| Brain-computer interface(BCI)is a human-computer interaction system that allows direct interaction between the brain and the external environment or devices without relying on peripheral nerves and muscle systems.Motor Imagery(MI)is a widely used spontaneous electroencephalogram(EEG)paradigm,which has broad application prospects in rehabilitation medicine,robot control,and other fields.However,from the perspective of the development status of MIBCI,there are still problems such as low decoding accuracy,low information transmission rate,and the need for long-term training.The existence of these problems leads to the application of MI-BCI being severely limited.The fusion of multi-sensory stimulation and MI paradigm has been proven to improve the performance of users in MI.However,the mechanism of the impact of tactile stimulation on MI is not yet clear,and existing research has yielded mixed results.Moreover,the application of tactile paradigm in EEG control systems is relatively limited,and it is still in the theoretical research stage.In order to explore the effect of vibrotactile stimulation on MI paradigm,find more efficient and robust methods for applying vibration stimulation to enhance the overall performance of MI paradigm,and explore the application of vibration stimulation in braincontrolled robot systems,a large number of experiments and research have been conducted in this thesis.The specific work is as follows:Firstly,aiming at the problems of interference with imagery and fatigue in the current application of vibration stimulation in MI-BCI,a concept of closed-loop application of vibration stimulation based on real-time EEG phase information has been proposed.The phase information of brain oscillations is a basic feature of neural processing.Different phase stages of brain oscillations represent different states of neural population excitability.Therefore,it can be used as an indicator to evaluate brain excitability.Hence,a creal-time EEG phase prediction algorithm based on fast Fourier transform(FFT)has been proposed and its effectiveness has been tested too.Based on this algorithm,a phase-dependent closed-loop vibration stimulation system was constructed,and comparative experiments were designed to compare the overall performance of MI paradigm under different vibration stimulation conditions.The experimental results showed that closed-loop vibration stimulation was more effective than no stimulation and open-loop stimulation conditions in improving the overall performance of non-dominant hand imagery and rest paradigms,with average offline classification accuracy increased by 9%and 4%,respectively.This also verifies the effectiveness of the closed-loop stimulation method and the system.Secondly,in practical applications,the MI paradigm often requires users to imagine limbs on both sides separately to enrich the number of commands.However,due to the spontaneous nature of MI paradigm,vibration stimulation cannot be applied only on one side of the imagined movement.There are few studies on tactile-MI hybrid paradigms,and there has not yet appeared a hybrid system with natural combination method and outstanding overall performance.Therefore,a hybrid BCI paradigm has been proposed that combines the two-handed MI paradigm with bilateral closed-loop vibration stimulation.Comparative experiments were also designed under open-loop stimulation and no-stimulation conditions to compare the real-time enhancement effect of tactile stimulation on MI paradigm.The experimental results showed that closed-loop vibration stimulation significantly enhanced the activation of bilateral sensory-motor cortex and functional connectivity of related brain areas during MI,and the offline classification accuracy increased by about 7% compared with no-stimulation condition.It further proved the effectiveness of phase-dependent closed-loop vibration stimulation method in improving overall performance of MI paradigm.Thirdly,MI paradigm has great potential application value in the field of motor function rehabilitation.To explore the application effect of vibration stimulation in rehabilitation training,a hand rehabilitation training system based on pneumatic soft gloves which controlled by MIBCI has been designed.The system allow users to control the hand rehabilitation robot through intention to complete rehabilitation training.An EEG experiment was designed to compare the effect of adding vibration stimulation on real-time brain activity during MI control and rehabilitation training periods.Some healthy subjects and stroke patients were recruited for functional near-infrared spectroscopy(fNIRS)experiment to explore the impact and application prospects of adding vibration stimulation on rehabilitation training.The results indicated that vibration stimulation can help users generate stronger tactile input during rehabilitation training,effectively enhancing the activation of relevant brain regions,and improving the effectiveness of rehabilitation training.Finally,a teleoperation robot system based on hybrid control of EEG and eye movement csignals has been designed.Users can complete target grasping tasks by controlling a mechanical arm remotely through visual image feedback.The control commands are provided by the left and right hand MI paradigm,and eye movement gaze information is used to switch control modes and provide auxiliary commands.Vibration stimulation is added to the control system as a training aid and control feedback method.The results of comparative experiments show that the addition of vibration stimulation improves the classification performance of MI training models,thereby improving the robustness and control efficiency of the control system.This thesis proposes new tactile stimulation methods and multimodal hybrid enhancement methods.Improved the overall performance of traditional MI and tactile-MI paradigms,explored the application of tactile stimulation in brain-controlled robot systems,enhanced the overall efficiency of control systems.By introducing tactile stimulation,the application scenarios of BCI have been expanded,provided a new approach to the challenges and problems faced by BCI development.These achievements will lay a solid foundation for in-depth research and extensive application of hybrid BCI. |
PDF Full Text Request