In an evolving health landscape, emerging research continues to highlight concerns that could impact everyday wellbeing. Here’s the key update you should know about:
Using one of the largest lifespan MRI datasets to date, researchers uncovered the four critical moments when the brain’s wiring reorganizes, revealing how our neural networks grow, stabilize, and eventually decline as we age.
Study: Topological turning points across the human lifespan. Image credit: PeopleImages/Shutterstock.com
In a study published in Nature Communications, researchers mapped structural topological development across the human lifespan.
As we move through life, the brain’s structure and function change in distinct ways. Its topology, the complex patterns that define how neural connections are organized, also evolves with age. It is associated with behavioral, mental health, and cognitive outcomes. Studies have noted significant differences in structural topology related to lifespan development and individual differences.
Structural connectivity patterns link to behavior and health
In the present study, researchers mapped structural topological development across the human lifespan. They used diffusion imaging data from nine datasets in a cross-sectional sample of neurotypical individuals. Twelve graph theory metrics, i.e., measures of integration, segregation, and centrality, were calculated for topological analysis. Centrality metrics were subgraph centrality (the weighted sum of all closed walks of a node) and betweenness centrality (the fraction of shortest path lengths passing through the node).
Segregation metrics captured how the brain’s network breaks into distinct, densely connected clusters. These included:
- Modularity (how well the network separates into non-overlapping groups of nodes)
- S-core (the largest subnetwork defined by connection strength, which tended to increase more steadily with age)
- K-core (the largest subnetwork defined by node degree)
- Clustering coefficient (how often a node’s neighbors are also connected to each other)
- core–periphery structure (the organization of nodes into a dense core and a sparse periphery)
- Local efficiency (how efficiently neighboring nodes can communicate through short paths)
Integration metrics included
- Global efficiency (average inverse shortest path length)
- Strength (sum of edge weights)
- Characteristic path length (the average shortest path length of the network)
- Small-worldness (the ratio between the clustering coefficient and the characteristic path length).
The age distribution ranged from 0 to 90 across the datasets. Topological analyses were performed on networks thresholded to a fixed density (10 %) to enable consistent comparisons across ages.
Lifespan metric trajectories
There were significant fluctuations in global efficiency across the lifespan, peaking at 29 years before declining to a minimum at 90 years. The average network strength showed a significant linear increase, peaking at 90 years.
Characteristic path length and small-worldness exhibited inverse patterns to global efficiency. Modularity showed significant fluctuations across the lifespan, with a minimum at 31 years and a maximum at 90 years. Meanwhile, the core/periphery structure showed greater fluctuations than modularity, reaching a minimum at 55 years and a maximum at 20 years.
S-core increased in a relatively linear manner, reaching a minimum at 12 years and a maximum at 90 years. K-core showed no significant changes across age. Local segregation measures, such as the clustering coefficient and local efficiency, increased more linearly with age, reaching a maximum at 90 years. Betweenness centrality also fluctuated throughout the lifespan, while subgraph centrality showed a more linear increase.
Many topological measures were highly correlated, conveying redundant and unique characteristics. Therefore, the dimensionality of the data was reduced to explore non-linear changes in lifespan topology using manifold learning. Significant age-predicted metrics were used to construct manifolds. In total, 968 uniform manifold approximations and projections (UMAPs) were created to capture global- and local-level information.
Subsequently, manifolds were used to ascertain turning points. Major turning points were identified around the ages of 9, 32, 66, and 83, which defined five epochs of life. Changes across epochs were assessed using Pearson correlations to determine the relationships between topological metrics and age. Regularized least absolute shrinkage and selection operator (LASSO) models were used to identify the metric(s) driving these relationships.
Turning points define five distinct developmental epochs
Epoch 1 ranged from zero to nine years, covering infancy through childhood. Eight measures showed significant correlations in this epoch. The clustering coefficient was the strongest topological predictor of age. A decline in global integration characterized topological development in this epoch. At the end of epoch 1, the factor driving the age-topology relationship changed from clustering coefficient to small-worldness.
There was also a shift in direction, with the network beginning to show increasing integration. The second epoch ranged from nine to 32 years, covering late childhood through early adulthood. All topological measures in this epoch were significantly correlated with age. In general, local-level and strength-based segregation increased, while global modularity decreased. In epoch 2, small-worldness was the top predictor of age.
Many directional changes were observed at the end of epoch 2, with a shift towards lower integration, higher betweenness centrality, and higher modularity. The factor driving the relationship with age also changed from small-worldness to local efficiency.
Epoch 3 ranged from 32 to 66 years, spanning three decades of adulthood, with 10 measures showing significant correlations. It was characterized by increases in segregation, decreases in integration, and minimal changes in centrality. The largest predictor of age in epoch 3 was local efficiency. There were no significant directional changes at the end of epoch 3; the factor driving the age-topology relationship changed from local efficiency to modularity.
Epoch 4, which ranged from 66 to 83 years, marked a transition from adulthood to early aging. Only four metrics were significantly correlated with age in epoch 4, characterized by distinct modularity changes, increasing centrality, and decreasing integration. Modularity was identified as the strongest predictor of age in epoch 4. Further, no significant changes in directionality were observed at the end of epoch 4, with the driving topological metric shifting from modularity to subgraph centrality.
The last epoch ranged from 83 to 90 years, extending from late aging to the maximum age studied. In epoch 5, only subgraph centrality was significantly correlated with age, which was also the strongest predictor of age. Findings in this final epoch should be interpreted cautiously due to reduced statistical power in the oldest age group.
Five lifespan phases capture non-linear brain development
The findings highlight complex, non-linear changes in topological development that occur throughout the lifespan. The results illustrate a pattern of increased network segregation and a decrease in the age-topology relationship during later years.
The analyses revealed four major turning points at ages 9, 32, 66, and 83, marking distinct phases of topological development with their own age-related patterns. Extensive robustness checks across harmonization procedures, density thresholds, and manifold parameters support the stability of these findings.
