Motor imagery method and system based on myoelectricity monitoring and storage medium
By combining EEG data and electromyography data to judge the data type in the motor imagination task, the problem of insufficient accuracy and difference in the effectiveness of motor imagination training in brain-computer interface technology is solved, and more efficient and personalized motor imagination training is achieved.
Patent Information
- Application Number
- CN202411428627.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-14
- Publication Date
- 2025-02-14
AI Technical Summary
The existing brain-computer interface technology has problems such as insufficient accuracy of training models and differences in training effects in the application of sports imagination, resulting in distrust of patients and poor training results.
By synchronously collecting EEG data and EEG data in performing sports imagination tasks, using EEG data to determine whether EEG data is limb-driven data or purely imaginative data, and then counting the number of times the user reaches the target state and the first target category , adjust the training method according to the comparison results.
It improves the accuracy and personalization of sports imagination training, enhances the training effect, and allows trainers to formulate more appropriate training plans based on the actual situation of each user.
Smart Images
Figure CN119442061A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of electroencephalogram (EEG) data processing, and specifically relates to a motor imagery method, system and storage medium based on electromyography monitoring. Background Art
[0002] With the vigorous development of brain-computer interface technology, brain-computer related applications have spread to various fields. For example, in the medical field, non-invasive EEG devices are used to collect brain waves and convert brain wave data into electroencephalograms (EEGs) for analysis, diagnosis or prevention of related lesions; and some specific paradigms and algorithms can be implanted for rehabilitation treatment, scientific and technological exploration, and even game entertainment.
[0003] At present, there are still some problems in the application of brain-computer interface in motor imagery. First, it is impossible to collect enough data in one training for machine learning training model, so the accuracy of the training model will not be very accurate, which will lead to unsatisfactory results of online training, causing patients to have negative thoughts such as disbelief and unwillingness to do it. Second, the training effect is not exactly the same for different groups of people, and it is necessary to find methods suitable for different groups of people to improve the training effect. Summary of the invention
[0004] In order to solve the above-mentioned technical problems existing in the prior art, the embodiment of the present application provides a motor imagery method based on electromyography monitoring to improve the effect of motor imagery training.
[0005] This application adopts the following technical solutions.
[0006] In a first aspect, the present application provides a motor imagery method based on electromyography monitoring, comprising: Synchronously collect EEG data and EMG data of users when they are performing motor imagery tasks; According to the electromyographic data, determining whether the electroencephalographic data is limb-driven data or pure imagination data; When it is determined that the EEG data is the limb-driven data, it is determined that the user has reached the target state in the motor imagery task, and the number of times the target state is reached is increased by one; When it is determined that the EEG data is the pure imagination data, the trained pure imagination model is used for classification judgment, and if the output result is the first target category, the number of times the first target category is reached is increased by one; After the motor imagery task is completed, the number of times the target state is reached is compared with the number of times the first target category is reached to obtain a comparison result, so as to perform motor imagery training on the user according to the comparison result.
[0007] In some implementations, judging whether the EEG data is limb-driven data or pure imagination data based on the EMG data includes: Taking the average amplitude of the electromyographic data within a set time period, if the average amplitude is greater than half of the difference between the first reference value and the second reference value, determining that the electroencephalographic data is limb-driven data, otherwise determining that the electroencephalographic data is pure imagination data; The first reference value is the average amplitude of the electromyographic data within the set time period when the user's limb performs the training action with maximum force, and the second reference value is the average amplitude of the electromyographic data within the set time period when the user's limb is relaxed.
[0008] In some embodiments, the motor imagery method further comprises pre-training the pure imagery model, comprising: Synchronously collect EEG data and EMG data of users when they are performing motor imagery tasks; According to the electromyographic data, determining whether the electroencephalographic data is pure imagination data; The EEG data and its corresponding category determined to correspond to the pure imagination data are input into the first classification model to be trained, and the first classification model is trained to obtain the pure imagination model.
[0009] In some implementations, comparing the number of times the target state is reached with the number of times the first target category is reached, and performing motor imagery training on the user according to the comparison result, includes: When the ratio of the number of times the first target category is reached to the number of times the target state is reached is greater than one, the user is subjected to motor imagination training using pure running imagination; when the ratio of the number of times the first target category is reached to the number of times the target state is reached is less than one, the user is subjected to motor imagination training using running imagination combined with limb-assisted movement.
[0010] In some embodiments, the motor imagery method further includes: when it is determined that the EEG data is limb-driven data, a limb-driven model obtained through training is used for classification judgment, and if the output result is a second target category, the EEG data is determined to be excellent data.
[0011] In some embodiments, the motor imagery method further includes pre-training the limb-driven model, including: Synchronously collect EEG data and EMG data of users when they are performing motor imagery tasks; According to the electromyographic data, determining whether the electroencephalographic data is limb-driven data; The EEG data and its corresponding category determined to correspond to the limb-driven data are input into the second classification model to be trained, and the second classification model is trained to obtain the limb-driven model.
[0012] In a second aspect, the present application provides a motor imagery system based on electromyography monitoring, comprising: A data acquisition module, used to synchronously collect EEG data and EMG data of the user when performing motor imagery tasks; A data classification module, used for determining whether the EEG data is limb-driven data or pure imagination data according to the EMG data; A counting module, for determining that the user has reached the target state in the motor imagery task when the data segmentation module determines that the EEG data is the limb-driven data, and adding one to the number of times the target state has been reached; and for performing classification judgment using the pure imagination model obtained through training, when the data segmentation module determines that the EEG data is the pure imagination data, and if the output result is the first target category, adding one to the number of times the first target category has been reached; The comparison and evaluation module is used to compare the number of times the target state is reached with the number of times the first target category is reached after the motor imagery task is completed to obtain a comparison result, so as to perform motor imagery training on the user according to the comparison result.
[0013] In some embodiments, the system further comprises a training module, wherein the training module is used to pre-train the pure imagination model, including: Synchronously collect EEG data and EMG data of users when they are performing motor imagery tasks; According to the electromyographic data, determining whether the electroencephalographic data is pure imagination data; The EEG data and its corresponding category determined to correspond to the pure imagination data are input into the first classification model to be trained, and the first classification model is trained to obtain the pure imagination model.
[0014] In some embodiments, the system also includes a high-quality data determination module, which is used to use the trained limb-driven model to perform classification judgment when the EEG data is determined to be the pure imagination data. If the output result is the second target category, the EEG data is determined to be high-quality data.
[0015] In a third aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program, characterized in that when the computer program is executed, it can implement a motor imagery method based on electromyography monitoring as provided in any possible implementation method of the first aspect.
[0016] Compared with the prior art, the motor imagery method based on electromyography monitoring provided in the embodiment of the present application can effectively distinguish whether the electromyography data is generated by actual limb movement or purely by imagination by synchronously collecting the EEG data and electromyography data of the user in the motor imagery task, thereby improving the accuracy of the motor imagery training process. According to the judgment results of the electromyography data, the number of times the user reaches the target state and the first target category can be counted respectively. This personalized data recording method enables the trainer to formulate a more personalized training plan based on the actual situation of each user.
[0017] Compared with the prior art, the motor imagery system based on electromyography monitoring provided in the embodiment of the present application has the same technical effects as the above-mentioned motor imagery method based on electromyography monitoring, and will not be elaborated on. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The drawings described herein are for explanation purposes only and are not intended to limit the scope of the present application in any way. In addition, the shapes and proportional dimensions of the components in the drawings are only for illustration purposes and are used to help understand the present application. They do not specifically limit the shapes and proportional dimensions of the components of the present application. Under the guidance of the present application, those skilled in the art can select various possible shapes and proportional dimensions to implement the present application according to specific circumstances. In the drawings: Figure 1 A schematic diagram of a flow chart of a motor imagery method based on electromyography monitoring provided in an embodiment; Figure 2 A schematic diagram of a classification model training process in a motor imagery method based on electromyography monitoring provided in an embodiment; Figure 3 This is a schematic diagram of the myoelectric data waveform within a set time period when the user's limbs are relaxed in the embodiment; Figure 4 It is a schematic diagram of the electromyographic data waveform when the user's limbs perform the training action with maximum force within a set time period in the embodiment; Figure 5 A schematic diagram of the framework of a motor imagery system based on electromyography monitoring provided in an embodiment. DETAILED DESCRIPTION
[0019] In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.
[0020] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more features.
[0021] The embodiment provides a motor imagery method based on electromyography monitoring, including: synchronously collecting EEG data and electromyography data of a user when performing a motor imagery task; judging whether the EEG data is limb-driven data or pure imagination data based on the electromyography data; when the EEG data is judged to be limb-driven data, determining whether the user has reached a target state during the imagination task, and increasing the number of times the target state has been reached by one; when the EEG data is judged to be pure imagination data, performing classification judgment using a pure imagination model obtained through training, and if the output result is a first target category, increasing the number of times the first target category has been reached by one; after the motor imagery task is completed, comparing the number of times the target state has been reached with the number of times the first target category has been reached to obtain a comparison result, so as to perform motor imagery training on the user based on the comparison result.
[0022] like Figure 1 As shown, in the embodiment, an electromyographic monitoring device (such as an electromyographic sensor) can be used as a monitoring device and attached to the muscle part of the affected limb of the user (patient) for electromyographic collection. At the same time, the user wears an electroencephalogram cap to enter the motor imagery paradigm. After entering the online motor imagery stage, the electromyographic monitoring device is used to record the electromyographic data of the affected limb activity during each motor imagery.
[0023] According to the electromyographic data, it is determined whether the electroencephalographic data is limb-driven data or pure imagination data, including: taking the average amplitude NC of the electromyographic data within the set time when the user performs the motor imagination task, if the average amplitude is greater than half of the difference between the first reference value NB and the second reference value NA, the expression is: NC>(NB-NA) / 2; It is determined that the EEG data is limb-driven data; wherein, the first reference value NB is the average amplitude of the electromyographic data within a set time period when the user's limbs perform the training action with maximum force, and the second reference value NA is the average amplitude of the electromyographic data within a set time period when the user's limbs are relaxed.
[0024] When the EEG data is judged to be limb-driven data, it means that this motor imagery training tends to the limb force state, which means that the patient's training is effective and is directly judged as the target state (that is, the limb drive is correct). The number of times the target state (correct limb drive) is reached is increased by one.
[0025] Otherwise, if the electromyographic data of this training does not meet the indicator, that is, the average amplitude of the electromyographic data within the set time is less than or equal to half of the difference between the first reference value and the second reference value, the expression is: NC≤(NB-NA) / 2; It is determined that the EEG data is pure imagination data, and the EEG data is input into the pure imagination model. When the result output by the pure imagination model is the first target category, it is judged that the pure imagination is correct, and the number of times the first target category (pure imagination correct) is reached is increased by one.
[0026] In the embodiment, after an online motor imagery training is completed, the number of times the target state is reached (correct limb movement) can be compared with the number of times the first target category (correct pure imagination) is reached, such as the number of correct pure imaginations is C1, and the number of correct limb-driven imaginations is C2. If C1:C2>1, it means that the patient has an advantage in pure imagination, and can better stimulate the patient's nerve repair and improve muscle memory through better guided imagination and wearing auxiliary mechanical equipment, thereby improving the patient's training results. When C1:C2<1, it means that the patient has an advantage in assisting motor imagery by driving limb movement, and improves brain control by awakening muscle memory through limb-assisted movement. The brain can be more flexibly strained to the degree of muscle contraction and relaxation under different impedances through better guided imagination and electrical stimulation impedance training, thereby improving the training effect.
[0027] The embodiments of the present application can further strengthen the training of imagination ability for users who perform well in pure imagination data; and for users with more limb movement data, the training content combining actual movement and imagination can be increased, so that the trainer can develop a more personalized training plan based on the actual situation of each user.
[0028] In some embodiments, Figure 1 As shown, the motor imagery method also includes: when it is determined that the EEG data is limb-driven data, a limb-driven model obtained through training is used for classification and judgment, and if the output result is the second target category, the EEG data is determined to be excellent data.
[0029] Excellent data can provide positive feedback for motor imagery training, indicating that when users perform motor imagery tasks, their EEG activity patterns are highly matched with the high-quality or expected patterns in the model, which can motivate users to continue to maintain good performance and increase their enthusiasm for participating in motor imagery training. Excellent data can also train and optimize models to improve the accuracy and generalization ability of model classification.
[0030] In some embodiments, the combination of electromyographic data and limb-driven imagination data can be further used to evaluate the patient's rehabilitation status through training. For example, after a course of treatment or stage of rehabilitation, the first reference value NB can be remeasured, recorded as NBn, that is, the average amplitude of the electromyographic data within the set time when the user's limb uses the maximum force to perform the training action. If NBn>NB, it means that after the stage of rehabilitation training, the patient's maximum force to achieve limb training movements has exceeded the original force, thereby judging that the patient's affected limb has recovered. Then the motor imagination training method is enhanced or improved.
[0031] In the embodiment, the pure imagination model and the limb-driven model are trained in an offline stage. Figure 2 As shown in the preparatory stage before offline training, the electromyographic data (waveform as shown in Figure 3 The average amplitude of the limbs in the current stage of the patient's training action is NA, and the maximum force of the preset duration (such as 4 seconds) reached by the patient's limbs during the corresponding training action process (waveform as shown in Figure 4 The average amplitude recorded is NB.
[0032] The training process begins in offline mode, and the EEG data and EMG data of the user during the motor imagery task are synchronously collected; if the average amplitude of the EMG data detected by the patient during the preset duration (4 seconds) of the guidance process of a single training process is NC.
[0033] During the training process, the patient can be prompted by running animations or voice to imagine the affected limb exerting force, and the EEG data can be divided. If NC ≤ (NB-NA) / 2, it means that this training tends to be in a state where the limbs do not exert force, which means that the collected EEG data is pure imagination data. If the data detected by electromyography is NC> (NB-NA) / 2, it means that this training tends to be in a state where the limbs exert force, and the corresponding EEG data is divided into limb-driven data. After the offline data collection is completed, the pure imagination data and limb-driven data are respectively input into two classification models for training, and finally a pure imagination model and a limb-driven model are obtained.
[0034] In some embodiments, the classification model may be determined by using SVM (support vector machine) for feature classification and using CSP co-space mode for feature extraction, thereby obtaining the features of the imagined left and right hands for binary classification.
[0035] This application combines EEG signals with EMG signals and uses data fusion to increase the richness of motor imagery features, further improving the accuracy and generalization performance of the model. This application determines the proportion of pure imagery data and limb-driven imagery data to further evaluate the patient's recovery status.
[0036] With the same inventive concept as the motor imagery method based on electromyography monitoring provided in the above embodiment, the present application embodiment also provides a motor imagery system based on electromyography monitoring, such as Figure 5 As shown, it includes: a data acquisition module, a data division module, a counting module and a comparison and evaluation module.
[0037] Among them, the data acquisition module is used to synchronously collect the EEG data and EMG data of the user when performing motor imagery tasks.
[0038] The data division module is used to determine whether the EEG data is limb-driven data or pure imagination data based on the EMG data.
[0039] The counting module is used to determine whether the user has reached the target state in the motor imagery task when the data segmentation module determines that the EEG data is limb-driven data, and increase the number of times the target state is reached by one; when the data segmentation module determines that the EEG data is pure imagination data, the pure imagination model obtained through training is used for classification judgment. If the output result is the first target category, the number of times the first target category is reached is increased by one.
[0040] The comparison and evaluation module is used to compare the number of times the target state is reached with the number of times the first target category is reached after the motor imagery task is completed to obtain a comparison result, so as to perform motor imagery training on the user according to the comparison result.
[0041] In some embodiments, the system further comprises a training module, and the training module is used to pre-train the pure imagination model, including: synchronously collecting EEG data and EMG data of the user when performing a motor imagination task; Based on the electromyographic data, determine whether the EEG data is pure imagination data; The EEG data and its corresponding category determined to be pure imagination data are input into the first classification model to be trained, and the first classification model is trained to obtain a pure imagination model.
[0042] The motor imagery system based on electromyography monitoring provided in the embodiment of the present application can more easily observe the effectiveness of the motor imagery process through electromyography data, thereby improving the accuracy and the patient's willingness to train; by comparing the number of correct pure imagination times and the time driven by the limbs, the patient's effort state can be observed and the rehabilitation progress can be evaluated.
[0043] In some embodiments, the system also includes a high-quality data determination module, which is used to use the trained limb movement model to perform classification judgment when the EEG data is judged to be pure imagination data. If the output result is the second target category, the EEG data is determined to be high-quality data.
[0044] The embodiment of the present application also provides a computer-readable storage medium storing a computer program. When the computer program is executed, the above-mentioned motor imagery method based on electromyography monitoring can be implemented.
[0045] It should be noted that in the present application, the emphasis of the schemes described in various embodiments is different, but the various embodiments are also somewhat interrelated. When understanding the scheme of the present application, the various embodiments can refer to each other.
[0046] The above is a detailed introduction to the motor imagination method, system and storage medium based on electromyography monitoring provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the concept of the present application and should not be understood as limiting the scope of protection of the present application.
Claims
1. A motor imagery method based on electromyography monitoring, characterized in that: include: Synchronously collect EEG data and EMG data of users when they are performing motor imagery tasks; According to the electromyographic data, determining whether the electroencephalographic data is limb-driven data or pure imagination data; When it is determined that the EEG data is the limb-driven data, it is determined that the user has reached the target state in the motor imagery task, and the number of times the target state is reached is increased by one; When it is determined that the EEG data is the pure imagination data, the trained pure imagination model is used for classification judgment, and if the output result is the first target category, the number of times the first target category is reached is increased by one; After the motor imagery task is completed, the number of times the target state is reached is compared with the number of times the first target category is reached to obtain a comparison result, so as to perform motor imagery training on the user according to the comparison result.
2. The motor imagery method based on electromyography monitoring according to claim 1, characterized in that: Judging whether the EEG data is limb-driven data or pure imagination data based on the EMG data includes: Taking the average amplitude of the electromyographic data within a set time period, if the average amplitude is greater than half of the difference between the first reference value and the second reference value, determining that the electroencephalographic data is limb-driven data, otherwise determining that the electroencephalographic data is pure imagination data; The first reference value is the average amplitude of the electromyographic data within the set time period when the user's limb performs the training action with maximum force, and the second reference value is the average amplitude of the electromyographic data within the set time period when the user's limb is relaxed.
3. The motor imagery method based on electromyography monitoring according to claim 1, characterized in that: The motor imagery method further includes pre-training the pure imagery model, including: Synchronously collect EEG data and EMG data of users when they are performing motor imagery tasks; According to the electromyographic data, determining whether the electroencephalographic data is pure imagination data; The EEG data and its corresponding category determined to correspond to the pure imagination data are input into the first classification model to be trained, and the first classification model is trained to obtain the pure imagination model.
4. The motor imagery method based on electromyography monitoring according to claim 1, characterized in that: The method further comprises: comparing the number of times the target state is reached with the number of times the first target category is reached, and performing motor imagery training on the user according to the comparison result, including: When the ratio of the number of times the first target category is reached to the number of times the target state is reached is greater than one, the user is subjected to motor imagination training using pure running imagination; when the ratio of the number of times the first target category is reached to the number of times the target state is reached is less than one, the user is subjected to motor imagination training using running imagination combined with limb-assisted movement.
5. The motor imagery method based on electromyography monitoring according to claim 1, characterized in that: The motor imagery method also includes: when it is determined that the EEG data is the limb-driven data, a limb-driven model obtained through training is used for classification judgment, and if the output result is the second target category, the EEG data is determined to be excellent data.
6. The motor imagery method based on electromyography monitoring according to claim 5, characterized in that: The motor imagery method further includes pre-training the limb driving model, including: Synchronously collect EEG data and EMG data of users when they are performing motor imagery tasks; According to the electromyographic data, determining whether the electroencephalographic data is limb-driven data; The EEG data and its corresponding category determined to correspond to the limb-driven data are input into the second classification model to be trained, and the second classification model is trained to obtain the limb-driven model.
7. The motor imagery system based on electromyography monitoring is characterized by: include: A data acquisition module, used to synchronously collect EEG data and EMG data of the user when performing motor imagery tasks; A data classification module, used for determining whether the EEG data is limb-driven data or pure imagination data according to the EMG data; A counting module, for determining that the user has reached the target state in the motor imagery task when the data segmentation module determines that the EEG data is the limb-driven data, and adding one to the number of times the target state has been reached; and for performing classification judgment using the pure imagination model obtained through training, when the data segmentation module determines that the EEG data is the pure imagination data, and if the output result is the first target category, adding one to the number of times the first target category has been reached; The comparison and evaluation module is used to compare the number of times the target state is reached with the number of times the first target category is reached after the motor imagery task is completed to obtain a comparison result, so as to perform motor imagery training on the user according to the comparison result.
8. The motor imagery system based on electromyography monitoring according to claim 7, characterized in that: The system further comprises a training module, which is used to pre-train the pure imagination model, including: Synchronously collect EEG data and EMG data of users when they are performing motor imagery tasks; According to the electromyographic data, determining whether the electroencephalographic data is pure imagination data; The EEG data and its corresponding category determined to correspond to the pure imagination data are input into the first classification model to be trained, and the first classification model is trained to obtain the pure imagination model.
9. The motor imagery system based on electromyography monitoring according to claim 7, characterized in that: The system also includes a high-quality data determination module, which is used to use the trained limb-driven model to perform classification judgment when the EEG data is determined to be the pure imagination data. If the output result is the second target category, the EEG data is determined to be high-quality data.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed, the motor imagery method based on electromyography monitoring as described in any one of claims 1 to 6 can be implemented.