An MRI scan may reveal more about dementia risk than doctors can see with the naked eye.
A new study suggests that when a person’s brain looks older than their actual age, their chance of developing dementia in the future may be higher.
Researchers in the Netherlands used computer-based tools to estimate the biological-looking age of the brain from MRI scans. They then compared this estimated brain age with each person’s real age.
The difference between the two numbers is known as the brain-predicted age difference, or brain-PAD. For example, if a 65-year-old person’s MRI resembles the brain of an average 70-year-old, the brain-PAD would be about five years older.
Brain aging is not exactly the same for everyone. Genetics, blood vessel health, lifestyle, disease and other factors can influence the brain’s structure as people grow older.
Dementia develops when diseases or damage to the brain gradually interfere with memory, thinking and the ability to manage everyday life. Alzheimer’s disease is the most common cause, but dementia can result from several different conditions.
One challenge is identifying people at high risk before major problems appear. Doctors can look for shrinkage in certain parts of the brain on MRI scans, but early changes can be subtle and difficult to judge visually.
The new study examined whether a computer-generated estimate of brain age could provide additional information. Researchers used records from patients at the Alzheimer Center Amsterdam who had been followed through major Dutch memory research programs between 2002 and 2024.
The analysis included 799 people. Among them, 412 had subjective cognitive decline, meaning they felt that their memory or thinking had worsened even though standard tests remained normal.
Another 387 participants had mild cognitive impairment, or MCI. These people had measurable difficulties with memory or thinking, but their problems were not severe enough to meet the definition of dementia.
The researchers used MRI scans, cognitive test results and the standard visual assessments doctors make when looking for brain shrinkage. They also used two computer systems to estimate each participant’s brain age.
Participants were then followed to see who remained stable and who eventually developed dementia. Across the whole group, a brain that appeared older than the person’s actual age was associated with greater risk.
For every additional year that the brain appeared older, the risk of developing dementia increased by about 6%. Even after the researchers considered what doctors could already see using traditional MRI rating methods, each extra year was linked to an independent increase in risk of about 4%.
The brain-age approach appeared particularly useful for people with subjective cognitive decline. In this group, each additional year of older-looking brain age was associated with about a 9% higher relative risk of later dementia.
The researchers also found a potentially reassuring pattern. People with subjective cognitive decline whose brains appeared at least 2.6 years younger than their real age had a 91% to 99% chance of remaining free of dementia during follow-up periods ranging from two to 10 years.
For these patients, the computer-generated brain age provided more useful information than standard visual MRI ratings. That could be important because subjective cognitive decline may occur very early, when structural changes are too small for doctors to easily see.
The results were different for people who already had mild cognitive impairment. In that group, estimating brain age added relatively little beyond the information available from standard MRI assessments.
The study was conducted by researchers in the Netherlands using patients from the Amsterdam Dementia Cohort and SCIENCe project at Amsterdam UMC. The findings were published in 2026 in Neurology, the medical journal of the American Academy of Neurology.
The study has an important strength: it examined real patients seen in a specialist memory setting rather than relying only on healthy volunteers. It also compared the computer method directly with the visual techniques already used by doctors.
However, brain age is not a diagnosis. An older-looking brain does not mean that a person will definitely develop dementia, and a younger-looking brain cannot guarantee that dementia will never occur.
The research also needs to be confirmed in other hospitals and more diverse groups of patients before the method can become a routine clinical tool. Different scanners, patient populations and computer systems could affect the estimates.
Overall, the findings suggest that automated brain-age measurements may be most valuable at the earliest stage of memory concern. Used alongside cognitive tests and doctors’ assessments, they could eventually help identify which patients need closer monitoring while reassuring some people whose risk appears lower.
