Hierarchical Flows of Human Cortical Activity
Ongoing brain activity unfolds as structured spatiotemporal patterns across the cortex, yet quantifying the direction and strength of this propagation on the folded cortical sheet is challenging within and across individuals. We introduce geodesic cortical flow, a surface-based optical-flow framework that estimates millisecond-resolved surface-tangent propagation fields from source-imaged magnetoencephalography (MEG) data. In resting-state MEG from 608 healthy adults, spontaneous propagation was anisotropic and bidirectionally aligned with the principal unimodal-to-transmodal functional gradient: slow activity (1-13 Hz) was biased toward upstream propagation from sensory to association cortex, whereas beta activity (13-30 Hz) was biased toward downstream propagation in the opposite direction. Across adulthood, this balance shifted toward weaker upstream slow propagation and stronger downstream beta propagation. Propagation strength, indexed by kinetic energy of the cortical flow, followed a robust posterior-to-anterior gradient and, within frontoparietal cortex, higher kinetic energy was associated with better fluid intelligence after adjustment for age. Kinetic-energy dynamics further identified stable-state dwell times that tracked regional neuronal timescales. Together, these findings establish geodesic cortical flow as a geometry-informed framework for quantifying frequency-resolved cortical propagation and its variation across aging and cognition.