Multiscale EEG network alterations in adolescents with depressive symptoms: static–dynamic connectivity, topological features, and exploratory scalp-level network hubs

Background Adolescence is a critical period for brain network development and a vulnerable stage for the emergence of depressive symptoms Although altered brain connectivity has been reported in depression, the multiscale organization of resting-state EEG networks in adolescents with depressive symptoms (ADS) remains insufficiently understood. In particular, conventional network analyses are often constrained by fixed thresholding strategies and single-level observations, which may obscure scale-specific and topology-related abnormalities. Methods Eyes-closed resting-state EEG data were collected from 96 adolescents with depressive symptoms and 154 healthy controls. A data-driven adaptive thresholding strategy was used to construct subject-specific brain functional networks. Static functional connectivity, dynamic functional connectivity, and higher-order topological features were analyzed across macro-, meso-, and micro-scale levels. A multidimensional abnormality convergence matrix was further constructed to explore the spatial correspondence between network abnormalities and exploratory scalp-level neuromodulation hubs. Results ADS showed scale-dependent abnormalities in brain network organization. At the macro- and meso-scale levels, abnormalities were mainly characterized by theta-band static hyperconnectivity, particularly involving central-region networks, together with reduced dynamic transition flexibility involving prefrontal systems. Micro-scale analysis further localized these alterations to specific connections and nodes. Static connectivity abnormalities were mainly observed along the midline prefrontal–central axis, including Fp1/Fz–FCz connections, whereas dynamic instability was mainly found in right prefrontal–occipital pathways, including Fp2/F4–Oz connections. Higher-order topological analysis indicated vulnerability of right prefrontal nodes, particularly Fp2 and F4. The convergence matrix showed spatial overlap between abnormal network features and the approximate scalp projections of prefrontal neuromodulation sites, while also identifying FCz and Oz as additional nodes that may warrant further investigation. Conclusion This study provides a multiscale EEG network characterization of adolescents with depressive symptoms. The findings suggest that ADS-related network alterations may involve coexisting compensatory connectivity changes, reduced dynamic flexibility, and focal topological vulnerability. These results may help refine network-level features and provide hypothesis-generating computational references for future studies exploring individualized neuromodulation strategies in adolescents with depressive symptoms.