Statistical considerations for assessing cognition and neuropathology associations in preclinical Alzheimer's disease

ABSTRACT Analysis of the associations between the neuropathology of Alzheimer's disease (AD) and cognition has become a major area of investigation as both basic and clinical researchers have turned their attention toward identifying the factors underlying the onset of preclinical AD. Here we provide a conceptual overview of statistical approaches for analyzing associations between cognition and AD neuropathology in the context of the prodromal AD. The review will discuss a variety of statistical approaches, their application to various clinical pathological variables, and research questions, as well as the importance of accounting for and including interaction terms in statistical models. The overview presented here will introduce data analysts and statisticians to the nomenclature of AD neuropathology and provide relevant background information regarding the nature of cognitive and neuropathological data generated in the investigation of preclinical AD. In addition, we will introduce a number of statistical approaches that researchers who specialize in AD neuropathology may utilize in their studies. For both audiences, this review will provide an applied statistical framework to draw from for future research.