Abstract
<title>Abstract</title> <p>Background Adolescence is a developmental stage in which depressive and anxiety manifestations often emerge and intensify quickly. Although girls are reported to experience higher risk for mood-related problems, it is not yet clear whether observed sex differences arise solely from differences in symptom levels or also from sex-specific patterns of symptom linkage within mood disorders. Methods We performed a retrospective, cross-sectional analysis of 1410 adolescents aged 13 to 18 years with at least one clinical diagnosis of a mood disorder. Participants completed the Symptom Checklist-90, from which twenty-eight clinically meaningful symptom indicators were chosen as network nodes. We estimated sex-specific symptom networks using regularized partial correlations. Sex-related discrepancies were tested with the Network Comparison Test. To characterize symptom importance, we calculated strength, expected influence, and bridge expected influence. Estimation reliability and stability were examined using nonparametric bootstrapping, including edge-weight confidence intervals and case-dropping assessment. We further conducted sensitivity analyses through subsampling to reduce the impact of unequal group sizes. Given significant age differences between sex groups, age was additionally included as a covariate in the Network Comparison Test to account for potential developmental confounding. Results After multiple-comparison correction, the Network Comparison Test did not indicate clear sex differences in overall network configuration or global strength. Consistent with this, Network Comparison Test showed no evidence of sex differences in global strength invariance or network structure invariance (Supplementary Table 3). In addition, edge-level discrepancies from the Top10 by absolute difference were not significant after FDR correction (Supplementary Table 4). Therefore, any observed sex-specific ordering of core or bridge symptoms should be interpreted as exploratory and based on point estimates rather than FDR-significant edge differences. Across both sexes, SCL59 (thoughts of death) showed the largest bridge expected influence in the estimated networks, suggesting a potentially important interconnectivity-related cognitive linkage within the symptom system. Conclusions In this clinical adolescent sample, symptom network structures were broadly comparable between sexes, but important core and bridging features showed exploratory sex-related patterns. These results may support clinicians in combining shared target symptoms with sex-informed assessment priorities when evaluating risk and guiding intervention planning.</p>