How routine eye test could predict Alzheimer’s

One of the most fascinating implications of research is the discovery of ways to flag the risk for serious medical conditions with ordinary exams. For instance, think of an eye exam that can predict dementia decades before memory slips into the picture?

That’s the promise behind a new study led by Ashleigh Barrett-Young of the University of Otago and Aaron Reuben of the University of Virginia, with a team spanning Duke University, King’s College London, the Singapore National Eye Centre, and more. Published the Journal of Alzheimer’s Disease, their work follows members of New Zealand’s remarkable Dunedin Study, a population-representative birth cohort that has been tracked carefully from birth into midlife, and asks the question: can the health of the retina at age 45 already reflect who is on a higher trajectory for Alzheimer’s disease and related dementias later on?

The eye as a window to the brain

You may have heard this phrase before, but here its implications are literal. The retina is neural tissue, directly wired to the brain. Better still, it’s easy and safe to image with tools common in community clinics and even retail optometry: optical coherence tomography (which maps the retina’s layers) and fundus photography (which pictures the tiny arterioles and venules that perfuse those layers). Because the retina shares both structure and blood-supply logic with the brain, researchers have spent years investigating whether its appearance can stand in for what’s happening in our heads. Reviews over the past few years have summarized evidence that retinal layer thinning and microvascular changes track with Alzheimer’s and mild cognitive impairment in older adults, strengthening the case that the eye mirrors brain health.

What’s been less clear is how early these retinal “tells” show up, and whether they’re meaningful in midlife, long before clinical symptoms. The Dunedin Study is uniquely positioned to answer that. This cohort, which includes 1,037 babies born in Dunedin, New Zealand, in 1972–73, has been followed for more than five decades with extraordinary care; at the age-45 wave, 94% of living cohort members took part, giving researchers a rare, nearly complete portrait of health at midlife.

What the researchers measured

The team focused on two broad categories of retinal features. The first one is Neuronal layers. Think of these as the retina’s wiring, the ganglion cells and their axons that carry visual information along the optic nerve. In many aging studies, thinner layers have been associated with neurodegeneration. The second one is Microvessels, tiny arterioles and venules that feed the retina. Their calibers (or diameters) subtly shift with cardiovascular and inflammatory health. Narrower arterioles and wider, “sluggish” looking venules have long been associated with high blood pressure, diabetes, obesity, and systemic inflammation, exactly the metabolic and inflammatory currents that also shape brain aging.

Rather than waiting decades for who does or doesn’t develop dementia, the team linked each participant’s retinal measurements to established midlife dementia-risk indices. If you’ve never heard of these, picture them as evidence-based checklists: add points for risk factors (like hypertension or physical inactivity), subtract for protective ones (like education or active lifestyle), and you get a score that predicts later dementia.

The headline finding

By age 45, the retina is already whispering a story about dementia risk. In this midlife cohort, people with narrower retinal arterioles and wider venules tended to have higher dementia-risk scores across all the major indices. The retina’s microvascular health, in other words, aligned robustly with the kinds of risk that public-health researchers and clinicians already track.

The retina’s neuronal layers told a subtler tale. Thinner layers were only modestly linked with higher risk, detectable, but not nearly as strong as the vessel signals. And when the team unpacked their “homegrown” benchmark into its ten domains (things like cardiometabolic risk, lifestyle, sensory function, inflammation, socioeconomic context), the vascular measures, especially wider venules, spread their associations across many different types of risk, from blood pressure and cholesterol to sleep, mood, and overall perceived health. Conversely, the neuronal measures mostly tracked with cardiometabolic risk alone.

Why that makes sense

If you imagine dementia risk as a river fed by many tributaries, two major streams are vascular and inflammatory health. The retina’s venules seem unusually sensitive to these forces: population studies show that wider venules co-travel with systemic inflammation and metabolic strain. That’s consistent with the broader idea that keeping blood vessels (and the immune signals that swirl around them) healthy in midlife is crucial for protecting the brain later.

The neuronal-layer result is just as interesting. It hints that retinal thinness at 45 may be capturing a narrower slice of risk (largely metabolic/vascular) rather than the whole Alzheimer’s spectrum, at least at this age. In older adults, thinning is more clearly tied to cognitive impairment and dementia diagnoses; midlife may be simply too early for those neural signatures to loom large outside of cardiometabolic pathways.

At the moment, this is not a new diagnostic test that says, “You will get Alzheimer’s.” Retinal measures are risk signals, not fate. And this study, being an observational work, can’t prove cause-and-effect. Also, the participants are only in their mid-40s; we’ll still need follow-ups into their 60s and 70s to see how these early eye signals map onto actual diagnoses.

But...

If you wear glasses, you already know how routine eye photos and scans are. That’s why this line of research is exciting: it leverages tools that are already everywhere. Over the past few years, researchers have even trained deep-learning systems to detect Alzheimer’s from retinal photographs alone, raising the possibility of scalable, first-line risk triage in optometry settings.

The public-health case is straightforward. If a quick eye image can tell you, at 45, that your vascular and inflammatory “weather” is trending in the wrong direction, you can do something about it while it still matters most. For instance, you can manage blood pressure, move more, eat better, quit smoking, treat sleep apnea, monitor hearing, address depression, and stay socially connected. Those are precisely the modifiable levers reflected in the risk indices used here, and they’re the levers the Lancet Commission and others point to for preventing or delaying dementia.

And the momentum is building. Large collaborations in the UK and elsewhere are assembling massive eye-image datasets linked to health records to refine retinal risk models. The vision (pun intended) is to make routine eye checks a gateway to brain-health screening, not to diagnose dementia, but to nudge earlier, smarter prevention.

Additionally, eye exams help track the whole-body factors that ultimately protect the brain: blood-vessel health and inflammation. Think of retinal imaging as an early-warning dashboard light. It doesn’t tell you exactly what’s wrong under the hood; it tells you it’s time to look at factors like your blood pressure, your sleep, your cholesterol, your activity, your hearing, your mood, and your connections to other people.

If you want to learn more, the original article titled "Measures of retinal health successfully capture risk for Alzheimer's disease and related dementias at midlife" on Journal of Alzheimer’s Disease at https://doi.org/10.1177/1387287725132.