Dimension Reduction in Observational Studies of Alzheimer’s Disease
Sarah Weinstein, MS
December 13, 2022
Alzheimer’s disease (AD) is a progressive brain disorder, affecting memory and cognitive function in over 6 million Americans. With an aging population, the public health and economic burden of AD continues to grow, leading to devastating consequences for patients and caregivers. In recent years, a number of large-scale research initiatives have led to new insights into the etiology and progression of AD by using neuroimaging. But some key statistical challenges remain in neuroimaging-based association studies and predictive models of AD. Specifically, given that these studies are often observational by design, confounders and other nuisance variables can pose threats to their generalizability and interpretability. In this talk, I will discuss recent work involving dimension reduction with built-in adjustment for nuisance variables to address these issues.