AI model reveals regional differences in human brain aging

Researchers have developed a deep learning system capable of estimating the apparent biological age of individual brain regions using structural magnetic resonance imaging (MRI). Published in the Proceedings of the National Academy of Sciences (PNAS), the study, led by N. N. Chaudhari, moves beyond traditional models that provide only a single, overall brain age estimate.

By analyzing nearly 15,000 MRIs from healthy adults aged 19 to 100, the AI model creates regional maps of aging. This allows for the identification of anatomical differences within the same individual. A key finding is a notable asymmetry: the right hemisphere tends to show higher estimated biological age values compared to the left hemisphere.

This methodology enables scientists to distinguish between areas showing accelerated or decelerated aging, which might be obscured by a single whole-brain average. The spatial representation of these results provides a more nuanced understanding of how aging processes are distributed across the central nervous system.