Overview
- The Stanford‑led study, published Wednesday, Nov. 4 in Science Advances, trained the Alzheimer’s Resilience‑Vulnerability Connectome (ARC) on resting‑state fMRI from 180 SuperAgers and 110 people with Alzheimer’s and distinguished the groups with roughly 85% accuracy.
- Researchers then applied ARC to more than 1,300 cognitively unimpaired adults aged 55 to 70 and found that higher, more Alzheimer’s‑like scores predicted faster cognitive decline, greater amyloid‑β accumulation, and higher rates of conversion to mild cognitive impairment or Alzheimer’s over follow‑up periods of about five to 12 years.
- The ARC score matched or outperformed standard risk measures such as demographics, APOE genetics, and amyloid/tau biomarkers, and combining ARC with existing markers improved prediction of who will decline.
- Authors caution the ARC score is not a diagnostic test and say further replication is needed while they plan intervention trials, including cognitive training and targeted transcranial magnetic stimulation, to test whether shifting connectivity can change outcomes.
- The study used SuperAgers as a resilience benchmark to reveal circuit‑level patterns of vulnerability and resilience in the functional connectome, offering a targetable signal that could guide earlier prevention but requiring standardization, ethical risk‑disclosure rules, and confirmatory trials before clinical use.