Mapping Individual Differences in the Topological Landscape of Naturalistic Brain Dynamics

Naturalistic stimuli elicit rich subjective experiences through adaptive neural coordination. However, how inherent behavioral traits shape individual neural dynamics in naturalistic settings remains unclear. Here, we introduce a computational framework, STIM, to systematically capture individual differences in brain dynamics while watching diverse movie stimuli. By leveraging Topological Data Analysis, STIM generates a robust group-level dynamical landscape of brain latent states, mapping individual-specific divergence into global topology and local geometry. Applying STIM to large-sample movie fMRI datasets, we found that inter-individual variation in global topology exhibits a center-periphery gradient in the landscape. This gradient significantly explains individual fluid intelligence from a dual perspective, highlighting the importance of both adaptability and diversity of neural dynamics. At the fine-grained narrative level, individual local geometry attributes are associated with context-specific psychological traits beyond cognition. Furthermore, STIM reveals how the dynamical landscape evolves across neurodevelopment and exhibits abnormalities in psychiatric disorders such as autism. In summary, the STIM framework has the potential to transform rich naturalistic stimuli with brain recording into neural ‘probes’ to measure individual differences in cognition and mental health.