Ergodic breakdown of brain dynamics under chronnectome complexity across cognitive states and age

Brain Dynamics Classifier achieves 97.4% accuracy for eight cognitive conditions in the Human Connectome Project young adult cohort and 92.2% in the aging cohort.

• Why it matters: It addresses how chronnectome complexity affects whether information distinguishing rest from task is captured by time-averaged connectivity or by non-ergodic trajectory structure.
• What they did: The method embeds dynamic functional connectivity with a graph neural network and learns trajectory vector fields with a neural ordinary differential equation for multi-class classification across conditions and age groups.
• The result: Ablation shows trajectory geometry contributes more under higher complexity (e.g., Working Memory) and can be absent under lower complexity (e.g., Relational Reasoning), while aging reduces ergodicity in resting dynamics and links higher fluid intelligence to more ergodic rest.

AI summary of the abstract · may contain errors · not medical advice · read the paper.