Individualized prediction of heroin cue-induced craving using task-based EEG functional connectivity

Area of Science:

• Neuroscience
• Addiction Research
• Biomarker Discovery

Background:

• Cue-induced craving is a primary driver of addiction and relapse.
• Heterogeneity in craving poses a challenge for effective, individualized interventions.
• Objective, quantifiable neurobiological biomarkers for craving are currently lacking.

Purpose of the Study:

• To develop an individualized neurobiological biomarker for cue-induced craving in heroin use disorder (HUD).
• To utilize task-based electroencephalography (EEG) and functional connectivity (FC) for predicting craving severity.
• To assess the prognostic value of the biomarker for intervention outcomes.

Main Methods:

• Employed task-based EEG to capture neural signatures of craving in HUD patients.
• Developed an individualized FC-based prediction model using β-band power envelope connectivity (PEC).
• Validated the model's predictive accuracy and prognostic value, including after neuromodulation (iTBS).

Main Results:

• Identified β-band PEC as a reliable biomarker for estimating individual craving severity.
• The PEC-based model effectively predicted craving levels even after brain stimulation-induced FC changes.
• Baseline β-band PEC strongly predicted craving score improvements following L-DLPFC and PCu iTBS.
• The model demonstrated generalizability across different cue-induced craving tasks.

Conclusions:

• EEG FC features, specifically β-band PEC, can predict individual cue-induced craving levels.
• This individualized predictive model shows promise for assessing intervention outcomes in precision medicine.
• The findings support the advancement of digital biomarker-driven approaches for addiction treatment.