Resting brain connectivity score flags Alzheimer's risk in cognitively healthy adults

Patterns of resting brain activity in cognitively healthy middle-aged and older adults can indicate who is likely to develop Alzheimer's disease and who is likely to stay sharp, according to a Stanford Medicine-led study. The researchers built a score from functional MRI data and found that it predicted later decline at least as well as established biomarkers. The work was published on 4 November in Science Advances.

Two extremes as a reference

Studies of so-called super agers usually compare them with other older adults who test as cognitively typical. Yu Zhang, assistant professor of psychiatry and behavioural sciences and senior author, argues that this design is flawed. People who look typical on standard tests can differ widely beneath the surface, and some may already show early amyloid-beta deposition or other signs of brain ageing.

The team instead contrasted super agers with a sharply defined opposite group: people of similar age with a diagnosis of Alzheimer's disease. Super agers were defined as people aged 75 or over with memory comparable to that of a typical 45-year-old and normal executive function for their age. The researchers drew on imaging from several large projects, including the Open Access Series of Imaging Studies, the Harvard Aging Brain Study and the Alzheimer's Disease Neuroimaging Initiative.

Building the ARC score

Using resting-state functional MRI from 180 super agers and 110 people with Alzheimer's disease, the team identified patterns of underactivity and overactivity across brain regions, known collectively as the functional connectome, that separated the two groups. From these they derived the Alzheimer's resilience-vulnerability connectome (ARC) score. It distinguished super agers from patients with roughly 85% accuracy.

The score was then applied to more than 1,300 people aged 55 to 70 with no sign of cognitive impairment. Their scores fell along a continuum, with lower values resembling super agers and higher values resembling Alzheimer's patients.

Prediction years ahead

At the time of scanning, high and low scorers did not differ clearly in cognition. Over follow-up, they diverged. Longitudinal data varied between datasets, reaching up to 12 years but typically covering five to six. People with more Alzheimer's-like scores declined faster on tests of memory, attention and executive function, and were more likely to develop amyloid-beta plaques or progress to mild cognitive impairment or Alzheimer's disease. Those with the most super-ager-like scores declined most slowly.

According to lead author Kanhao Zhao, a postdoctoral scholar, the score matched or outperformed demographics, amyloid-beta and tau biomarkers and genetics in predicting conversion, and did so earlier. Combining it with existing biomarkers improved prediction further. ARC scores also correlated to some degree with brain age, though the authors describe ARC as more specific to resilience and vulnerability in older people.

Limitations and next steps

The score is not diagnostic and is intended as a guide for monitoring cognitive status. Zhang stresses that it is a snapshot rather than a fixed fate: the share of people with super-ager-like scores falls with age, for instance after brain injury or the onset of neuropsychiatric conditions. Follow-up was also relatively short in most of the datasets.

The longer-term aim is to use the resilient pattern as a target for intervention. Transcranial magnetic stimulation and cognitive training are possible approaches, but neither has been tested for cognitive decline in this way, and it is unknown whether shifting connectivity would alter disease course. Zhang's team is planning a trial to measure ARC scores before and after cognitive training.

Researchers from Children's National Hospital, the University of Texas at Austin, Lehigh University, Penn State College of Medicine, New York University Grossman School of Medicine and the University of California, Los Angeles, also contributed.

Source: Brain activity patterns may predict Alzheimer's years before memory slips (Medical Xpress, 5 November 2026).

Image: AI-generated illustration.