The electroencephalograms (EEG) are design to measure the electrical activity of the brain, commonly called brain waves. The brain waves are the summation of nature depolarizations in the brain due to stimuli from the five senses as well as from the thought process.

In EEG recordings the eye-ball movements are a notorious “artefact” making automated analysis difficult. The “signature” of eye-ball movement and other artefacts can nevertheless be readily recognised by nurses and doctors. The project is to try to understand how a classification algorithm can be developed that would automatically identify periods where EEG recordings are dominated by artefacts and which periods of EEG recording represent valid data for diagnostic purposes.

The project has the following aspects-literature survey on identification and removal of eye-ball motion and other artefacts in EEG recordings. – Implementation of several published algorithms in MATLAB. – Evaluation of these algorithms against previously unpublished data. The ideal outcome would be a novel suite of algorithms that would work more reliably compared to existing methods, and perhaps could be as reliable as the experts.

Chang-Hsun Alex Yang, Jian Peng Gao, Andrea Varsavsky, Jason Yang Yang, Professor Iven Mareels