Multimodal Psychophysiological Assessment of Craving in Patients With Alcohol Dependence During Virtual Reality Cue Exposure: Exploratory Single-Arm Clinical Study

Abstract Background Craving is a diagnostic criterion and predictor of relapse in patients with alcohol dependence (AD) and is induced in cue exposure therapy (CET) to prepare patients for real-life risk situations. The benefits of virtual reality (VR) as an innovative tool within treatment for this highly prevalent disorder include increased practicability, standardization, and personalization of CET. Accurate measurement of craving is essential to develop effective virtual reality cue exposure (VR-CE) scenarios. Despite being relevant for diagnostics and therapy, the psychophysiological reaction to alcohol cues and its relationship to subjective craving has not been sufficiently examined. Objective This study aimed to investigate the induction of subjective craving, its physiological correlates, and their relationship in patients with AD during an innovative VR-CE paradigm, including 2 alcohol-associated risk scenarios (bar and living room) and a neutral baseline scenario. Methods Craving was analyzed by measuring physiological reactions (electrodermal activity, including nonspecific skin conductance responses NS-SCR and skin conductance level SCL; heart rate variability HR(V); brightness-corrected pupil diameter BCPD; and respiration rate RR) and subjective craving (visual analog scale) in 61 patients with AD. Linear mixed-effects models were conducted to estimate the effects of VR-CE. Correlations between subjective and physiological craving parameters were analyzed using Spearman correlations. Results Results showed that alcohol-associated VR scenarios had significant effects on subjective craving ( β values between 9.48, 95% CI 6.02‐12.95 and 15.93, 95% CI 12.47‐19.40, with moderate effect sizes between d =0.56 and d =0.72) and on NS-SCR frequency ( β values between 0.97, 95% CI 0.14‐1.80 and 3.06, 95% CI 2.23‐3.89, with small-to-large effect sizes between d =0.31 and d =0.91), BCPD ( β values between 0.03, 95% CI 0.01‐0.06 and 0.05, 95% CI 0.02‐0.07, with small effect sizes between d =0.21 and d =0.29), and RR ( β values between 0.67, 95% CI 0.06‐1.28 and 1.66, 95% CI 1.05‐2.26, with small effect sizes between d =0.22 and d =0.45), but not on SCL and HR(V). Correlation analyses showed significant but weak correlations between subjective craving and electrodermal activity (SCL: r =0.20, P =.04; NS-SCR frequency: r =0.21, P =.03). Conclusions This study showed that subjective and physiological craving (NS-SCR, BCPD, and RR but not SCL and HRV) can be successfully induced by VR-CE in patients with AD. These outcomes add to research on the induction of craving using a wide range of psychophysiological and subjective parameters. Furthermore, this study expands the still-limited research on the relationship between subjective and psychophysiological craving in patients with AD. In the long term, this study informs the development of effective VR-CET, which could, by further building on psychophysiological parameters, lead to biofeedback VR-CET as an innovative treatment option for patients with AD.