Resting brain scans of cognitively normal adults aged 55 to 70 appear to carry early signals about whether they will later develop Alzheimer's disease, according to research led by Stanford Medicine and published Nov. 4 in Science Advances. The team, led by senior author Yu Zhang, PhD, and lead author Kanhao (Alex) Zhao, PhD, used functional MRI, which measures brain activity while a person lies still in a scanner, to look at the way different brain regions were underactive or overactive relative to one another.
The researchers started with two well-defined extremes: adults 75 and older who had been diagnosed with Alzheimer's, and "super agers" whose memory matched people decades younger. Comparing 180 super agers with 110 people with Alzheimer's, they identified distinct connectivity patterns that separated the groups and built a scoring system, the Alzheimer's resilience-vulnerability connectome or ARC score, that distinguished the two groups with roughly 85% accuracy. They then applied it to more than 1,300 adults aged 55 to 70 with no signs of cognitive impairment.
At the time of scanning, people with higher and lower scores did not show clear cognitive differences. Over the follow-up period, which was mostly five to six years and at most 12 years in some datasets, the picture changed. Those with more Alzheimer's-like scores showed faster declines on tests of memory, attention and executive function, while those with the most super-ager-like scores declined most slowly. The researchers also reported that the ARC score predicted conversion to mild cognitive impairment or Alzheimer's at least as well as existing biomarkers such as amyloid-beta and tau, and that combining it with those markers improved prediction.
The significance lies in timing. Amyloid and tau are usually detectable only at relatively late stages, while this approach looked at a signal in people with no apparent impairment. Zhang's team also hopes the pattern could eventually be a target. The authors suggest that interventions such as transcranial magnetic stimulation or cognitive training might one day shift brain activity toward the resilient end, though that has not been tested for cognitive decline, and it is not yet known whether changing the pattern would alter disease course. The group plans a trial examining ARC scores before and after cognitive training.
The researchers are clear that the score is not diagnostic and is not destiny. Zhang noted that it is a snapshot, and that the proportion of people with super-ager-like scores tends to decrease with age, particularly after brain injuries or neuropsychiatric conditions.