Science

Brain activity score predicts Alzheimer's risk up to 12 years early

51 views

How the score works

The study started with two groups at opposite ends of ageing: people in their 80s and older who perform as well as adults decades younger, known as super agers, and patients with Alzheimer's disease. The Stanford-led team compared patterns of brain activity between them using scans and machine learning, then built a single number that places any brain on a line between the two.

They called it the Alzheimer's resilience-vulnerability connectome score, or ARC. It told the two extremes apart with roughly 85% accuracy. The point was not diagnosis, since the patients already had the disease, but measurement: a way to describe how far a normal brain sits from either pattern.

Twelve years of follow-up

The researchers then applied the score to cognitively normal adults. At the start, people with high and low scores performed the same on cognitive tests. The difference appeared later. Those whose activity looked more like Alzheimer's patients were more likely to decline, to build up amyloid-beta plaques, and to be diagnosed during follow-up that ran as long as 12 years. Those closer to the super ager pattern tended to hold their abilities.

Combining the score with existing biomarkers improved prediction further, the authors report. Lead author Kanhao (Alex) Zhao, a postdoctoral scholar in psychiatry, said the model performed as well as or better than existing measures and did so earlier. The paper appears in Science Advances.

What it could change

Alzheimer's trials have long struggled with timing. Drugs that clear amyloid have shown modest effects, and one explanation is that they arrive after too much damage. A cheap scan-based score that flags risk a decade ahead would give researchers a way to enrol the right people sooner, and give clinicians a reason to talk about prevention while it can still matter.

It would not, on its own, tell a patient what to do. The score is a risk marker, not a cause, and the study is observational. Senior author Yu Zhang, an assistant professor of psychiatry, noted that cognitively normal people are biologically varied, which is why the team compared them against both extremes rather than a single healthy average. Independent replication in larger, more diverse groups is the next test.

Source: Stanford Medicine / Science Advances