PROCESSING OF ELECTROENCEPHALOGRAM SIGNALS CONSIDERING THE DIPOLE NATURE OF BRAIN SIGNALS AND PROSPECTS OF NEURAL ODE MODELING

Abstract

This study proposes an approach to eye state classification (open/closed) based on 64-channel EEG data using independent component analysis (ICA) decomposition with component validation through equivalent dipole fitting. An original loss function with L2 regularization and ellipsoidal anatomical constraint based on a three-dimensional brain model was developed for dipole fitting. Among the ICA components, sources with characteristic alpha, mu, beta, and gamma rhythms were identified, confirming the physiological interpretability of the method. Neural network classification achieved 97% accuracy on the training set and 90% on the test set. Additionally, the perspectives of applying Neural Ordinary Differential Equations (Neural ODE) for modeling cleaned ICA signals are discussed as a promising direction for future research.