Magnetic field noise elimination method based on limb movement analysis

By constructing a brain magnetic noise model based on limb movement analysis and using regression algorithms to eliminate noise interference in brain magnetic detection, the problem of magnetic field noise caused by limb movement was solved, and the effect of brain magnetic detection was improved.

CN115245334BActive Publication Date: 2024-11-26BEIJING QUANMAG HEALTHCARE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211107381.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-09-13
Publication Date
2024-11-26
Estimated Expiration
2042-09-13

AI Technical Summary

Technical Problem

In wearable magnetoencephalography (MEG) detection, magnetic field noise interference caused by limb movement seriously affects the detection effect of MEG signals, and existing technologies are difficult to eliminate effectively.

Method used

By constructing a brain magnetic noise model based on limb movement analysis, the model is trained and its parameters are optimized using a regression algorithm. Based on limb movement data and spatial magnetic field distribution data, noise components in the brain magnetic data are removed.

Benefits of technology

It effectively eliminates magnetic field noise interference caused by limb movement, improves the convenience and operability of magnetoencephalography (MEG) detection, and achieves sensitive and accurate dynamic MEG detection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115245334B_ABST
    Figure CN115245334B_ABST
Patent Text Reader

Abstract

A method for eliminating magnetic field noise based on limb movement analysis comprises: obtaining spatial magnetic field distribution data of a magnetoencephalogram acquisition environment and a limb movement data set and an original magnetoencephalogram data set of a subject in the magnetoencephalogram acquisition environment; using the spatial magnetic field distribution data, the original magnetoencephalogram data set and the limb movement data set as input quantities, inputting the magnetoencephalogram noise model to obtain the magnetoencephalogram noise data set; removing the noise data contained in the original magnetoencephalogram data set by using the magnetoencephalogram noise data set to obtain denoised magnetoencephalogram data; solving the problem that magnetic field noise generated by movement interferes with magnetoencephalogram detection results when a wearable magnetoencephalogram device is used for magnetoencephalogram detection by analyzing and filtering the magnetoencephalogram noise signal; calculating and filtering the magnetoencephalogram noise based on limb movement analysis to improve the convenience and operability of magnetoencephalogram detection; and providing a basis for realizing sensitive and accurate dynamic magnetoencephalogram.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of computer technology, and in particular to a method for eliminating magnetic field noise based on limb movement analysis. Background Art

[0002] Magnetoencephalography (MEG) is a non-invasive brain function, brain imaging and brain disease diagnosis technology that measures the magnetic field generated by neural currents outside the head to infer brain activity inside the head. Atomic magnetometers (OPMs), which have gradually matured in recent years, have been developed as detection units for MEG because of their high sensitivity and small size.

[0003] The intensity of the EEG signal is on the order of 100fT, the noise floor of OPM is on the order of 10-20fT, and the environmental background of EEG signal detection is generally required to be on the order of 1nT. In traditional EEG acquisition, the head and the detector array are in a static position, and a relatively low magnetic field environment is created for EEG detection by magnetic shielding and magnetic compensation of the area.

[0004] The use of OPM makes wearable EEG detection possible, but even in a commercial double-layer magnetic shielding room, the gradient of the spatial magnetic field in the shielding area reaches 1-10nT / m, that is, a 1cm movement of the head and the detector will bring 10000fT of magnetic field interference, which is much greater than the EEG signal strength, and then transfers the information of the spatial magnetic field gradient to the signal time series, drowning out the effective EEG activity signal. Summary of the invention

[0005] In view of the above-mentioned limitations, the present invention proposes a method for eliminating magnetic field noise based on limb motion analysis. By modeling the noise generated by the analysis of human limb motion and eliminating the noise through the noise model, the magnetic field noise generated by the brain magnetic signal due to the movement of the detector is eliminated.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A method for eliminating magnetic field noise based on limb movement analysis, the method comprising the following steps:

[0008] Step 1, obtaining the spatial magnetic field distribution data of the brain magnetic acquisition environment and the limb movement data set and original brain magnetic data set of the test subject in the brain magnetic acquisition environment;

[0009] Step 2, using the spatial magnetic field distribution data, the original brain magnetic data set and the limb movement data set as input quantities, inputting the brain magnetic noise model to obtain the brain magnetic noise data set;

[0010] Step 3: Remove the noise data contained in the original MEG data set through the MEG noise data set to obtain denoised MEG data.

[0011] The brain magnetic noise model is constructed in the following way:

[0012] The brain magnetic noise model is constructed in the following way:

[0013] Acquire a spatial magnetic field distribution data set, wherein the spatial magnetic field distribution data set includes spatial magnetic field distribution data under a plurality of magnetic field environments;

[0014] Acquire a limb movement sample data set, wherein the limb movement sample data set includes limb movement data under a plurality of limb movement states;

[0015] Acquire a number of brain magnetic noise sample data sets corresponding to the limb movement states, wherein the brain magnetic noise sample data sets are magnetic field noise data sets that do not include original brain magnetic signals;

[0016] Based on the regression algorithm, the spatial magnetic field distribution data set, the limb movement sample data set and the brain magnetic noise sample data set are used to train the regression model and optimize the parameters to obtain the brain magnetic noise model.

[0017] Compared with the prior art, the present invention has the following advantages:

[0018] (1) By analyzing and filtering the EEG noise signal, the problem of magnetic field noise generated by movement interfering with the EEG detection results when using wearable EEG devices for EEG detection is solved;

[0019] (2) Calculation and filtering of EEG noise based on limb movement analysis improves the convenience and operability of EEG detection;

[0020] (3) The EEG noise model is used to analyze and filter the EEG noise generated by movement, providing a basis for achieving sensitive and accurate dynamic EEG.

[0021] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the following specifically cites a preferred embodiment and describes it in detail with the accompanying drawings as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 The present invention is a flowchart of a method for eliminating magnetic field noise based on limb movement analysis according to an embodiment of the present invention.

[0023] Figure 2 The figure is a schematic diagram of the structure of a magnetoencephalometer involved in an embodiment of the present invention.

[0024] Figure 3 The present invention is a schematic diagram of the structure of a magnetic field noise elimination device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The following is an explanation of the implementation of the present invention by specific specific embodiments. People familiar with the technology can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are 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 the field without creative work are within the scope of protection of the present invention. In order to further understand the present invention, the present invention is further described in detail below in conjunction with the best embodiment.

[0026] The following is an explanation of the terms involved in the embodiments of the present invention:

[0027] OPM: Optically pumped atomic magnetometer, a magnetometer made by using circularly polarized light to excite the gas atom system in the magnetic field to be measured to produce the difference in the number of particles between its Zeeman sub-energy levels, thereby observing the magnetic resonance effect; OPM is mainly used to measure weak magnetic fields.

[0028] The implementation of the present invention is described in detail below in conjunction with the aforementioned terms:

[0029] The invention point of the present invention is to provide a magnetic field noise elimination method based on limb movement analysis, which is used to analyze and eliminate the brain magnetic noise generated by limb movement during brain magnetic detection;

[0030] refer to Figure 1 , the method comprises the following steps:

[0031] Step 1, obtaining the spatial magnetic field data of the brain magnetometry acquisition environment; obtaining the limb movement data and brain magnetometry data of the subject in the brain magnetometry acquisition environment once every preset acquisition time interval T0, and excluding invalid data to form a limb movement data set and an original brain magnetometry data set respectively;

[0032] The spatial magnetic field data is a magnetic field matrix composed of magnetic field intensity values ​​at various positions in space;

[0033] The limb motion data set is composed of motion data of a plurality of action recording points corresponding to a plurality of acquisition times; the motion data is a displacement vector;

[0034] The brain magnetic data set is composed of brain magnetic data of a number of brain magnetic detection points corresponding to a number of acquisition times; the brain magnetic data is the magnetic field intensity at the brain magnetic detection points;

[0035] Step 2, using the spatial magnetic field data, the original brain magnetic data set and the limb movement data set as input quantities, inputting the brain magnetic noise model to obtain the brain magnetic noise data set;

[0036] The EEG noise data set is composed of EEG noise data of several EEG detection points corresponding to several acquisition times; the EEG noise data is the noise magnetic field intensity at the EEG detection points;

[0037] Step 3, removing the noise data contained in the original MEG data set through the MEG noise data set to obtain denoised MEG data;

[0038] The EEG noise data set includes a number of acquisition times that match the original EEG data and noise magnetic field intensity values ​​of a number of EEG detection points corresponding to each acquisition time.

[0039] The brain magnetic noise model is constructed in the following way:

[0040] Acquire a spatial magnetic field distribution data set, wherein the spatial magnetic field distribution data set includes spatial magnetic field distribution data under a plurality of magnetic field environments;

[0041] Acquire a limb movement sample data set, wherein the limb movement sample data set includes limb movement data under a plurality of limb movement states;

[0042] The limb movement data is composed of a number of sample collection times and limb displacement data corresponding to the sample collection times; the limb displacement data is the displacement vector of a number of body part data collection points;

[0043] Acquire a number of magnetoencephalogram (MEG) noise sample data sets corresponding to the limb movement states, wherein the MEG noise sample data sets are sets of magnetic field noise data that do not include original MEG signals of the brain;

[0044] The EEG noise sample data set consists of several sample collection times and magnetic field noise data sets corresponding to the sample collection times; the magnetic field noise data set is the noise magnetic field intensity values ​​of several EEG detection points.

[0045] Based on the regression algorithm, the spatial magnetic field distribution data set, the limb movement sample data set and the brain magnetic noise data set are used to train the regression model and optimize the parameters to obtain the brain magnetic noise model.

[0046] As an embodiment, the limb movement sample dataset is obtained by:

[0047] Select several test personnel of different body shapes to perform several preset actions in a magnetic shielding room;

[0048] Record the motion position data of each action recording point when performing a number of preset actions according to a preset recording interval to obtain an original motion position sample data set;

[0049] The position movement matrix of several recorded time points compared with the previous recorded time point is calculated, and the limb movement sample data set is obtained through screening and combination.

[0050] The position data are the spatial coordinates of each action recording point;

[0051] The original motion position sample data set includes a plurality of original motion position sample data subsets P; an original motion position sample data subset is used to record the limb motion data generated when a certain test person performs a certain preset action.

[0052] The original motion position sample data subset P can be expressed as: P=[P1, P2, ..., P n ];

[0053] Where P i , (i=1,2,……,n) is the position matrix at the recording time point i, and the position matrix is ​​a matrix composed of the position data, which is expressed as follows:

[0054]

[0055] in, P iX , P iY , P iZ are the coordinate matrices along the X, Y, and Z axes at the recording time point i, P iX , P iY , P iZ The elements of are the coordinate values ​​of each record point along the X, Y, and Z axes; the specific expressions are as follows:

[0056] , , .

[0057] The limb movement sample data set includes several limb movement sample data subsets Q;

[0058] The limb movement sample data subset Q can be expressed as: Q=[Q1,Q2,……,Q n ];

[0059] Where Q i , (i=1,2,……,n) is the position movement matrix at the recording time point i, expressed as follows:

[0060]

[0061] in, Q iX , Q iY , Q iZ are the position movement matrices along the X, Y, and Z axes at the recording time point i, Q iX , Q iY , Q iZ The elements of are the displacements of each recording point along the X, Y, and Z axes compared to the recording time point i-1.

[0062] i The calculation method is:

[0063]

[0064] Where P i is the position matrix at the time point i, P i-1 is the position matrix at the previous recording time point of recording time point i.

[0065] As an embodiment, the action recording points are composed of body action recording points and head action recording points;

[0066] The head movement recording point is set on the brain magnetoencephalogram instrument, and the specific position is determined according to the position of the detector on the brain magnetoencephalogram instrument;

[0067] Generally, an action recording point is set at each detector position; during the actual model training process, the action recording points can be added, removed, or moved according to the data processing effect to improve the effectiveness of the data recorded by the action recording points.

[0068] As an embodiment, the magnetic brain noise sample data set is obtained by:

[0069] When the tester performs a number of preset actions, the noise magnetic field data of each EEG recording point is recorded at a preset recording time interval to obtain an original EEG noise sample data set;

[0070] The original magnetic brain noise sample data is screened to obtain the magnetic brain noise sample data.

[0071] The noise magnetic field data is collected by a magnetoencephalometer worn by the tester; when obtaining the magnetoencephalogram noise sample, the distance between the OPM detector on the magnetoencephalometer and the human brain is set to a preset distance D0.

[0072] It should be noted that when the distance between the OPM detector on the magnetoencephalometer and the human brain is set to the preset distance D0, the magnetic field strength of the human brain will attenuate at the OPM detector to a level that does not affect the noise magnetic field acquisition. At this time, the magnetic field strength data collected by the magnetoencephalometer is the noise magnetic field data.

[0073] The original magnetoencephalogram noise sample data set includes several original magnetoencephalogram noise sample data subsets M0; an original magnetoencephalogram noise sample data subset is used to record the magnetoencephalogram noise data generated when a certain test person performs a certain preset action.

[0074] The original EEG noise sample data subset M0 is composed of EEG noise matrices corresponding to several recording time points, that is, M0=[ M 01 , M 02 ,……, M 0j ],

[0075] Among them, M 0j , (j=1,2,…,k) is the EEG noise matrix at the recording time point j, expressed as follows:

[0076]

[0077] Among them, M 0j The elements of are the magnetic field intensity values ​​at each brain magnetic recording point.

[0078] The magnetic brain noise sample data includes a plurality of magnetic brain noise sample data subsets M;

[0079] The magnetic brain noise sample data subset M is obtained by eliminating invalid recorded data (including empty data and abnormal data) in M0, and the elimination rule is: if an element in M ​​contains invalid recorded data, the element is discarded.

[0080] After the magnetoencephalogram noise sample data is screened, the corresponding data needs to be eliminated from the limb movement sample data set according to the recording time corresponding to the eliminated data, to ensure that the number of elements in the magnetoencephalogram noise sample data set is equal to that of the limb movement sample data set and the sequence numbers match.

[0081] It should be noted that the magnetoencephalogram noise sample data set and the limb movement sample data set should be obtained synchronously, that is, the limb movement sample data set should be obtained at the same time as the magnetoencephalogram noise sample data set to ensure that the recording time points of the two are consistent.

[0082] As an embodiment, the regression model is constructed in the following manner:

[0083] Taking the limb movement sample data set and the spatial magnetic field distribution data set as input samples, and taking the brain magnetic noise sample data set as target output samples;

[0084] Selecting a regression algorithm and performing regression analysis on the input sample using the selected regression algorithm to obtain a predicted output sample;

[0085] The output value of the preset objective function is determined based on the labeled output sample and the predicted output sample, and the parameters of the regression algorithm are adjusted based on the output value until the preset model index is met, and the model information is output.

[0086] The objective function is in the form of:

[0087]

[0088] in, J(f) is regularization, which is used to describe the complexity of the regression model; L ( y i , f ( x i )) is the cost function, which is as follows:

[0089]

[0090] in, x i is the input value, y i is the target output value, f ( x i ) is the predicted output value;

[0091] J(f) It can be implemented using the L1 norm (Lasso regression) or the L2 norm (Ridge regression).

[0092] The regression algorithm is implemented by a combination of one or more algorithms selected from general linear regression, generalized linear regression, stepwise regression, ridge regression, Lasso regression, elastic net regression, and neural network.

[0093] The above regression algorithms are all existing mature technical means. Those skilled in the art can successfully implement the above regression algorithms according to the description of the embodiments, and will not be described in detail here.

[0094] As an embodiment, the limb movement data set and the limb movement sample data set required for constructing the magnetoencephalographic noise model are obtained by a combination of one or more methods including motion sensors, video acquisition cameras, and depth sensing cameras.

[0095] The methods of obtaining limb movement data using motion sensors, video acquisition cameras and depth-sensing cameras are all conventional technical means, which can be successfully implemented by those skilled in the art through the description of the above embodiments and will not be elaborated here.

[0096] As an embodiment, the original EEG data set is obtained by a magnetoencephalometer, and the magnetoencephalometer acquires EEG signals based on a quantum detector (OPM detector) with quantum manipulation technology as its core.

[0097] The quantum detector is a magnetic field sensitive device, which consists of an alkali metal gas chamber, a light source, a heating plate, an optical element, a photoelectric sensor, a modulation coil and a photoelectric detector. It is characterized by its small size, easy to wear, and can work normally at room temperature.

[0098] refer to Figure 2 In addition to a number of OPM detectors 4, the brain magnetometer also includes one or more structures of a brain magnet cap 1 with a slot, a slot 2, and a limiting sleeve 3.

[0099] When the brain magnetic cap is made of a rigid material, the method of the present invention can eliminate the brain magnetic noise generated by movement without any modification.

[0100] When the brain magnetic cap is made of a flexible material, in step 1, the data of the original brain magnetic data set must include detector position data in addition to the brain magnetic data, and the detector position data is the position coordinates of each detector;

[0101] At the same time, when training the noise model, the input samples include a detector position data set in addition to the limb movement sample data set; the detector position data set is composed of a plurality of the detector position data.

[0102] The brain magnetic acquisition environment is composed of one or more structures including a magnetic shielding room, a spatial magnetic compensation coil, and a magnetic interference-free data acquisition device; the magnetic interference-free data acquisition device is composed of one or more structures including the motion sensor, the video acquisition camera, and the depth sensing camera.

[0103] As an embodiment, the specific calculation method of denoising the brain magnetic data in step 3 is:

[0104]

[0105] Among them B i (0) is the i-th original MEG data set element, representing the i-th original MEG data matrix; B i (N) is the i-th element of the MNE data set, representing the i-th MNE noise matrix; B i (0) With Bi (N) is a homogeneous matrix, the specific form is as follows:

[0106]

[0107] where u ij (0) is the original magnetic field strength at position ij;

[0108]

[0109] where u ij (N) is the magnetic field intensity of brain magnetic noise at position ij.

[0110] As an example, Figure 3 The device shown includes a brain magnetic signal collection module, a motion signal collection module, a noise analysis and filtering module, and a model training and optimization module;

[0111] The EEG signal collection module is used to collect and process raw EEG signals and EEG noise sample data;

[0112] The motion signal acquisition module is used to collect and process limb motion signals and limb motion sample data;

[0113] The noise analysis and filtering module is used to perform noise analysis and filtering on the original EEG signals collected by the EEG signal collection module according to the EEG noise model to obtain real EEG signals;

[0114] The model training and optimization module is used to train and optimize the magnetic brain noise model.

[0115] As an embodiment, the method of the present invention may be implemented in software and / or a combination of software and hardware, for example, may be implemented using an application specific integrated circuit (ASIC), a general purpose computer or any other similar hardware device.

[0116] The method described in the present invention can be implemented in the form of a software program, and the software program can be executed by a processor to implement the steps or functions described above. Similarly, the software program (including related data structures) can be stored in a computer-readable recording medium, such as a RAM memory, a magnetic or optical drive or a floppy disk and the like.

[0117] In addition, some steps or functions of the method described in the present invention may be implemented by hardware, for example, as a circuit that cooperates with a processor to execute each step or function.

[0118] In addition, a part of the method described in the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present application through the operation of the computer. The program instruction for calling the method described in the present invention may be stored in a fixed or removable recording medium, and / or transmitted through a data stream in a broadcast or other signal-bearing medium, and / or stored in a working memory of a computer device that runs according to the program instruction.

[0119] As an embodiment, the present invention also provides a device, which includes a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the device is triggered to run the methods and / or technical solutions based on the aforementioned multiple embodiments.

[0120] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0121] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0122] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of the technical solution of the present invention.

Claims

1. A method for eliminating magnetic field noise based on limb movement analysis, characterized in that: The method comprises the following steps: Step 1, obtaining the spatial magnetic field distribution data of the brain magnetic acquisition environment and the limb movement data set and original brain magnetic data set of the test subject in the brain magnetic acquisition environment; Step 2, using the spatial magnetic field distribution data, the original brain magnetic data set and the limb movement data set as input, inputting the brain magnetic noise model to obtain the brain magnetic noise data set; Step 3: Remove the noise data contained in the original MEG data set through the MEG noise data set to obtain denoised MEG data.

2. The method according to claim 1, characterized in that The brain magnetic noise model is constructed in the following way: Acquire a spatial magnetic field distribution data set, wherein the spatial magnetic field distribution data set includes spatial magnetic field distribution data under a plurality of magnetic field environments; Acquire a limb movement sample data set, wherein the limb movement sample data set includes limb movement data under a plurality of limb movement states; Acquire a number of brain magnetic noise sample data sets corresponding to the limb movement states, wherein the brain magnetic noise sample data sets are magnetic field noise data sets that do not include original brain magnetic signals; Based on the regression algorithm, the spatial magnetic field distribution data set, the limb movement sample data set and the brain magnetic noise sample data set are used to train the regression model and optimize the parameters to obtain the brain magnetic noise model.

3. The method according to claim 2, characterized in that The limb movement sample dataset is obtained by: Select several personnel of different body types to perform several preset actions in a magnetic shielding room; Record the motion position data of each action recording point when performing a number of preset actions according to a preset recording interval to obtain an original motion position sample data set; The position data is the spatial coordinates of each action recording point; the original motion position sample data set is composed of a number of recording time points and the position matrix corresponding to each recording time point; the position matrix is ​​a matrix composed of the position data; The position movement matrix of several recorded time points compared with the previous recorded time point is calculated, and the limb movement sample data set is obtained through screening and combination.

4. The method according to claim 2, characterized in that: The magnetic brain noise sample dataset is obtained in the following way: When the tester performs a number of preset actions, the noise magnetic field data of each EEG recording point is recorded at a preset recording time interval to obtain an original EEG noise sample data set; The original magnetic brain noise sample data is screened to obtain magnetic brain noise sample data; The original EEG noise sample data set consists of a number of recording time points and the EEG noise matrix corresponding to each recording time point; the EEG noise matrix is ​​a matrix composed of the noise magnetic field data, and the matrix elements are set according to the position of the EEG recording point.

5. The method according to claim 2, characterized in that: The regression model is trained by: Taking the limb movement sample data set and the spatial magnetic field distribution data set as input samples, and taking the brain magnetic noise sample data set as target output samples; Selecting a regression algorithm and performing regression analysis on the input sample using the selected regression algorithm to obtain a predicted output sample; The output value of the preset objective function is determined based on the labeled output sample and the predicted output sample, and the parameters of the regression algorithm are adjusted based on the output value until the preset model index is met, and the model information is output.

6. The method according to claim 2, characterized in that The regression algorithm is implemented by a combination of one or more algorithms selected from general linear regression, generalized linear regression, stepwise regression, ridge regression, Lasso regression, elastic net regression, and neural network.

7. The method according to claim 1, characterized in that The limb movement data set is obtained by a combination of one or more methods including motion sensors, video acquisition cameras, and depth sensing cameras.

8. The method according to claim 1, characterized in that The original brain magnetic data set is obtained by a brain magnetic instrument, which is based on a quantum detector with quantum manipulation technology as the core to achieve brain magnetic signal acquisition; The brain magnetic field acquisition environment is composed of one or more structures of a magnetic shielding room, a spatial magnetic compensation coil, and a magnetic interference-free data acquisition device.

9. A device for eliminating brain magnetic noise, characterized in that: The device includes a magnetoencephalographic signal collection module, a motion signal collection module, a noise analysis and filtering module, and a model training and optimization module; The EEG signal collection module is used to collect and process raw EEG signals and EEG noise sample data; The motion signal acquisition module is used to collect and process limb motion signals and limb motion sample data; The noise analysis and filtering module is used to perform noise analysis and filtering on the original EEG signals collected by the EEG signal collection module according to the EEG noise model to obtain real EEG signals; specifically, the spatial magnetic field data of the EEG collection environment is obtained, and the spatial magnetic field data, the original EEG data set and the limb movement data set are used as input, and the EEG noise model is input to obtain the EEG noise data set, and the noise data contained in the original EEG data set is removed by the EEG noise data set to obtain the denoised EEG data; The model training and optimization module is used to train and optimize the magnetic brain noise model.

10. An electronic device, characterized in that: The electronic device comprises a memory and a processor; a computer program is stored in the memory, and the computer program is configured to run any method in claims 1-8; and the processor is configured to run the computer program to execute any method in claims 1-8.