A hybrid method for muscle artifact removal from EEG signals
Qiang Chen1, Yingying Li2, Xiaohui Yuan3
1Department of Biomedical Engineering, Hefei University of Technology, Hefei, Anhui 230009, China; Department of Computer Science and Engineering, University of North Texas, Denton, TX 76203, USA.
Background:
Electroencephalogram (EEG) signals may be contaminated with muscle artifacts that are usually difficult to be removed.
New Method:
In this article, a new hybrid method for suppressing muscle artifacts is proposed. Our method leverages variational mode decomposition (VMD) and canonical correlation analysis (CCA) algorithms. Each channel of EEG is decomposed into intrinsic mode functions (IMFs) with VMD to achieve an extended data set that contains more channels than the original data set. The potential artifact components are decomposed by CCA for further isolation.
Results:
The proposed method is evaluated with semi-simulation and real contaminated EEG signals. The results show that the performance of removing artifacts for VMD-CCA exceeds the comparison methods.
Comparison With Existing Methods:
Regardless of the number of EEG channels and the signal-to-noise ratio of the signal, the VMD-CCA approach is superior to the existing methods. As the number of EEG channels decreases, the average de-artifact effects of VMD-CCA and the comparison approaches are basically the same, but the randomness increases.
Conclusions:
The VMD-CCA method can effectively isolate muscle artifacts in EEG in case of multiple channels or few channels.
