Combining Residual U-Net and Data Augmentation for Dense Temporal Segmentation of Spike Wave Discharges in Single-Channel EEG

TL;DR

This study develops and enhances a deep learning model, AugUNet1D, for automated detection of spike-wave discharges in EEG, improving cross-subject generalization through residual connections and data augmentation, and compares it favorably to existing methods.

Contribution

The paper introduces AugUNet1D, a novel residual U-Net based model with data augmentation for better EEG spike detection across subjects, and demonstrates its superior performance over existing algorithms.

Findings

AugUNet1D outperforms other classifiers and the Twin Peaks algorithm.

Data augmentation improves cross-subject generalization.

Pretrained and untrained versions of AugUNet1D are publicly available.

Abstract

Manual annotation of spike-wave discharges (SWDs), the electrographic hallmark of absence seizures, is labor-intensive for long-term electroencephalography (EEG) monitoring studies. While machine learning approaches show promise for automated detection, they often struggle with cross-subject generalization due to high inter-individual variability in seizure morphology and signal characteristics. In this study we compare the performance of 15 machine learning classifiers on our own manually annotated dataset of 961 hours of EEG recordings from C3H/HeJ mice, including 22,637 labeled SWDs and find that a 1D U-Net performs the best. We then improve its performance by employing residual connections and data augmentation strategies combining amplitude scaling, Gaussian noise injection, and signal inversion during training to enhance cross-subject generalization. We also compare our method, named AugUNet1D, to a recently published time- and frequency-based algorithmic approach called "Twin Peaks" and show that AugUNet1D performs better on our dataset. AugUNet1D, pretrained on our manually annotated data or untrained, is made public for other users.