Functional electrical stimulation mirror training method for movement disorders based on electromyography recognition
Through the combination of electromyography recognition and functional electrical stimulation of healthy gestures, the affected limbs are driven to undergo synchronous movement training, solving the problem of insufficient activation of brain motor sensory cortex in the existing technology, and achieving more effective recovery of motor disorders , and the improvement of training effect was verified through the EEG signal evaluation index.
Patent Information
- Application Number
- CN202111665127.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-31
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2041-12-31
AI Technical Summary
Although mirror training and functional electrical stimulation methods are effective in the rehabilitation of motor dysfunction, they have shortcomings in activate the activity of brain motor sensory cortex, and lack effective evaluation indicators, making it difficult to fully reflect the training effect.
Through the electromyography of the healthy side gesture, an exoskeleton robot is used to drive the affected limbs to conduct synchronous motion training, and combined with functional electrical stimulation, the affected muscles are mirrored, and the μ rhythm time domain event correlation of the EEG signal is used to synchronize. The absolute value of the difference in eigenvalues is used as an evaluation index to evaluate the remodeling effect of sensorimotor cortex.
Compared with traditional methods, the combination of electromyography recognition and functional electrical stimulation significantly improves the control ability of the brain motor sensory cortex, provides more effective rehabilitation effects, and verifies the training effect from the perspective of neuroplasticity through EEG analysis. improve.
Smart Images

Figure CN114469641B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of pattern recognition, and relates to a myoelectric recognition and functional electrical stimulation technology, and in particular to a training method in which an exoskeleton robot drives the affected limb to perform healthy and affected bilateral synchronous movements through myoelectric recognition of healthy-side gestures. And according to the healthy-side myoelectric recognition gesture movements, mirror training of functional electrical stimulation is performed on the corresponding muscles on the affected side. By calculating the absolute value E of the difference between the time-domain event-related desynchronization eigenvalues of the EEG signal μ rhythm in the left and right motor control areas of the cerebral cortex as an evaluation index of the remodeling effect of the sensorimotor cortex, it indicates whether the patient's damaged motor control area has been improved through mirror training. Compared with the traditional exoskeleton robot mirror training method based on myoelectric recognition, it is verified that the mirror training + functional electrical stimulation method has a better effect. Background Art
[0002] Mirror therapy (MT) is a commonly used rehabilitation therapy. Usually, a mirror is placed between the patient's arms or legs to encourage the patient to imagine the normal movement of the affected limb through the movement of the healthy limb, or to complete the same movement of the affected limb as the healthy limb with external mechanical assistance. In this way, the patient's brain motor sensory cortex can be stimulated to reshape. A more commonly used and effective method for rehabilitation training of movement disorders is functional electrical stimulation (FES). Through a pre-set current pulse sequence, the patient's motor disorder limb muscles are stimulated from the outside, which can not only assist the movement of the affected limb, but also improve the transmission of sensory information from the muscle spindles and tendon spindles of the corresponding action muscle groups to the nerve center, and promote the remodeling of the sensorimotor cortex. Motor imagery can activate the movement-related cerebral cortex like actual movement, promote the reorganization or reconstruction of the sensorimotor cortex function, and is therefore used in active rehabilitation training and treatment of patients with limb paralysis caused by brain injury. With the miniaturization of electromyographic acquisition sensors, gesture recognition technology based on surface electromyography (sEMG) has become one of the research hotspots in the field of gesture recognition due to its natural interaction method. At present, with the development of acquisition technology and processing technology, the recognition accuracy of gestures can reach more than 90%. Electroencephalogram (EEG) is the comprehensive effect of the electrical activity of the cerebral cortex neuron group and the result of the joint activity of many neurons. Event Related Desynchronization (ERD) and Event Related Synchronization (ERS) phenomena are the result of the resonance of a large number of neuronal activities on physiological electrical signals, indicating the interaction between the neurons corresponding to a certain frequency band of EEG and related neurons. Its ERD value can reflect the activity of the nerves in the cortical area. Therefore, the healthy side action can be identified through the gesture surface electromyography signal, and the exoskeleton robot on the affected side can be controlled to complete the same action as the healthy side according to the recognition result. At the same time, FES is applied to the corresponding muscles on the affected side, which is feasible to further promote the rehabilitation of movement disorders, and the rehabilitation effect can be verified by the ERD characteristics of the cortical EEG before and after training. Summary of the invention
[0003] In view of the deficiencies in the prior art, the present invention proposes a functional electrical stimulation movement disorder mirror training method based on electromyography recognition.
[0004] Functional electrical stimulation movement disorder mirror training method based on electromyography recognition, the method includes the following steps:
[0005] Step 1. According to the single variable principle, the subjects of different genders were randomly divided into two groups, A and B, with equal numbers of people. Group A used sEMG identification on the healthy side and the exoskeleton robot mechanically assisted the affected side to complete the mirror training; Group B used sEMG identification on the healthy side and the exoskeleton robot mechanically assisted the affected side, and at the same time, FES stimulation was given to the corresponding muscles of the affected side's moving limb; the above-mentioned healthy side sEMG collection point locations and Group B FES stimulation locations were selected according to different limb movements.
[0006] Step 2. Before the training of the two groups of subjects in step 1 begins, collect EEG signals from the left central C3 area and the right central C4 area when the two groups of subjects perform motor imagery of specific movements with both hands at the same time.
[0007] Step 3. Conduct mirror training on the subjects who have been divided into groups in step 1.
[0008] Step 4. After one training cycle of step 3, collect EEG signals from the left central C3 area and the right central C4 area when the two groups of subjects perform motor imagery of specific movements with both hands at the same time.
[0009] Step 5. Preprocess the EEG data collected in steps 2 and 4 by bandpass filtering in the frequency range of 8 to 12 Hz, and calculate the sensorimotor cortex motor control ability evaluation index E.
[0010] Step 6: Compare the E values calculated before and after the two training methods, and analyze the improvement effects of the two training methods on the motor control ability of the sensorimotor cortex.
[0011] Preferably, the specific steps of EEG data collection in steps 2 and 4 are:
[0012] 2-1. Collection of EEG data. The subjects performed motor imagery according to the video guidance, and collected EEG signals from leads C3 and C4. Specifically, the subjects performed motor imagery for 10 seconds and rested for 10 seconds, and a total of 10 sets of data were recorded.
[0013] As a preferred embodiment, the specific steps of step 3 mirror training are:
[0014] The affected side is fixed on the exoskeleton robot, and the exoskeleton robot drives the affected arm to complete the corresponding action; the electromyography acquisition device is placed on the healthy arm. During training, the subjects in group A make movements with both hands at the same time, and the healthy arm actually performs the action, while the affected side is not fully executed due to motor dysfunction. At this time, by collecting the sEMG signal of the healthy side, the corresponding gesture action is identified, and the exoskeleton robot is controlled to drive the affected side to perform the corresponding gesture action. The subjects in group B apply functional electrical stimulation to the muscles of the "action" on the affected side on the basis of group A. The subjects all perform mirror training at a fixed time every day, such as 9:00 am, each time lasting 20 minutes, once a day, 5 days a week, and 4 consecutive weeks.
[0015] As a preferred method, the calculation and analysis method of the evaluation index E is as follows:
[0016] After filtering, the mean of all EEG sampling values was calculated, and the mean of the squared differences was calculated;
[0017]
[0018] In formula (1), x ij is the value of the jth sampling in the i-th experiment, N is the number of experiments, A j is the mean of the j-th sampling difference squares, It is expressed as the average value of the jth sampling of all groups;
[0019] Calculate the energy rate of change ERD;
[0020]
[0021]
[0022] In formula (2), [r, s] is the resting interval, R is the mean of the squared differences of all sample values in the resting interval; in formula (3), [m, n] is the motor imagery interval.
[0023] Calculate the absolute value E of the difference between ERD of leads C3 and C4;
[0024] E i =|ERD c3i -ERD c4i | (4)
[0025] In formula (4), ERD c3i is the energy change rate of the i-th sampling of lead C3, ERD c4i is the energy change rate of the i-th sampling of lead C4; E i That is, it is an evaluation indicator of the improvement of sensorimotor cortical control ability in each group.
[0026] Preferably, the method further comprises performing a significant difference analysis using a t-test based on the evaluation criteria E calculated before and after the training of each group of subjects;
[0027] Specifically: calculate the P value between the two groups before and after mirror training. If P>0.05, the two groups of data have no statistical difference and can be regarded as the same group. iThe mean E was compared before and after the training within the group and between the groups. When comparing within the group, if E decreased before and after the mirror training, and there was a statistical difference, P<0.05, it means that this training method is effective in improving the control ability of the sensorimotor cortex. When comparing between groups, if after training, the evaluation index E of group B is significantly lower than the evaluation index E of group A, and the two groups of data have a statistical difference, P<0.05, it means that the training method of group B is more effective than the training method of group A in improving the control ability of the sensorimotor cortex.
[0028] Compared with the prior art, the present invention has the following effects: the training method uses the electromyography recognition of healthy-side gestures to drive the affected-side limbs to perform synchronous movements on both sides, and applies FES stimulation to the corresponding muscles on the affected side. Compared with the existing training methods, the present invention has a better effect on activating the activity of the motor sensory cortex of the brain and can more effectively improve the control ability of the sensory motor cortex. At the same time, the training effect can be evaluated from the perspective of neuroplasticity through EEG analysis, which is an equally convenient and effective way. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 A principle block diagram for the implementation of the present invention;
[0030] Figure 2 Collecting EEG signal position distribution map for the present invention example;
[0031] Figure 3 The experimental paradigm for the present invention example; DETAILED DESCRIPTION
[0032] The following is a detailed description of an example of the present invention in conjunction with the accompanying drawings: This example is implemented on the premise of the technical solution of the present invention, and a detailed implementation method and specific operation process are given.
[0033] like Figure 1 As shown, this embodiment includes the following steps:
[0034] Step 1: Before training, collect the EEG values of C3 and C4 leads of the two groups of subjects when they perform the "elbow flexion" exercise imagination on both sides. The acquisition position of C3 and C4 leads is as follows: Figure 2 The specific process is as follows:
[0035] The subjects sat in the middle of the armchair, and the display screen played a video of the "elbow flexion" action to guide the subjects to imagine the "elbow flexion" movement with both hands at the same time, and collected EEG signals from leads C3 and C4. Specifically, the subjects imagined the movement for 10 seconds and then rested for 10 seconds, and a total of 10 sets of data were recorded as the original data before the two groups of training.
[0036] Step 2: According to the single variable principle, the subjects were randomly divided into two groups, A and B, according to different genders, so that the number of subjects of the same gender in each group was equal. Group A used sEMG identification on the healthy side and the exoskeleton robot mechanically assisted the affected side to complete the training; Group B used sEMG identification on the healthy side and the exoskeleton robot mechanically assisted the affected side, and at the same time, FES stimulation was given to the corresponding muscles of the affected side's moving limb. The specific cycle and training method are as follows:
[0037] The affected side of the subject was fixed on the exoskeleton robot, and the exoskeleton robot drove the affected arm to complete the corresponding movements. The electromyography acquisition device was placed on the healthy arm. During training, the subject performed the "elbow flexion" movement with both hands at the same time. The healthy arm actually performed this movement, but the affected side did not perform the movement well enough due to motor dysfunction. At this time, the sEMG signal of the healthy side was collected, the corresponding gesture movement was identified, and the exoskeleton robot was controlled to drive the affected side to perform the corresponding gesture movement. Group B also applied functional electrical stimulation to the muscles of the "elbow flexion" movement on the affected side.
[0038] The subjects in group A received mechanical-assisted mirror training, and the subjects in group B received mechanical-assisted + FES stimulation mirror training. Both groups A and B underwent mirror training at a fixed time every day, such as 9:00 a.m., each time lasting 20 minutes, once a day, 5 days a week, and for 4 consecutive weeks.
[0039] Step 3: After the training is completed, the EEG values of the C3 and C4 leads of the two groups of subjects are collected in the same way as step 1, when the two groups of subjects do the "elbow flexion movement imagination" with both hands after the mirror training. The overall experimental process is as follows: Figure 3 shown.
[0040] Step 4, processing the data collected before and after the two groups of mirror training, calculating the control ability evaluation index E as shown in Table 1, and comparing and analyzing the results to verify that the functional electrical stimulation movement disorder mirror training method based on electromyography recognition of the present invention is effective;
[0041] Table 1
[0042]
[0043] The specific analysis is as follows:
[0044] According to Table 1, The average values of E before and after the training of groups A and B are shown in Figure 2. Comparison of the mirror images before and after the training of groups A and B. Before the training, the μ rhythm of the subjects in group A After four weeks of mechanical assisted mirror training Group B subjects had μ rhythm After four weeks of mechanical assistance + FES stimulation mirror training Obviously, the mechanical-assisted mirror training method in group A and the mechanical-assisted + FES stimulation mirror training method in group B are both effective in remodeling the sensorimotor cortex on the affected side. The comparison between groups A and B before and after training, the evaluation index E before training in groups A and B was analyzed by t-test for difference. It was found that there was no statistical difference between the two groups before training (P>0.05), and they can be regarded as the same group. After one cycle of different training, the evaluation index E of groups A and B improved, and group B had the best effect on remodeling the sensorimotor cortex in patients with unilateral motor dysfunction, as shown in Table 2. Based on this, the functional electrical stimulation movement disorder mirror training method based on electromyography recognition proposed in this paper is effective for patients with unilateral motor dysfunction, and the effect is better than the electromyography recognition mechanical-assisted mirror training method;
[0045] Table 2
[0046]
[0047]
Claims
1. A functional electrical stimulation movement disorder mirror training system based on electromyography recognition, characterized by: The system includes a data acquisition module, a training module and an evaluation module; The data acquisition module collects EEG signals of the left central C3 area and the right central C4 area when the subject's hands simultaneously perform motor imagery of specific movements before and after the training begins, and transmits them to the evaluation module; The training module performs mirror training on the subjects in groups A and B respectively: for the subjects in group A, sEMG recognition of the healthy side is used, and the exoskeleton robot is used to mechanically assist the affected side to complete the mirror training; for the subjects in group B, sEMG recognition of the healthy side is used, and the exoskeleton robot is used to mechanically assist the affected side, and FES stimulation is given to the corresponding muscles of the motor part of the affected side; The evaluation module pre-processes the EEG signals of the subjects collected before and after the training with a band-pass filter in the frequency range of 8 to 12 Hz, calculates the evaluation index E of the motor control ability of the sensorimotor cortex, and analyzes the improvement effect of different training methods on the motor control ability of the sensorimotor cortex according to the evaluation index E; The subject is guided to perform motor imagery through the video, and at the same time, the data acquisition module acquires EEG signals of C3 and C4 leads; The training module is used to fix the affected side of the subject on the exoskeleton robot, and drive the affected arm to complete the corresponding action through the exoskeleton robot; place an electromyography acquisition device on the healthy arm; during training, the subjects in group A make movements with both hands at the same time, and the healthy arm actually performs the movement, while the affected side does not perform the movement well enough due to motor dysfunction; at this time, the corresponding gesture movement is identified by collecting the sEMG signal of the healthy side, and the exoskeleton robot is controlled to drive the affected side to perform the corresponding gesture movement; the subjects in group B apply functional electrical stimulation to the muscles of the affected side on the basis of group A; The evaluation module calculates the mean of all EEG sampling values and the mean of the squared differences; In formula (1), x ij is the value of the jth sampling in the i-th experiment, N is the number of experiments, A j is the mean of the j-th sampling difference squares, It is expressed as the average value of the jth sampling of all groups; Calculate the energy rate of change ERD; In formula (2), [r, s] is the resting interval, R is the mean of the squared difference of all sample values in the resting interval; in formula (3), [m, n] is the motor imagery interval; Calculate the absolute value E of the difference between ERD of leads C3 and C4; E i =|ERD c3i -ERD c4i | (4) In formula (4), ERD c3i is the energy change rate of the i-th sampling of lead C3, ERD c4i is the energy change rate of the i-th sampling of lead C4; E i That is, it is an evaluation indicator of the improvement of sensorimotor cortical control ability in each group.
2. The functional electrical stimulation movement disorder mirror training system based on electromyography recognition according to claim 1, characterized in that: The evaluation module uses a t-test to perform significant difference analysis based on the evaluation criteria E calculated before and after training for each group of subjects; Specifically: calculate the P value between the two groups before and after mirror training. If P>0.05, the two groups of data have no statistical difference and can be regarded as the same group; according to the E calculated before and after the two groups of mirror training i The mean E was compared before and after training within the group and between groups respectively; when comparing within the group, if E decreased before and after mirror training and there was a statistical difference, P<0.05, it means that this training method is effective in improving the control ability of the sensorimotor cortex; when comparing between groups, if after training, the evaluation index E of group B was significantly smaller than the evaluation index E of group A, and the two groups of data had a statistical difference, P<0.05, it means that the training method of group B is more effective than the training method of group A in improving the control ability of the sensorimotor cortex.
