Unpacking the bidirectional nexus between suicide risk and NSSI addiction: evidence from in silico intervention and Bayesian network analysis

Background Individuals with non-suicidal self-injury (NSSI), particularly those with addiction-like features, frequently exhibit comorbid depressive symptoms. However, the symptom-level associations between these conditions remain unclear. Methods A total of 387 clinical participants with NSSI were included. Symptom networks were constructed using the binary Ising model and Gaussian Graphical Model (GGM) based on the Ottawa Self-Injury Inventory (OSI) and the 24-item Hamilton Depression Rating Scale (HAMD-24). Bridge centrality, NodeIdentifyR (NIRA), and Bayesian network analyses were performed to identify bridge symptoms, simulate interventions, and explore potential directional relationships. Results Suicide (HAMD3) emerged as the strongest bridge symptom linking depressive symptoms and NSSI addiction across network models. NIRA identified HAMD3 as the most effective intervention target, while Bayesian analysis suggested a potential bidirectional relationship between suicide risk and NSSI addiction. Conclusions Suicide-related symptoms may represent the key intervention target in NSSI addiction, highlighting the importance of integrating suicide risk assessment and intervention into clinical care.