Central and bridge nodes in depression, anxiety, and resilience networks across adolescent latent profiles: a latent profile and network analysis

Resilience is closely associated with adolescent mental health, but patterns of its co-occurrence with depression and anxiety symptoms remain unclear. This study combined latent profile analysis (LPA) and network analysis to characterize symptom-resilience profiles and examine networks within the assigned profiles. A cross-sectional survey was conducted among 33,821 adolescents aged 10–19 years in Wuhan, China, between October and December 2025. After data cleaning, 33,230 participants were included. LPA identified symptom-resilience profiles based on depressive and anxiety symptoms and resilience dimensions. Participants were assigned to their most likely profile using posterior probabilities, and networks were estimated separately within each assigned profile. Centrality, bridge expected influence, exploratory pairwise network comparisons, and network stability were examined. A three-profile solution was retained, comprising High Resilience-Low Distress (HR-LD), Moderate Resilience-Moderate Distress (MR-MD), and Low Resilience-High Distress (LR-HD). Exploratory network comparisons yielded significant network structure and global strength invariance tests between HR-LD and both other profiles, whereas no individual edge comparison remained significant after Holm-Bonferroni correction, and several strong edges were similar across profiles. Based on within-network expected influence rankings, anxiety-related nodes were most prominent in the HR-LD and MR-MD profiles, whereas Depressed mood was the most central node in the LR-HD profile. Emotional control was the resilience bridge node most strongly connected negatively with the symptom community across all three profiles. On the symptom side, Anhedonia was the strongest negative bridge node in HR-LD, whereas Thoughts of death or self-harm was the strongest in MR-MD and LR-HD. In the present study, a solution comprising three profiles was retained, reflecting a gradient in distress and resilience, with descriptive variation observed in the patterns within each network. The findings provide a descriptive, cross-sectional characterization of depression, anxiety, and resilience across these profiles.