Mapping Cortical Flexibility in Infancy Using Multiscale Functional Connectivity Gradients

Motivation: Extending gradient mapping to dynamic FC is limited by temporal harmonization challenges; novel, efficient metrics are needed to study cortical flexibility development during infancy. Goal(s): We develop multiscale dynamic FC gradients capturing information across timescales, reducing parameter sensitivity, and quantify cortical flexibility during infancy using a large infant dataset. Approach: Using multiscale sliding windows, we computed and aligned connectivity gradients, then quantified cortical flexibility by averaging 1 - cosine similarity of adjacent gradient vectors per vertex. Results: Our computed flexibility maps show meaningful change concentrated in the first year of life. Flexibility undergoes differential development across resting state networks during infancy and childhood. Impact: This abstract presents a novel method for multiscale dynamic functional connectivity gradients, reducing parameter sensitivity, and introduces an efficient metric for cortical flexibility, significantly advancing understanding of cortical development in infancy through analysis of extensive high-quality fMRI data.