Hello all,
I am very excited to announce our bioRxiv preprint, “Computational characterization of metacognitive ability in subjective decision-making”! You can find the publication here:
https://www.biorxiv.org/content/10.1101/2025.05.23.655775v1
The work in this paper utilized Stan to fit models of metacognition and, where applicable, subjective utility, to choice and confidence behavior from perceptual and value-based decision-making tasks. The purpose was to evaluate how stable, or trait-like, metacognitive ability is across different tasks. Our contribution was in exploring this idea using a computational modeling pipeline that can approximate and utilize internal decision variables.
Our findings, evaluated by exploring posteriors over correlations in metacognitive ability across tasks through the application of a copula within these models of metacognition, revealed strong relationships in both the perceptual and value-based domains, as well as some evidence of cross-domain relationships.
We provide the code for the models, as well as sample data, on a GitHub repository:
This code was developed with the generous guidance of the Stan community, especially Stan’s developers, and we are delighted to share the fruits of that development with the Stan community. We thank you for all of your help!
If you wish to use the code contained in the publication GitHub, please cite the paper preprint on bioRxiv until its publication in a peer-reviewed journal.