ODEBrain: Continuous-Time EEG Graph for Modeling Dynamic Brain Networks
This work is significant for neuroscientific research and clinical applications by providing a more accurate and robust method for modeling continuous brain dynamics from EEG data, which could lead to better understanding of brain function and improved diagnostic tools.
The paper addresses the challenge of modeling continuous brain dynamics from EEG data, which conventional methods discretize, leading to prediction errors and an inability to capture instantaneous, nonlinear characteristics. ODEBRAIN, a Neural ODE-based framework, is proposed to model continuous latent dynamics by integrating spatio-temporal-frequency features into spectral graph nodes. This approach significantly improves EEG dynamics forecasting, demonstrating enhanced robustness and generalization.
Modeling neural population dynamics is crucial for foundational neuroscientific research and various clinical applications. Conventional latent variable methods typically model continuous brain dynamics through discretizing time with recurrent architecture, which necessarily results in compounded cumulative prediction errors and failure of capturing instantaneous, nonlinear characteristics of EEGs. We propose ODEBRAIN, a Neural ODE latent dynamic forecasting framework to overcome these challenges by integrating spatio-temporal-frequency features into spectral graph nodes, followed by a Neural ODE modeling the continuous latent dynamics. Our design ensures that latent representations can capture stochastic variations of complex brain states at any given time point. Extensive experiments verify that ODEBRAIN can improve significantly over existing methods in forecasting EEG dynamics with enhanced robustness and generalization capabilities.