KNOWLEDGE-BASED ENHANCEMENT OF HUMAN EEG SIGNALS

The human electroencephalogram (EEG) contains useful diagnostic information on a variety of neurological disorders. However, like all biomedical signals, the EEG is very susceptible to a variety of large-signal contaminations or artefacts which reduce its clinical usefulness. The paper discusses the use of knowledge-based techniques to overcome the limitations of an adaptive signal-processing method developed for real-time ocular artefact removal. Knowledge-based techniques are used to recognise and subsequently classify the pathological waves and ocular artefacts, based on a knowledge of their characteristics and the heuristics used by EEG experts. This makes it possible to remove the artefacts from the EEG signal only when it is necessary and with minimal distortion of any diagnostic information in the EEG. Preliminary results are presented to illustrate the advantages of the new approach.