Preliminary investigation of between-network connectivity and craving during early alcohol abstinence

BackgroundAlcohol use disorder (AUD) is a chronic condition characterized by compulsive drinking and high relapse risk. Craving in early abstinence is a strong predictor of relapse, yet its underlying neurobiological mechanisms remain unclear. Guided by Menons Triple Network Model (TNM) of psychopathology, this study investigates whether altered connectivity between the salience (SN), default mode (DMN), and central executive (CEN) networks --previously implicated in alcohol-related behaviours -- underlies craving during early abstinence.

MethodsA final cohort of 27 individuals with AUD recruited from an inpatient alcohol withdrawal program completed resting-state fMRI scans on day 1 of withdrawal and 18 days later. Additionally, 17 healthy controls underwent fMRI at two sessions spaced two weeks apart. Craving was assessed in the AUD group at both timepoints using the obsessive thoughts subscale of the Obsessive Compulsive Drinking Scale (OCDS). Functional connectivity between brain networks was computed by referencing each individuals between-network connectivity to normative models derived from large-scale reference data to generate scores reflecting their deviations from normative values.

Proposed analysesPlanned analyses will leverage large-scale lifespan normative models to test associations between patient deviation scores in SN-DMN connectivity and craving during acute withdrawal, along with longitudinal associations during abstinence. Exploratory analyses will assess correlations between craving and connectivity of other network pairs of the TNM.

ConclusionsThis report aims to identify functional neurobiological markers of craving during early abstinence in AUD employing normative models. Findings may advance understanding of relapse vulnerability and inform personalized interventions targeting large-scale brain network dysfunctions in AUD. This submission corresponds to Level 3 of the Peer Community In (PCI) Registered Report bias-control taxonomy: data were collected and pre-processed prior to hypothesis formulation, but key variables (subject-level values) have not been observed and no statistical analyses have been performed.