Brain scan score flags Alzheimer’s risk years before memory loss

Stanford Medicine researchers have built a score from resting brain scans that can pick out older adults likely to develop Alzheimer's disease years before any memory trouble shows up, according to a study published in Science Advances.

The tool, called the Alzheimer's resilience-vulnerability connectome score, or ARC score, ignores physical markers such as plaque buildup and brain shrinkage. It focuses instead on how different regions of the brain talk to one another.

To create it, the team contrasted scans from two extremes: patients diagnosed with Alzheimer's and so-called super agers, older people whose memory rivals that of adults decades younger. The resulting score told the two groups apart with roughly 85% accuracy.

The researchers then ran the model on more than 1,300 adults aged 55 to 70, all of whom performed normally on memory tests. Their scans still diverged, with some resembling super agers and others looking closer to Alzheimer's patients.

Participants were tracked for as long as 12 years, though most were followed for five to six. Those whose brain activity looked Alzheimer's-like lost memory and thinking ability faster, built up greater amyloid deposits and faced higher odds of eventually receiving the diagnosis.

In spotting early decline, the ARC score matched or beat standard risk measures, including amyloid and tau biomarkers and genetic testing. Lead author Kanhao Zhao, a postdoctoral scholar in psychiatry, said the model can flag risk at an earlier point, and that pairing it with existing biomarkers sharpens predictions further.

Timing is the core problem in Alzheimer's care. Physical damage often starts decades before patients notice lapses, and by then treatments do little.

The researchers cautioned that the scan method is not ready for doctors to use by itself to diagnose the disease. Further studies must confirm its accuracy and weigh it against other emerging tests.

Senior author Yu Zhang, an assistant professor of psychiatry and behavioral sciences, said mapping the resilient brain could guide future therapies, including noninvasive brain stimulation, by showing both where to intervene and what change to aim for.