Uncovering the Heterogeneity of Neurodegeneration Trajectories in Alzheimer's Disease Using a Network Guided Reaction-Diffusion Model

项目总结/文摘

Project Summary/Abstract Alzheimer’s disease (AD) is a heterogeneous, multifactorial neurodegenerative disorder. Due to the multiplicity of clinical symptoms, standard neuropsychological assessments inadequately reflect the underlying pathophysiological mechanisms, which renders a significant gap between neurobiological examinations of AD pathology and clinical diagnoses. Mounting evidence shows that AD is caused by the build-up of two abnormal proteins, beta-amyloid and tau. Over time, these AD-related neuropathological burdens begin to spread throughout the brain, which results in the characteristic progression of symptoms in AD. Although striking efforts have been made to investigate the neurobiological factors behind the acquisition of amyloid (A), protein tau (T), and neurodegeneration [N] biomarkers, a system-level understanding of how these neuropathological burdens promote neurodegeneration and why AD exhibits characteristic progression is still largely elusive. In this study, we will combine the power of systems biology and network neuroscience to disentangle the heterogeneous trajectories of cognitive decline in AD population by understanding the dynamic interaction and diffusion process of AT[N] biomarkers from an unprecedented amount of longitudinal neuroimaging data. The backbone of this project is our recently developed network guided reaction-diffusion model that characterizes not only the interaction of AT[N] biomarkers at each brain region but also their propagation pattern across the brain networks using PDEs (partial differential equations). Given its promising results in predicting the evolution of AT[N] biomarkers, we will further develop our current PDE-based model by incorporating spatiotemporal-adaptive mechanistic pathways of AT[N] biomarkers. Then, we will investigate the system behaviors that steer the trajectory of cognitive decline in Aim 1. After that, we will develop a novel deep learning approach to stratify aging brains into a set of fine-grained categories (aka. subtypes) with distinct neurobiological underpinnings, where individuals within the same subtype are expected to have very similar trajectories of cognitive decline. We will evaluate the novel population stratification result using the longitudinal imaging data from the ADNI database in Aim 2. The success of this project will allow us to have a new understanding of the neurodegeneration process in the cognitive continuum spectrum. This is an important step because slowing down this spread at an early stage might prevent or halt the symptoms of AD.