Rapid acquisition of near optimal stopping using Bayesian by-products

Abstract

Prior to any decision, animals (including humans) must stop deliberating. Both accumulating evidence for too little or too long can be costly. In contrast to accounts of decision making, accounts of stopping do not typically claim that animals use Bayesian posteriors. Considering a generic perceptual decision making task we show that, under approximation, only two variables are relevant to the question of when to stop evidence accumulation; time and the maximum posterior probability. We explored the rate at which stopping rules are learned using deep neural networks as model learners. A network which only used time and the maximum posterior probability learned faster than any other network considered. Therefore, such an approach may be highly adaptive, and animals may be able to reuse the same neural machinery they use for decisions for stopping. These results suggest that Bayesian inference may be even more important for animals than previously thought.