Efficient EEG motion artifact elimination framework for ambulatory epileptic seizure detection application

Area of Science:

• Neuroscience
• Signal Processing
• Biomedical Engineering

Background:

• Motion artifacts are a significant challenge in ambulatory electroencephalogram (EEG) monitoring, often degrading signal quality and hindering accurate neurological assessments.
• Effective removal of motion artifacts is crucial for reliable seizure detection and analysis of brain activity.

Purpose of the Study:

• To develop and evaluate a novel artifact removal technique for ambulatory EEG data.
• To enhance the accuracy and reliability of seizure detection by improving EEG signal quality.

Main Methods:

• A two-level Singular Spectrum Analysis (SSA) decomposition was combined with a Relative Total Variation (RTV) filter for motion artifact removal.
• The proposed method decomposes EEG signals into frequency bands, filters artifact-corrupted sub-bands using RTV, and reintegrates denoised components.

Main Results:

• The proposed algorithm demonstrated superior performance over existing methods in metrics such as ΔSNR, η, MAE, and PSNR.
• The technique effectively removed motion artifacts while preserving essential neurological information in the EEG data.
• Pre-processing EEG data with the proposed method led to substantial improvements in seizure detection accuracy and reliability.

Conclusions:

• The developed SSA and RTV-based framework offers an effective solution for motion artifact removal in ambulatory EEG monitoring.
• This noise reduction technique significantly enhances the utility of EEG data for both signal analysis and clinical applications like seizure detection.