EDGeNet: Electroencephalography Denoising Efficient Network for Fast Artifact Removal.

Area of Science:

• Neuroscience
• Signal Processing
• Artificial Intelligence

Background:

• Electroencephalography (EEG) is crucial for brain activity analysis but suffers from artifacts.
• Traditional artifact removal methods are computationally intensive and lack real-time efficiency.
• Automated, efficient artifact removal is needed for advanced EEG applications.

Purpose of the Study:

• To develop a deep learning framework for automated EEG denoising and artifact removal.
• To ensure the framework is efficient for real-time deployment.
• To evaluate the model's performance using standard metrics.

Main Methods:

• A deep learning-based framework was designed for simultaneous EEG denoising and artifact removal.
• The model was evaluated using metrics including relative-root-mean-square error (RRMSE), structural similarity index measure (SSIM), and correlation (CC).
• Performance was compared against state-of-the-art methods.

Main Results:

• The model achieved average temporal and spectral RRMSE of 0.214 and 0.217, respectively.
• Average SSIM and CC were recorded at 0.964 and 0.963, demonstrating high signal fidelity.
• The proposed model uses 295 times fewer parameters than prior methods while effectively removing various artifacts.

Conclusions:

• The deep learning framework offers an efficient and effective solution for real-time EEG artifact removal.
• The model's reduced parameter count and high performance indicate its potential for practical clinical and research applications.
• The developed framework advances automated signal processing in neuroscience.