THE ARISING · Longevity & Healthspan · Long reads
Brain-Age Scans May Mostly Measure Blood Vessel Health, Not Neurons. Our Desk Rates It 67 Percent
MRI brain age is becoming a trial endpoint. Our research desk proposes that after 60 it mainly tracks small-vessel health, which would change how drugs are judged. The evidence so far is mixed.
Some longevity clinics now sell a number called your brain age, computed from an MRI. Researchers increasingly use the same number to judge whether a treatment slows aging. Nobody fully agrees on what it actually measures.
A brain-age model is trained on thousands of scans from people of known age. It then estimates the age of a new brain. The difference from your real age is the brain age gap, and a bigger gap tracks with worse health.
Our research desk proposes that in adults over 60, most of that gap reflects small blood vessels: stiffening, silent small-vessel disease and poor blood-flow control. True neuron loss, it argues, contributes less.
The prediction follows. Improving vascular health should shrink the gap within one to two years. A drug that protects neurons without touching vessels should leave it unchanged. The desk puts the odds near 67 percent that vascular measures explain more of the gap's change than neuronal measures.
The vascular link is real. In the Maastricht Study of 4,881 middle-aged adults, published in Brain Communications in 2024, fluid spaces, gray matter, white matter hyperintensities and microbleeds best predicted the gap. The last two are classic markers of small-vessel disease.
A UK Biobank analysis of 18,817 adults in eBioMedicine in 2025 linked better scores on the American Heart Association's Life's Essential 8 with a younger white-matter brain age. That study was cross-sectional, so it shows association, not cause.
But the other side has weight too. Cole's 2020 multimodal UK Biobank model, in Neurobiology of Aging, relied mostly on gray-matter volume and white-matter microstructure. Its gap tracked stroke, diabetes, smoking and alcohol. Gray-matter volume partly reflects neuronal loss, so the signal is a blend.
The best randomized data come from SPRINT MIND. Our desk credited the MRI result to Williamson and colleagues; that 2019 paper reported cognitive outcomes. The MRI substudy appeared separately in JAMA in 2019. Intensive blood pressure control slowed white matter lesion growth, 0.92 versus 1.45 cubic centimeters over about four years.
There is a twist the desk did not mention. Total brain volume fell slightly more in the intensive arm, by 3.7 cubic centimeters. A volume-heavy brain-age model could therefore read the vascular winners as slightly older. That is exactly why this hypothesis is worth testing rather than assuming.
If brain age is mostly a vascular reading, it becomes a useful dial for blood pressure, fitness and metabolic care. It would also be a poor judge of drugs aimed at neurons or amyloid, which might fail on this measure while still working.
• Treat a brain-age report as a rough signal, not a diagnosis; ask which model and scan type produced it. • Keep blood pressure at your physician's target; it is the best-tested way to protect small vessels. • Build regular aerobic exercise and keep blood sugar and lipids in range. • If an MRI shows white matter changes or microbleeds, review vascular risk with your doctor.
The hypothesis assumes models weight vascular features heavily, and published models differ. No study has yet tested whether change in vascular measures or change in cortical thickness better predicts change in the gap. The 67 percent figure is the desk's judgment, not a measured quantity.
• A UK Biobank repeat-scan analysis comparing vascular change and cortical thinning as predictors of gap change. • A brain-age reanalysis of SPRINT MIND scans. • Whether trial sponsors begin reporting which features drive their brain-age endpoint.