How incarceration influences learning from reward and punishment: Longitudinal evidence from computational modelling and EEG.
Imprisonment intends to modify future behaviour partly through punishment. However, little is known about how incarceration alters reinforcement-learning and decision-making. This preregistered longitudinal study investigated behavioural and neural changes in reward and punishment processing during incarceration and release. Newly-incarcerated prisoners (Incoming; n = 61), newly-released individuals (Outgoing; n = 25), and community controls (n = 42) completed a probabilistic reinforcement-learning task at baseline and four months later. Hierarchical Bayesian computational modelling estimated reward and punishment sensitivities, learning rates, and lapse responding. EEG measured feedback-related negativity (FRN), feedback-locked frontal midline theta (FMθ), and delta activity during outcome evaluation. Compared to controls, incoming prisoners exhibited reduced reward sensitivity and learning rate at follow-up, whereas both incoming and outgoing participants showed greater punishment learning. Only controls showed the expected reduction in punishment learning rate across sessions. Behaviourally, controls made more optimal choices than prisoners at baseline, although outgoing participants improved to the same level as controls after release. Neurally, loss–win FRN differentiation strengthened across sessions in all groups. Delta activity was enhanced in both prisoner groups following reward and punishment, whereas FMθ was elevated overall in incoming prisoners. Computational and neurophysiological findings jointly indicate that incarceration is associated with alterations in reinforcement learning, characterized by reduced reward learning, sustained reliance on punishment, and altered neural feedback processing. This shift toward vigilance for negative outcomes may be adaptive within the restrictive prison environment, yet a reliance on punishment information also present after release could shape decision-making during community re-entry and could undermine successful reintegration.
Authors
- Senne Braem (ORCID: https://orcid.org/0000-0002-2619-8225)
- Victoria Rambaud (ORCID: https://orcid.org/0009-0003-0418-7618)
- Louis Favril (ORCID: https://orcid.org/0000-0002-9519-7824)
- Émilie A. Caspar (ORCID: https://orcid.org/0000-0003-1415-2032)
- Ilke Veeckman (ORCID: https://orcid.org/0000-0002-8799-3564)
- Tom Vander Beken
Institutions
- Ghent University (BE)
Publication Details
- Journal
- PsyArXiv (OSF Preprints)
- Published
- 2026-11-14
- DOI
- https://doi.org/10.31234/osf.io/nvf8h
- Primary Topic
- Neural and Behavioral Psychology Studies
- Type
- preprint