Feature extraction and classification of Event-related EEG based on Kolmogorov entropy

In this paper, we propose a novel method with Kolmogorov entropy to extract the feature of event-related EEG. The results show that the Kolmogorov entropy can effectively quantify the dynamic process of event-related desynchronization/synchronization (ERD/ERS) time course of EEG in relation to hand movement imagination. The relative increase and decrease of Kolmogorov entropy could be an indicator of ERD/ERS. To testify the validity of Kolmogorov entropy measure, the method is tested on five human subjects for feature extraction to classify the left- and right-hand motor imagery by Support Vector Machine (SVM) classifier. An average accuracy and mutual information of 91.5% and 0.5374 are obtained, which highly outperforms the commonly used method of AR model. The results confirm that Kolmogorov entropy analysis may improve accuracy for classification of motor imagery tasks, and have applications to performance improvement of brain-computer interface (BCI) systems.