Neural operator-based digital twins for modeling amyloid-$β$ and tau propagation and treatment optimization in Alzheimer's disease
📄 arXiv:2606.25185 · 📥 PDF · 2026-06-23 · math-ph
Authors: Xiaofeng Xu [arXiv · scholar] , Tingting Dan [arXiv · scholar] , Zifan Zhou [arXiv · scholar] , Bin Li [arXiv · scholar] , Guorong Wu [arXiv · scholar] , Wenrui Hao [arXiv · scholar]
🕰 Orloj analysis
Tento článek představuje datově řízený rámec pro konstrukci digitálních dvojčat pro modelování šíření amyloid-β a tau proteinů u Alzheimerovy choroby pomocí reakčně-difúzní dynamiky. Využívá učení operátorů k odvození neznámých nelineárních mechanismů z klinických dat a následně formuluje problém optimálního řízení pro personalizované terapeutické strategie.
💡 Práce nabízí inovativní kombinaci strojového učení a matematického modelování pro personalizovanou medicínu, což má vysokou potenciální hodnotu pro diagnostiku a léčbu neurodegenerativních onemocnění.
Categories:
INF-6
EXP-7
MET-4
MET-1
MET-2
INF-5
✓ falsifiable, modest_claims
⚠ Sparse, noisy, and heterogeneous clinical data present a significant challenge for robust inference., Validation of inferred nonlinear aggregation mechanisms is crucial and not fully detailed in the abstract., Lack of explicit mention of code or data availability.
📄 Abstract
Accurately predicting the spatiotemporal evolution of amyloid-$β$ and tau proteins at the individual level is critical for improving the diagnosis and treatment of Alzheimer's disease. We consider the problem of constructing patient-specific digital twins that model the propagation of these biomarkers on the cortical surface using reaction--diffusion dynamics. A major challenge is that the underlying nonlinear aggregation mechanisms are unknown and must be inferred from sparse, noisy, and heterogeneous longitudinal PET imaging data. To address this, we develop a data-driven framework that learns biomarker dynamics directly from clinical observations. The approach combines operator learning with reduced-order representations to infer governing equations of disease progression from data. Using this framework, we achieve predictive accuracies of 87\% for amyloid-$β$ and 81\% for tau. Building on the learned dynamics, we further formulate a PDE-constrained optimal control problem to design personalized therapeutic strategies that regulate pathological protein propagation. By integrating data-driven dynamical modeling with treatment optimization, the proposed digital twin framework provides an interpretable and predictive platform for understanding disease progression and enabling precision interventions in neurodegenerative disorders.