Dynamic Network Analysis of Cognitive Attacks Using Meta-Network Modeling and Community Detection Algorithms

Cognitive attacks on Social media such as propaganda activities are a sophisticated form of psychological manipulation that targets the mental processes of individuals or groups to influence their perceptions, decisions, and behaviors. This paper utilizes dynamic network analysis to study propaganda and non-propaganda social media content in the Middle East. We transformed textual features of propaganda and non-propaganda news headlines and the corresponding comments on social media into meta-networks and applied various centrality measures and community detection algorithms to them. We then compared the structural characteristics of each network. Our findings reveal significant differences in the centrality and community structures. Our network analysis of propaganda and non-propaganda content has revealed distinct structural differences, with propaganda networks exhibiting lower topic centrality and more fragmented topic structures. This indicates a strategic focus on targeted messaging rather than broad-based engagement. The higher centrality values in non-propaganda networks reveal a more interconnected and diverse information flow. The findings of our research are significant for both practitioners and policymakers. Our results inform the development of more effective counter-propaganda strategies. Efforts can be made to disrupt echo chambers and introduce alternative narratives that challenge propaganda content.