Anxiety is associated with idiosyncratic neural event boundaries during naturalistic perception

Anxiety is associated with altered patterns of attention and memory, but its relationship to how individuals segment and structure continuous experiences remains underexplored. This study investigates the relationship between anxiety and neural event segmentation during naturalistic perception using computational modeling of fMRI data. Leveraging a large open-access dataset and methods including Greedy State Boundary Search (GSBS) and Hidden Markov Models (HMM), we compared event segmentation patterns between healthy adults with high versus low anxiety, as measured by the Hospital Anxiety and Depression Scale (HADS-A). We found that individuals with higher anxiety exhibited idiosyncratic event boundaries: anxious individuals had less boundary alignment to both non-anxious participants and other anxious individuals, particularly in frontoparietal control and dorsal attention networks. These results suggest that anxiety is associated with differences in narrative processing, possibly reflecting disrupted top- down integration. This association underscores the potential relationship between individual differences in anxiety symptoms and variability in the neural organization of experience, and highlights directions for future intervention research.