Abstract

Fast and accurate decisions are fundamental for adaptive behaviour. Theories of decision-making posit that evidence in favour of different choices is gradually accumulated until a critical value is reached. It remains unclear, however, which aspects of the neural code get updated during evidence accumulation. Here we investigated whether evidence accumulation relies on a gradual increase in the precision of neural representations of sensory input (i.e., a decrease in neural uncertainty). Healthy human volunteers (N = 32, 24 females) discriminated global motion direction over a patch of moving dots, and their brain activity was recorded using electroencephalography. Time-resolved neural uncertainty was estimated using multivariate feature-specific analyses of brain activity. Behavioural measures were modelled using iterative Bayesian inference either on its own (i.e., the full model), or by swapping free model parameters with neural uncertainty estimates derived from brain recordings. The neurally-restricted model was further refitted using randomly shuffled neural uncertainty. The full model and the unshuffled neural model yielded very good and comparable fits to the data, whereas the shuffled neural model yielded worse fits. Taken together, the findings reveal that the brain relies on reducing neural uncertainty during decision-making. They also provide neurobiological support for Bayesian inference as a fundamental computational mechanism in support of decision making.

Significance Statement The computational and neural mechanisms of decision making have attracted much attention across many fields, including philosophy, economics, psychology and biology. Theories of decision-making posit that evidence in favour of different choices is gradually accumulated until a critical value is reached. This study provides empirical support that evidence accumulation relies on gradual decrease of uncertainty of neural sensory representations. The findings support and extend the idea that the brain implements iterative Bayesian inference to navigate the environment in an adaptive manner which offers a general and unifying mathematical framework for cognition.