Lightweight Transformer for EEG Classification via Balanced Signed Graph Algorithm Unrolling
This work addresses EEG classification for epilepsy diagnosis, offering a more interpretable and parameter-efficient approach, though it appears incremental as it builds on existing graph-based and transformer methods.
The authors tackled the problem of classifying EEG signals to differentiate epilepsy patients from healthy subjects by building lightweight transformer-like neural nets via unrolling a spectral denoising algorithm on balanced signed graphs. Their method achieved classification performance comparable to representative deep learning schemes while using dramatically fewer parameters.
Samples of brain signals collected by EEG sensors have inherent anti-correlations that are well modeled by negative edges in a finite graph. To differentiate epilepsy patients from healthy subjects using collected EEG signals, we build lightweight and interpretable transformer-like neural nets by unrolling a spectral denoising algorithm for signals on a balanced signed graph -- graph with no cycles of odd number of negative edges. A balanced signed graph has well-defined frequencies that map to a corresponding positive graph via similarity transform of the graph Laplacian matrices. We implement an ideal low-pass filter efficiently on the mapped positive graph via Lanczos approximation, where the optimal cutoff frequency is learned from data. Given that two balanced signed graph denoisers learn posterior probabilities of two different signal classes during training, we evaluate their reconstruction errors for binary classification of EEG signals. Experiments show that our method achieves classification performance comparable to representative deep learning schemes, while employing dramatically fewer parameters.