Research On Decoding Technology Of Adjacent Joint Motion Intention Based On Hybrid EEG Paradigm

The Brain-computer interface(BCI)based on motor imagery(MI)can directly translate the user's subjective motion intention into a control command,which not only provides new daily communication for patients with motor impairment,but also provide a new type of rehabilitation treatment for stroke patients.However,the current conventional MI-BCI can only distinguish the motion intentions between the distant limbs(such as distinguishing the left and right upper limbs),and there is still a lack of high-confidence research for the imaginary motion recognition of the adjacent limbs.In recent years,a new hybrid paradigm combining MI and Steady-state somatosensory evoked potential(SSSEP)has been proposed,which can improve the content of identifiable information related to motor imagery by increasing the carrier frequency band,thereby effectively improving the recognition performance of MI-BCI and also provides a new idea for improving MI-BCI recognition of adjacent joint motion imaging.Therefore,this paper uses the MI+SSSEP hybrid paradigm to explore the recognition of multiple adjacent joint motion imaging in the ipsilateral upper limb,and constructs an online system to complete the identification verification with high credibility.In this work,the separability of the motor imagery of the finger,wrist,elbow and shoulder joints in the ipsilateral upper limbs under the MI+SSSEP paradigm was first explored.The results of offline EEG data show that the separability of the above four joint motor imagery is clear under the hybrid paradigm.The recognition accuracy of the four-joint based on the hybrid paradigm is significantly higher than the chance selection.The experimental results show that the hybrid paradigm can be applied to the recognition of multiple adjacent joint motor imagery in the ipsilateral limb.Furthermore,the hybrid paradigm design was optimized in the study,and the online recognition system of the finger-elbow joint motor imagery of the ipsilateral limb was constructed,and the online BCI experiment of 16 subjects was carried out.The results show that all the subjects can use the finger and elbow motion to output different BCI instructions through the online system.The online average recognition accuracy under the hybrid paradigm reached 76.4%,which is significantly higher than the traditional MI paradigm by about 11%.Studies have shown that the MI+SSSEP paradigm can effectively improve the recognition effect of adjacent joints in the ipsilateral limbs.Finally,the motor imagery recognition of the ipsilateral thumb and index finger based on the hybrid paradigm was further tried in the study.The results of online experiments show that the motor imagery recognition of the ipsilateral two fingers can also obtain the recognition accuracy that significantly higher than the chance selection.The subjects with the highest recognition accuracy in the study reached 81.67% of the two fingers motor imagery distinction,which showing the potential of the hybrid paradigm in the recognition of fine limb motions such as fingers.In this paper,a series of studies on the decoding of the ipsilateral limb adjacent joint motor imagery under the hybrid paradigm are carried out,and the online recognition of the multi-joint motor imagery of the ipsilateral limb is initially realized,which provides a certain technical basis for improving the decoding performance of MI-BCI.