EEG signal analysis and artifact removal by wavelet transform

Abstract

Electroencephalogram (EEG) is a non-invasive method to collect brain signals from human’s scalp. EEG signals are located in low frequency range and relatively small. The amplitude of these signals are approximately 50μV with the maximum amplitude is about 100μV. Therefore, there are number of sources such as power line, EOG or ECG can extremely interfere EEG signals. Detection and elimination of artifacts plays an important role to acquire clean EEG signals to analyze and detect brain activities. Besides, the extraction of important components in recorded EEG required fast and reliable algorithm to process mix of data set. In this paper, we will demonstrate EEG acquisition from EEG Exea Ultra system of Bitmed, analyze and compare signals from volunteers in relaxation mode and contaminated EEG signals with eye blinks. We then design filters to remove powerlines and baseline noise from acquired signals. With the aim to assess the feasibility of Wavelet transform technique to identify feature in recorded EEG signals, we carried out Wavelet transform and applied threshold method to detect and remove artifacts in EEG signals with eye blinks. We achieved PSNR of original signals and wavelet filtered signal that approximately 17,7810 dB. Our preliminary results show that wavelet can be utilized as automatically detection tools for artifacts and event-related potentials and in applications require real-time processing of EEG signals.