Abstract:
|
Functional exposures are commonly measured longitudinally. But the current joint models involve only scalar variables. We propose a functional joint model (FJM) that consists of a longitudinal regression model with longitudinal functional exposure (high dimensional MRI) and a survival model for event time. We also develop methods for model-based personalized dynamic predictions of future outcome trajectories and risks of target events. Our proposed model is motivated and applied to the Alzheimer's Disease Neuroimaging Initiative (ADNI), a motivating clinical study to determine the relationships among the clinical, cognitive, imaging, genetic and biochemical biomarkers characteristics of the entire spectrum of Alzheimer's disease (AD).
|