Scientists have found a way to distinguish older people with exceptionally sharp memory from patients with Alzheimer's disease by their pattern of brain activity. The same marker, it turns out, may indicate a hidden risk in outwardly healthy older adults. The work was published in the journal Science Advances.
Alzheimer's disease is a progressive condition in which nerve cells die and memory, thinking and the ability to care for oneself are affected. Amyloid plaques (sticky clumps of protein between cells) and tau tangles (twisted fibers inside cells) accumulate in the brain. At the other extreme are "SuperAgers": a small group of older people whose episodic memory, that is, the ability to recall events and recently learned information, remains at the level of middle-aged people.
Between these extremes are millions of older adults with no visible memory problems, whose brains age differently. To capture these differences, researchers led by Kanhao Zhao and Yu Zhang of Stanford University School of Medicine studied the functional connectome, the network of coordinated activity across different brain regions. It is measured with resting-state functional MRI: a person lies in the scanner for several minutes without performing tasks while the machine tracks fluctuations in blood oxygen levels. Regions where signals change synchronously are considered connected.
Unlike earlier work, the authors first compared the extreme groups. They combined data from open databases and compared 180 "SuperAgers" (aged at least 75 whose memory is better than that of an average person under 65) with 110 Alzheimer's patients of the same age (median 80 years). A machine learning model trained on the synchrony of 100 cortical regions separated these groups with 80% accuracy. On an independent sample (64 "SuperAgers" and 79 Alzheimer's patients), accuracy was 71%.
In "SuperAgers," the frontal and parietal regions responsible for attention and self-control are more strongly synchronized. In Alzheimer's patients, activity is elevated in the right insula (a deep region associated with emotions and bodily sensations) and in the anterior parts of the temporal lobes. In older adults with average memory, the measures occupied an intermediate position.
The model was then applied to healthy older adults without cognitive impairment: 818 people in the main group and 497 in the replication group. About one in four (about 26%) had an activity pattern resembling the Alzheimer's one. At baseline, the two subgroups differed little on memory tests and everyday skills, although the "Alzheimer's-like" one was on average older. But over years of follow-up, its overall clinical rating, memory, attention and executive functions deteriorated faster, and the share of those who progressed to mild cognitive impairment (an intermediate stage) or dementia was higher. The difference persisted after adjustment for age, baseline scores and amyloid and tau levels. Such people also accumulated amyloid faster, especially in the frontal and temporal regions.
The authors emphasize that this is an observational study, so it does not prove that changes in brain connections cause the disease. Correlation, a statistical association, does not by itself indicate causation. Accuracy of 71% on an independent sample means the signal is reproducible at the group level, but the spread is too large to use the method as an individual diagnosis. Limitations: the model relies only on resting-state MRI, without accounting for white matter structure and vascular changes, and the definition of "SuperAgers" was adapted for age 75.
Nevertheless, according to independent experts, the work shows that functional brain connections carry warning signs that standard tests miss. In the future, such a measure could help select people into risk groups for preventive drug trials.