A new artificial intelligence model has been developed to assess the rate at which a patient’s brain is aging. According to researchers from USC Leonard Davis School of Gerontology, this tool could revolutionize our understanding and treatment approaches for cognitive decline and dementia.
This pioneering method utilizes magnetic resonance imaging (MRI) scans in a non-invasive manner to track changes in brain health over time. The correlation between faster brain aging and an increased risk of cognitive impairment highlights the significance of this development, as noted by Andrei Irimia, associate professor at USC Leonard Davis School of Gerontology.
“This is a novel measurement that could change the way we monitor brain health both in research settings and clinical practices,” said Irimia. “Knowing how fast an individual’s brain ages can be highly valuable.”
Irimia, who serves as the senior author of this study published on February 24, 2025 in Proceedings of the National Academy of Sciences, emphasizes that biological age differs from chronological age. Two people may share the same birthdate but have vastly different biological ages due to their overall health and cellular functioning levels.
Traditional methods for measuring biological age often rely on blood samples to evaluate epigenetic aging and DNA methylation, which are crucial in determining gene roles within cells. However, these techniques fall short when it comes to accurately reflecting the brain’s specific conditions as they don’t directly measure methylation or other aging-related processes occurring inside the brain.
Previous studies by Irimia’s team had shown potential for using MRI scans to non-invasively assess biological brain age. The earlier model leveraged AI analysis to compare a patient’s brain anatomy with data derived from MRIs of thousands of individuals, spanning various ages and cognitive health statuses.
However, this method had its limitations as it could only estimate the difference between a person’s brain age and their chronological age but couldn’t pinpoint when or how rapidly aging occurred. A more accurate solution was required to fully understand these dynamics.
The new approach introduces a three-dimensional convolutional neural network (3D-CNN) designed to provide a clearer picture of brain aging over time. Developed in collaboration with Paul Bogdan from the USC Viterbi School of Engineering, this model utilized more than 3,000 MRI scans from cognitively healthy adults for training and validation.
Differing from traditional cross-sectional methods that analyze one scan at a particular point in time, this longitudinal method compares baseline and follow-up MRIs taken from the same individual. As such, it can identify specific neuroanatomical changes linked to accelerated or decelerated brain aging more precisely.
The 3D-CNN model generates interpretable “saliency maps,” indicating which parts of the brain are most influential in determining age-related processes. Applying this tool to groups including cognitively healthy individuals and those with Alzheimer’s disease revealed correlations between rates of brain aging and cognitive function test results, suggesting its potential as an early biomarker for neurocognitive decline.
Furthermore, Irimia explained that the new model has a broader scope by identifying variations in regional brain aging. This could offer deeper insights into how different pathologies develop within various parts of the brain. Additionally, sex-specific differences in brain aging rates might explain why men and women are differently susceptible to neurodegenerative diseases like Alzheimer’s.
Irimia is also optimistic about using this new model to identify individuals with faster-than-normal brain aging even before they exhibit any signs of cognitive decline. This could prove crucial for developing preventive strategies as current drugs targeting Alzheimer’s have shown limited efficacy, possibly due to delayed treatment initiation when extensive pathology already exists in the brain.
“I believe measures like these will be immensely useful in forecasting Alzheimer’s risk and helping us develop potential preventative treatments,” he stated.
In summary, this innovative AI-driven approach offers a significant leap forward in understanding and addressing cognitive decline. It not only provides deeper insights into brain aging but also opens new avenues for diagnosing and treating neurodegenerative disorders before symptoms become apparent.
