Novel approach to remove Electrical Shift and Linear Trend artifact from single channel EEG

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electroencephalogram (EEG) signals are vital for Brain-Computer Interfacing (BCI).
• EEG signals are susceptible to artifacts, hindering accurate brain function assessment.
• Electrical Shift and Linear Trend (ESLT) artifacts are common issues in EEG data.

Purpose of the Study:

• To develop and evaluate a new algorithm for eliminating ESLT artifacts from EEG signals.
• To improve the quality of EEG data for BCI applications.
• To provide an effective artifact removal method without requiring initial calibration or large datasets.

Main Methods:

• The proposed algorithm utilizes Singular Spectrum Analysis (SSA) to decompose EEG signals into frequency components.
• Enhanced local Polynomial Approximation-based Total Variation (EPATV) filtering is applied to isolate and remove artifactual components.
• The denoised signal is reconstructed by combining the filtered components with the remaining signal parts.

Main Results:

• The algorithm demonstrated high effectiveness in removing ESLT artifacts across three databases.
• Achieved a highest averaged Correlation Coefficient (CC) of 0.9534.
• Exhibited a highest averaged Signal to Noise Ratio (SNR) of 10.2208dB, lowest averaged Relative Root Mean Square Error (RRMSE) of 0.2787, and averaged Mean Absolute Error (MAE) in the alpha band of 0.0557.

Conclusions:

• The developed algorithm is effective in extracting ESLT artifacts from EEG signals.
• The method offers a viable solution for real-time EEG artifact removal in BCI.
• The approach is efficient, requiring minimal calibration and data volume.