Brain magnetic signal preprocessing method, device and equipment

Abstract

The invention relates to the technical field of brain-computer interfaces, and discloses a brain magnetic signal preprocessing method, device and equipment, and the method comprises the steps: carrying out the multi-strategy filtering of an original brain magnetic signal, and fusing a filtering result, and obtaining a fused filtering signal; the multi-strategy filtering comprises at least one filtering mode suitable for a linear system and at least one filtering mode suitable for a nonlinear system; performing frequency decomposition and feature extraction based on the fused filtering signal to obtain a signal modal feature; and performing signal reconstruction according to the signal modal characteristics meeting the relevancy condition to obtain a reconstructed brain magnetic signal. The method has the beneficial effects that adaptive noise elimination is performed on the brain magnetic signals through multi-strategy filtering, so that the advantages of different filtering modes are combined, the influence of dynamic attributes of the brain magnetic signals on signal classification is reduced, the quality of the brain magnetic signals is effectively improved, the classification effect and the classification accuracy of the brain magnetic signals are improved, and the classification efficiency of the brain magnetic signals is improved. And more accurate brain magnetic signal characteristics are provided for subsequent research.

Description

Technical Field

[0001] The present application relates to the field of brain-computer interface technology, and in particular to a method, device and equipment for preprocessing brain magnetic signals. Background Art

[0002] Brain-Computer Interface (BCI) refers to the technology of deciphering specific intentions from brain activity signals and converting them into machine instructions to control external devices. Based on BCI technology, users can use the event-related potential (ERP) triggered by motor imagery (MI) as an active neural activity feature to be used as a control signal in the active-BCI system to control external devices. However, due to the high impedance of the brain skull, it is not conducive to the passage of high-frequency signals, resulting in a large signal attenuation in the collected EEG signals in actual situations. And as a scalar field, EEG signals will cause the loss of spatial features in EEG signals when tracing or inverting EEG signals, which will affect the study of brain neural activity. In contrast, EEG signals will not be affected and attenuated when penetrating the brain skull, and EEG signals, as vector fields, are superior to EEG signals in terms of temporal and spatial resolution. Therefore, EEG signals can replace EEG signals and be applied to BCI technology, and BCI systems using EEG signals are also emerging in an endless stream.

[0003] In related technologies, the signal quality of magnetic brain signals still needs to be improved. Summary of the invention

[0004] The present application provides a method, device and equipment for preprocessing EEG signals, which perform adaptive noise elimination on EEG signals through multi-strategy filtering, thereby combining the advantages of different filtering methods, reducing the impact of the dynamic properties of EEG signals on signal classification, and effectively improving the quality of EEG signals.

[0005] In order to achieve the above objectives, the main technical solutions adopted in this application include: In a first aspect, an embodiment of the present application provides a method for preprocessing a magnetic brain signal, the method comprising: Performing multi-strategy filtering on the original magnetoencephalogram signal and fusing the filtering results to obtain a fused filtered signal; wherein the multi-strategy filtering includes at least one filtering method applicable to a linear system and at least one filtering method applicable to a nonlinear system; Performing frequency decomposition and feature extraction based on the fused filtered signal to obtain signal modal features; The original magnetoencephalogram signal is reconstructed according to the signal modal features that meet the correlation condition to obtain a reconstructed magnetoencephalogram signal.

[0006] The method for preprocessing EEG signals proposed in the embodiment of the present application combines the filtering method applicable to linear systems and the filtering method applicable to nonlinear systems, uses a multi-strategy filtering method to adaptively eliminate noise on the original EEG signals, and reconstructs the signal based on the obtained fused filtered signal, thereby obtaining a reconstructed EEG signal after preprocessing. The reconstructed EEG signal can be used as the signal basis for EEG signal classification, and provides accurate EEG signal characteristics for brain-computer interface technology research. Compared with the related art, the present application combines a variety of filtering methods, processes the stationary noise in the original EEG signal by linear filtering, and processes the complex nonlinear signal in the original EEG model by nonlinear filtering, reduces the impact of nonlinear interference in the original EEG signal, effectively improves the filtering effect, and improves the robustness of the EEG signal. The obtained filtering results are fused to obtain a fused filtered signal, so that the advantages of different filtering methods can be combined to improve the quality of the EEG signal, thereby improving the classification effect and classification accuracy of the EEG signal.

[0007] Optionally, the original EEG signal is subjected to multi-strategy filtering, and the filtering results are fused to obtain a fused filtering signal, including Performing linear system filtering on the original brain magnetic signal to obtain a linear filtered signal; Performing nonlinear system filtering on the original brain magnetic signal to obtain a nonlinear filtered signal; The linear filter signal and the nonlinear filter signal are fused to obtain the fused filter signal.

[0008] Optionally, performing linear system filtering on the original magnetoencephalogram signal to obtain a linear filtered signal includes: updating the estimated weight of the sample reference value corresponding to the original magnetoencephalogram signal according to the fusion filter signal at the previous moment; Obtaining a motion artifact noise matrix of the original magnetoencephalogram signal using the sample reference value and the estimated weight of the updated sample reference value; The original magnetoencephalogram signal is subjected to linear system filtering according to the motion artifact noise matrix to obtain the linear filtering signal.

[0009] Optionally, the original brain magnetometry signal includes a radial signal and a tangential signal; and performing nonlinear system filtering on the original brain magnetometry signal to obtain a nonlinear filtered signal includes: According to the fusion filtering signal of the radial signal at the first moment and the external stimulation input signal at the second moment, a predicted state of the radial signal at the current moment is obtained as a radial predicted state; wherein the first moment and the second moment are both before the current moment; Obtaining a predicted state of the tangential signal at a current moment according to a fused filter signal of the tangential signal at a first moment and an external stimulus input signal at a second moment as a tangential predicted state; The radial prediction state and the tangential prediction state are component-coupled to obtain the coupled prediction state of the original magnetoencephalogram signal; and the coupled prediction state is state-observed to obtain the nonlinear filtering signal.

[0010] Optionally, fusing the linear filtering signal and the nonlinear filtering signal to obtain the fused filtering signal includes: According to the signal-to-noise ratio of each of the linear filtering signal and the nonlinear filtering signal, a preset weight of the linear filtering signal is adjusted to obtain a fusion weight of the linear filtering signal; The linear filter signal and the nonlinear filter signal are weightedly calculated according to the fusion weight to obtain the fusion filter signal.

[0011] Optionally, reconstructing the original magnetoencephalogram signal according to the signal modality characteristics satisfying the correlation condition to obtain the reconstructed magnetoencephalogram signal includes: Performing information entropy detection on the signal modal features to obtain the correlation of the signal modal features; Screening the signal modal features according to the correlation condition to obtain signal reconstruction features that meet the correlation condition; Signal reconstruction is performed according to the signal reconstruction feature to obtain the reconstructed electroencephalographic signal.

[0012] Optionally, the method further comprises: The reconstructed magnetic brain signal is classified to obtain a magnetic brain signal classification result.

[0013] Optionally, classifying the reconstructed magnetic brain signal to obtain a magnetic brain signal classification result includes: The reconstructed magnetoencephalogram signal and the corresponding label set are input into a deep learning neural network model to classify the reconstructed magnetoencephalogram signal and output the classification result of the magnetoencephalogram signal; wherein the deep learning neural network model is trained by the reconstructed magnetoencephalogram signal before the current moment.

[0014] In a second aspect, an embodiment of the present application provides a device for preprocessing electroencephalographic signals, the device comprising: A fusion filtering module, used for performing multi-strategy filtering on the original magnetoencephalogram signal and fusing the filtering results to obtain a fused filtered signal; wherein the multi-strategy filtering includes at least one filtering method applicable to a linear system and at least one filtering method applicable to a nonlinear system; A feature extraction module, used for performing frequency decomposition and feature extraction based on the fused filtered signal to obtain signal modal features; The signal reconstruction module is used to reconstruct the original magnetoencephalogram signal according to the signal modal characteristics that meet the correlation condition to obtain a reconstructed magnetoencephalogram signal.

[0015] In a third aspect, an embodiment of the present application provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the method described in any one of the above embodiments by executing the computer instructions.

[0016] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which computer instructions are stored, and the computer instructions are used to enable a computer to execute any one of the methods in the above embodiments.

[0017] In a fifth aspect, an embodiment of the present application provides a computer program product, including computer instructions, and the computer instructions are used to enable a computer to execute any one of the methods in the above embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 A step diagram of a method for preprocessing electroencephalographic signals provided in an embodiment of the present application; Figure 2a In the embodiment of this application, Location Schematic diagram of signal modal characteristics obtained by component; Figure 2b In the embodiment of this application, Location Schematic diagram of signal modal characteristics obtained by component analysis; Figure 3 A diagram showing the steps of obtaining a fused filter signal in an embodiment of the present application; Figure 4 A diagram showing the steps of filtering a linear system in an embodiment of the present application; Figure 5 A diagram showing the steps of filtering a nonlinear system in an embodiment of the present application; Figure 6 This is a step diagram of signal fusion in an embodiment of the present application; Figure 7 A diagram showing the steps of obtaining a reconstructed brain magnetic signal in an embodiment of the present application; Figure 8a In the embodiment of this application, Location Schematic diagram of the reconstructed EEG signal obtained by the component; Figure 8b In the embodiment of this application, Location Schematic diagram of the reconstructed EEG signal obtained by the component; Fig. 9 A module diagram of a device for preprocessing electroencephalographic signals provided in an embodiment of the present application; Fig.10 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution 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 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 those skilled in the art without creative work are within the scope of protection of the present application.

[0021] Brain-Computer Interface (BCI) refers to the technology of deciphering specific intentions from brain activity signals and converting them into machine instructions to control external devices. Based on BCI technology, users can use the event-related potential (ERP) triggered by motor imagery (MI) as an active neural activity feature to be used as a control signal in the active-BCI system to control external devices.

[0022] Among the current related technologies, the brain activity signals used by brain-computer interface technology are usually the user's EEG signals. EEG signals are non-invasive signals with high temporal resolution, which can reflect changes in brain activity in real time, facilitating real-time control and interaction. Compared with other brain activity signals, EEG signal devices are low-cost and easy to use, suitable for long-term wear and real-time monitoring.

[0023] However, in reality, the brain skull has high impedance, which is not conducive to the passage of high-frequency signals. During the signal acquisition process, EEG signals will experience significant signal attenuation when they penetrate the skull to reach the electrode. In addition, EEG signals are a scalar field. When conducting in-depth research on brain nerve activity, if the EEG signals are traced or inverted, the spatial features in the EEG signals will be lost, which will affect the research on brain nerve activity.

[0024] With the development of atomic magnetometers, EEG signals are widely used as a brain activity signal. Compared with EEG signals, EEG signals will not be affected and attenuated when penetrating the skull, and as a vector field, EEG signals are superior to EEG signals in terms of temporal and spatial resolution. Therefore, EEG signals can replace EEG signals in brain-computer interface technology, and brain-computer interface systems using EEG signals are emerging in an endless stream.

[0025] However, since the atomic magnetometer is susceptible to low-frequency disturbances, when measuring brain magnetic signals using an atomic magnetometer, the brain magnetic signals will inevitably be contaminated by physiological artifacts and non-physiological artifacts. Especially in real-time closed-loop brain-machine interface (BMI) applications, effective means are needed to process the artifact noise that appears in the brain magnetic signals.

[0026] In related technologies, the means of removing artifacts from EEG signals include Independent Component Analysis (ICA) and Principal Components Analysis (PCA). However, the calculation process of ICA and PCA is highly complex, and the filtering effect obtained is not ideal. In the actual brain-computer interface working environment, it is difficult for related technologies to provide a comprehensive and sufficient method to classify all characteristic components in EEG signals, which is also a defect of the current brain-computer interface system based on atomic magnetometers.

[0027] Based on the above problems, the present application provides a method, device and equipment for preprocessing EEG signals, the method comprising: performing multi-strategy filtering on the original EEG signals, and fusing the filtering results to obtain a fused filtered signal; the multi-strategy filtering includes at least one filtering method suitable for linear systems and at least one filtering method suitable for nonlinear systems; performing frequency decomposition and feature extraction based on the fused filtered signal to obtain signal modal characteristics; performing signal reconstruction according to the signal modal characteristics that meet the correlation conditions to obtain a reconstructed EEG signal.

[0028] The method for preprocessing EEG signals provided in this application combines the filtering method applicable to linear systems with the filtering method applicable to nonlinear systems, uses a multi-strategy filtering method to adaptively eliminate noise from the original EEG signals, and reconstructs the signals based on the obtained fusion filtered signals, thereby obtaining the reconstructed EEG signals after preprocessing. The reconstructed EEG signals can serve as the signal basis for EEG signal classification, and provide accurate EEG signal characteristics for brain-computer interface technology research.

[0029] Compared with the related art, this application combines multiple filtering methods, processes the stationary noise in the original brain magnetic signal by linear filtering, and processes the complex nonlinear signal in the original brain magnetic model by nonlinear filtering, reducing the impact of nonlinear interference in the original brain magnetic signal, effectively improving the filtering effect, and improving the robustness of the brain magnetic signal. The obtained filtering results are fused to obtain a fused filtered signal, so that the advantages of different filtering methods can be combined to improve the quality of the brain magnetic signal, thereby improving the classification effect and classification accuracy of the brain magnetic signal.

[0030] The method for preprocessing brain magnetic signals provided in this specification can be applied to classify brain magnetic signals collected by an atomic magnetometer. The type of atomic magnetometer can be any of the magnetometers such as superconducting quantum interference device (SQUID) magnetometer, optically pumped atomic magnetometer (OPM), coherent placement trapped (CPT) atomic magnetometer, non-self-selected exchange relaxation (SERF) magnetometer or nonlinear magneto-optical rotation (NMOR) atomic magnetometer. When the present application is applied to different types of atomic magnetometers, certain adaptive modifications are required. It is understandable that after adaptive modification, the present application can also be used to filter and classify brain magnetic signals collected by other devices, or other signals other than brain magnetic signals.

[0031] According to an embodiment of the present application, an embodiment of a method for preprocessing magnetic brain signals is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0032] In this embodiment, a method for preprocessing brain magnetic signals is provided, which can be used for the brain magnetic signals collected by the atomic magnetometer. Figure 1 As shown, the method includes: S100. Perform multi-strategy filtering on the original magnetoencephalogram signal, and fuse the filtering results to obtain a fused filtered signal; wherein the multi-strategy filtering includes at least one filtering method applicable to a linear system and at least one filtering method applicable to a nonlinear system.

[0033] S200. Perform frequency decomposition and feature extraction based on the fused filtered signal to obtain signal modal features.

[0034] S300. Reconstruct the original magnetoencephalogram signal according to the signal modal features that meet the correlation condition to obtain a reconstructed magnetoencephalogram signal.

[0035] The raw brain magnetic signal may be a signal obtained by measuring a user who is performing motor imagery using an atomic magnetometer. There may be multiple atomic magnetometers, which are respectively placed at different positions on the user's head, and each atomic magnetometer may contain multiple data channels. For example, there may be five atomic magnetometers, and the positions of the atomic magnetometers may follow the 10-20 international standard leads, and may be respectively placed at , , , and in which is the reference position, for and The area between for and In the area between , each position contains two data channels, and the data channels respectively represent the radial magnetic field component and the tangential magnetic field component of the position.

[0036] Multi-strategy filtering can be a filtering process that includes two or more different filtering methods, including at least one filtering method applicable to linear systems and at least one filtering method applicable to nonlinear systems. Among them, the filtering method applicable to linear systems can be one or more of Kalman filtering, H∞ filtering, minimum mean square filtering, recursive least square filtering, and minimum mean square error filtering. The filtering method applicable to nonlinear systems can be one or more of particle filtering, adaptive filtering, Gaussian process regression filtering, neural network filtering, and fuzzy filtering.

[0037] It should be noted that the present application can be repeatedly performed during the real-time process of collecting EEG signals, and the real-time collected EEG signals can be pre-processed so that they can be used for real-time classification of EEG signals to quickly and accurately obtain characteristic component signals corresponding to the motor imagination task currently performed by the user.

[0038] Specifically, the original brain magnetic signals collected by the atomic magnetometer are subjected to multi-strategy filtering, and the filtering method used can be selected based on the actual scene needs and the characteristics of the filtering method itself. After multi-strategy filtering, the filtering results of various filtering methods can be obtained, and the filtering results are fused to obtain a fused filtering signal, so that the advantages of different filtering methods can be combined to further improve the filtering effect of the original brain magnetic signals.

[0039] It should be noted that in the present application, the original brain magnetic signals are filtered by a filtering method suitable for linear systems, thereby removing the stationary noise in the original brain magnetic signals and effectively reducing the interference of noise on the original brain magnetic signals. In the present application, the original brain magnetic signals are also filtered by a filtering method suitable for nonlinear systems, so that the nonlinear behavior and nonlinear characteristics in the original brain magnetic signals can be effectively processed, the impact of nonlinear interference is reduced, and the quality and robustness of the original brain magnetic signals are improved. Compared with the technical solutions in the related art that only use a single filtering method, the present application combines the filtering results obtained by different filtering methods. While utilizing the denoising ability of the linear filtering method, it also utilizes the ability of the nonlinear filtering method to target nonlinear interference in the signal, thereby comprehensively improving the signal quality of the original brain magnetic signals from different angles, and providing a high-quality signal data foundation for subsequent signal reconstruction and signal classification processes.

[0040] Exemplarily, the filtering method applicable to linear systems may be H∞ filtering, and the filtering method applicable to nonlinear systems may be particle filtering. H∞ filtering has strong robustness and adaptability, less dependence on system models, and a wide range of applications. Particle filtering is applicable to nonlinear systems and non-Gaussian systems, can handle complex problems in high-dimensional state spaces, has high adaptability, and does not require an accurate system model. In the case of using only H∞ filtering, the filtering process cannot handle the nonlinear behavior and nonlinear characteristics in the original brain magnetic signal, resulting in limited filtering effect. In the case of using only particle filtering, due to the high computational complexity of particle filtering, the amount of calculation in the filtering process will be too large. By combining the filtering results of H∞ filtering and particle filtering, the linear part and the nonlinear part in the original brain magnetic signal can be processed separately, the filtering effect is improved, and the robustness and signal quality of the original brain magnetic signal are improved.

[0041] Furthermore, the fused filter signal is used as a signal basis, and the fused filter signal is frequency decomposed to extract the part related to the brain activity signal from the fused filter signal. Exemplarily, the extracted signal frequency range can be between 8Hz and 30Hz, wherein the alpha wave signal between 8Hz and 12Hz represents the brain activity signal of the user in a sober and quiet state, and the beta wave signal between 12Hz and 30Hz represents the brain activity signal of the user in a busy state with active thinking. By collecting signals within these two ranges, the user's intention can be accurately identified, thereby obtaining effective signal support.

[0042] The frequency decomposition process is specifically described. The frequency decomposition method can be a discrete wavelet transform (DWT) algorithm. The discrete wavelet transform algorithm is used to perform frequency decomposition on the fused filter signal, and the fused filter signal is decomposed into multiple narrowband signals as the initial frequency band signal. In this embodiment, according to the characteristics of the brain magnetic signal and the power frequency interference characteristics, the discrete wavelet transform algorithm uses the db4 wavelet as the wavelet basis function, and the downsampling multiple is 2, the number of wavelet decomposition layers is 2.

[0043] Specifically explain the process of feature extraction. Further, the method of modal feature extraction can be an empirical mode decomposition algorithm (EMD). The process of extracting modal features from the initial frequency band signal includes: according to the maximum points of the initial frequency band signal, the maximum points are connected by a cubic spline curve to obtain an upper envelope. Correspondingly, according to the minimum points of the initial frequency band signal, the minimum points are connected by a cubic spline curve to obtain a lower envelope. The modal features of the initial frequency band signal are calculated according to the average value of the upper envelope and the lower envelope, and the form thereof can be expressed according to the following formula: in, is the average value of the upper envelope and the lower envelope; Represents the initial modal characteristics of the initial frequency band signal. After obtaining the initial modal characteristics, determine whether it meets the intrinsic mode function (IMF) condition. The intrinsic mode function conditions may include: 1. In the entire data segment, the number of extreme points and the number of zero-crossing points must be equal or the difference cannot exceed one at most; 2. At any time, the average value of the upper envelope formed by the local maximum point and the lower envelope formed by the local minimum point is zero, that is, the upper envelope and the lower envelope are locally symmetrical with respect to the time axis. In addition to the above conditions, the corresponding intrinsic mode function conditions can also be selected according to the actual scenario needs.

[0044] If the initial modal characteristics of the initial frequency band signal If the intrinsic mode function condition is met, the initial mode feature As the first modal feature If the initial modal characteristics of the initial frequency band signal If the intrinsic mode function condition is not met, the initial mode feature As the basic data, repeat the above steps of obtaining the upper envelope and the lower envelope and obtaining the modal characteristics until the initial modal characteristics that meet the conditions of the intrinsic modal function are obtained as the first modal characteristics .

[0045] In order to calculate other modal features, the first modal feature is removed from the initial frequency band signal to obtain the intermediate frequency band signal, which is expressed according to the following formula: in, is the intermediate frequency band signal. The above steps of obtaining the upper envelope and the lower envelope to obtain the modal characteristics and removing the modal characteristics from the intermediate frequency band signal for updating are performed until the updated intermediate frequency band signal becomes a monotonic function. At this time, the signal modal characteristics can be obtained: And the residual, the form of the residual is expressed according to the following formula: in, is the residual; To repeat the above steps The intermediate frequency band signal after that.

[0046] It should be noted that the discrete wavelet transform algorithm and the empirical mode decomposition algorithm are combined in this application, and the discrete wavelet transform algorithm provides multi-scale signal decomposition capabilities, extracts signals of different frequency components, and adaptively decomposes and extracts features of the extracted signals through the empirical mode decomposition algorithm to obtain characteristic components in the signal. By combining the two methods, the performance of this application in blind source separation is improved, so that the key signal modal features in the brain magnetic signal can still be separated under the interference of large power frequency noise, providing more accurate signal features for brain magnetic signal classification.

[0047] by For example, refer to Figure 2a to Figure 2b As shown, Figure 2a Shown by Location The six signal modal characteristics and residuals obtained by the components are Figure 2b Shown by Location The six signal modal characteristics and residuals obtained by the components. Figure 2a It can be seen that for The signal modal characteristics obtained by the component, the signal modal characteristics , and The correlation between is higher than that of other signal modal features. Figure 2b It can be seen that for The signal modal characteristics obtained by the component, the signal modal characteristics , and The correlation between them is higher than that of other signal modal features.

[0048] Furthermore, the original EEG signal is reconstructed according to the extracted signal to obtain a reconstructed EEG signal. Through signal reconstruction, the interference of irrelevant components in the original EEG signal is removed, and only effective information is extracted, thereby improving the quality of the EEG signal and enhancing the expression effect of key characteristic components in the EEG signal, so as to be used in the subsequent EEG signal classification process and improve the accuracy of EEG signal classification.

[0049] The method for preprocessing EEG signals provided in this embodiment combines a filtering method suitable for a linear system with a filtering method suitable for a nonlinear system, uses a multi-strategy filtering method to adaptively eliminate noise from the original EEG signals, and reconstructs the signals based on the obtained fusion filtered signals, thereby obtaining a reconstructed EEG signal after preprocessing. The reconstructed EEG signal can be used as the signal basis for EEG signal classification, providing accurate EEG signal characteristics for brain-computer interface technology research.

[0050] Compared with the related art, this application combines multiple filtering methods, processes the stationary noise in the original brain magnetic signal by linear filtering, and processes the complex nonlinear signal in the original brain magnetic model by nonlinear filtering, reducing the impact of nonlinear interference in the original brain magnetic signal, effectively improving the filtering effect, and improving the robustness of the brain magnetic signal. The obtained filtering results are fused to obtain a fused filtered signal, so that the advantages of different filtering methods can be combined to improve the quality of the brain magnetic signal, thereby improving the classification effect and classification accuracy of the brain magnetic signal.

[0051] Reference Figure 3 As shown, as an embodiment of the present application, the original brain magnetic signal is subjected to multi-strategy filtering, and the filtering results are fused to obtain a fused filtering signal, including S110. Perform linear system filtering on the original brain magnetic signal to obtain a linear filtered signal.

[0052] S120. Perform nonlinear system filtering on the original brain magnetic signal to obtain a nonlinear filtered signal.

[0053] S130. Fusing the linear filtering signal and the nonlinear filtering signal to obtain a fused filtering signal.

[0054] Specifically, the filtering methods used in the multi-strategy filtering include at least one filtering method applicable to linear systems and at least one filtering method applicable to nonlinear systems. The present application uses a filtering method applicable to a linear system to perform linear system filtering on the original brain magnetic signal to remove the stationary noise in the original brain magnetic signal and obtain a linear filtered signal; and uses a filtering method applicable to a nonlinear system to perform nonlinear system filtering on the original brain magnetic signal to reduce the impact of nonlinear interference in the original brain magnetic signal and obtain a nonlinear filtered signal. It can be understood that a variety of different filtering methods can eliminate noise from the original brain magnetic signal from different angles. After fusing multiple filtered signals, the advantages of different filtering methods can be combined to improve the filtering effect, effectively improve the signal quality of the brain magnetic signal, and then provide an accurate signal basis for the classification of brain magnetic signals.

[0055] Furthermore, the linear filter signal and the nonlinear filter signal may be fused by weighted calculation. In this case, the fused filter signal may be expressed according to the following formula: in, is the fusion filter signal; is a linear filtered signal; is the nonlinear filtering signal; The preset weights.

[0056] Reference Figure 4 As shown, as an embodiment of the present application, the original brain magnetic signal is subjected to linear system filtering to obtain a linear filtered signal, including: S112. Update the estimated weight of the sample reference value corresponding to the original magnetoencephalographic signal according to the fused filter signal at the previous moment.

[0057] S114. Obtain the motion artifact noise matrix of the original magnetoencephalogram signal using the sample reference value and the estimated weights of the updated sample reference value.

[0058] S116. Perform linear system filtering on the original brain magnetic signal according to the motion artifact noise matrix to obtain a linear filtered signal.

[0059] Specifically, the linear system filtering method can be H∞ filtering. In the process of performing linear system filtering on the original EEG signal, the original EEG signal is analyzed channel by channel using the H∞ filter, and state update and noise filtering are performed according to the corresponding H∞ adaptive rule.

[0060] Furthermore, there is a corresponding H∞ adaptive rule with time-varying weight assumption for H∞ filter design, which is expressed according to the following formula: in, is the estimated weight vector of sample reference values; To measure the sample point A reference value vector of ; is the original brain magnetic signal; is the noise matrix of motion artifacts; is the EEG signal after H∞ filtering to remove motion artifacts; is the noise covariance matrix, in, is the energy from the disturbance to the output estimation error - the bound on the energy gain; for The prior information of the noise covariance matrix at the moment is updated according to the following formula: in, is the prior information about how fast the weight changes over time, is the identity matrix; The closer it is to 1, the closer the behavior of the H∞ filter is to the optimal filter. In this case, it can ensure that the system maintains stable performance under various degrees of interference.

[0061] It should be noted that in the H∞ filtering process, the estimated weight vector of the sample reference value is updated according to the fused filter signal at the previous moment of the current moment. By utilizing the fused filter signal, the advantages of the nonlinear system filtering method are combined, the effect of H∞ filtering is improved, and the signal quality is improved. The present application performs linear system filtering on the original brain magnetic signal, thereby filtering the environmental noise, instrument noise, and possible stationary noise that may exist in the original brain magnetic signal, thereby reducing the impact of noise interference on the original brain magnetic signal and subsequent research and analysis.

[0062] Reference Figure 5 As shown, as an embodiment of the present application, the original brain magnetic signal includes a radial signal and a tangential signal; the original brain magnetic signal is subjected to nonlinear system filtering to obtain a nonlinear filtering signal, including: S122. According to the fusion filter signal of the radial signal at the first moment and the external stimulation input signal at the second moment, obtain the predicted state of the radial signal at the current moment as the radial predicted state; wherein the first moment and the second moment are both before the current moment.

[0063] S124. Obtain a predicted state of the tangential signal at the current moment according to the fused filter signal of the tangential signal at the first moment and the external stimulation input signal at the second moment, as the tangential predicted state.

[0064] S126. Component-couple the radial prediction state and the tangential prediction state to obtain the coupled prediction state of the original magnetoencephalogram signal; and perform state observation on the coupled prediction state to obtain a nonlinear filtering signal.

[0065] Specifically, the nonlinear system filtering method can be a particle filter. In the process of performing nonlinear system filtering on the original brain magnetic signal, the particle filter is used to analyze the original brain magnetic signal channel by channel to estimate the state of the original brain magnetic signal, and the original brain magnetic signal is gradually updated according to the state transfer model and the observation model.

[0066] Furthermore, the state transfer model and observation model of the particle filter are expressed according to the following formula: in, for The estimated value of the state at any measurement point at any time; for The radial magnetic field component of any measuring point at the moment The state estimation result after multi-strategy filtering and fusion of filtering results is: for The tangential magnetic field component of any measuring point at the moment The state estimation result after multi-strategy filtering and fusion of filtering results; for The radial magnetic field component of any measuring point at the moment The external stimulus input for motor imagery, for The tangential magnetic field component of any measuring point at the moment The external stimulus input received for motor imagery, if there is no external stimulus input, neither of them will be represented; is the coupling coefficient; for The process noise in the radial direction of any measurement point at the moment is: for The process noise in the tangent direction of any measurement point at that moment; for The measured value of any measuring point at that moment; yes The observation noise at any measurement point at that moment; and is the weight coefficient.

[0067] It should be noted that due to the actual brain reaction mechanism, when users perform motor imagery tasks, the characteristic signals of brain activity signals generally appear within 600ms or 1000ms after the external stimulus input, so there is delayed feedback in motor imagery tasks. In the state transition model of particle filtering, by designing the first moment and the second moment, nonlinear delay is introduced into the state transition model to describe the brain reaction mechanism in actual situations. It can be understood that the first moment, that is, The moment is related to the sampling rate of the original EEG signal; the second moment, i.e. The timing is related to the time when the external stimulus input occurs.

[0068] Furthermore, the cerebral cortex has magnetic field signals in different directions. The magnetic field components in the radial direction and the tangential direction of the cerebral cortex can be measured by atomic magnetometer. There is signal coupling between these two magnetic field components, which affects the signal intensity, clarity and spatial distribution characteristics of the two magnetic field components. In the state transition model of the particle filter, by designing the coupling coefficient , thereby characterizing the signal coupling behavior between the radial magnetic field component signal and the tangential magnetic field component signal, making the state transition model more consistent with the state transition situation in actual situations.

[0069] It should also be noted that, similar to H∞ filtering, in the particle filtering process, both the state transfer model and the observation model are updated according to the fused filter signal of the previous moment. By utilizing the fused filter signal, the advantages of H∞ filtering and particle filtering are combined, the state prediction effect of particle filtering is improved, and thus the signal quality is improved.

[0070] The present application processes the nonlinear components in the original EEG signals by performing nonlinear system filtering on the original EEG signals, thereby reducing the impact of nonlinear interference in the original EEG signals, effectively improving the filtering effect, and improving the robustness of the EEG signals.

[0071] Specifically explain the process of particle filtering. First, initialize the particles and the corresponding weight values, and generate random samples. Particles , the initial weight value of each particle for , where the number of particles is one percent of the total number of samples. The weight value of each particle is updated, and the update formula is expressed according to the following formula: in, is the particle at the previous moment The weight value of For a given particle state The observed probability density of .

[0072] Secondly, the updated particle weight values ​​are normalized. The normalized particle weight values ​​are expressed according to the following formula: According to the normalized particle weight value pair as the current moment The state estimation result at the moment is updated, and its update formula is expressed according to the following formula: .

[0073] Reference Figure 6 As shown, as an embodiment of the present application, a linear filtering signal and a nonlinear filtering signal are fused to obtain a fused filtering signal, including: S132. According to the signal-to-noise ratios of the linear filtering signal and the nonlinear filtering signal, the preset weight of the linear filtering signal is adjusted to obtain a fusion weight of the linear filtering signal.

[0074] S134. Perform weighted calculation on the linear filter signal and the nonlinear filter signal according to the fusion weight to obtain a fusion filter signal.

[0075] Specifically, the adjustment formula of the preset weight is expressed according to the following formula: in, represents the signal-to-noise ratio of the linearly filtered signal, Indicates the signal-to-noise ratio of the nonlinear filtering signal. The higher the signal-to-noise ratio, the higher the signal quality. By increasing the weight corresponding to the signal with a higher signal-to-noise ratio, the fusion effect is improved to adapt to the noise changes in the actual scene, thereby improving the signal quality of the fusion filtering signal.

[0076] Reference Figure 7 As shown, as an embodiment of the present application, the original magnetic brain signal is reconstructed according to the signal modality characteristics that meet the correlation condition to obtain the reconstructed magnetic brain signal, including: S232. Perform information entropy detection on the signal modal features to obtain the correlation of the signal modal features.

[0077] S234. Filter the signal modal features according to the correlation condition to obtain the signal reconstruction features that meet the correlation condition.

[0078] S236. Reconstruct the signal according to the signal reconstruction feature to obtain a reconstructed brain magnetic signal.

[0079] Specifically, information entropy is calculated between signal modal features obtained from the same magnetic field component, and the correlation between any two signal modal features is determined based on the obtained result. It is understandable that in other embodiments, correlation calculation can also be performed in other ways to select signal modal features that meet the conditions.

[0080] Furthermore, the form of reconstructed EEG signals is expressed according to the following formula: in, To reconstruct brain magnetic signals; is the number of signal modal features that satisfy the correlation condition, exemplarily, It can be 3.

[0081] by For example, refer to Figure 8a to Figure 8b As shown, Figure 8a Shown Location The signal waveform of the component after different steps, Figure 8b Shown Location The signal waveform of the component after different steps. It can be seen that Quantity and The original EEG signal of the component is greatly affected by noise, and the characteristic components cannot be distinguished. After multi-strategy filtering processing, it can be seen that even when the amplitude of the power frequency interference is significantly higher than the amplitude of the original EEG signal, the motion artifacts in the original EEG signal can still be effectively removed, and a good noise removal effect can be achieved. After signal reconstruction processing, the reconstructed EEG signal only contains characteristic components related to the motor imagery task, and the signal quality is high, providing a high-quality signal basis for subsequent EEG signal classification. In this embodiment, the accuracy of EEG signal classification based on the reconstructed EEG signal is improved by about five percentage points compared with the accuracy of a single algorithm.

[0082] As an embodiment of the present application, the method further includes: S400. Classify the reconstructed magnetoencephalogram signal to obtain a magnetoencephalogram signal classification result.

[0083] Specifically, signal classification is performed based on the reconstructed EEG signals to distinguish signal components belonging to different intentions and actions in the EEG signals, and characteristic component signals of the reconstructed EEG signals are obtained as EEG signal classification results. It can be understood that the characteristic component signals of the reconstructed EEG signals represent different action intentions of the user, and the usability of the EEG signals is enhanced by classifying the EEG signals, which improves the control effect of the brain-computer interface technology on external devices, and can also provide accurate EEG signal characteristics for subsequent brain-computer interface technology research.

[0084] As an embodiment of the present application, the reconstructed magnetic brain signal is classified to obtain a magnetic brain signal classification result, including: S410. Input the reconstructed EEG magnetic signals and the corresponding label set into the deep learning neural network model to classify the reconstructed EEG magnetic signals and output the EEG magnetic signal classification results; wherein the deep learning neural network model is trained by the reconstructed EEG magnetic signals before the current moment.

[0085] Specifically, the brain magnetic signal acquisition process can be divided into three stages, in which the first and second stages respectively train the deep learning neural network model according to the user's left and right hand movement imagination data, and the user's brain magnetic signals are not classified at this time. During the training process of the first and second stages, the user receives an instruction signal and performs corresponding movement imagination according to the instruction signal. The user's brain magnetic signals at this time are collected by an atomic magnetometer, and the brain magnetic signals and the corresponding label sets are used as training data to train the deep learning neural network model. In the third stage, the reconstructed brain magnetic signals are classified in real time using the trained deep learning neural network model to obtain the characteristic component signals in the user's brain magnetic signals, which are output as the brain magnetic signal classification results.

[0086] Accordingly, please refer to Fig. 9 The present application provides a device for preprocessing electroencephalographic signals, the device comprising: The fusion filtering module 910 is used to perform multi-strategy filtering on the original EEG signal and fuse the filtering results to obtain a fused filtered signal; wherein the multi-strategy filtering includes at least one filtering method applicable to a linear system and at least one filtering method applicable to a nonlinear system.

[0087] The feature extraction module 920 is used to perform frequency decomposition and feature extraction based on the fused filtered signal to obtain signal modal features.

[0088] The signal reconstruction module 930 is used to reconstruct the original magnetoencephalogram signal according to the signal modal characteristics that meet the correlation condition to obtain a reconstructed magnetoencephalogram signal.

[0089] In some optional implementations, the fusion filter module 910 includes: The linear filtering unit is used to perform linear system filtering on the original magnetoencephalographic signal to obtain a linear filtered signal.

[0090] The nonlinear filtering unit is used to perform nonlinear system filtering on the original brain magnetic signal to obtain a nonlinear filtered signal.

[0091] The signal fusion unit is used to fuse the linear filter signal and the nonlinear filter signal to obtain a fused filter signal.

[0092] In some optional implementations, the linear filtering unit includes: The weight updating subunit is used to update the estimated weight of the sample reference value corresponding to the original magnetoencephalographic signal according to the fusion filtering signal at the previous moment.

[0093] The noise determination subunit is used to obtain the motion artifact noise matrix of the original brain magnetic signal by using the sample reference value and the estimated weight of the updated sample reference value.

[0094] The noise filtering subunit is used to perform linear system filtering on the original brain magnetic signal according to the motion artifact noise matrix to obtain a linear filtered signal.

[0095] In some optional embodiments, the original brain magnetic signal includes a radial signal and a tangential signal; and the nonlinear filtering unit includes: The radial prediction subunit is used to obtain the predicted state of the radial signal at the current moment as the radial prediction state based on the fused filter signal of the radial signal at the first moment and the external stimulation input signal at the second moment; wherein the first moment and the second moment are both before the current moment.

[0096] The tangential prediction subunit is used to obtain the predicted state of the tangential signal at the current moment as the tangential prediction state according to the fusion filter signal of the tangential signal at the first moment and the external stimulation input signal at the second moment.

[0097] The component coupling subunit is used to component-couple the radial prediction state and the tangential prediction state to obtain the coupled prediction state of the original magnetoencephalogram signal; and to perform state observation on the coupled prediction state to obtain a nonlinear filtering signal.

[0098] In some optional embodiments, the signal fusion unit includes: The weight adjustment subunit is used to adjust the preset weight of the linear filtering signal according to the signal-to-noise ratio of the linear filtering signal and the nonlinear filtering signal, so as to obtain the fusion weight of the linear filtering signal.

[0099] The weighted calculation subunit is used to perform weighted calculation on the linear filter signal and the nonlinear filter signal according to the fusion weight to obtain a fusion filter signal.

[0100] In some optional implementations, the signal reconstruction unit includes: The correlation calculation subunit is used to perform information entropy detection on the signal modal features to obtain the correlation of the signal modal features.

[0101] The correlation screening subunit is used to screen the signal modal features according to the correlation conditions to obtain signal reconstruction features that meet the correlation conditions.

[0102] The signal reconstruction subunit is used to reconstruct the signal according to the signal reconstruction feature to obtain a reconstructed brain magnetic signal.

[0103] In some optional embodiments, the device further comprises: The signal classification module is used to classify the reconstructed magnetoencephalogram signal and obtain the magnetoencephalogram signal classification result.

[0104] In some optional embodiments, the signal classification module includes: The characteristic component signal classification unit is used to input the reconstructed EEG magnetic signal and the corresponding label set into the deep learning neural network model to classify the reconstructed EEG magnetic signal and output the EEG magnetic signal classification result; wherein the deep learning neural network model is trained by the reconstructed EEG magnetic signal before the current moment.

[0105] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0106] The magnetoencephalographic signal preprocessing device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0107] See also Fig.10 , Fig.10It is a structural diagram of a computer device provided by an embodiment of the present application. As shown in the figure, the computer device includes: one or more processors 10, a memory 20, and an interface for connecting various components, including a high-speed interface and a low-speed interface. Each component uses different buses to communicate with each other, and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig.10 A processor 10 is taken as an example.

[0108] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0109] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0110] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0111] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0112] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0113] The embodiment of the present application also provides a computer-readable storage medium. The above method according to the embodiment of the present application can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0114] The embodiment of the present application provides a computer program product, which includes computer instructions, which are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method of any embodiment of the present application.

[0115] Although the embodiments of the present application are described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present application, and such modifications and variations are all within the scope defined by the appended claims.

[0116] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0117] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0118] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0119] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0120] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0121] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0122] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or 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, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0123] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0124] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

[0125] Although the embodiments of the present application have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present application, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for preprocessing brain magnetic signals, characterized in that: The method comprises: Performing multi-strategy filtering on the original magnetoencephalogram signal and fusing the filtering results to obtain a fused filtered signal; wherein the multi-strategy filtering includes at least one filtering method applicable to a linear system and at least one filtering method applicable to a nonlinear system; Performing frequency decomposition and feature extraction based on the fused filtered signal to obtain signal modal features; The original magnetoencephalogram signal is reconstructed according to the signal modal features that meet the correlation condition to obtain a reconstructed magnetoencephalogram signal.

2. The method according to claim 1, characterized in that The original EEG signal is subjected to multi-strategy filtering, and the filtering results are fused to obtain a fused filtering signal, including Performing linear system filtering on the original brain magnetic signal to obtain a linear filtered signal; Performing nonlinear system filtering on the original brain magnetic signal to obtain a nonlinear filtered signal; The linear filter signal and the nonlinear filter signal are fused to obtain the fused filter signal.

3. The method according to claim 2, characterized in that The performing linear system filtering on the original brain magnetic signal to obtain a linear filtered signal includes: updating the estimated weight of the sample reference value corresponding to the original magnetoencephalogram signal according to the fusion filter signal at the previous moment; Obtaining a motion artifact noise matrix of the original magnetoencephalogram signal using the sample reference value and the estimated weight of the updated sample reference value; The original magnetoencephalogram signal is subjected to linear system filtering according to the motion artifact noise matrix to obtain the linear filtering signal.

4. The method according to claim 2, characterized in that: The original brain magnetic signal includes a radial signal and a tangential signal; the nonlinear system filtering of the original brain magnetic signal to obtain a nonlinear filtered signal includes: According to the fusion filtering signal of the radial signal at the first moment and the external stimulation input signal at the second moment, a predicted state of the radial signal at the current moment is obtained as a radial predicted state; wherein the first moment and the second moment are both before the current moment; Obtaining a predicted state of the tangential signal at a current moment according to a fused filter signal of the tangential signal at a first moment and an external stimulus input signal at a second moment as a tangential predicted state; The radial prediction state and the tangential prediction state are component-coupled to obtain the coupled prediction state of the original magnetoencephalogram signal; and the coupled prediction state is state-observed to obtain the nonlinear filtering signal.

5. The method according to claim 2, characterized in that: The fusing the linear filter signal and the nonlinear filter signal to obtain the fused filter signal includes: According to the signal-to-noise ratio of each of the linear filtering signal and the nonlinear filtering signal, a preset weight of the linear filtering signal is adjusted to obtain a fusion weight of the linear filtering signal; The linear filter signal and the nonlinear filter signal are weightedly calculated according to the fusion weight to obtain the fusion filter signal.

6. The method according to claim 1, characterized in that The reconstructing the original magnetoencephalogram signal according to the signal modality characteristics satisfying the correlation condition to obtain the reconstructed magnetoencephalogram signal comprises: Performing information entropy detection on the signal modal features to obtain the correlation of the signal modal features; Screening the signal modal features according to the correlation condition to obtain signal reconstruction features that meet the correlation condition; Signal reconstruction is performed according to the signal reconstruction feature to obtain the reconstructed electroencephalographic signal.

7. The method according to any one of claims 1 to 6, characterized in that The method further comprises: The reconstructed magnetic brain signal is classified to obtain a magnetic brain signal classification result.

8. The method according to claim 7, characterized in that The step of classifying the reconstructed magnetic brain signals to obtain magnetic brain signal classification results includes: The reconstructed magnetoencephalogram signal and the corresponding label set are input into a deep learning neural network model to classify the reconstructed magnetoencephalogram signal and output the classification result of the magnetoencephalogram signal; wherein the deep learning neural network model is trained by the reconstructed magnetoencephalogram signal before the current moment.

9. A device for preprocessing brain magnetic signals, characterized in that: The device comprises: A fusion filtering module, used for performing multi-strategy filtering on the original magnetoencephalogram signal and fusing the filtering results to obtain a fused filtered signal; wherein the multi-strategy filtering includes at least one filtering method applicable to a linear system and at least one filtering method applicable to a nonlinear system; A feature extraction module, used for performing frequency decomposition and feature extraction based on the fused filtered signal to obtain signal modal features; The signal reconstruction module is used to reconstruct the original magnetoencephalogram signal according to the signal modal characteristics that meet the correlation condition to obtain a reconstructed magnetoencephalogram signal.

10. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method according to any one of claims 1 to 8 by executing the computer instructions.

Images