Removal of ocular artifacts from the electroencephalogram signal flow using median filtering

Abstract

An electroencephalogram is the easiest way to record brain activity. In medicine, an electroencephalogram is an indispensable tool for diagnosing, together with the biochemical analysis, brain and neurodegenerative diseases, mental disorders and can be used for different tasks like developing complex diagnostic systems and the operation of brain-computer interfaces. However, interference appears when recording brain activity signals, called artifacts, which leads to incorrect reading of the electroencephalogram results. Removing an artifact is a complex process. It is important to correctly determine its type and location for correct removal from the flow of electroencephalogram signals. Otherwise, some of the necessary information may be lost. This paper proposes using the median filtering method to remove ocular artifacts due to its ability to suppress weakly correlated noise. The results of mathematical modeling showed that the removal of ocular artifacts by the proposed method gives a better result than the method based on the use of the Daubechies Wavelet. These results will allow median filtering to preprocess the complex diagnostic systems' signals and brain-computer interface signals.