Brain states recur across diverse narrative contexts during longitudinal viewing

Area of Science:

• Neuroscience
• Cognitive Science
• Dynamic Systems

Background:

• The brain's dynamic activity patterns during naturalistic experiences like storytelling are not fully understood.
• It remains unclear whether the brain's repertoire of recurring states is fixed or adaptable to new content.

Purpose of the Study:

• To investigate the stability and content-dependency of the dynamic brain-state repertoire during prolonged narrative engagement.
• To determine if an individual's brain-state repertoire is a stable property or is reshaped by diverse narrative experiences.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to record brain activity in six individuals watching the TV series Friends (approx. 54 hours per person).
• A sticky hierarchical Dirichlet process hidden Markov model was applied to discover recurring whole-brain activity patterns (brain states) without pre-specifying their number.
• Functional connectivity similarity and resting-state network boundaries were analyzed to understand state transitions.

Main Results:

• Each individual's brain visited approximately 45 dynamic states, forming a continuous recurrence gradient from highly frequent to episode-specific states.
• Brain state recurrence was heterogeneous, with most states remaining content-eligible rather than being tied to the scanning structure.
• The discovered brain-state repertoire and its recurrence ordering during Friends transferred to other social-narrative films, but attenuated for non-narrative stimuli.

Conclusions:

• The dynamic brain-state repertoire is a stable, individual-specific property, not fundamentally altered by new experiences.
• Narrative content modulates the moment-to-moment occupancy of brain states but does not change the underlying repertoire.
• This suggests a consistent neural framework for processing diverse narrative content across individuals.