UCLA

Electroencephalography (EEG) is a technique that is used to non-invasively monitor the electrical activity of the brain. Although the EEG device is supposed to record only cerebral activity, it also records artifacts, which are recorded activities that are not of cerebral origin. These artifacts include motion artifacts and stimulation artifacts. Artifacts corrupt the EEG signals and prevent the device from being used successfully. In order to remove the artifacts in real-time, an artifact cancellation system that utilizes adaptive filters is proposed. Adaptive filters can self-adjust the transfer function, giving them the ability to self learn and change filter parameters to adapt to different signal characteristics. Multiple adaptive filter algorithms were tested in the artifact cancellation system in Matlab and Simulink, including Least Mean Squares (LMS) algorithms and Recursive Least Squares (RLS) algorithms. The RLS algorithm has a faster convergence time but is more computationally demanding than the LMS algorithm.