The metacognitive control of decisions predicts whether and how mice override their default policy

Decisions permeate every aspect of our lives (what to eat, where to live, etc) but decision-making policies seem to vary tremendously across time and across individuals. Rather than processing all decision-relevant information, we often rely on fast habitual and/or intuitive decision policies, which can lead to irrational biases and errors. Yet, we dont always follow the fast and negligent lead of habits or intuitions. So what determines whether we engage cognitive control and override our default responses? A possibility is that engagement of cognitive control optimizes a cost-benefit trade-off. In this work, we rely on Markov Decision Processes to derive the optimal control allocation policy in standard decision-making tasks, under arbitrary default preferences. Our working hypothesis is that decision confidence serves as the benefit term of this allocation problem, hence the "metacognitive" nature of decision control. Importantly, we provide behavioural evidence that the ensuing model accurately predicts whether and how mice override their default policy in the context of a repeated perceptual decision-making task. This opens an alleyway for assessing the brain circuits that operate the arbitration between default and controlled decision processes, in rodent models of both healthy and neuropsychiatric conditions.