B1331
Title: Modeling longitudinal trajectories of neuropsychological and neuroimaging brain changes
Authors: John Kornak - University of California, San Francisco (United States) [presenting]
Abstract: The hypothetical 2010 Jack model attempts to describe the timeline for which different biomarkers change in Alzheimer's disease and has sparked much discussion and subsequent research into disease trajectory modelling. Understanding this temporal ordering of effects in dementia and other neurological diseases/illnesses would have major benefits for both individual-level prediction and clinical trial design. Some Bayesian nonlinear mixed effects methods (with inhomogeneous variance) are presented aimed at estimating normalized cognitive test scores and imaging measures for individuals that appropriately account for demographic and other factors. These normalized scores are subsequently used in an application that models the temporal path of brain biomarker changes (cognition, imaging and otherwise) in frontotemporal dementia, the goal being to determine potential differences in trajectories across genetic subtypes.