Meta-Reinforcement Learning reconciles surprise, value, and control in the anterior cingulate cortex.

The role of the dorsal anterior cingulate cortex (dACC) in cognition is a frequently studied yet highly debated topic in neuroscience. Most authors agree that the dACC is involved in either cognitive control (e.g., voluntary inhibition of automatic responses) or monitoring (e.g., comparing expectati...

Full description

Saved in:
Bibliographic Details
Main Authors: Tim Vriens, Eliana Vassena, Giovanni Pezzulo, Gianluca Baldassarre, Massimo Silvetti
Format: Article
Language:English
Published: Public Library of Science (PLoS) 2025-04-01
Series:PLoS Computational Biology
Online Access:https://doi.org/10.1371/journal.pcbi.1013025
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850193571986014208
author Tim Vriens
Eliana Vassena
Giovanni Pezzulo
Gianluca Baldassarre
Massimo Silvetti
author_facet Tim Vriens
Eliana Vassena
Giovanni Pezzulo
Gianluca Baldassarre
Massimo Silvetti
author_sort Tim Vriens
collection DOAJ
description The role of the dorsal anterior cingulate cortex (dACC) in cognition is a frequently studied yet highly debated topic in neuroscience. Most authors agree that the dACC is involved in either cognitive control (e.g., voluntary inhibition of automatic responses) or monitoring (e.g., comparing expectations with outcomes, detecting errors, tracking surprise). A consensus on which theoretical perspective best explains dACC contribution to behaviour is still lacking, as two distinct sets of studies report dACC activation in tasks requiring surprise tracking for performance monitoring and cognitive control without involving surprise monitoring, respectively. This creates a theoretical impasse, as no single current account can reconcile these findings. Here we propose a novel hypothesis on dACC function that integrates both the monitoring and the cognitive control perspectives in a unifying, meta-Reinforcement Learning framework, in which cognitive control is optimized by meta-learning based on tracking Bayesian surprise. We tested the quantitative predictions from our theory in three different functional neuroimaging experiments at the basis of the current theory crisis. We show that the meta-Reinforcement Learning perspective successfully captures all the neuroimaging results by predicting both cognitive control and monitoring functions, proposing a solution to the theory crisis about dACC function within an integrative framework. In sum, our results suggest that dACC function can be framed as a meta-learning optimisation of cognitive control, providing an integrative perspective on its roles in cognitive control, surprise tracking, and performance monitoring.
format Article
id doaj-art-e47fcf2520b7446aa9624cacd4e82d65
institution OA Journals
issn 1553-734X
1553-7358
language English
publishDate 2025-04-01
publisher Public Library of Science (PLoS)
record_format Article
series PLoS Computational Biology
spelling doaj-art-e47fcf2520b7446aa9624cacd4e82d652025-08-20T02:14:15ZengPublic Library of Science (PLoS)PLoS Computational Biology1553-734X1553-73582025-04-01214e101302510.1371/journal.pcbi.1013025Meta-Reinforcement Learning reconciles surprise, value, and control in the anterior cingulate cortex.Tim VriensEliana VassenaGiovanni PezzuloGianluca BaldassarreMassimo SilvettiThe role of the dorsal anterior cingulate cortex (dACC) in cognition is a frequently studied yet highly debated topic in neuroscience. Most authors agree that the dACC is involved in either cognitive control (e.g., voluntary inhibition of automatic responses) or monitoring (e.g., comparing expectations with outcomes, detecting errors, tracking surprise). A consensus on which theoretical perspective best explains dACC contribution to behaviour is still lacking, as two distinct sets of studies report dACC activation in tasks requiring surprise tracking for performance monitoring and cognitive control without involving surprise monitoring, respectively. This creates a theoretical impasse, as no single current account can reconcile these findings. Here we propose a novel hypothesis on dACC function that integrates both the monitoring and the cognitive control perspectives in a unifying, meta-Reinforcement Learning framework, in which cognitive control is optimized by meta-learning based on tracking Bayesian surprise. We tested the quantitative predictions from our theory in three different functional neuroimaging experiments at the basis of the current theory crisis. We show that the meta-Reinforcement Learning perspective successfully captures all the neuroimaging results by predicting both cognitive control and monitoring functions, proposing a solution to the theory crisis about dACC function within an integrative framework. In sum, our results suggest that dACC function can be framed as a meta-learning optimisation of cognitive control, providing an integrative perspective on its roles in cognitive control, surprise tracking, and performance monitoring.https://doi.org/10.1371/journal.pcbi.1013025
spellingShingle Tim Vriens
Eliana Vassena
Giovanni Pezzulo
Gianluca Baldassarre
Massimo Silvetti
Meta-Reinforcement Learning reconciles surprise, value, and control in the anterior cingulate cortex.
PLoS Computational Biology
title Meta-Reinforcement Learning reconciles surprise, value, and control in the anterior cingulate cortex.
title_full Meta-Reinforcement Learning reconciles surprise, value, and control in the anterior cingulate cortex.
title_fullStr Meta-Reinforcement Learning reconciles surprise, value, and control in the anterior cingulate cortex.
title_full_unstemmed Meta-Reinforcement Learning reconciles surprise, value, and control in the anterior cingulate cortex.
title_short Meta-Reinforcement Learning reconciles surprise, value, and control in the anterior cingulate cortex.
title_sort meta reinforcement learning reconciles surprise value and control in the anterior cingulate cortex
url https://doi.org/10.1371/journal.pcbi.1013025
work_keys_str_mv AT timvriens metareinforcementlearningreconcilessurprisevalueandcontrolintheanteriorcingulatecortex
AT elianavassena metareinforcementlearningreconcilessurprisevalueandcontrolintheanteriorcingulatecortex
AT giovannipezzulo metareinforcementlearningreconcilessurprisevalueandcontrolintheanteriorcingulatecortex
AT gianlucabaldassarre metareinforcementlearningreconcilessurprisevalueandcontrolintheanteriorcingulatecortex
AT massimosilvetti metareinforcementlearningreconcilessurprisevalueandcontrolintheanteriorcingulatecortex