ABSTRACT
Objective:To propose a motor imagery paradigm with regular rhythm to improve the accuracy of classification in brain-computer interface based on motor imagery. Methods:Ten untrained healthy subjects were asked to image the movement guided by the picture accompanied with regular rhythm. The common spatial pattern feature extraction algorithm and Fisher classifier were used for classification, and compared with the traditional paradigm and the motor imagery paradigm accompanying with irregular rhythm. Results:More desynchronization was found as motor imagery with regular rhythm, while the classification accuracy improved. Conclusion:Motor imagery with regular rhythm gives a new idea for active rehabilitation training based on brain-computer interface.
ABSTRACT
Objective:To propose a motor imagery paradigm with regular rhythm to improve the accuracy of classification in brain-computer interface based on motor imagery. Methods:Ten untrained healthy subjects were asked to image the movement guided by the picture accompanied with regular rhythm. The common spatial pattern feature extraction algorithm and Fisher classifier were used for classification, and compared with the traditional paradigm and the motor imagery paradigm accompanying with irregular rhythm. Results:More desynchronization was found as motor imagery with regular rhythm, while the classification accuracy improved. Conclusion:Motor imagery with regular rhythm gives a new idea for active rehabilitation training based on brain-computer interface.