A Frequency Division Based Approach for EMG Artifact Minimization from Single Channel EEG

The recorded electroencephalogram (EEG) are often contaminated by electromyogenic (EMG) artifacts. This inhibits the further analysis/processing of EEG signal, for corrupting EMG signal has an overlapping frequency spectrum and a higher amplitude than the original EEG signal. This paper proposes an algorithm to retrieve EEG signal from corrupted EEG by dividing it into several frequency bandwidths. EEMD is used to decompose frequency divided EEG. From the decomposed mode functions, the muscle artifact components are identified and discarded. This approach effectively helps in minimizing the muscle artifact and retaining the cerebral activities. The performance is evaluated using parameters such as relative root mean square error (RRMSE) and correlation coefficient (CC) by implementing the proposed algorithm on simulated dataset.