Neural Synchronization Landscapes Reveal Altered Structure–Function Coupling in Neurodegenerative Diseases

Tianmu Hu, Jiaqi Ding, Guorong Wu, Tingting Dan

Keywords: Neurodegenerative diseases, Kuramoto model, SC-FC coupling, fMRI

Abstract:

Modern neuroimaging technologies enable the study of structural connectivity (SC) and functional connectivity (FC) in vivo. However, due to the distinct biological underpinnings of SC and FC, understanding how the altered coupling mechanism is associated with the progression of neurodegeneration remains a challenge in the neuroscience field. Drawing inspiration from the rich neural dynamics captured by the Kuramoto model, we introduce a brain-inspired neural network, coined KM-Net, to explain cognitive behavior from neuroimages, which is rooted in the neuroscience principle of oscillatory synchronization.
Given that disrupted synchronization in neural oscillations closely underlines neurodegenerative diseases, we further extend KM-Net to an explainable deep model in the arena of disease early diagnosis.
By capturing the emergence of synchronized FC patterns from the underlying SC architecture, our approach provides a biologically informed representation for the dynamical system of functional fluctuations. We validate our novel computational framework through extensive experiments on a diverse set of neuroimaging cohorts, demonstrating its effectiveness in characterizing cognition-relevant brain fingerprint and disease-specific imaging biomarkers.
Together, promising results indicate the potential of neural synchronization modeling for advancing computational neuroscience and improving the understanding of neurodegenerative diseases.

Primary Area: applications to neuroscience & cognitive science

Submission Number: 13658