Synthetic ALS-EEG Data Augmentation for ALS Diagnosis Using Conditional WGAN with Weight Clipping
This work addresses data scarcity and class imbalance for ALS diagnosis using EEG, but it is incremental as it applies an existing generative method to a specific medical domain without demonstrating concrete performance gains.
The paper tackled the problem of scarce and imbalanced EEG data for Amyotrophic Lateral Sclerosis (ALS) diagnosis by generating synthetic ALS EEG signals using a Conditional Wasserstein Generative Adversarial Network (CWGAN), resulting in realistic synthetic samples that mimic real patterns and could improve classifier training.
Amyotrophic Lateral Sclerosis (ALS) is a rare neurodegenerative disease, and high-quality EEG data from ALS patients are scarce. This data scarcity, coupled with severe class imbalance between ALS and healthy control recordings, poses a challenge for training reliable machine learning classifiers. In this work, we address these issues by generating synthetic EEG signals for ALS patients using a Conditional Wasserstein Generative Adversarial Network (CWGAN). We train CWGAN on a private EEG dataset (ALS vs. non-ALS) to learn the distribution of ALS EEG signals and produce realistic synthetic samples. We preprocess and normalize EEG recordings, and train a CWGAN model to generate synthetic ALS signals. The CWGAN architecture and training routine are detailed, with key hyperparameters chosen for stable training. Qualitative evaluation of generated signals shows that they closely mimic real ALS EEG patterns. The CWGAN training converged with generator and discriminator loss curves stabilizing, indicating successful learning. The synthetic EEG signals appear realistic and have potential use as augmented data for training classifiers, helping to mitigate class imbalance and improve ALS detection accuracy. We discuss how this approach can facilitate data sharing and enhance diagnostic models.