Scientists Map Five Distinct Phases of Brain Rewiring Across the Human Lifespan

Understanding how the brain's structural organization changes across the entire human lifespan—from birth to old age—has been limited by studies focusing on narrow age windows and single metrics. This large-scale study addressed that gap by assembling diffusion MRI data from nine datasets spanning ages 0 to 90 years, totaling 4,216 participants, to comprehensively map how brain network topology evolves.

The team calculated 12 graph theory metrics covering network integration (global efficiency, characteristic path length, small-worldness), segregation (modularity, clustering coefficient, local efficiency, core-periphery structure), and centrality (betweenness and subgraph centrality). Networks were harmonized across datasets and analyzed both with variable density and at a controlled 10% density threshold to ensure fair topological comparisons unbiased by raw connectivity differences. Age-predicted metric values were then projected into low-dimensional manifold spaces using Uniform Manifold Approximation and Projection (UMAP) to capture the multivariate trajectory of topological change.

Four major topological turning points emerged at approximately ages 9, 32, 66, and 83, dividing the lifespan into five epochs. In childhood (0–9), networks shift from dense but weak connections toward increased integration. Adolescence (9–32) is characterized by rising global efficiency, peaking at age 29, alongside increasing core-periphery structure peaking around age 20. Adulthood (32–66) marks a transition from peak efficiency toward rising modularity and local clustering. Early aging (66–83) and late aging (83+) are defined by continued loss of global integration, pronounced increases in modularity, betweenness centrality, and local efficiency, and a shift toward sparser but stronger remaining connections.

The study also found that raw network density follows a non-linear trajectory—highest at birth and around age 30, lowest around ages 10 and 80+—while average node strength increases nearly linearly across the lifespan. This dissociation between density and strength reveals that aging brains become sparser but retain and strengthen key connections. Each topological epoch has a distinct directional signature in manifold space, meaning the brain reorganizes itself along qualitatively different dimensions during each life phase, not merely more or less of the same trajectory.

These findings underscore that brain network development cannot be adequately described by simple inverted-U models or single inflection points. The UMAP manifold approach successfully detected population-level phase transitions that would be invisible when examining individual metrics in isolation, offering a powerful framework for future studies linking topological epochs to cognitive trajectories, mental health, and neurodegenerative risk.