Research Of Virtual Rehabilitation System Based On EEG-EMG

Virtual rehabilitation is a new technology in the field of stroke rehabilitation, which can help patients complete specific tasks with strong fidelity and immersive enjoyably, so more and more domestic and foreign scholars get the favour of it. However, the existing virtual rehabilitation training methods mainly focus on the interestingness of human-computer interaction, and it is difficult to reflect the individual initiative and adaptability of patients, which may cause the fatigue or injury of muscle during training and make it the limit of virtual rehabilitation promotion. In view of the above problems, the electroencephalogram(EEG) and electromyography(EMG) signals which can reflect the motion pattern and fatigue of patients were introduced into the virtual rehabilitation training in this paper. Features of EEG and EMG were explored to realize the control and feedback adjustment of virtual scene. The specific works are as follows:Firstly, this article introduced the development of virtual reality technology in the field of rehabilitation, and analyzed the present researches of the virtual rehabilitation training based on EEG and EMG. The shortcomings of the current virtual rehabilitation training were summarized, and then the research content about the virtual rehabilitation system based on EEG and EMG was determined. Consulting the need of rehabilitation training for stroke patients during convalescence, the nature of stroke and characteristics of patients with stroke were introduced, and the solution of virtual rehabilitation training system was designed.Secondly, to address the fatigue characteristics of body and brain region corresponding to sports during rehabilitation exercise, brain fatigue index analysis method was proposed to effectively depict the fatigue state of corresponding brain movement area, and mean power frequency(MPF) analysis method was applied to evaluate body movement state of fatigue. These methods were verified in this paper.Thirdly, since the existing linear analysis method can’t effectively depict the nonstationarity and nonlinearity of the EMG, a method named auto permutation entropy(APE) is put forward to quantitatively describe the internal dynamic and coupling features of EMG in this article. Experiments were carried out for acquiring the EMG data of elbow joint under different bending angles. The validity of the proposed method is verified by the analysis and comparison with the pre-existing EMG features.Finally, the virtual rehabilitation system based on EEG and EMG feedback was built. 4 stroke patients and 4 healthy subjects were entrolled in our experiments and EEG and EMG were simultaneously recorded. EMG features were used to control the virtual rehabilitation training scene in order to complete tasks of the specified rehabilitation training. While, the integrated features of EEG and EMG were used to adjust the difficulty level of the virtual scene. The system was verified by comparison experiment between healthy people and patients, and compared with the traditional virtual rehabilitation training system which come from its own without regulating functions in difficulty level.