Interacting with volatile environments stabilizes hidden-state inference and its brain signatures

Aurélien Weiss,Valérian Chambon,Jan Drugowitsch,Junseok K Lee,Valentin Wyart

doi:10.1038/s41467-021-22396-6

Abstract

Making accurate decisions in uncertain environments requires identifying the generative cause of sensory cues, but also the expected outcomes of possible actions. Although both cognitive processes can be formalized as Bayesian inference, they are commonly studied using different experimental frameworks, making their formal comparison difficult. Here, by framing a reversal learning task either as cue-based or outcome-based inference, we found that humans perceive the same volatile environment as more stable when inferring its hidden state by interaction with uncertain outcomes than by observation of equally uncertain cues. Multivariate patterns of magnetoencephalographic (MEG) activity reflected this behavioral difference in the neural interaction between inferred beliefs and incoming evidence, an effect originating from associative regions in the temporal lobe. Together, these findings indicate that the degree of control over the sampling of volatile environments shapes human learning and decision-making under uncertainty.

Highlights

Making accurate decisions in uncertain environments requires identifying the generative cause of sensory cues, and the expected outcomes of possible actions
Accurate decision-making in uncertain environments requires identifying the generative cause of observed stimuli, and the expected consequences of one’s own actions
By comparing cue-based and outcomebased inference in tightly matched conditions, we show that interacting with uncertain evidence—rather than observing the same evidence—increases the perceived stability of volatile environments

Summary

Introduction

Making accurate decisions in uncertain environments requires identifying the generative cause of sensory cues, and the expected outcomes of possible actions. Multivariate patterns of magnetoencephalographic (MEG) activity reflected this behavioral difference in the neural interaction between inferred beliefs and incoming evidence, an effect originating from associative regions in the temporal lobe Together, these findings indicate that the degree of control over the sampling of volatile environments shapes human learning and decision-making under uncertainty. Canonical “sequential-sampling” models of sensory evidence accumulation are cast in terms of a continuous random walk process spanning hundreds of milliseconds[3,4], whereas “reinforcement learning” models of action valuation rely on discrete updates of expected outcomes over much longer timescales[5,6] Another challenge for a direct comparison between the two types of inference comes from the large differences in the experimental paradigms developed to study perceptual (cue-based) decisions and reward-guided (outcome-based) decisions. We obtained converging behavioral and neural evidence that interacting with uncertain information stabilizes hidden-state inference, as if humans perceive volatile environments as more stable when interacting with uncertain outcomes than when observing uncertain cues

Methods

Results

Conclusion