Single-trial classification of an EEG-based brain computer interface using the wavelet packet decomposition and cepstral analysis

[摘要] ENGLISH ABSTRACT: Brain-Computer Interface (BCI) monitors brain activity by using signalssuch as EEG, EcOG, and MEG, and attempts to bridge the gap betweenthoughts and actions by providing control to physical devices that range fromwheelchairs to computers. A crucial process for a BCI system is feature extraction,and many studies have been undertaken to find relevant informationfrom a set of input signals.This thesis investigated feature extraction from EEG signals using twodifferent approaches. Wavelet packet decomposition was used to extract informationfrom the signals in their frequency domain, and cepstral analysis wasused to search for relevant information in the cepstral domain. A BCI was implementedto evaluate the two approaches, and three classification techniquescontributed to finding the effectiveness of each feature type.Data containing two-class motor imagery was used for testing, and the BCIwas compared to some of the other systems currently available. Results indicatethat both approaches investigated were effective in producing separablefeatures, and, with further work, can be used for the classification of trialsbased on a paradigm exploiting motor imagery as a means of control.

[发布机构] Stellenbosch University