A method for selecting neural regulation targets based on decoding brain-muscle interaction information

Abstract

The present invention discloses a method for selecting neural control targets based on brain-muscle interaction information decoding. The method includes synchronously collecting EEG data and electromyographic data corresponding to the EEG leads and electromyographic leads of a subject during a motor control task; forming a first weight set with the corresponding eigenvalues ​​in a preset first EEG lead number set according to the numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data; forming a second weight set with the corresponding eigenvalues ​​in a preset second EEG lead number set according to the numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and electromyographic data; calculating lead weights based on the first weight set and the second weight set, and selecting the lead with the largest weight as the neural control target brain area. The present invention can effectively improve the accuracy of neural control target selection.

Description

Technical Field

[0001] The present invention relates to the technical field of digital signal processing, and in particular to a method for selecting neural regulation targets based on brain-muscle interaction information decoding. Background Art

[0002] The accuracy of target selection for neuromodulation in an organism significantly impacts the outcome of the regulation. However, the localization of brain targets for neuromodulation is mostly based on structural information of the brain, such as the calibration and alignment of regulatory targets using brain structural images generated by magnetic resonance imaging (MRI). However, the functional divisions of the brain vary greatly from individual to individual, and cannot be fully characterized by structural information. Furthermore, different control functions of the brain correspond to different brain regions of interest. Therefore, localization based solely on structural information limits further improvement in the effectiveness of neuromodulation. Furthermore, the brain's control function is also reflected in the transmission of neural signals to muscles, causing muscle activation and contraction to produce movement. Movement, in turn, generates sensations that are fed back to the brain and act on the next motor control. Therefore, the interactive information between the brain and muscles will more comprehensively characterize the brain's actual control function. Incorporating this interactive information of brain-muscle interaction into the selection of neuromodulation targets will greatly improve the effectiveness of target selection.

[0003] Existing methods for neuromodulation target selection: Patent CN111657947A selects and calibrates targets based on MRI images, while Patent CN116250809A selects and calibrates target regions in biological samples based on photoacoustic signals. These inventions, however, lack the ability to reference brain function information when selecting target regions. Therefore, existing neuromodulation target selection techniques fail to consider this critical factor. Summary of the Invention

[0004] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0005] To this end, the present invention proposes a method for selecting neural regulation targets based on decoding of brain-muscle interaction information, and proposes a targeted brain area calibration technology based on EEG, EMG and their correlation analysis to improve the accuracy of neural regulation target selection.

[0006] Another object of the present invention is to propose a neural regulation target selection system based on brain-muscle interaction information decoding.

[0007] To achieve the above objectives, the present invention proposes a method for selecting neural regulation targets based on brain-muscle interaction information decoding, comprising:

[0008] Synchronously collect the EEG data and EMG data corresponding to the EEG leads and EMG leads of the subjects during the motor control task;

[0009] According to the numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data, corresponding eigenvalues ​​in a preset first EEG lead number set are combined into a first weight set;

[0010] According to the numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and the EMG data, corresponding eigenvalues ​​in the preset second EEG lead number set are combined into a second weight set;

[0011] Lead weights are calculated based on the first weight set and the second weight set, and the lead with the largest weight is used as the target brain area for neural regulation.

[0012] The method for selecting neural regulation targets based on brain-muscle interaction information decoding according to an embodiment of the present invention may also have the following additional technical features:

[0013] In one embodiment of the present invention, according to the numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data, the corresponding eigenvalues ​​in the preset first EEG lead number set are formed into a first weight set, including:

[0014] All brain activation eigenvalues ​​are calculated based on the EEG data of all EEG leads, and are arranged in ascending order according to the lead number to form a brain activation eigenvector;

[0015] Arranging the features greater than zero in the brain activation feature vector from large to small according to their feature values ​​to obtain a first numerical sorting result;

[0016] Selecting a first preset number of brain activation features from the first numerical sorting result, and forming a first EEG lead number set from the EEG lead numbers corresponding to the first preset number of brain activation features;

[0017] The brain activation feature value corresponding to each lead in the first EEG lead number set is combined into a first weight set.

[0018] In one embodiment of the present invention, according to the numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and the EMG data, the corresponding eigenvalues ​​in the preset second EEG lead number set are formed into a second weight set, including:

[0019] Calculating brain-muscle interaction feature values ​​based on the EEG data of all EEG leads and the EMG data of all EMG leads to obtain a brain-muscle interaction feature matrix;

[0020] Arranging all brain-muscle interaction features in the brain-muscle interaction feature matrix from large to small according to numerical values ​​to obtain a second numerical sorting result;

[0021] selecting a second preset number of brain-muscle interaction features from the second numerical sorting result, and forming a second EEG lead number set from the EEG lead numbers corresponding to the second preset number of brain-muscle interaction features;

[0022] The brain-muscle interaction feature value corresponding to each lead in the second EEG lead number set is combined into a second weight set.

[0023] In one embodiment of the present invention, calculating lead weights based on the first weight set and the second weight set, and using the lead with the largest weight as the target brain region for neuromodulation, includes:

[0024] Obtaining an intersection weight set based on the first weight set and the second weight set;

[0025] Linearly adding the lead weights in the intersection weight set corresponding to the first EEG lead number set and the second EEG lead number set to obtain a linear sum weight value;

[0026] The lead with the largest weight among the linear sum weight values ​​is used as the target brain area for neuromodulation.

[0027] In one embodiment of the present invention, linearly adding the lead weights in the intersection weight set corresponding to the first EEG lead number set and the second EEG lead number set to obtain a linear sum weight value further includes:

[0028] Respectively obtaining lead coefficients of corresponding leads in the first EEG lead number set and the second EEG lead number set;

[0029] respectively obtaining weight values ​​of corresponding leads in the first EEG lead number set and the second EEG lead number set; and,

[0030] A linear sum weight value of the corresponding lead is calculated based on the lead coefficient and the weight value of the corresponding lead.

[0031] To achieve the above objectives, the present invention further proposes a neural regulation target selection system based on brain-muscle interaction information decoding, comprising:

[0032] The EEG and myoelectric data acquisition module is used to synchronously acquire EEG data and EMG data corresponding to the EEG leads and EMG leads of the subject during the motor control task;

[0033] a brain activation information extraction module, configured to group corresponding eigenvalues ​​in a preset first EEG lead number set into a first weight set based on numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data;

[0034] a brain-muscle interaction information extraction module, configured to group corresponding eigenvalues ​​in a preset second EEG lead number set into a second weight set according to numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and the EMG data;

[0035] The targeted brain region identification module is used to calculate the lead weights based on the first weight set and the second weight set, and to use the lead with the largest weight as the targeted brain region for neural regulation.

[0036] The neural regulation target selection method and system based on brain-muscle interaction information decoding of the embodiment of the present invention calculates the lead weight based on the extraction of brain activation characteristics and brain-muscle interaction characteristics to use the lead with the largest weight as the neural regulation target brain area. The present invention can effectively improve the accuracy of neural regulation target selection.

[0037] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0039] Figure 1 is a flowchart of a method for selecting neural regulation targets based on brain-muscle interaction information decoding according to an embodiment of the present invention;

[0040] Figure 2 is a schematic diagram of experimental data collection according to an embodiment of the present invention;

[0041] Figure 3 4 is a structural diagram of a neural regulation target selection system based on brain-muscle interaction information decoding according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0043] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0044] The following describes a method and system for selecting neural regulation targets based on brain-muscle interaction information decoding according to an embodiment of the present invention with reference to the accompanying drawings.

[0045] Figure 1 is a flowchart of a method for selecting neural regulation targets based on brain-muscle interaction information decoding according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0046] S1, synchronously collect the EEG data and EMG data corresponding to the EEG leads and EMG leads of the subjects in the movement control task.

[0047] It can be understood that the present invention randomly selects a subject for data collection, and the subject completes a specific motion control task under the instruction of the task prompter. The motion control task should include a specific motion task and a rest task without motion task.

[0048] In one embodiment of the present invention, N leads of EEG data and M leads of EMG data of the subject throughout the entire task are synchronously collected for the extraction of brain activation features and brain-muscle interaction features.

[0049] It is understandable that the subject wears an EEG acquisition device or equipment on his head, and the acquisition device can achieve the acquisition effect in the existing technology, so there is no specific limitation.

[0050] It is understandable that the task prompter selected in the present invention may be a computing terminal, which may be a notebook computer.

[0051] Specifically, the subjects completed the motor task of alternating flexion and relaxation of the left upper limb under the visual prompt of the task prompter. The motor control task should include the motor task of flexion of the left upper limb and the resting task of relaxation of the left upper limb. The subjects' 64-lead EEG data and 1-lead EMG data were collected synchronously throughout the task for the extraction of brain activation features and brain-muscle interaction features. Figure 2 shown.

[0052] S2, according to the numerical sorting results of all brain activation eigenvalues ​​obtained based on the EEG data, corresponding eigenvalues ​​in a preset first EEG lead number set are combined into a first weight set.

[0053] In one embodiment of the present invention, all brain activation feature values ​​are calculated based on the EEG data of all EEG leads, and are arranged from small to large according to the lead numbers to form a brain activation feature vector; the features greater than zero in the brain activation feature vector are arranged from large to small according to the feature values ​​to obtain a first numerical sorting result; a first preset number of brain activation features in the first numerical sorting result are selected, and the EEG lead numbers corresponding to the first preset number of brain activation features are formed into a first EEG lead number set; the brain activation feature values ​​corresponding to each lead in the first EEG lead number set are formed into a first weight set.

[0054] Specifically, the steps of extracting brain activation features of the present invention may be:

[0055] First, based on the EEG signal x(t) collected by a certain EEG lead, the brain activation characteristic value corresponding to the lead is calculated. The calculation method is as shown in formula (1):

[0056]

[0057] Where SpecR represents the mean value of the power spectrum density of a certain EEG signal x(t) when no motor task is performed, and SpecT represents the mean value of the power spectrum density of a certain EEG signal x(t) when a motor task is performed.

[0058] In practical applications, the present invention can traverse the EEG data of all N EEG leads, calculate N spec values, and arrange them from small to large according to the lead number to form an N*1 brain activation feature vector Spec.

[0059] In one embodiment of the present invention, based on the N*1 brain activation feature vector Spec, the features that are greater than zero are taken, and the first numerical sorting result is obtained by arranging the features from large to small according to the feature value. The top k% of the features are taken, which in the embodiment of the present invention can be a first preset number of brain activation features, and the EEG lead numbers corresponding to the top k% of the features are combined into a first EEG lead number set, i.e., set E1, and the feature values ​​corresponding to each lead in set E1 are combined into a first weight set, i.e., weight set W1.

[0060] Specifically, the present invention can traverse all 64 EEG leads, calculate 64 spec values, and arrange them in ascending order according to the lead number to form a 64*1 activation vector Spec. The present invention can take the top 20% of features, and form a set E1 with the EEG lead numbers corresponding to the top 20% of features. The feature values ​​corresponding to each lead in set E1 form a first weight set W1.

[0061] S3, according to the numerical sorting results of all brain-muscle interaction eigenvalues ​​obtained based on the EEG data and the EMG data, corresponding eigenvalues ​​in the preset second EEG lead number set are combined into a second weight set.

[0062] In one embodiment of the present invention, brain-muscle interaction feature values ​​are calculated based on the EEG data of all EEG leads and the EMG data of the EMG leads to obtain a brain-muscle interaction feature matrix; all brain-muscle interaction features in the brain-muscle interaction feature matrix are arranged from large to small according to numerical values ​​to obtain a second numerical sorting result; a second preset number of brain-muscle interaction features in the second numerical sorting result are selected, and the EEG lead numbers corresponding to the second preset number of brain-muscle interaction features are combined into a second EEG lead number set; the brain-muscle interaction feature values ​​corresponding to each lead in the second EEG lead number set are combined into a second weight set.

[0063] Specifically, the specific steps of extracting brain-muscle interaction information of the present invention may include:

[0064] First, select any EEG signal x(t) and any EMG signal y(t) to calculate the brain-muscle interaction information features. The calculation method is as shown in formula (2):

[0065]

[0066] in, Represents the power spectrum of the observed signal, t c represents the instantaneous time of the reference observation point, f c represents a fixed observation frequency point, μ1 and μ2 represent the time delay of the EEG signal x(t) and the EMG signal y(t) observed from the reference observation point, respectively.

[0067] Secondly, the EEG signals of all N EEG leads and the EMG signals of M EMG leads are traversed to calculate the brain-muscle interaction eigenvalue, that is, The brain-muscle interaction features are arranged from large to small according to the numerical value to obtain a second numerical sorting result, and the first r% of features are taken, which are the second preset number of brain-muscle interaction features in the embodiment of the present invention, and the EEG lead numbers corresponding to the first r% of features are combined into a second EEG lead number set, i.e., set E2, and the eigenvalues ​​corresponding to each lead in set E2 are combined into a second weight set, i.e., weight set W2.

[0068] Specifically, the present invention can traverse all 64 EEG leads and 1 EMG lead to calculate The brain-muscle interaction features are arranged from large to small according to their numerical values, and the top 20% of the features are taken. The EEG lead numbers corresponding to the top 20% of the features are grouped into a set E2, and the eigenvalues ​​corresponding to each lead in the set E2 are grouped into a weight set W2.

[0069] S4, calculating lead weights based on the first weight set and the second weight set, and using the lead with the largest weight as the target brain area for neural regulation.

[0070] In one embodiment of the present invention, an intersection weight set is obtained based on the first weight set and the second weight set; the lead weights in the corresponding intersection weight set in the first EEG lead number set and the second EEG lead number set are linearly added to obtain a linear sum weight value; and the lead with the largest weight in the linear sum weight value is used as the target brain area for neural regulation.

[0071] The method for calculating the linear sum weight value in an embodiment of the present invention is: respectively obtaining the lead coefficients of the corresponding leads in the first EEG lead number set and the second EEG lead number set; respectively obtaining the weight values ​​of the corresponding leads in the first EEG lead number set and the second EEG lead number set; and, calculating the linear sum weight value of the corresponding lead based on the lead coefficients and weight values ​​of the corresponding lead.

[0072] Specifically, the intersection weight set W of the first weight set W1 and the second weight set W2 is taken, and the lead weights corresponding to W in the first EEG lead number set E1 and the second EEG lead number set E2 are linearly added. The calculation method of the linear addition weight value w of a certain lead is as shown in formula (3):

[0073] w=t1*e1+t2*e2 (3)

[0074] Among them, e1 is the weight value of a lead in E1, t1 is the coefficient corresponding to the lead, e2 is the weight value of a lead in E2, and t2 is the coefficient corresponding to the lead.

[0075] It can be understood that the determination of the above-mentioned conduction coefficient is based on the specific control task type in step 1, and the coefficient range is 0 to 1. Optionally, an embodiment of the present invention defines an EEG conduction coefficient vector T, where vector T is an N*1 vector (one embodiment of the present invention can define vector T as a 64*1 vector), and the coefficient range of each lead that constitutes T is 0 to 1. If the subject's motor control task is left limb movement, the EEG conduction coefficient of the sensorimotor area on the right cerebral hemisphere is set to be greater than the EEG conduction coefficient of the sensorimotor area on the left cerebral hemisphere, which is greater than the EEG conduction coefficient of other areas of the whole brain; if the subject's motor control task is right limb movement, the EEG conduction coefficient of the sensorimotor area on the left cerebral hemisphere is set to be greater than the EEG conduction coefficient of the sensorimotor area on the right cerebral hemisphere, which is greater than the EEG conduction coefficient of other areas of the whole brain.

[0076] In the embodiment of the present invention, the EEG conduction coefficient of the sensorimotor area of ​​the right cerebral hemisphere is set to 0.8, the EEG conduction coefficient of the sensorimotor area of ​​the left cerebral hemisphere is set to 0.5, and the EEG conduction coefficient of other areas of the whole brain is set to 0.3.

[0077] Therefore, the lead point with the largest weight after linear addition is selected as the target brain area for neuromodulation.

[0078] According to the method for selecting neuroregulatory targets based on brain-muscle interaction information decoding according to an embodiment of the present invention, EEG data and EMG data corresponding to the EEG leads and EMG leads of the subject in the motor control task are synchronously collected; according to the numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data, the corresponding eigenvalues ​​in the preset first EEG lead number set are formed into a first weight set; according to the numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and EMG data, the corresponding eigenvalues ​​in the preset second EEG lead number set are formed into a second weight set; the lead weights are calculated based on the first weight set and the second weight set, and the lead with the largest weight is used as the target brain area for neuroregulation. Therefore, the targeted brain area calibration technology based on EEG, EMG and their correlation analysis proposed by the present invention can effectively improve the accuracy of neuroregulatory target selection.

[0079] In order to implement the above embodiment, Figure 3 As shown, this embodiment also provides a neural regulation target selection system 10 based on brain-muscle interaction information decoding, and the system 10 includes:

[0080] The EEG and myoelectric data acquisition module 100 is used to synchronously acquire EEG data and EMG data corresponding to the EEG leads and EMG leads of the subject during the motor control task;

[0081] The brain activation information extraction module 200 is configured to form a first weight set of corresponding eigenvalues ​​in a preset first EEG lead number set according to the numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data;

[0082] The brain-muscle interaction information extraction module 300 is configured to form a second weight set 400 of corresponding eigenvalues ​​in a preset second EEG lead number set according to the numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and the EMG data;

[0083] The target brain region identification module is used to calculate the lead weights based on the first weight set and the second weight set, and use the lead with the largest weight as the target brain region for neural regulation.

[0084] Furthermore, the brain activation information extraction module 200 is further configured to:

[0085] Traverse the EEG data of all EEG leads to calculate all brain activation eigenvalues, and arrange them in ascending order according to the lead number to form a brain activation eigenvector;

[0086] Arrange the features greater than zero in the brain activation feature vector from large to small according to their feature values ​​to obtain a first numerical sorting result;

[0087] Selecting a first preset number of brain activation features from the first numerical sorting result, and forming a first EEG lead number set from the EEG lead numbers corresponding to the first preset number of brain activation features;

[0088] The brain activation feature value corresponding to each lead in the first EEG lead number set is combined into a first weight set.

[0089] Furthermore, the brain-muscle interaction information extraction module 300 is also used to:

[0090] Traversing the EEG data of all EEG leads and the EMG data of all EMG leads to calculate the brain-muscle interaction eigenvalues ​​to obtain the brain-muscle interaction feature matrix;

[0091] Arrange all brain-muscle interaction features in the brain-muscle interaction feature matrix from large to small according to their values ​​to obtain a second numerical sorting result;

[0092] Selecting a second preset number of brain-muscle interaction features from the second numerical sorting result, and forming a second EEG lead number set from the EEG lead numbers corresponding to the second preset number of brain-muscle interaction features;

[0093] The brain-muscle interaction feature value corresponding to each lead in the second EEG lead number set is combined into a second weight set.

[0094] Furthermore, the above-mentioned target brain region identification module 400 includes:

[0095] A weight set intersection unit, configured to obtain an intersection weight set based on the first weight set and the second weight set;

[0096] A weighted linear summing unit, configured to linearly sum the lead weights in the corresponding intersection weight set in the first EEG lead number set and the second EEG lead number set to obtain a linear sum weight value;

[0097] The neuromodulation targeting unit is used to select the lead with the largest weight in the linear sum weight value as the neuromodulation targeting brain area.

[0098] Furthermore, the weighted linear summing unit is also used to:

[0099] Respectively obtain the lead coefficients of the corresponding leads in the first EEG lead number set and the second EEG lead number set;

[0100] Obtaining weight values ​​of corresponding leads in the first EEG lead number set and the second EEG lead number set respectively; and,

[0101] A linear sum weight value of the corresponding lead is calculated based on the lead coefficient and the weight value of the corresponding lead.

[0102] According to the neural control target selection system based on brain-muscle interaction information decoding of an embodiment of the present invention, the EEG data and EMG data of the subjects' leads in the motor control task are synchronously collected; according to the numerical sorting results of all brain activation eigenvalues ​​obtained based on the EEG data, the corresponding eigenvalues ​​in the preset first EEG lead number set are combined into a first weight set; according to the numerical sorting results of all brain-muscle interaction eigenvalues ​​obtained based on the EEG data and the EMG data, the corresponding eigenvalues ​​in the preset second EEG lead number set are combined into a second weight set; the lead weights are calculated based on the first weight set and the second weight set, and the lead with the largest weight is used as the neural control target brain area. Therefore, the targeted brain area calibration technology based on EEG, EMG and their correlation analysis proposed by the present invention can effectively improve the accuracy of neural control target selection.

[0103] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0104] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

Claims

1. A method for selecting neural regulation targets based on brain-muscle interaction information decoding, characterized in that: include: Synchronously collect the EEG data and EMG data corresponding to the EEG leads and EMG leads of the subjects during the motor control task; According to the numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data, corresponding eigenvalues ​​in a preset first EEG lead number set are combined into a first weight set; According to the numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and the EMG data, corresponding eigenvalues ​​in the preset second EEG lead number set are combined into a second weight set; Calculating lead weights based on the first weight set and the second weight set, and using the lead with the largest weight as the target brain region for neural regulation; According to the numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data, the corresponding eigenvalues ​​in the preset first EEG lead sequence number set are formed into a first weight set, including: All brain activation eigenvalues ​​are calculated based on the EEG data of all EEG leads, and are arranged in ascending order according to the lead number to form a brain activation eigenvector; Arranging the features greater than zero in the brain activation feature vector from large to small according to their feature values ​​to obtain a first numerical sorting result; Selecting a first preset number of brain activation features from the first numerical sorting result, and forming a first EEG lead number set from the EEG lead numbers corresponding to the first preset number of brain activation features; The brain activation feature value corresponding to each lead in the first EEG lead number set is combined into a first weight set; According to the numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and the EMG data, the corresponding eigenvalues ​​in the preset second EEG lead number set are formed into a second weight set, including: Calculating brain-muscle interaction feature values ​​based on the EEG data of all EEG leads and the EMG data of all EMG leads to obtain a brain-muscle interaction feature matrix; Arranging all brain-muscle interaction features in the brain-muscle interaction feature matrix from large to small according to numerical values ​​to obtain a second numerical sorting result; selecting a second preset number of brain-muscle interaction features from the second numerical sorting result, and forming a second EEG lead number set from the EEG lead numbers corresponding to the second preset number of brain-muscle interaction features; The brain-muscle interaction feature value corresponding to each lead in the second EEG lead number set is combined into a second weight set; Calculating lead weights based on the first weight set and the second weight set, and using the lead with the largest weight as a target brain region for neural regulation, includes: Obtaining an intersection weight set based on the first weight set and the second weight set; Linearly adding the lead weights in the intersection weight set corresponding to the first EEG lead number set and the second EEG lead number set to obtain a linear sum weight value; The lead with the largest weight among the linear sum weight values ​​is used as the target brain area for neural regulation; The method further comprises: Based on the EEG signal x(t) collected by the EEG lead, the brain activation characteristic value corresponding to the lead is calculated. The calculation method is as follows: Where SpecR represents the mean value of the power spectrum density of a certain EEG signal x(t) when no motor task is performed, and SpecT represents the mean value of the power spectrum density of a certain EEG signal x(t) when a motor task is performed; Select any EEG signal x(t) and any EMG signal y(t) to calculate the brain-muscle interaction information features. The calculation method is as shown in formula (2): in, Represents the power spectrum of the observed signal, t c represents the instantaneous time of the reference observation point, f c represents a fixed observation frequency point, μ1 and μ2 represent the time delay of the EEG signal x(t) and the EMG signal y(t) observed from the reference observation point, respectively.

2. The method according to claim 1, characterized in that Linearly adding the lead weights in the intersection weight set corresponding to the first EEG lead number set and the second EEG lead number set to obtain a linear sum weight value, further comprising: Respectively obtaining lead coefficients of corresponding leads in the first EEG lead number set and the second EEG lead number set; respectively obtaining weight values ​​of corresponding leads in the first EEG lead number set and the second EEG lead number set; and, A linear sum weight value of the corresponding lead is calculated based on the lead coefficient and the weight value of the corresponding lead.

3. A neural regulation target selection system based on brain-muscle interaction information decoding, characterized in that: include: The EEG and myoelectric data acquisition module is used to synchronously acquire EEG data and EMG data corresponding to the EEG leads and EMG leads of the subject during the motor control task; a brain activation information extraction module, configured to group corresponding eigenvalues ​​in a preset first EEG lead number set into a first weight set based on numerical sorting results of all brain activation eigenvalues ​​calculated based on the EEG data; a brain-muscle interaction information extraction module, configured to group corresponding eigenvalues ​​in a preset second EEG lead number set into a second weight set according to numerical sorting results of all brain-muscle interaction eigenvalues ​​calculated based on the EEG data and the EMG data; a targeted brain region identification module, configured to calculate lead weights based on the first weight set and the second weight set, and select the lead with the largest weight as the targeted brain region for neural regulation; The brain activation information extraction module is further used to: All brain activation eigenvalues ​​are calculated based on the EEG data of all EEG leads, and are arranged in ascending order according to the lead number to form a brain activation eigenvector; Arranging the features greater than zero in the brain activation feature vector from large to small according to their feature values ​​to obtain a first numerical sorting result; Selecting a first preset number of brain activation features from the first numerical sorting result, and forming a first EEG lead number set from the EEG lead numbers corresponding to the first preset number of brain activation features; The brain activation feature value corresponding to each lead in the first EEG lead number set is combined into a first weight set; The brain-muscle interaction information extraction module is further used to: Calculating brain-muscle interaction feature values ​​based on the EEG data of all EEG leads and the EMG data of all EMG leads to obtain a brain-muscle interaction feature matrix; Arranging all brain-muscle interaction features in the brain-muscle interaction feature matrix from large to small according to numerical values ​​to obtain a second numerical sorting result; selecting a second preset number of brain-muscle interaction features from the second numerical sorting result, and forming a second EEG lead number set from the EEG lead numbers corresponding to the second preset number of brain-muscle interaction features; The brain-muscle interaction feature value corresponding to each lead in the second EEG lead number set is combined into a second weight set; The targeted brain region identification module includes: a weight set intersection unit, configured to obtain an intersection weight set based on the first weight set and the second weight set; a weighted linear summing unit, configured to linearly sum the lead weights in the first EEG lead number set and the second EEG lead number set corresponding to the lead weights in the intersection weight set to obtain a linear summed weight value; a neuromodulation targeting unit, configured to select the lead with the largest weight among the linear summed weight values ​​as a neuromodulation targeting brain region; The system is also used for: Based on the EEG signal x(t) collected by the EEG lead, the brain activation characteristic value corresponding to the lead is calculated. The calculation method is as follows: Where SpecR represents the mean value of the power spectrum density of a certain EEG signal x(t) when no motor task is performed, and SpecT represents the mean value of the power spectrum density of a certain EEG signal x(t) when a motor task is performed; Select any EEG signal x(t) and any EMG signal y(t) to calculate the brain-muscle interaction information features. The calculation method is as shown in formula (2): in, Represents the power spectrum of the observed signal, t c represents the instantaneous time of the reference observation point, f c represents a fixed observation frequency point, μ1 and μ2 represent the time delay of the EEG signal x(t) and the EMG signal y(t) observed from the reference observation point, respectively.

4. The system according to claim 3, characterized in that The weighted linear summing unit is further used for: Respectively obtaining lead coefficients of corresponding leads in the first EEG lead number set and the second EEG lead number set; Respectively obtaining weight values ​​of corresponding leads in the first EEG lead number set and the second EEG lead number set; as well as, A linear sum weight value of the corresponding lead is calculated based on the lead coefficient and the weight value of the corresponding lead.

Images