Uncovering Structural-Functional Coupling Alterations for Alzheimer's Diseases

Tingting Dan, Guorong Wu

Keywords: Brain structure-functional coupling, Imaging biomarkers, Alzheimer's diseases.

TL;DR: Tailored a physics-guided graph neural network (GNN), which can predict self-organized functional fluctuations and generate a novel biomarker for early detection of neurodegeneration through altered SC-FC coupling.

Abstract: A confluence of neuroscience and clinical studies suggests that disrupted structural connectivity (SC) and functional connectivity (FC) in the brain is an early signs of neurodegenerative diseases. However, current methods lack the neuroscience foundation to understand how these altered coupling mechanisms contribute to cognitive decline. To address this issue, we spotlight a neural oscillation model that characterizes the behavior of neural oscillators coupled via nerve fibers throughout the brain. Tailored a physics-guided graph neural network (GNN), which can predict self-organized functional fluctuations and generate a novel biomarker for early detection of neurodegeneration through altered SC-FC coupling. Our method outperforms conventional coupling methods, providing higher accuracy and revealing the mechanistic role of coupling alterations in disease progression. We evaluate the biomarker using the ADNI dataset for Alzheimer's disease diagnosis.