ensemble of hues makes subsequent ensembles appear less variable in hue. We also found that this effect partially generalized to the perception of orientation variance-adaptation to highly variable color ensembles made subsequent ensembles also appear less variable in orientation. This is a novel demonstration of adaptation aftereffects reflecting processing of visual ensemble information across domains. The results could imply a neural mechanism encoding visual variance that is not selective to the domain from which the variance signal is derived. This mechanism may form the basis for cross-domain visual comparisons, and may play a role in predictive coding, enabling the brain to calibrate to the complexity of the visual environment. (PsycINFO Database Record (c) 2020 APA, all rights reserved).