352. Baseline-dependent normalization of salience–control connectivity after right DLPFC rTMS in alcohol use disorder: pilot support for mechanistically motivated enrichment

Abstract

Abstract Background Treatment response to repetitive transcranial magnetic stimulation (rTMS) in alcohol use disorder (AUD) is heterogeneous, and group-level neuroimaging effects are often inconsistent. We hypothesized that apparent inconsistency reflects baseline-dependent salience/attention-control circuit dynamics. Aims & Objectives To test whether changes in ventral attention network (VAN) - frontoparietal network (FPN) functional connectivity (FC) after right dorsolateral prefrontal cortex (DLPFC) rTMS depend on baseline coupling, and to evaluate a mechanistic enrichment signal for future sham-controlled trials. Method In an open-label pilot, six adults with AUD (mean age 46.2 +/- 11.5 years) completed resting-state functional MRI and clinical assessments at Baseline, Post-treatment, and 1-month follow-up (1M). High-frequency rTMS targeted the right DLPFC (F4, international 10-20 system): 10 Hz, 110% resting motor threshold, and 3,000 pulses/session (30,000 total; 10 sessions). VAN-FPN FC was computed using Fisher z-transformed correlations. Clinical outcomes included the Alcohol Use Disorders Identification Test (AUDIT) and craving assessed using a visual analog scale (VAS). Results AUDIT decreased from 31.3 +/- 7.2 at Baseline to 28.0 +/- 9.4 at 1M (Wilcoxon p = 0.042). Group-mean VAN-FPN FC change was not significant; however, individual responses were strongly baseline-dependent. Baseline VAN-FPN FC correlated negatively with change in FC (Post - Baseline) (Spearman rho = -0.94, p = 0.005; n = 6). Participants with baseline hypersynchrony (VAN-FPN > 0; n = 4) showed the largest decoupling, with directionally consistent effects at 1M (rho = -0.83, p = 0.042). Craving VAS showed high inter-individual variability across timepoints and did not consistently track with FC change. Discussion & Conclusions Right DLPFC rTMS was associated with reduced hazardous drinking risk and a baseline-dependent normalization pattern in salience/attention-control coupling, preferentially down-regulating VAN-FPN hypersynchrony. The dissociation between circuit decoupling and variable craving supports FC as a mechanistic biomarker. Findings are hypothesis-generating but motivate mechanistically motivated enrichment by stratifying participants on baseline VAN-FPN coupling in future sham-controlled trials.