hide
Free keywords:
-
Abstract:
Perceptual multistability has been studied for centuries using diverse approaches. Insights derived range from core principles of information processing such as perceptual inference, to high-level concerns such as visual awareness. The dominant computational explanations of perceptual multistability are based on “analysis by synthesis”. However, they 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 formulate perceptual multistability in terms of dynamic, value-based, choice, employing the formalism of a partially observable Markov decision process (POMDP). We use binocular rivalry as an example, considering different explicit and implicit sources of reward and punishment for each percept (Safavi & Dayan, Neuron 2022). The resulting values are time-dependent and influenced by novelty as a form of exploration. The solution of the POMDP is the optimal perceptual policy. This optimal policy can replicate and explain several characteristics of binocular rivalry: 1. It reproduces apparently spontaneous random switches, with approximately gamma-distributed dominance periods (two hallmarks of perceptual multistability). 2. It captures modulation by reward. 3. It explains the rich temporal dynamics of perceptual switching rates. To our knowledge, this model is unique in explaining the last two observations. In sum, our value-based decision-theoretic model of perceptual multistability not only synergizes with previous models, but also explains elusive aspects. The latter might link to the differences reported between psychiatric and healthy populations that exactly concern the temporal dynamics of perception (e. g., differences in switching rate).
