Feature Extraction of Electroencephalogram (EEG) Signal - A Review

This paper presents a review on signal analysis method for feature extraction of electroencephalogram (EEG) signal. It is an important aspect in signal processing as the result obtained will be used for signal classification. A good technique for feature extraction is necessary in order to achieve robust classification of signal. Considering several techniques have been implemented for extracting features in EEG signal, we only highlight the most commonly used for schizophrenia. The techniques are Hilbert-Huang transform, Principal Component Analysis, Independent Component Analysis and Local Discriminant Bases. Despite of their drawbacks, they can be applied which depends on the aim of a research, parameters and the data collected. Nevertheless, these techniques can be modified so that the new algorithm can overcome the shortcomings of original algorithm or algorithm beforehand. The modified Local Discriminant Bases algorithm is introduced in the present paper as another powerful adaptive feature extraction technique for EEG signal which is not reported elsewhere in investigating schizophrenia.