The online metacognitive control of decisions

• Published in: bioRxiv
• Main Authors: Benon, Juliette, Lee, Douglas G, Hopper, William, Morgan Verdeil, Pessiglione, Mathias, Vinckier, Fabien, Bouret, Sebastien, Rouault, Marion, Lebouc, Raphael, Pezzulo, Giovanni, Schreiweis, Christiane, Burguiere, Eric, Daunizeau, Jean
• Format: Paper
• Language: English
• Published: Cold Spring Harbor Cold Spring Harbor Laboratory Press 15.02.2024
• Subjects: Cognition; Computational neuroscience; Decision making; Information processing; Metacognition; Resource allocation
• DOI: 10.1101/2023.01.02.522463

Difficult decisions typically involve mental effort, which scales with the deployment of cognitive (e.g., mnesic, attentional) resources engaged in processing decision-relevant information. But how does the brain regulate mental effort? A possibility is that the brain optimizes a resource allocation problem, whereby the amount of invested resources balances its expected cost (i.e. effort) and benefit. Our working assumption is that subjective decision confidence serves as the benefit term of the resource allocation problem, hence the "metacognitive" nature of decision control. Here, we present a computational model for the online metacognitive control of decisions or oMCD. Formally, oMCD is a Markov Decision Process that optimally solves the ensuing resource allocation problem under agnostic assumptions about the inner workings of the underlying decision system. We demonstrate how this makes oMCD a quasi-optimal control policy for a broad class of decision processes, including -but not limited to- progressive attribute integration. We disclose oMCD's main properties (in terms of choice, confidence and response time), and show that they reproduce most established empirical results in the field of value-based decision making. Finally, we discuss the possible connections between oMCD and most prominent neurocognitive theories about decision control and mental effort regulation.Competing Interest StatementThe authors have declared no competing interest.Footnotes* The revised manuscript accounts for the review comments that were raised by two independent reviewers.