Removal of Artifacts In EEG Signals Using Sign Based LMS Adaptive Filtering Techniques

The electroencephalogram (EEG) has been used for many years to help doctors diagnose and assess conditions affecting the brain. An electroencephalogram is a specialized diagnostic tool for monitoring brain activity electrically. Human brain anomalies and responses to stimuli can be deduced from EEG recordings. The presence of extraneous signals, known as artefacts, can cause distortion in EEG recordings. Since these aberrations introduce spikes that can be mistaken for neurological rhythms and since the properties of practically all biomedical signals vary with time, it is challenging to evaluate the EEG. Therefore, in order to guarantee an accurate analysis and diagnosis, noise and unwanted signals must be removed from the EEG. The identification of artefacts, or any potential deviation that has a source other than the brain, is a critical issue in this field. Artifacts can originate from a variety of sources, including the undesired movement of the eyes and eyelids, breathing, bodily movements, power line interference, muscular, cardiac activity, etc. The computational complexity is the primary subject of extensive research into adaptive filtering algorithms. The goal is to create a variety of novel ANC's for denoising EEG signals while taking into account computing complexity, Signal-to-Noise Ratio, Mis-adjustment, and convergence difficulties. This is done by taking into account three types of adaptive algorithms: The Least Mean Square (LMS) Adaptive algorithm and its several weight update variations.