Prediction error-related activity predicts orbitofrontal reorganization linked to behavioral policy adjustment

Cognitive flexibility requires rapid updating of internal representations when action–outcome contingencies change. Yet how feedback on a single trial reshapes the distributed activity patterns that implement these internal representations remains poorly understood. Using functional magnetic resonance imaging (fMRI) during probabilistic reversal learning, we show that a Q-learning model with separate learning rates outperforms a standard model and reveals stronger learning from positive than negative reward prediction errors (RPE).

Reversal increases prediction error magnitude and induces rapid multivoxel reorganization in medial temporal and adjacent subcortical regions. During early reversal, RPE-related activity in right striatal and medial orbitofrontal regions predicts subsequent changes in right lateral orbitofrontal cortex (lOFC) representations; these effects occur only after nonrewarded outcomes. Moreover, greater representational change in lOFC is associated with weakening of the action policy, and this pattern change in right lOFC is related to reward-stay behavior.

Together, these findings link lOFC updating to behavioral policy adjustment during reversal learning. This work was supported by the following funding sources: Research Center for Brain Cognition and Human Development, Guangdong, China (Project number 2024B0303390003) to B.A.W.; Guangdong Basic and Applied Basic Research Foundation to B.A.W. (Project number 2025A1515010766); Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) to B.P. (Project number 122679504 and PL602/6-1), and the FoRUM grant F971-2019 and F1081N-2023, medical faculty of the Ruhr-University Bochum to B.P. These authors contributed equally: Burkhard Pleger, Bin A.

School of Psychology, Center for Studies of Psychological Application, Guangdong Key Laboratory of Mental Health and Cognitive Science, Ministry of Education Key Laboratory of Brain Cognition and Educational Science, South China Normal University, Guangzhou, China Chaofeng Qin, Yuanting Zhao, Wenli Fu, Xiaomin Chen & Bin A. Wang Department of Neurology, BG University Hospital Bergmannsheil, Ruhr-University Bochum, Bochum, Germany The authors declare no competing interests. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material.

If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/. Qin, C., Zhao, Y., Fu, W. et al.