MRI brain age and structure point to earlier dementia signals

At Amsterdam UMC, researchers asked a computer to do what the eye can miss: estimate how old a patient’s brain looks on an MRI scan. Among 799 people seen through the Amsterdam Dementia Cohort and SCIENCe project, each year that the brain appeared older than the patient’s actual age was associated with a 6% higher risk of later dementia.

The measure, called brain-predicted age difference, or brain-PAD, was calculated with the AI tools brainageR and cNeuro. It added information beyond standard visual MRI ratings, which doctors use to judge brain shrinkage. After those ratings were included, every additional year of brain-PAD still carried a 4% increase in dementia risk.

The signal was clearest in the 412 patients with subjective cognitive decline: people who reported memory problems but had normal test scores. For them, each extra year of apparent brain age raised the relative risk of progression to dementia by 9%. Patients whose brains looked at least 2.6 years younger than their actual age had a 91% to 99% chance of remaining dementia-free over the next 2–10 years. In the 387 people with mild cognitive impairment, brain age added little beyond conventional visual ratings.

A separate team at the University of Oslo found an earlier biological trace. By following healthy individuals with repeated MRI scans over nearly two decades, researchers detected, in those who later developed plaques, structural brain changes at least 7 years before amyloid plaques became visible on amyloid-PET, the scan currently used to identify early plaque buildup. The finding suggests that brain processes may begin before today’s most sensitive imaging can show plaques, although the researchers say those changes might either contribute to plaque accumulation or arise from another process.

So what changes in practice? Not a diagnosis delivered by software tomorrow. The Amsterdam result points to an automated MRI measure that could help memory clinics identify which people with subjective cognitive decline need closer follow-up, while the Oslo result gives drug researchers a window before plaque detection. Both studies remain research findings: Amsterdam’s team says clinical validation is still needed, and Oslo researchers say more work is required before the meaning of the early structural changes is clear.