Research from Stanford Medicine has developed a method to analyze brain activity patterns that may indicate a person’s risk for cognitive decline and Alzheimer’s disease years before memory symptoms appear.
The study, published in Science Advances in November 2026, focused on identifying early markers by comparing the brain scans of two distinct groups of older adults.
The research team began by analyzing functional MRI (fMRI) scans from 180 individuals aged 75 and older with exceptionally strong memory, often referred to as “super agers,” whose memory capacity was comparable to that of a typical 45-year-old.
They compared these scans to those from 110 people of the same age group who had been diagnosed with Alzheimer’s disease. The scans measured brain activity while participants were at rest, revealing how different regions of the brain coordinated with one another.
This analysis identified distinct patterns of connectivity in various brain regions, including the frontal, temporal, and insular areas, which separated the two groups with about 85% accuracy. Based on these differences, the researchers developed a scoring system called the Alzheimer’s Resilience and Vulnerability Connectome (ARC) score.
This score was designed to place an individual’s brain activity on a spectrum, with lower scores reflecting patterns more similar to the resilient super agers and higher scores indicating patterns more aligned with those of Alzheimer’s patients.
To test the score’s predictive power, the researchers applied it to brain scans from more than 1,300 people aged 55 to 70 who showed no signs of cognitive decline or memory problems at the time of their scan.
Although their cognitive test results were similar, the ARC scores revealed a wide range of brain activity patterns. Long-term follow-up, which extended up to 12 years with an average of five to six years, showed that these patterns were predictive of future outcomes.
Individuals whose brain activity scores were higher, resembling the Alzheimer’s group, tended to experience a faster rate of decline on memory and thinking tests. They were also more likely to receive a subsequent diagnosis of mild cognitive impairment or Alzheimer’s disease.
Conversely, those with lower scores, whose brain patterns were closer to the super agers, generally experienced a slower rate of cognitive decline and had a lower risk of developing the disease.
According to the researchers, the ARC score demonstrated predictive ability comparable to or better than some existing measures. When combined with other established Alzheimer’s biomarkers, such as protein levels, its predictive accuracy improved further. The study suggests that these brain activity differences can emerge long before two people with similar memory test results begin to follow different cognitive paths.
While the findings offer a new tool for identifying at-risk individuals for research studies and potential early intervention, the ARC score is not yet ready for clinical diagnosis.
The researchers noted that it remains unknown whether interventions like cognitive training or transcranial magnetic stimulation could alter these brain patterns to change a person’s disease trajectory. The team plans to investigate if such methods can shift the ARC score and potentially build cognitive resilience.
The research was supported by funding from the National Institutes of Health, the Alzheimer’s Association, and other foundations. The original study was published by Stanford Medicine.
