Functional Connectivity of the Orbitofrontal Cortex Predicts Cocaine Relapse: Protective and Risk Circuits, Individual Differences, and Neuromodulation Implications.

BACKGROUND: Substance use disorder (SUD) remains a significant public health problem as it is highly refractory, with low treatment commitment and high relapse rates. The human orbitofrontal cortex (OFC), a brain region involved in decision making, valuation, and reward processing, has been implicated in multiple neuropsychiatric diseases and is considered one of the potential therapeutic targets for SUD treatment using noninvasive neuromodulation. However, the OFC is a large, heterogeneous brain area, and the precise location within the OFC that could serve as an efficacious stimulation target is unknown. Moreover, exploring all possible loci within the OFC with clinical trials is operationally not feasible.

METHODS: Using Cox regression-based prediction modeling, we examined resting-state functional connectivity across 20 OFC regions of interest (ROIs) of treated patients with cocaine use disorder. We aimed to identify loci and their functional circuits that predict cocaine relapse, quantified by area under the receiver operating characteristic curve (AUC) with cross-validation in a main cohort and external validation in an independent cohort.

RESULTS: We identified 2 left lateral OFC ROIs whose connectivity predicted relapse with high accuracy (cross-validated AUC of 0.896 and externally validated AUC of 0.780). Each ROI exhibited a pair of protective and risk circuits that contributed distinctively toward relapse.

CONCLUSIONS: The predictive OFC ROIs and their functional circuits identified provide system-level insights into drug relapse, may help reveal specific disease subtypes, and offer potential neuromodulation targets to be tested in future clinical trials.