hide
Free keywords:
-
Abstract:
Perceptual multistability -- when two or more percepts alternate or switch in response to a single ambiguous sensory input -- has been studied for centuries using diverse approaches, and has afforded insights into a diverse set of cognitive functions (e.g., perceptual inference, attention, visual awareness). However, prominent theories of the phenomenon fail to account for the crucial role played by value, e.g., with percepts paired with reward dominating for longer periods than unpaired ones. We reformulate perceptual multistability in terms of a decision process, employing the formalism of a partially observable Markov decision process (POMDP). In our model, information accrues about perceived stimuli, which duly provide reward as a dynamic form of time-varying aesthetic value. The solution of the POMDP is the optimal perceptual policy which replicates and explains several elusive aspects of rivalry. In sum, our value-based decision-making account of perceptual multistability synergizes with previous models and also offers a more comprehensive treatment of computational and algorithmic facets of this phenomenon.