Resilient hubs, shifting links: The brain's network architecture during deceptive behavior

Deception is one of the most enigmatic human behaviors, and understanding its neural mechanisms remains a central challenge in neuroscience. Investigating brain dynamics during deception provides insights into cognitive control and social processing. In this study, we analyzed the architecture of brain region connections, focusing on rich, feeder, and local links, using EEG signals from 22 participants performing a visual task under instructed deception. Functional connectivity matrices were computed with phase lag index, and binary networks were constructed across multiple thresholds. Results showed that brain networks are organized around stable hub regions, which remain largely unchanged during deception. However, deceptive behavior disrupted peripheral connections, with feeder and local links showing reductions relative to truthful responding, suggesting reallocation of cognitive resources. Although hub regions were consistent, hubs in the truthful condition exhibited higher connectivity than during deception in both binary and weighted networks across all percentages of rich nodes (p < 10^-6), indicating more efficient hub-to-hub integration during truthful processing. Analysis of link deviations across thresholds and rich-node percentages revealed a dynamic pattern. At low percentages, rich links were stable while local links varied most. As percentages increased, rich links showed the largest deviations, local links the smallest, and feeder links maintained intermediate variability. These results indicate that deception initially disrupts local processing, while communication among core hubs is affected as network engagement grows. These findings highlight the importance of investigating the architecture of brain region connections during complex social behaviors, showing how stable hubs and flexible peripheral links support cognitive processing.