Electroconvulsive therapy (ECT) produces widespread effects across the brain and remains the most effective intervention for severe major depressive disorder (MDD). Despite its clinical efficacy, how ECT reshapes the global organization of functional brain networks is not fully understood. Traditional node-based connectomic analyses may miss large-scale reconfiguration patterns. The authors applied an edge-centric connectomics approach, using normalized entropy as a topological measure, to characterize reorganization of functional connectomes at multiple scales before and after ECT.
The study analyzed longitudinal resting-state fMRI from two cohorts. The primary cohort included 80 patients with MDD and 75 healthy controls; an independent validation cohort comprised 30 MDD patients. Edge-centric analyses were used to derive normalized entropy measures that reflect the distribution and diversity of edge-level interactions across the functional connectome. Changes in these topological metrics were assessed at baseline and following ECT. The study additionally related topological shifts to clinical symptom changes and performed multiscale spatial correlations with neurotransmitter receptor maps and transcriptomic profiles. Finally, baseline edge-centric features were tested in a machine learning framework to predict ECT treatment response.
At baseline, before ECT, patients with MDD exhibited a heterogeneous pattern of edge-centric topology compared with healthy controls. Specifically, the study reports increased normalized entropy in subcortical networks and decreased entropy in the dorsal attention and sensorimotor networks (SMN). These baseline differences indicate that MDD is associated with both elevated and reduced edge-level diversity across distinct functional subsystems, suggesting complex large-scale reorganization in the disorder.
Following ECT, the principal longitudinal change identified by the authors was a further reduction in SMN entropy. This selective reorganization of the sensorimotor network was observed in the primary cohort and was replicated in the independent validation cohort of 30 MDD patients. The authors emphasize that ECT was linked to selective SMN rearrangement rather than a global normalization of the baseline connectomic abnormalities identified in MDD.
SMN reorganization after ECT was significantly associated with improvements in specific depressive symptoms. The abstract reports this clinical association but does not specify which symptoms or the magnitude of these correlations in detail. The association suggests that topological changes within the sensorimotor network may relate to symptomatic benefit following ECT, supporting functional relevance for the observed connectomic shifts.
The study performed multiscale spatial decoding to place the observed topological shifts in a broader biological context. The authors report that the spatial pattern of SMN reorganization aligned with broad distributions of monoaminergic receptors. In addition, transcriptomic analyses implicated gene expression signatures related to neuroplasticity and to particular cell types. These multiscale correspondences support a link between ECT-induced network reconfiguration and underlying molecular and cellular substrates, although details of the receptor maps and specific gene sets or cell types are not described in the abstract.
Baseline edge-centric features were entered into a machine learning framework to predict ECT treatment response. According to the abstract, these edge-centric predictors outperformed conventional fMRI metrics and demonstrated partial cross-site generalizability. The reporting indicates improved predictive utility for edge-level connectomic measures relative to standard node-based features, but the abstract does not provide model performance statistics, feature importance details, or exact cross-site accuracy values.
The authors conclude that ECT is associated with selective reorganization of the sensorimotor network rather than a straightforward normalization of baseline abnormalities in MDD. They propose that edge-centric connectomics, combined with multiscale biological annotation (receptor maps and transcriptomics), offers a promising framework for elucidating therapeutic mechanisms and for developing predictive biomarkers in MDD. The work is reported as a preprint and has not been peer reviewed; the clinical trial registration is ChiCTR2400093114. Ethical approval was obtained from Shandong Daizhuang Hospital. The authors state that all data produced in the study are available upon reasonable request. The abstract does not include detailed numerical results, statistical values, exact symptom correlations, or machine learning performance metrics; those details are not reported in the source abstract and would require consulting the full manuscript or supplementary materials for confirmation.