Mitigation of ocular artifacts for EEG signal using improved earth worm optimization-based neural network and lifting...

Area of Science:

• Biomedical Engineering
• Signal Processing
• Neuroscience

Background:

• Electroencephalogram (EEG) signals are frequently corrupted by artifacts, particularly ocular artifacts, which are involuntary and difficult to eliminate.
• Existing methods for artifact removal often struggle with the complex overlap between ocular artifacts and true EEG signals.
• The presence of these artifacts significantly impacts the accuracy of EEG-based analyses and diagnoses.

Purpose of the Study:

• To develop and implement a novel, intelligent method for the effective detection and removal of ocular artifacts from EEG signals.
• To enhance the quality and reliability of EEG data for clinical and research applications.
• To propose an integrated system combining optimized artifact detection and removal techniques.

Main Methods:

• EEG signal decomposition using 5-level Discrete Wavelet Transform (DWT) and Empirical Mean Curve Decomposition (EMCD).
• Feature extraction including kurtosis, variance, Shannon's entropy, and first-order statistics for artifact detection.
• Machine learning-based detection using a Neural Network (NN) trained with an improved Earth Worm Optimization Algorithm (EWA), termed Dual Positioned Elitism-based EWA (DPE-EWA).
• Artifact removal employing optimized Lifting Wavelet Transform (LWT) with filter coefficients optimized by the DPE-EWA.

Main Results:

• The proposed method successfully detects and removes ocular artifacts, significantly improving EEG signal quality.
• The integration of DPE-EWA optimized NN and LWT demonstrates enhanced performance in artifact management.
• The developed model offers a robust solution for mitigating the impact of ocular artifacts on EEG data.

Conclusions:

• The proposed novel method provides a promising approach for accurate ocular artifact detection and effective removal in EEG signals.
• The optimized machine learning and wavelet transform techniques offer a significant advancement in EEG signal processing.
• This work contributes to improving the reliability of EEG data for various neuroscience and clinical applications.