Several covariates were included to account for potential confounding factors. These include the site of imaging, as well as scanning date, head size, and resting state functional MRI motion, which was further split by site, as recommended by UKB. 24 We also incorporated covariates related to participant demographics and lifestyle commonly reported as risk factors for AD. These include genetic sex, body mass index (BMI), years of education, and income, as well as lifetime history of daily smoking, lifetime history of pack years smoked, drinking status, and past year alcohol consumption (drinks per week). 25 To account for population structure, we included the first ten genetic principal components (PCs) computed by UKB using principal component analysis (PCA). Total gray matter volume, which decreases with age, was also added as a covariate to allow us to assess differences in hippocampal volume relative to the rest of the brain.
