Network analysis for the symptom of depression with Children's Depression Inventory in a large sample of school-aged children

Abstract Background: Depressive disorders have various symptom presentations, which may have complex and dynamic interactions. This study aimed to investigate the network structures underlying the symptoms and to identify the central symptoms of depression in school-aged children. Methods: Participants were a large community sample of elementary school children aged 6 to 12 years (N = 10,233). To assess the depressive symptoms, we utilized the Children's Depression Inventory (CDI). We binarized the scores on the CDI to generate a symptom network using the eLasso method, based on the Ising model. We evaluated the centralities in individual symptoms using the network centrality indices and the associations between symptoms. Results: Of the symptoms, loneliness, self-hatred, school dislike, and low self-esteem were the most central symptoms in the network of depressive symptoms. School work difficulty–school performance decrement, sadness–crying, self-hatred–negative body image, low self-esteem–fight, anhedonia–school dislike, sadness–loneliness, self-deprecation–school work difficulty, and school dislike–lack of friendship had significantly higher edge weight than most edges. The estimated network between the symptoms was robust to stability and accuracy tests. Limitations: Participants were not clinical but community samples, who show lower level of symptoms. Conclusion: The present symptom network analysis provided important insights on various interconnectivities between individual symptoms in childhood depression and on the central symptoms. In addition, our findings presented both similarities and differences with a previous Western study, thus, warranting future cross-cultural studies.