Altered EEG markers of reward learning during abstinence in alcohol dependence: A probabilistic reversal learning study.
Source: PubMed, NCBI / U.S. National Library of Medicine
This study investigated neurophysiological and behavioural adaptations in reward learning and decision making which may contribute to the development and persistence of alcohol use disorder. 20 abstinent alcohol dependent participants (mean abstinence: 20 months, range 1-76) and 26 healthy controls completed an electroencephalography (EEG) probabilistic reversal learning paradigm. Reinforcement learning modelling, event-related potentials (ERPs), and unsupervised machine learning (tensor decomposition) were used to characterise spatiotemporal patterns of reward valuation. Behavioural performance and learning were comparable between groups. In alcohol dependent compared to healthy control participants, feedback related negativity was reduced for positive and negative outcomes. No group differences were observed in the feedback P3; however, substantial variability was present within the alcohol dependent group, with longer abstinence associated with decreased P3. Tensor decomposition revealed early centro-frontal hyperactivity linked to alcohol dependence and associated with early abstinence. Findings suggest altered neural processing of reward learning in alcohol dependence, with indications of neurophysiological adaptation over prolonged abstinence. Data-driven tensor decomposition identified clinically meaningful EEG markers of reward valuation. We provide mechanistic insights into neural adaptations associated with abstinence and present proof-of-concept EEG-based mark
Abstract
This study investigated neurophysiological and behavioural adaptations in reward learning and decision making which may contribute to the development and persistence of alcohol use disorder. 20 abstinent alcohol dependent participants (mean abstinence: 20 months, range 1-76) and 26 healthy controls completed an electroencephalography (EEG) probabilistic reversal learning paradigm. Reinforcement learning modelling, event-related potentials (ERPs), and unsupervised machine learning (tensor decomposition) were used to characterise spatiotemporal patterns of reward valuation. Behavioural performance and learning were comparable between groups. In alcohol dependent compared to healthy control participants, feedback related negativity was reduced for positive and negative outcomes. No group differences were observed in the feedback P3; however, substantial variability was present within the alcohol dependent group, with longer abstinence associated with decreased P3. Tensor decomposition revealed early centro-frontal hyperactivity linked to alcohol dependence and associated with early abstinence. Findings suggest altered neural processing of reward learning in alcohol dependence, with indications of neurophysiological adaptation over prolonged abstinence. Data-driven tensor decomposition identified clinically meaningful EEG markers of reward valuation. We provide mechanistic insights into neural adaptations associated with abstinence and present proof-of-concept EEG-based markers of alcohol dependence that merit further longitudinal evaluation.