Approximately 30% of older adults exhibit the neuropathological features of Alzheimer’s disease without signs of cognitive impairment. Yet, little is known about the genetic factors that allow these potentially resilient individuals to remain cognitively unimpaired in the face of substantial neuropathology.

This study explored the genetic variants associated with resilience to Alzheimer’s disease, leveraging cognitive and pathology data harmonized across two datasets with autopsy measures of amyloid (ROS/MAP and ACT) and two datasets with amyloid PET (A4 and ADNI). Resilience phenotypes were calculated using a latent variable modeling approach, modeling better than expected cognition for a given level of amyloid pathology, following the method originally published in Hohman, et al, 2017 (PMID: 27743375). Cognitive resilience accounted for amyloid levels, age, and sex. Global cognitive resilience further incorporated education with cognitive resilience. Genotype data in each study underwent standard quality control and imputation onto the HRC reference panel (genome build 37). GWAS on both resilience phenotypes were performed including all samples as well as subsetted to only cognitively normal samples to evaluate resilience at the preclinical stage of disease.