A trust-aware computational framework for mental health risk amplification and propagation in online social networks

Computational social systems are facing big challenges because of the rapid dissemination of inaccurate information, evidenced by the growing number of negative impacts on individual and collective mental health. Although extensive research has investigated the dissemination of false information and the assessment of the mental health of individuals separately, limited attention has been given to investigating the role of trust and psychological vulnerability on the likelihood that individuals suffer from psychological harm as a result of engaging with false information in an online community. This lack of research can be addressed by developing computational models to understand and minimize the psychological harm caused by false information across communities. To address this challenge, a model for mental-health risk propagation due to trust-awareness is needed for examining the spread of mental health risk caused by misinformation in online social networks. The framework provides a Mental Health Risk Amplification Index (MHRAI), which quantifies individual risk using three categories of variables: exposure to misinformation, the influence that trust has on an individual’s risk level, and how vulnerable a person is to experiencing mental health problems. These risk estimates are combined into a trust-weighted propagation process to observe how mental health risk propagates through the network over time. In addition, this work considers an intervention mechanism at the community level to assess the possible effects of targeting risk reduction within high-risk communities. The framework was evaluated on five representative synthetic network topologies together with platform-inspired Twitter-like, Facebook-like, and Reddit-like network models to assess its robustness across diverse social network structures. The results demonstrate that trust acts as a significant amplification mechanism in misinformation-induced mental health risk propagation and that the proposed community-based intervention effectively reduces community-level risk across different network topologies. The proposed framework provides mechanistic insights into the formation, amplification, and mitigation of misinformation-induced mental health risk through simulation-based computational modeling. The reported findings are derived from controlled simulation experiments and should not be interpreted as validated predictions of real-world mental health outcomes. Future work will focus on validating the framework using longitudinal social-media datasets and dynamic trust-aware network models.