Understanding the Selective Mechanism of Tau Propagation in Alzheimer's Disease Through Multi-layer Spreading Pathways

Abstract Alzheimer’s disease (AD) is characterized by the propagation of tau aggregates throughout the brain. Tremendous efforts have been made to the development of tau biomarker, aimed at advancing our current understanding of AD etiology and facilitating early diagnosis in routine clinical practice. Since current pathology imaging technologies only provide a spatial mapping of tau pathology, many biomarker studies have to link stereotypical patterns of tau accumulation to the complex pathophysiological mechanism of disease progression. Recently, we have made significant progress to characterize dynamic process of tau spreading in a principled potential energy transport model that allows us to uncover the region-to-region spreading pathways of tau aggregates from the longitudinal tau-PET scans. However, our current method is limited to a single-layer scenario to the extent that the spreading pathways follow either the wiring topology of white matter fiber bundles (referred to as "air-based") or the folding patterns on the brain cortex (referred to as "road-based"), despite the fact that air-based and road-based transport might collaboratively contribute to the whole-brain tau propagation. To address this limitation, we will develop a multi-layer neural transport equation to jointly model the spreading of tau aggregates throughout the brain network and along the brain cortex in Aim 1. Furthermore, we introduce the notion of optimal control to characterize the interaction between air-based and road-based tau propagation, which allows us to uncover a system-level understanding of the selective mechanism in multi-layer tau propagation in Aim 2. By capitalizing on large scale of existing neuroimaging data, we will explore a set of open neuroscience questions regarding tau propagation in Aim 2, including dynamic coupling patterns, region vulnerability, and system criticality. Meanwhile, we will evaluate the clinic value of our multi-layer model in predicting the future tau accumulation and the risk of developing AD. The success of this project has the potential to offer valuable insights into AD etiology, thereby enhancing our understanding of AD progression and refining treatment strategies. Moreover, the computational tools can be readily applied to other biomarker studies in neurodegenerative diseases. This initiative lays the groundwork for future research into advanced diagnosis and tailored personalized treatment for AD.