Abstract
Cognitive aging is characterized by normative change involving multiple neuropsychological domains. For some individuals, age-related decline crosses a threshold signaling the beginning of pathological decline. A wealth of potential predictors of negative cognitive aging have been proposed, including cerebrospinal fluid (CSF) and blood-based biomarkers of known neurological pathologies, novel biomarkers of the central and peripheral nervous systems, structural and functional magnetic resonance imaging, diffusion tensor imaging, arterial spin labeling, and a range of cardiovascular, metabolic, and neuroinflammatory indicators. This study sought to integrate the wide number of biometrics available by leveraging advancements in Bayesian machine learning for the purpose of identifying the most salient predictors of cognitive trajectories across multiple neuropsychological domains. Results suggest baseline neuropsychological functioning, global hypometabolism, morphometry, and CSF biomarkers are consistent predictors of decline. Additionally regional volumetrics and indicators of pathology burden demonstrated specificity to trajectories of memory and executive functions.