Scientists at the University of Southern California have developed an artificial intelligence system that can show how different parts of the brain age at different speeds.
Instead of giving a person’s brain one overall “age,” the new method creates a detailed map showing which areas look younger or older than expected.
The research was led by Associate Professor Andrei Irimia of the USC Leonard Davis School of Gerontology.
The study was published in the Proceedings of the National Academy of Sciences, or PNAS, and used brain scans from thousands of adults across a very wide age range.
Brain aging is a normal part of life. As people grow older, the brain can slowly change in size and structure, but those changes do not happen in exactly the same way in every person or in every part of the brain.
Researchers have therefore become interested in the idea of “brain age.” This involves comparing a person’s brain scan with scans from many healthy people to estimate whether the brain looks typical for that person’s actual age.
Earlier approaches often produced just one number for the whole brain. A 70-year-old person, for example, might be told that their brain looks more like that of a 75-year-old, but that single number cannot show where the biggest differences are located.
The USC team wanted a more detailed picture. They trained a deep-learning AI model using MRI scans from 14,748 cognitively healthy adults between 19 and 100 years old.
MRI uses a strong magnetic field to create detailed pictures of structures inside the body without using X-rays. The researchers drew their scans from six major research datasets, including the UK Biobank, the Human Connectome Project and the Alzheimer’s Disease Neuroimaging Initiative.
After learning what healthy brains tend to look like at different ages, the AI could examine very small areas of a new brain scan. It then estimated how old each area appeared compared with what would normally be expected at the person’s real age.
The researchers tested the model on more than 1,900 additional people. This group included healthy adults as well as people with mild cognitive impairment and Alzheimer’s disease.
In healthy adults, some parts of the brain consistently appeared older than others. The frontal and temporal areas, which support abilities such as memory, planning and decision-making, tended to look older than areas toward the back of the brain involved in vision and spatial processing.
The researchers also found a small difference between the two sides of the brain. The right side generally appeared slightly older than the left, and this pattern did not seem to depend on whether a person was right-handed or left-handed.
More striking differences appeared in people with memory and thinking problems. People with mild cognitive impairment or Alzheimer’s disease showed unusually advanced aging in areas that are known to be affected early by Alzheimer’s.
These areas included the hippocampus, which is essential for forming memories, and the amygdala, which is involved in emotion and memory. Several deeper brain structures also appeared older than expected.
The team then compared these local brain-age estimates with results from tests of memory and thinking. People whose important brain regions looked older generally performed worse, with some of the strongest relationships appearing in people with Alzheimer’s disease.
This matters because Alzheimer’s disease begins changing the brain years before severe dementia becomes obvious. A tool that identifies unusual regional aging could eventually help researchers detect concerning patterns earlier or follow how disease changes over time.
The maps might also help scientists test treatments. If an experimental therapy is designed to protect a particular brain region, researchers could potentially examine whether that area continues to age unusually quickly or begins to look more typical.
However, the new AI system is not ready to diagnose patients in everyday clinics. The researchers noted that it was developed mainly using high-quality research MRI scans, and it still needs testing in broader and more diverse patient groups.
Another limitation is that much of the information came from people scanned at one point in time. Following the same people for many years would provide stronger evidence about whether an older-looking region can predict future memory decline or progression to Alzheimer’s disease.
Overall, the study offers an important improvement over the idea that the entire brain can be described with one age number. Its strongest contribution is showing that brain aging is highly regional and that these local differences are connected with cognitive health.
The findings are promising, but an older-looking brain area should not yet be treated as proof that someone will develop dementia. More long-term research is needed before these detailed AI maps can become reliable tools for individual medical decisions.
If you care about health, please read studies about how Mediterranean diet could protect your brain health, and the best time to take vitamins to prevent heart disease.
For more health information, please see recent studies that olive oil may help you live longer, and vitamin D could help lower the risk of autoimmune diseases.
