Authors :
Qiangqiang Liu, MD – Ruijin Hospital, Shanghai Jiao Tong University School of Medicine
Junquan zhang, PhD – University of Electronic Science and Technology of China
Honghao Han, PhD – University of Electronic Science and Technology of China
Presenting Author: Jiwen Xu, PhD – Shanghai Ruijin Hospital
Rong Li, PhD – University of Electronic Science and Technology of China
Rationale:
Deep brain stimulation (DBS) targeting the anterior thalamic nucleus (ANT) and centromedian nucleus (CM) is a promising therapy for drug-resistant epilepsy (DRE), yet their distinct neuromodulatory mechanisms remain incompletely understood. This study employs preoperative, DBS-OFF to DBS-ON transitions resting-state functional magnetic resonance imaging (rs-fMRI) to elucidate how ANT and CM DBS differentially modulate neural networks, with a focus on shared and target-specific effects on the default mode network (DMN) and sensorimotor systems.
Methods:
A retrospective observational study at Ruijin Hospital included 21 patients with drug-resistant epilepsy: 7 with bilateral temporal lobe epilepsy (BTLE) receiving ANT-DBS and 14 with generalized epilepsy receiving CM-DBS. All patients underwent rs-fMRI scans preoperatively and at 6 months postoperatively in both DBS-OFF and DBS-ON states, with analyses including amplitude of low-frequency fluctuation (ALFF), functional connectivity (FC), and degree centrality (DC) to assess whole-brain changes. Statistical comparisons used paired t-tests, with demographic and clinical data collected to contextualize findings.
Results:
Baseline ALFF abnormalities revealed target-specific pathologies, with BTLE patients showing hippocampal hyperactivity and generalized patients exhibiting cerebellar-insular hyperactivity. The pre/postcentral gyrus hypoactivity in both groups reflects a common sensorimotor integration deficit in DRE, suggesting these regions act as "bottlenecks" in seizure generalization. Both DBS targets demonstrated convergent modulation of the DMN-cerebellar-temporal axis, enhancing activity in the cingulate cortex, precuneus, and cerebellum while suppressing temporal lobe connectivity. However, distinct mechanisms emerged: ANT-DBS primarily reduced FC in limbic-cerebellar pathways and increased DC in parietal-cerebellar hubs, indicating inhibitory disconnection of hyperexcitable networks. In contrast, CM-DBS enhanced FC in frontoparietal regions and reduced DC in frontal-cingulate areas, reflecting excitatory recruitment of executive control networks. Longitudinal post-DBS ALFF increases in DMN regions (e.g., precuneus) suggested sustained neuroplastic adaptations.
Conclusions:
ANT and CM DBS achieve seizure reduction through divergent pathways—ANT via limbic network inhibition and CM via frontoparietal network enhancement—while converging on DMN stabilization and cerebellar recruitment. These findings highlight DMN engagement as a unified biomarker for DBS efficacy and support personalized target selection based on individual epileptogenic network profiles. Future studies should integrate ALFF, FC, and DC metrics to optimize DBS programming and validate mechanisms in vivo.
Funding:
Shanghai Municipal Health Commission (No. 202340291), Guangci Innovation Plan of Shanghai Ruijin Hospital (Grant No. GCQH2023073) and National Natural Science Foundation of China (Nos. 82372085 and 62333003).