Using Passive Sensing to Isolate a Biosignature for Craving Among Individuals in Early Alcohol Use Disorder Recovery

Craving is one of the most robust proximal predictors of both treatment dropout and relapse during early recovery for alcohol use disorder (AUD). Craving can onset rapidly, and the ability to accurately detect or modulate cravings varies significantly within- and between-persons. Because of this, craving management is a central feature of most current AUD treatment models. However, impaired interoceptive awareness, which is common in early AUD recovery, can stymie individuals' capacity to implement craving management strategies. Measurement and detection of craving through passive biosensor monitoring could support novel just-in-time adaptive interventions that can mitigate this vulnerability. This study sought to characterize the biosignatures of craving through the measurement of heart rate variability (HRV) using a combination of ambulatory electrocardiogram (ECG) monitoring and ecological momentary assessment (EMA) in individuals in the first year of a current AUD recovery attempt (N = 40, observations = 400). A multilevel negative binomial regression analysis was conducted to estimate correlations between craving and HRV. Results demonstrated that, at the within-person level, high-frequency HRV (HF-HRV; IRR = 1.01, p = 0.013) and SD1 (IRR = 0.97, p = 0.020) were associated with craving at the same moment. Findings highlight HRV as a viable target for passive monitoring in interventions aimed at supporting craving mitigation in early AUD recovery, particularly at the momentary level.