Detection and Removal of EOG Artifact from EEG Signal using Fuzzy Logic and Wavelet Transform

Abstract

The EEG signals (electroencephalogram) are one of the most used biosignals in diagnosis of neurological disorders and technology like brain computer interface (BCI). The main problem in the use of EEG signals is that are often contaminated by physiological and non-physiological artifacts. EOG (electrooculogram) artifact is one of the first origins of artifacts in EEG recordings. It severely affects the EEG signal and modifies the shape of the neural activity that drive the BCI system affecting then its accuracy and performance. This paper presents an original algorithm to detect and remove EOG artifacts from EEG signals. In the proposed method, the statistical parameters extracted from the signal are used as the inputs of a fuzzy inference system to take a decision if an epoch of EEG is contaminated by the EOG or not. Thereafter, the level-5 stationary wavelet transform (SWT) is applied on the detected contaminated epochs to remove the artifacts without bringing much distortion to the neural activity. The performance of the developed algorithm is demonstrated in two scenarios: semi-simulated artifactual EEG data and fully real artifactual EEG data. The presented results show that our algorithm outperforms some existing algorithms.