AUTHOR=Le Chengjie , Hou Tianyu , Zhu Yifan , Zhang Yaoyin , Ling Jianjie , Li Zixian , Sun Benteng , Ma Ting , Ye Chenfei 

TITLE=Dynamic brain network reconfiguration following rTMS in males with cocaine use disorder

JOURNAL=Frontiers in Human Neuroscience

VOLUME=Volume 19 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2025.1603888

DOI=10.3389/fnhum.2025.1603888

ISSN=1662-5161

ABSTRACT=Cocaine use disorder (CUD) is characterized by cortico-striatal circuit dysregulation and high relapse rates, with repetitive transcranial magnetic stimulation (rTMS) emerging as a potential neuromodulatory intervention. This study investigates rTMS-induced dynamic brain network reconfigurations in 30 CUD patients using longitudinal resting-state fMRI from the SUDMEX-TMS cohort. Applying Leading Eigenvector Dynamics Analysis (LEiDA) to phase-locking states, we identified four metastable network configurations mapped to canonical resting-state networks. Post-rTMS analyses revealed selective modulation of visual network (VIS)-dominant states, showing increased duration and occupancy, alongside reduced self-transition probabilities in frontoparietal control network (FPCN) states after rTMS therapy. Temporal dynamics of these states correlated with subjective craving intensity: increased duration of the VIS-dominant state was associated with lower craving severity (CCQ-N) post-treatment. These findings suggest that increased VIS metastability strengthened bottom-up sensory gating that attenuates drug-cue salience through perceptual desensitization. Although FPCN-state self-transition decreased significantly following stimulation, it was not directly linked to craving improvement, indicating a potentially supportive but nonspecific role in perceptual recalibration. Together, these dynamic markers highlight the relevance of network-level flexibility in mediating rTMS treatment efficacy for cocaine addiction. By establishing dynamic network state reconfiguration as a mechanism linking rTMS to symptom evolution, this work provides a framework for optimizing neuromodulation protocols and developing neurodynamics-dependent biomarkers in addiction therapeutics.