Functional connectome through the human life span

Abstract

The functional connectome of the human brain represents the fundamental network architecture of neural activity, but its normative growth trajectory over the life course remains unknown. Here, we aggregate the largest, quality-controlled multimodal neuroimaging dataset across 119 global sites, including 33,809 task-free fMRI and structural MRI scans of 32,328 individuals aged from 32 postmenstrual weeks to 80 years old. The lifespan growth charts of the connectome are quantified at the whole cortex, system, and regional levels using generalized additive models for location, scale, and shape. We find critical inflection points in the nonlinear growth trajectories of the whole-brain functional connectome, notably peaking in the fourth decade of life. After establishing the first fine-grained, lifespan-spanning suite of system-level brain atlases, we generate person-specific parcellation maps and further elucidate distinct timelines of maturation for functional segregation within various subsystems. We identify a spatiotemporal gradient axis that governs the life-course growth of regional connectivity, transitioning from primary sensory cortices to higher-order association regions. Using the connectome-based normative model, we demonstrate substantial individual heterogeneities at the network level in patients with autism spectrum disorder and patients with major depressive disorder, respectively. Our findings shed light on the functional connectome9s life-course evolution, serving as normative references for understanding network growth principles of the human brain and assessing individual variations of patients with neuropsychiatric conditions.