Understanding the Spreading Mechanism of Tau Propagation in Alzheimers Disease Through a Multi-Layer Transport

Alzheimer's disease (AD) is characterized by the progressive spread of tau protein aggregates throughout the brain. Significant advances have been made in modeling this process using a potential energy transport framework, where the overarching goal is to elucidate the “prion-like” pathways of tau propagation from longitudinal tau-PET scans. However, current methods are limited to single-layer models, where tau spreading is constrained to either white matter fiber pathways (“air-based”) or the cortical folding patterns of the brain (“road-based”). This simplification overlooks the possibility that both air-based and road-based transport mechanisms might collaboratively drive the whole-brain tau propagation. To fill this gap, we present a novel multi-layer neural transport equation that models the simultaneous spreading of tau aggregates through both the brain's network and its cortical surface. Specifically, we incorporate the concept of optimal control to shape the interaction between air-based and road-based tau propagation, yielding a system-level perspective on the selective mechanism that steers the multi-layer tau spreading. We evaluate the clinical utility of our multi-layer model on the ADNI dataset by predicting future tau accumulation and assessing AD risk. Experimental results demonstrate that our approach outperforms other computational methods in both tau accumulation prediction and disease diagnosis. This highlights the potential of our model for future advancements in early diagnosis and personalized therapeutics for AD.