Multi-class Classification of EEG Spectral Data for Artifact Detection.

Electroencephalographic signals are known to be highly sensitive to various types of noise originating from external and internal sources. External sources are usually related to experiment conditions whereas internal artifacts are usually generated by the persons examined. Internal artifacts, in contrast to external ones, are characterised by non-stationarity, which results in higher detection complexity. Typical internal artifacts are related to eye blinking, eye movement and muscle activity.