Linking electric field pattern to clinical efficacy of transcranial electrical stimulation in substance use disorder:...

Transcranial electrical stimulation (tES) studies in substance use disorders (SUDs) have shown promise in reducing drug craving and consumption. Effect sizes reported in previous studies provide a quantitative measure of the efficacy of this intervention. However, the specific brain regions where electric field strength is most strongly associated with therapeutic effect sizes remain unclear. To address this gap, we quantified the efficacy of tES in reducing craving and consumption using Hedges' g and employed a combined meta-analysis and electric field modeling (meta-modeling) framework to identify brain regions whose modeled E-field exposure was most strongly associated with these outcomes. Studies published up to April 2026 were identified through a systematic PubMed search. Trials reporting craving or consumption outcomes with measurable effect sizes (Hedges' g) were included. A total of 71 randomized clinical trials with 3840 total participants that received at least one session of active or sham tES were analyzed. We applied computational head modeling to simulate tES-induced E-fields across 360 anatomically realistic head models and 31 electrode montages. This produced 11,160 E-field maps across six SUD categories, which served as the basis for constructing brain-wide E-field-behavior correlation maps. The primary outcomes were correlations between cortical E-field strength and changes in drug craving and consumption, quantified as Hedges' g. tES significantly reduced craving (g = 0.34, 95% CI [0.23, 0.45], p < 0.001) and consumption (g = 0.22, 95% CI [0.09, 0.36], p < 0.001) compared to sham stimulation. Correlation meta-modeling identified the left dorsolateral prefrontal cortex (DLPFC) as the primary region associated with reductions in craving and consumption, with E-field strength in this region explaining 29% of the variance in treatment effects. In subgroup analyses, the dorsomedial prefrontal cortex accounted for an additional 7.8% of the variance. Decomposition of the normal E-field component further revealed that inward-directed current flow in these regions was positively associated with clinical improvement, whereas outward-directed current showed a negative association. These findings highlight the critical role of the left DLPFC and bilateral dorsomedial PFC, as well as E-field directionality, in modulating craving and consumption in SUDs. Integrating computational modeling into tES protocols may enable more personalized and effective neuromodulation strategies for addiction treatment.