Adaptive Noise Cancellation for the Electroencephalogram

Electroencephalography is a technique for recording the brain’s electrical activity – the electroencephalogram or EEG. It is an important procedure in the diagnosis of several brain disorders, as well being a valuable physiological tool for studies of normal brain function. However, the EEG is often contaminated by numerous artefacts such as eye-blinks, muscle activities, and eye-movements. This paper describes the Emotiv EEG Neuroheadset to record EEG time-series ensembles and correct the inevitable ocular artefact, the eye-blink. In order to classify and remove eye-blink artefacts from the EEG, a multilayer perceptron, feed-forward artificial neural network was used. In MATLAB, an multilayer perceptron network was trained to classify and remove the eye-blink artefact. Preliminary results show that eye-blink artefacts can be substantially reduced in the EEG when trained and used on a single subject. The generalised case to eyeblink removal is also addressed.