Abstract

Please join us for our fortnightly journal club online via zoom where two presenters will jointly present a topic together. The next topic is ‘Neurocomputational basis of confidence in self and others’ presented by Jakub Onysk and Yul Kang.

Zoom information: https://us02web.zoom.us/j/84958321096?pwd=dFpsYnpJYWVNeHlJbEFKbW1OTzFiQT09 Meeting ID: 849 5832 1096 Passcode: 506576

Summary: Computing confidence in one’s own and others’ decisions is crucial for success in many social situations. We first cover how people perform metacognitive assessment of the accuracy of their decision. We discuss a general Bayesian framework in which self-evaluation is cast as a “second-order” inference on a distinct system, computationally equivalent to inferring the performance of another actor. This model provides a unified account of confidence and error detection, how one’s own actions affect metacognitive judgments, and why people’s metacognition may sometimes be better or worse than task performance.

We then move on to show how humans may estimate confidence in other people. We discuss a recent account by Bang et al. (2022) which suggests that our confidence in others’ choices depends on the combination of our own independent estimate of the task difficulty and how good we think the other agent is at the task. To capture this dynamic, the authors propose a Bayesian Theory of Mind (ToM) computational model that predicts subjects’ responses better than accounts which rely only on either self-projection or ability tracking. Finally, the evidence of this computation is shown to implicate an interaction between regions of perceptual decision-making (human-LIP) and theory of mind (TPJ, dmPFC).

References:

Fleming, S. M. & Daw, N. D. Self-Evaluation of Decision-Making: A General Bayesian Framework for Metacognitive Computation. Psychol Rev 124, 91–114 (2017).

Bang, D., Moran, R., Daw, N. D. & Fleming, S. M. Neurocomputational mechanisms of confidence in self and others. Biorxiv 2021.03.05.434065 (2021) doi:10.1101/2021.03.05.434065.