Brain age prediction from structural and functional connectivity across the lifespan using connectome-based predictive modeling
This study utilizes connectome-based predictive modeling on structural and functional MRI data across the lifespan to demonstrate that chronological age can be accurately predicted by distributed connectivity patterns, revealing both shared and complementary information between modalities that is further enhanced through multimodal integration.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
The human brain is not a static organ; it is a living landscape that changes shape and function from the moment we are born until our final days. Scientists have long sought a way to measure this journey, creating what is known as a "brain age" model. This is a tool that looks at the brain's wiring and structure to estimate how old a person is, much like a doctor might estimate a tree's age by looking at its rings. When the brain looks older or younger than the person's actual years, it can signal that something is happening with their development or aging process. To understand these changes, researchers often look at two different types of maps. One map shows the physical roads of the brain, the white matter tracts that connect different regions, built up over years of growth. The other map shows the traffic on those roads, the patterns of activity that happen when the brain is resting and simply thinking. While we know both maps change as we age, it has remained unclear whether they tell the same story or if they offer different, complementary clues about how we grow and grow old.
A team of researchers at Yale University decided to explore this question by looking at the entire human lifespan, from childhood through old age. They gathered data from nearly 1,500 healthy people, using advanced brain scans to create detailed maps of both the brain's physical connections and its resting activity. They then used a computer method designed to find patterns in these maps to predict each person's age. The goal was to see if the physical structure of the brain or its functional activity was better at telling time, and whether combining both maps would give a clearer picture than looking at either one alone.
The researchers found that both the physical structure and the functional activity of the brain were excellent at predicting a person's age, but the strength of these predictions changed depending on the stage of life. The models worked best when looking at children and older adults, periods where the brain undergoes rapid and significant changes. In contrast, the models were less accurate for young adults, a time when the brain is relatively stable. This suggests that the brain's wiring and activity are most distinct during the years of rapid growth and the years of decline, making those periods easier to identify. Interestingly, the physical structure of the brain was a stronger predictor of age for older adults, while the functional activity was a stronger predictor for children. This indicates that as we move through life, the way our brain ages shifts from being driven by how its circuits are firing to how its physical roads are changing.
Despite these differences in which map worked best at different times, the two types of maps often agreed with each other. When the physical map predicted a person was older, the functional map usually did too. However, when the researchers looked closely at exactly which connections were driving these predictions, they found that the two maps were using different clues. The physical map relied heavily on connections in the deep, central parts of the brain and the cerebellum, while the functional map drew on a wider variety of networks, including those involved in movement and attention. This means that while both maps are telling us about the same thing—how old the brain is—they are reading different parts of the story.
To test if combining these two maps would help, the researchers built a new model that used both the physical structure and the functional activity at the same time. In most groups, this combined approach was more accurate than using either map alone. It was particularly helpful for young adults and for the group spanning the entire lifespan, where the combination of clues reduced the error in prediction. The combined model did not create a new, mysterious pattern; instead, it simply brought together the best parts of the physical and functional maps, confirming that they offer unique and valuable information that the other misses.
The study also looked at whether men and women showed different patterns in how their brains aged. While previous research has shown that male and female brains can look different, this study found that the ability to predict age was remarkably similar for both sexes. A model trained on men worked just as well on women, and vice versa. This suggests that the fundamental changes in brain connectivity that mark the passage of time are shared across sexes, even if the specific details of the brain's structure vary slightly.
Ultimately, this work shows that the brain's age is written in its connections, but the handwriting changes as we grow. By looking at both the physical roads and the traffic flowing through them, scientists can get a more complete picture of how we develop and age. The findings suggest that no single view of the brain is enough; to truly understand the timeline of human life, we must look at the brain from multiple angles, recognizing that the story of aging is told through a complex, distributed network of changes that vary from childhood to old age.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.