Online cue-based discrimination of left / right hand movement imagination

Brain Computer Interface (BCI) is a system in which people can interact with electronic devices without using any body movement but only the brain activity itself. In this system, the brain signals obtained from the scalp surface are analysed from the EEG records. Aiming the discrimination of left and right hand movement imaginations (2 classes), an online cue-based BCI system has been developed in this study. In the offline analysis, feature extraction is performed by using Distinctive-Sensitive Learning Vector Quantization and Time-Frequency Analysis methods and training model is created from these features. This model is used in online classification and the result is given as a feedback. Using these methods, the cross-validation accuracy of the offline system is found to be 87% which yields an online prediction accuracy of 97% on a single subject.