Journal
MEDICAL & BIOLOGICAL ENGINEERING & COMPUTING
Volume -, Issue -, Pages -Publisher
SPRINGER HEIDELBERG
DOI: 10.1007/s11517-023-02976-y
Keywords
Variational mode extraction (VME); Particle swarm optimization (PSO); Harris hawks optimization (HHO); Eye blink artifact; Regression analysis
Ask authors/readers for more resources
This research proposes an optimized VME-REG algorithm to detect eye blink artifacts in EEG recordings and tests it on multiple datasets, demonstrating improved performance compared to existing methods.
The electroencephalogram (EEG) recordings from the human brain are useful for detecting various brain syndromes. These recordings are typically contaminated by high amplitude eye blink artifacts, which leads to deliberate misinterpretation of the EEG signal. Recently, variational mode extraction (VME) has been used to detect eye blink artifacts. But, the VME performance is impacted by the balancing parameter and center frequency selection. Therefore, this research uses two metaheuristic algorithms, particle swarm optimization and Harris hawks optimization, to determine the optimal set of the VME parameters. In the proposed method, the optimized VME (OVME) extracts the desired mode to locate the eye blink artifactual intervals. Then, the regression analysis (REG) filters the identified artifactual intervals from short EEG data segments. The significance of the proposed OVME-REG algorithm is that it is adequate for determining the optimum values of the VME algorithm. The analysis is carried out on the CHB-MIT Scalp EEG, BCI Competition, and EEG motor movement/imagery datasets. The proposed OVME-REG method provides an improved performance for suppressing single and repeated eye blink artifacts as compared to the current approaches in terms of (a) high correlation coefficient (93.08%, 87.3%, 82.17%), respectively, (b) low value of RRMSE (0.379, 0.506, 0.502), respectively, (c) high SSIM (0.892, 0.842, 0.694), and (d) low computation time and better preservation of the EEG data.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available