IJCNN'09 Proceedings of the 2009 international joint conference on Neural Networks
| Hi-index | 0.00 |
Eye-blink is one of the most eminent artifacts in Brain-Computer Interface (BCI) systems. Only a small amount of eye blink can significantly affect the accuracy of EEG-based data classification. Automatic eye-blink artifact removal system is a crucial function of neurofeedback training systems. The eye-blink artifacts are inevitable and unpredictable due to subject's physical conditions. This factor makes the cognitive state classifier less reliable. One of the most prominent techniques employed to suppress this ocular artifact is Independent Component Analysis (ICA). This technique separates EEG signals into Independent Components (ICs) and discriminates ocular artifacts from brain signals. However, the source separation of ICA algorithm is imperfect. Frequently, the IC identified to be an artifact includes brain wave activity useful for cognitive state classification. The proposed method will elaborate on the IC with a novel wavelet transform analysis to extract the useful brain signals from the artifact component. Experimental results illustrate the performance and accuracy of the proposed removal algorithm of eye-blink artifacts. This removal methodology is implemented in NECTEC's Neurofeedback System for Attention Training which was tested in pre-trial sessions with 10 healthy subjects and 5 Mild Cognitive Impairment (MCI) patients at Chulalongkorn Hospital, Bangkok.