New Method Places Brains on SuperAger-Alzheimer's Spectrum

A new framework can identify features of brain function that are linked to Alzheimer's disease (AD) before people show memory loss. Unlike previous efforts, which have compared AD patients with cognitively unimpaired adults, the new approach instead compares AD patients with SuperAgers – people who retain unusually good cognition into late life – to set a more accurate endpoint on the continuum of cognitive aging possibilities. "These findings support the potential utility of [the tool] for early risk stratification among cognitively unimpaired older adults," Kanhao Zhao and colleagues write. Scientists have linked AD onset to the proteins amyloid-β and tau, as well as genetic risk factors. They have also discovered a subset of people called SuperAgers, whose cognitive skills remain relatively undiminished well into old age. Zhao et al. asked: how do the brains of each cohort differ? Can defining those as-yet unexplored differences help scientists identify AD before clinical symptoms appear? To answer these questions, the team developed the Alzheimer's Resilience-Vulnerability Connectome (ARC) machine learning framework. They trained and evaluated the model on resting-state fMRI data from 180 SuperAgers and 110 patients already diagnosed with AD, all at least 75 years old. They further demonstrated that the model generalized to an independent replication cohort of 143 individuals. Then, they applied ARC to examine the brains of 818 people in a new, slightly younger cohort with no signs of cognitive impairment – none of whom were already classified as a SuperAger. By measuring relevant features of brain function, ARC successfully identified where participants stood on a continuum of SuperAger to AD diagnosis. It showed similar performance in an independent cohort of 497 people. Within 5 and 15 years, depending on available data, people with brain function patterns that resembled AD showed faster cognitive decline, greater amyloid-β accumulation, and a higher likelihood of progression to clinical AD.