Social Connectedness Buffering the Interaction Between Problematic Internet Use and Psychopathological Symptoms Among Adolescents: Undirected and Bayesian Network Analyses

Background Problematic internet use (PIU) in adolescents often co-occurs with psychopathological symptoms. Social connectedness (SC) is a potential protective factor for both PIU and psychopathological symptoms, but how SC interacts with the PIU-psychopathology system remains unclear. Objective This study aimed to characterize the factor-level comorbidity network linking PIU with psychopathological symptoms among adolescents at high risk of PIU and to clarify how offline and online SC function in this system. We also applied Bayesian network structure learning to explore algorithm-preferred edge orientations. Methods This multicenter study included 9407 adolescents aged 12 to 18 years from 4 provinces in China. PIU, depression, anxiety, irritability, repetitive thoughts and behaviors, anger, aggression, and SC were assessed. We estimated an undirected factor-level symptom network and a directed acyclic graph derived from a Bayesian network structure to characterize associations and conditional dependency structures among SC, PIU, and 8 psychopathological symptoms. Results The integrative symptom network encompassing PIU, internalizing and externalizing symptoms, and SC showed great stability and accuracy. Anxiety, anger, and depression were identified as central bridging symptoms that reinforce the interplay between PIU and other psychopathological symptoms. Offline SC exhibited the highest number of negative edges and the highest bridge expected influence value, followed by online SC. In the exploratory directed acyclic graph, offline SC showed algorithm-selected outgoing arrows to PIU salience and to several internalizing and externalizing symptoms, including depression, anxiety, verbal aggression, and hostility. Conclusions PIU and psychopathological symptoms formed a closely interconnected factor-level network among adolescents in the high-risk PIU group. Offline SC occupied a prominent negative bridge position, whereas online SC showed a secondary negative network position and was positively connected with offline SC. These findings support further evaluation of offline SC as a candidate prevention and intervention target.