In a new study published in Nature Aging, researchers leveraged brain imaging from nearly 11,000 middle-aged and older adults to explore how biological “brain age” differs from chronological age—and how certain proteins in the blood might reveal this difference.
Using data from the UK Biobank, the team utilized an AI model trained on features of the brain’s structure—such as cortical thickness and white matter integrity—to estimate how old someone’s brain appears. When those AI-derived ages were compared with the participants’ actual ages, many individuals’ brains turned out to be biologically younger or older than expected, sometimes by several years.
Next, the scientists measured roughly 3,000 different proteins in blood plasma for about half of the study participants. Their key aim was to pinpoint which proteins best correlated with a “younger” or “older” brain age according to the AI model. Thirteen proteins ultimately stood out.
Many were tied to inflammation or to the maintenance of neural connections—both known to shift as we age. Two specific proteins, however, seemed especially powerful in flagging brain health. Brevican (BCAN), important for preserving the brain’s wiring, correlated with a slower aging process and lower risk for problems like dementia and stroke. By contrast, growth differentiation factor 15 (GDF15) is released in response to cellular damage and was linked to chronic inflammation and a higher risk of age-related diseases.
Importantly, changes in these key proteins weren’t linear across a person’s entire lifespan. Rather, the data highlighted three ages—57, 70, and 78—where protein levels spiked or shifted dramatically. These distinct phases suggest the brain undergoes multiple stages of aging, each dominated by different molecular processes, from metabolism and wound healing to inflammation.
If future work confirms these findings, a simple blood test could provide an early signal for accelerated brain aging or susceptibility to conditions like dementia, reducing reliance on expensive and less accessible brain scans. Although more research is needed—including validation in diverse populations and controlled experiments in animals—the study offers a new glimpse into how our brains age at the molecular level. Ultimately, identifying these protein “hallmarks” could pave the way for personalized treatments or lifestyle interventions that target brain aging before irreversible damage occurs.