Method, system, device and medium for collecting and denoising electroencephalogram signals

Abstract

The embodiments of this specification provide a method, system, device and medium for denoising collected EEG signals. The method includes obtaining and determining a signal transfer relationship of a measured object, wherein the signal transfer relationship represents a signal change relationship when an initial noise signal propagates between various positions of the measured object; determining a target noise signal based on the initial noise signal and the signal transfer relationship; and removing the target noise signal from the collected EEG signal from the measured object to obtain a noise-free target EEG signal.

Description

[0001] Divisional application

[0002] This application is a divisional application of Chinese patent application 202210659471.8, entitled “A method and system for generating simulated EEG signals”, filed on June 13, 2022. Technical Field

[0003] This specification relates to the field of electroencephalogram (EEG) signals, and in particular to a method, system, device, and medium for collecting and denoising EEG signals. Background Art

[0004] EEG signals are generated by amplifying and recording the weak bioelectricity of the human brain. They can be acquired using EEG acquisition equipment. EEG signals are widely used in research in psychology, neuroscience, psychiatry, and brain-computer interfaces. EEG signals primarily occur in the low- and ultra-low-frequency range, with frequencies primarily between 0.5 and 100 Hz and amplitudes ranging from 5 to 300 μV. Because EEG signals are so weak, they are easily interfered with and overwhelmed by noise. For example, this can be caused by the subject's physiological activity during acquisition, as well as by equipment noise. Therefore, EEG signals collected by EEG acquisition equipment require denoising to obtain noise-free signals.

[0005] In order to obtain noise-free EEG signals, it can be determined that the EEG signals of the subject are not subjected to denoising when the EEG signals are collected, especially to remove the noise brought by the EEG collection equipment itself during the collection process.

[0006] Therefore, it is necessary to propose a method for denoising collected EEG signals. The method can determine the signal change relationship when the noise signal propagates between various positions of the object under test when collecting EEG signals, and further determine the noise signal after propagation in the collected EEG signals based on the noise signal of the object under test that has not been propagated, thereby realizing denoising of the collected EEG signals. Summary of the Invention

[0007] One or more embodiments of the present specification provide a method for denoising collected EEG signals, characterized in that it includes: determining a signal transfer relationship of a measured object, wherein the signal transfer relationship represents a signal change relationship when an initial noise signal propagates between various positions of the measured object; determining a target noise signal based on the initial noise signal and the signal transfer relationship; removing the target noise signal based on the target noise signal and a collected EEG signal collected from the measured object to obtain a noise-free target EEG signal.

[0008] One or more embodiments of the present specification provide a system for denoising collected EEG signals, the system comprising: a first determination module for determining a signal transfer relationship of a measured object, wherein the signal transfer relationship represents a signal change relationship when an initial noise signal propagates between various positions of the measured object; a second determination module for determining a target noise signal based on the initial noise signal and the signal transfer relationship; and a signal processing module for obtaining a noise-free target EEG signal based on the target noise signal and the collected EEG signal from the measured object.

[0009] One or more embodiments of this specification provide a device for denoising collected EEG signals, comprising a processor, wherein the processor is configured to execute the method for denoising collected EEG signals as described in any one of claims 1 to 4.

[0010] One or more embodiments of this specification provide a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes them to implement the method for denoising collected EEG signals as described in any one of the above embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] This specification will be further described in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, like numbers represent like structures, wherein:

[0012] Figure 1 This is a schematic diagram of an application scenario of a system for collecting and denoising EEG signals according to some embodiments of this specification;

[0013] Figure 2 is an exemplary module diagram of a system for collecting and denoising EEG signals according to some embodiments of this specification;

[0014] Figure 3 is an exemplary flow chart of a method for denoising collected EEG signals according to some embodiments of this specification;

[0015] Figure 4 is an exemplary flow chart of obtaining signal transfer relationships according to some embodiments of this specification;

[0016] Figure 5 is a schematic diagram of a signal transfer model according to some embodiments of this specification;

[0017] Figure 6 is a schematic diagram of another signal transfer model according to some embodiments of this specification;

[0018] Figure 7is an exemplary flow chart of determining a target noise signal according to some embodiments of this specification. DETAILED DESCRIPTION

[0019] To more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly describes the drawings required for describing the embodiments. Obviously, the drawings described below are merely examples or embodiments of this specification. Those skilled in the art can apply this specification to other similar scenarios based on these drawings without inventive effort. Unless otherwise apparent from the context or otherwise noted, the same reference numerals in the figures represent the same structure or operation.

[0020] It should be understood that the terms "system," "device," "unit," and / or "module" used herein are a method for distinguishing different components, elements, parts, portions, or assemblies at different levels. However, if other terms can achieve the same purpose, the terms may be replaced by other expressions.

[0021] As used in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not refer to the singular but also include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0022] Flowcharts are used throughout this specification to illustrate the operations performed by systems according to embodiments of this specification. It should be understood that preceding or following operations do not necessarily need to be performed in exact order. Instead, the steps may be processed in reverse order or simultaneously. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0023] Figure 1 This is a schematic diagram of an application scenario of an EEG signal denoising system according to some embodiments of this specification.

[0024] like Figure 1 As shown, an application scenario 100 of the EEG signal denoising system may include a subject 110, an EEG acquisition device 120, a network 130, an initial noise signal 140, a server 150, a storage device 160, and a terminal device 170. The EEG signal denoising system may collect EEG signals and perform denoising processing by implementing the methods and / or processes disclosed in this specification.

[0025] The subject 110 may include a target subject who is ready to receive a test EEG signal. In some embodiments, the EEG acquisition device 120 may be placed on the subject 110 .

[0026] The EEG acquisition device 120 can be used to collect relevant signals from the subject 110. In some embodiments, the EEG acquisition device 120 can collect at least one of an EEG signal, an initial test signal, and / or an acquisition test signal from the subject 110, and transmit the collected signal to the server 150, the storage device 160, and / or the terminal device 170 via the network 130.

[0027] The network 130 may be used for transmission of information and / or data. In some embodiments, one or more components in the application scenario 100 (e.g., the server 150 and / or the storage device 160, etc.) may send information and / or data to another component in the application scenario 100 via the network 130.

[0028] Initial noise signal 140 may be a signal obtained from the collected EEG signal before the noise signal has been propagated. Initial noise signal 140 may be acquired from subject 110 using a signal acquisition device. In some embodiments, initial noise signal 140 may be transmitted via network 130 to server 150, storage device 160, and terminal device 170.

[0029] The server 150 can be used to manage resources and process data and / or information from at least one component of the application scenario 100 or an external data source (e.g., a cloud data center). The server 150 can communicate with the network 130, the storage device 160, and the terminal device 150 to implement various functions of the EEG signal denoising system. In some embodiments, the server 150 can receive the initial noise signal 140 and / or the data output by the EEG acquisition device 120 via the network 130 and process it. In some embodiments, the server 150 can output relevant data to the storage device 160 and the terminal device 170 via the network 130. In some embodiments, the server 150 can be a single server or a server group. In some embodiments, the server 150 can be locally connected to the network 130 or remotely connected to the network 130. In some embodiments, the server 150 can be implemented on a cloud platform.

[0030] The storage device 160 can store data and / or instructions. The data may include data related to the processor 110, the terminal device 170, the initial noise signal 140, etc. For example, the storage device 160 can store the initial noise signal 140. In some embodiments, the storage device 160 can store data and / or instructions used by the server 150 to execute or use to complete the exemplary methods described in this specification. In some embodiments, the storage device 160 can be connected to the network 130 to communicate with one or more components of the application scenario 100 (e.g., the EEG acquisition device 120, the server 150, and / or the terminal device 170). In some embodiments, the storage device 160 can be part of the server 150. In some embodiments, the storage device 160 may include a large-capacity memory, a removable memory, a volatile read-write memory, a read-only memory (ROM), etc., or any combination thereof. In some embodiments, the storage device 160 can be implemented on a cloud platform.

[0031] Terminal device 170 may refer to one or more terminals or software used by a user. A user may refer to a researcher. A user may view relevant data of various components in application scenario 100 through terminal device 170. In some embodiments, terminal device 170 may include a mobile device 170-1, a tablet computer 170-2, a laptop computer 170-3, or the like, or any combination thereof. In some embodiments, terminal device 170 may be fixed and / or mobile. For example, terminal device 170 may be directly installed on server 150, becoming part of server 150. For another example, terminal device 170 may be a mobile device, allowing a user to carry terminal device 170 to a location that is remote from server 150 and initial noise signal 140. Terminal device 170 may connect to and / or communicate with server 150 and storage device 160 via network 130.

[0032] It should be noted that application scenario 100 is provided for illustrative purposes only and is not intended to limit the scope of this specification. A person skilled in the art may make various modifications or variations based on the description of this specification. For example, application scenario 100 may further include a database. However, such variations and modifications do not deviate from the scope of this specification.

[0033] Figure 2 This is an exemplary module diagram of a system for collecting and denoising EEG signals according to some embodiments of this specification.

[0034] In some embodiments, the EEG signal denoising system 200 may include a first determination module 210 , a second determination module 220 and a signal processing module 230 .

[0035] The first determination module 210 is used to determine the signal transfer relationship of the object under test. The signal transfer relationship represents the signal change relationship when the initial noise signal propagates between various positions of the object under test. For more information about the initial noise signal, the object under test, and the signal transfer relationship, please refer to Figure 3 In some embodiments, the first determination module 210 may also be configured to input an initial test signal from each position of the subject under test, wherein the magnitude of the initial test signal is greater than the magnitude of the EEG signal of the subject under test; collect signals from other positions of the subject under test to obtain a collected test signal; and determine a signal transfer relationship based on the initial test signal and the collected test signal. For more information on the initial test signal and the collected test signal, please refer to Figure 4 and its related descriptions.

[0036] The second determination module 220 is used to determine the target noise signal based on the initial noise signal and the signal transfer relationship. For more information about the target noise signal, please refer to Figure 3 And its related description. In some embodiments, the signal transfer relationship includes multiple simple signal transfer relationships, and the simple signal transfer relationship characterizes the signal change relationship when the first simple noise signal propagates between various positions of the object under test, wherein the first simple noise signal is obtained by transforming the initial noise signal, and the second determination module 220 can also be used to transform the initial noise signal into multiple first simple noise signals of different frequencies and intensities; for each of the first simple noise signals, the first simple noise signal is processed based on the simple signal transfer relationship corresponding to the first simple noise signal to determine the second simple noise signal corresponding to the first simple noise signal; and each second simple noise signal is synthesized to obtain the target noise signal. For more information about simple signal transfer relationships, first simple noise signals, and second simple noise signals, please refer to Figure 6 and its related descriptions.

[0037] The signal processing module 230 is used to remove the target noise signal from the collected EEG signal of the subject to obtain a noise-free target EEG signal. Figure 3 and its related descriptions.

[0038] It should be noted that the above description of the EEG signal denoising system 200 and its modules is for convenience only and does not limit this specification to the scope of the embodiments. It is understandable that those skilled in the art, after understanding the principles of the system, may arbitrarily combine the modules or form a subsystem connected with other modules without deviating from the principles. In some embodiments, Figure 2The first determination module 210, second determination module 220, and signal processing module 230 disclosed herein may be different modules within a system, or a single module may implement the functions of two or more of the aforementioned modules. For example, the modules may share a storage module, or each module may have its own storage module. Such variations are within the scope of protection of this specification.

[0039] Figure 3 FIG3 is an exemplary flow chart of a method for collecting and denoising EEG signals according to some embodiments of this specification. In some embodiments, process 300 may be executed by server 150. Figure 3 As shown, process 300 may include the following steps:

[0040] Step 310 , determining the signal transfer relationship of the object under test, wherein the signal transfer relationship represents the signal change relationship when the initial noise signal propagates between various positions of the object under test. In some embodiments, step 310 may be performed by the first determination module 210 .

[0041] The object to be measured can refer to the object for which the signal transfer relationship needs to be determined. The object to be measured can be a biological object, such as the head area of ​​a human body. The object to be measured can also be a non-biological object, such as a virtual head model.

[0042] The initial noise signal may refer to the signal before the noise signal in the collected EEG signal is propagated. The collected EEG signal may refer to the EEG signal of the subject collected through the electrode points.

[0043] It should be understood that when collecting the EEG signals of the subject, there are many factors (for example, the operation of the EEG acquisition equipment or the physiological activities of the subject) that affect the acquisition process, so that the collected EEG signals contain a variety of noise signals. Common noise signals in the collected EEG signals can include non-biological artifact signals and biological artifact signals. Among them, non-biological artifact signals can be noise signals caused by the external environment, equipment, etc., and can include but are not limited to noise signals caused by EEG acquisition equipment, noise signals caused by mains interference, etc. Non-biological artifact signals can be obtained by querying relevant information. For example, the noise signals caused by the model of EEG acquisition equipment can be determined through the Internet. Biological artifact signals can be noise signals caused by the physiological activities of the subject, and can include but are not limited to electrocardiogram signals caused by heartbeats, electromyogram signals caused by muscle activity, and electrooculogram signals caused by eye movements, etc. Biological artifact signals can be collected and acquired from the subject using relevant acquisition equipment. Due to the differences between biological individuals, the biological artifact signals corresponding to different subjects should be different.

[0044] The signal transfer relationship may refer to the change in the initial noise signal when it propagates between various locations of the object under test. In some embodiments, based on the signal transfer relationship, the change in the signal at one location of the object under test when it propagates to another location can be determined. For example, based on the signal transfer relationship, the change in the signal from location 0 of the object under test can be determined. i An incoming signal S i Propagate to the position O of the object being measured j After that, it changes to signal S j In some embodiments, due to differences in physiological conditions and brain structures of the subjects, different subjects may have different corresponding signal transfer relationships. For example, the signal transfer relationship between the elderly and children is different. For another example, the signal transfer relationship between a 26-year-old woman and a 26-year-old man is also different.

[0045] In some embodiments, the signal transfer relationship can be represented as an N*N signal transfer matrix, where N is the number of electrodes of the object under test, and each electrode can represent a different position of the object under test. The elements in the signal transfer matrix can be real numbers, which are adjustment parameters between corresponding electrode points, and can be used to determine the transfer relationship when the signal propagates between corresponding electrode points. In some embodiments, the elements in the signal transfer matrix can also be conduction functions, which can characterize the influence of the corresponding electrode point on the electrode point noise signal. Each element in the signal transfer matrix can be a conduction function between two corresponding positions of the object under test. It should be understood that the conduction function between a certain position in the signal transfer matrix and the position itself is 1, indicating that the signal itself will not change when it is not propagating. In some embodiments, more information about obtaining the signal transfer matrix can be found in Figure 4 and its related descriptions.

[0046] In some embodiments, the signal transfer relationship can also be represented as a signal transfer model. The initial noise signal can be processed based on the signal transfer model to obtain the target noise signal. For more information about the signal transfer model, please refer to Figure 5 and its related descriptions.

[0047] In some embodiments, when the object to be measured is a virtual head model, the attenuation relationship of the signal at various positions in the virtual head model can be determined based on the material and shape of the virtual head model, thereby determining the signal transfer relationship of the virtual head model.

[0048] In some embodiments, an initial test signal may be input from a first position of the subject, wherein the order of magnitude of the initial test signal is greater than the order of magnitude of the EEG signal of the subject; a signal may be collected from a second position of the subject to obtain a collected test signal; and a signal transfer relationship may be determined based on the initial test signal and the collected test signal. For more information on the above embodiments, please refer to Figure 4 and its related descriptions.

[0049] Step 320 : Determine the target noise signal based on the initial noise signal and the signal transfer relationship. In some embodiments, step 320 may be performed by the second determination module 220 .

[0050] The target noise signal may refer to a noise signal in the collected EEG signal. In some embodiments, the target noise signal may be a noise set, which is a noise signal collected from various positions of the subject under test.

[0051] In some embodiments, after determining the initial noise signal, the basic features corresponding to each initial noise signal can be determined. The basic features corresponding to the initial noise signal may include but are not limited to the amplitude and variance of the signal. For each of the multiple electrode points, when collecting EEG signals, the signal collected by the electrode point can be dynamically evaluated, and by comparing the basic features of each collected signal with the basic features corresponding to each initial noise signal, it can be determined whether there is an initial noise signal in the signal collected by the electrode point. When there is an initial noise signal in the signal collected by the electrode point, the initial noise signal can be determined as the target noise signal in the EEG signal collected by the electrode point. At the same time, the initial noise signal collected by other electrode points after propagation can be determined through the signal transfer relationship, and the initial noise signal after propagation can be determined as the target noise signal in the EEG signals collected by other electrode points.

[0052] In step 330 , the target noise signal is removed from the collected EEG signal of the subject to obtain a noise-free target EEG signal. In some embodiments, step 330 may be performed by the signal processing module 230 .

[0053] The target EEG signal may refer to a noise-free EEG signal of the subject being measured. The target EEG signal may be used to analyze and diagnose brain diseases of the subject being measured.

[0054] In some embodiments, when the target noise signal collected by a certain electrode point is an initial noise signal that has not been propagated at the electrode point, the initial noise signal can be directly removed from the collected EEG signal collected by the electrode point to obtain the target EEG signal of the electrode point.

[0055] In some embodiments, when the target noise signal collected by a certain electrode point is the initial noise signal after propagation, the target EEG signal of the electrode point can be determined based on the following formula:

[0056]

[0057] Among them, S clean is the target EEG signal of the electrode point at position j, S obs is the collected EEG signal of the electrode point, h is the signal transfer matrix, h ij It represents the conduction function of the signal from position i to position j in the signal transfer matrix, S i is the initial noise signal originating from the electrode point at position i, and k is the number of initial noise signals.

[0058] In some embodiments, the initial noise signal at a certain location can be processed based on the signal transfer model to obtain the target noise signal in the collected EEG signals at other locations. The target noise signal in the collected EEG signals at other locations is then removed to obtain the target EEG signal at the corresponding location.

[0059] In some embodiments of this specification, by determining the signal transfer relationship when the signal propagates between different positions, the target noise signal after propagation can be accurately and quickly determined based on the initial noise signal and the signal transfer relationship, so that the target noise signal in the collected EEG signal can be removed to obtain a noise-free target EEG signal. Compared with the existing method of denoising the collected EEG signal, the cost is greatly reduced. At the same time, by determining the corresponding signal transfer relationship based on each measured object, the determined signal transfer relationship can be made more accurate and in line with the actual situation of the measured object. On the basis of accurately and completely removing the target noise signal in the collected EEG signal, the target EEG signal will not be removed.

[0060] Figure 4 is an exemplary flow chart of obtaining a signal transfer relationship according to some embodiments of this specification. In some embodiments, process 400 may be executed by the first determination module 210. Figure 4 As shown, process 400 may include the following steps:

[0061] Step 410: For each position of the subject being measured, input an initial test signal from the position, wherein the order of magnitude of the initial test signal is greater than the order of magnitude of the EEG signal of the subject being measured.

[0062] The initial test signal may refer to a test signal input to the object under test, and the initial test signal may be used to measure the relationship between changes in the test signal when it propagates between different electrode points. In some embodiments, the order of magnitude of the initial test signal is greater than the order of magnitude of the human brain electrical signal. For example, the order of magnitude of the human brain electrical signal is μV, and the order of magnitude of the initial test signal may be 3 orders of magnitude greater than the human brain electrical signal, that is, the order of magnitude of the selected initial test signal may be mV. In some embodiments, the initial test signal may include multiple signals of different frequencies and different intensities. In some embodiments, the initial test signal may be determined by pre-setting. For example, the initial test signal may be a pre-set pulse signal, and the order of magnitude of the pulse signal is greater than the order of magnitude of the brain electrical signal of the object under test. Each position of the object under test may refer to each electrode point of the object under test, and the initial test signal may be input at an electrode point of the object under test.

[0063] It should be understood that because the initial test signal is larger than the EEG signal of the subject being tested, the collected signal is less affected by the EEG signal during signal acquisition, ensuring the accuracy of the test results. Furthermore, the initial test signal should not exceed a signal level that the human body can tolerate to ensure the safety of the subject being tested. For example, the amplitude of the initial test signal may range from 20mV to 200mV.

[0064] Step 420: Collect signals from other locations of the object to obtain collected test signals.

[0065] Collecting a test signal may refer to collecting a signal at another location after propagation of the initial test signal input at the aforementioned location. The other location may be another electrode point. Correspondingly, collecting a test signal may include collecting signals at the other electrode point based on the electrode point at that location. For example, electrode points A to D may be provided on the object under test, where the location at which the initial test signal is input is electrode point A, and the other locations may refer to motor points B, C, and D. Correspondingly, signals may be collected from motor points B, C, and D to obtain a collected test signal.

[0066] It should be understood that the initial test signal will be affected by other electrode points when propagating between electrode points, so the collected test signal may be different from the initial test signal.

[0067] Step 430: Determine a signal transfer relationship based on the initial test signal and the collected test signal.

[0068] In some embodiments, modeling can be performed or various data analysis algorithms, such as regression analysis and discriminant analysis, can be used to analyze and process the initial test signal and the collected test signal to determine the signal transfer relationship.

[0069] In some embodiments, when the elements in the signal transfer matrix can be real numbers, each element in the matrix can be determined based on the following formula:

[0070] S j (t) = βs i (t) (2)

[0071] Among them, S j (t) represents the test signal collected by electrode point j, β is the adjustment parameter for the signal propagating from electrode point i to electrode point j, s i (t) represents the initial test signal input from electrode point i.

[0072] In some embodiments, the signal transfer relationship may include multiple simple signal transfer relationships. The simple signal transfer relationship may refer to the change relationship of a simple signal with relatively regular signal changes when it propagates between multiple positions of the object under test. In some embodiments, the input initial test signal may be transformed into multiple initial simple test signals of different frequencies and intensities, and the acquisition test signals collected at other electrode points may be transformed into multiple acquisition simple test signals of different frequencies and intensities; for each initial simple test signal, based on the initial simple test signal and its corresponding acquisition simple test signal, the simple signal transfer relationship corresponding to the initial simple test signal is determined; and each simple signal transfer relationship is synthesized to obtain a signal transfer relationship containing multiple simple signal transfer relationships. The initial simple test signal may be a simple signal obtained by transforming the initial test signal, and the acquisition simple test signal may be a simple signal obtained by transforming the acquisition test signal, wherein the signal transformation method may be Fourier transform. For example, the initial simple test signal and the acquisition simple test signal may both be sinusoidal wave signals.

[0073] In some embodiments, the initial test signal input at position i can be Fourier transformed to obtain a set S of initial simple test signals of M different phases and frequencies, wherein S is an N*M matrix and N is the number of electrode points. When no signal is input at other positions of the object under test and the initial test signal is not propagated, each element in the i-th row of S represents an initial simple test signal, and other positions in S can be 0, indicating that other positions do not contain signals. After the initial test signal is propagated, the collected test signals collected at other positions are Fourier transformed to obtain a set S′ containing multiple collected simple test signals, wherein S′ is also an N*M matrix, and each element in S′ represents a collected simple test signal collected at the corresponding electrode point, which is the signal after the corresponding initial simple test signal has been propagated. Then, calculations can be performed based on S and S′ to determine the signal transfer matrix H so that the following formula holds:

[0074] S′=HS (3)

[0075] The elements of H are all conduction functions, which represent the changing relationship of the signal when it propagates at the corresponding position. For example, h ij It can be an element in H, representing the conduction function of the signal propagating from position i to position j. In some embodiments, the conduction function can be expressed in various forms. As shown in the following formula, the conduction function h ij It is equivalent to an FIR (Finite Impulse Response, finite length unit impulse response) filter:

[0076] y[n]=b0x[n]+b1x[n-1]+…+b L x[nL] (4)

[0077]

[0078] Where y[n] is the collected simple test signal, x[n] is the initial simple test signal, n is the signal length, L is the number of parameters, which can be determined by presetting. When L≤M, the conduction function h can be determined by solving a multivariate linear equation system. ij All parameters in .

[0079] In some embodiments, the signal transfer model can also be trained based on the initial test signal and the collected test signal to obtain a trained signal transfer model. For more information about the signal transfer model, please refer to Figure 5 and its related descriptions.

[0080] It should be understood that the brain structure of the subject is nearly fixed, so even if the strengths of two signals differ, the signal transition relationships between locations within the subject should be the same. Therefore, some embodiments of this specification can determine the signal transition relationships within the subject by analyzing and processing the initial test signal and the acquired test signal. This signal transition relationship can also characterize the signal transition relationships between the initial noise signal as it propagates between various locations within the subject.

[0081] In some embodiments of this specification, for each object under test, the signal transfer relationship of the object under test can be quickly determined through the initial test signal and the collected test signal, thereby ensuring the accuracy of the signal transfer relationship and avoiding errors caused by individual differences in the objects under test during denoising.

[0082] Figure 5 is a schematic diagram of a signal transfer model according to some embodiments of this specification.

[0083] In some embodiments, the initial noise signal can be processed based on the signal transfer model to obtain the target noise signal. Figure 5 As shown, the input of the signal transfer model 530 may include an initial noise signal 510 and a position 520 of the initial noise signal, and the output may include a target noise signal 540. In some embodiments, the signal transfer model may include but is not limited to a deep neural network model, a support vector machine model, and the like.

[0084] In some embodiments, the signal transfer model can be obtained by training a machine learning model using training samples. Figure 5 As shown, the signal transfer model can be obtained by training the initial signal transfer model 550 using the training samples 560 and the labels 570, wherein the initial signal transfer model can be a signal transfer model without setting parameters. The training samples 560 can include the initial test signal and the location of the initial test signal, and the labels 570 can include the collected test signal. The method for obtaining the training samples and labels can be found in Figure 4 Multiple sets of training samples 560 with labels 570 are input into the initial signal transfer model 550. A loss function is constructed based on the output of the initial signal transfer model 550 and the labels 570. The parameters of the initial signal transfer model 550 are iteratively updated based on the loss function until a preset condition is met. Training is completed, and a trained signal transfer model 530 is obtained. The preset conditions may include, but are not limited to, the loss function being less than a threshold, convergence, or the training cycle reaching a threshold.

[0085] In some embodiments, the signal transfer model may include an object information embedding layer and a signal determination layer. The object information embedding layer is used to extract features from the shape and hair information of the object being measured to obtain head features. The signal determination layer is used to process the head features, the initial noise signal, and the position information of the initial noise signal to obtain the target noise signal. The shape information may include, but is not limited to, the shape and size of the object being measured, and the hair information may include, but is not limited to, the length, hardness, curvature, and oiliness of the hair of the object being measured. The shape and hair information of the object being measured can be pre-set and determined.

[0086] like Figure 6As shown, the signal transfer model 530 may further include an object information embedding layer 530-1 and a signal determination layer 530-2, which are connected in sequence. The input of the object information embedding layer 530-1 may include shape information 610 and hair information 620, and the output is head features 531. The input of the signal determination layer may include head features 531, an initial noise signal 510, and the position 520 of the initial noise signal, and the output may be a target noise signal 540. The object information embedding layer may be a naive Bayes model, and the signal determination layer may be a deep neural network model.

[0087] In some embodiments, the object information embedding layer can be obtained through training: the training samples may include the historical shape information and historical hair condition information of the test object, and the labels may include the historical head features of the aforementioned test object, wherein the test object may refer to the object whose data is used to train the object information embedding layer. Similar to the object being tested, the test object may be a biological or non-biological object. The aforementioned training samples and labels can be obtained by artificially expressing the relevant information of the test object. Multiple groups of training samples can be input into the initial object information embedding layer, and a loss function can be constructed based on the output and label of the initial object information embedding layer. The parameters of the initial object information embedding layer are iteratively updated based on the loss function until the preset conditions are met to obtain a trained object information embedding layer. The preset conditions may include but are not limited to the loss function being less than a threshold, convergence, or the training cycle reaching a threshold.

[0088] In some embodiments, a signal determination layer can be acquired through training: training samples may include historical head features, an initial test signal of a tested object, and the location of the initial test signal; labels may include collected test signals. The acquisition method of the aforementioned training samples and their labels can be referred to above in this specification. Multiple sets of training samples can be input into the initial signal determination layer, and a loss function can be constructed based on the output and labels of the initial signal determination layer. The parameters of the initial signal determination layer are iteratively updated based on the loss function until a preset condition is met, thereby obtaining a trained signal determination layer. The preset conditions may include, but are not limited to, a loss function being less than a threshold, convergence, or a training cycle reaching a threshold.

[0089] It should be understood that there are differences in the objects corresponding to the training data used to train the object information embedding layer and the signal determination layer. The trained object information embedding layer can extract the head features of all objects (including test objects and tested objects). The object information embedding layer is a common model for all tested objects, so the object information embedding layer obtained through training based on the relevant data of the test object can extract the head features of different tested objects. The trained signal determination layer is used to determine the signal transfer relationship of a certain tested object. As mentioned above, when the tested objects are different, their corresponding signal transfer relationships are also different. Therefore, the signal determination layer obtained through training based on a certain tested object can only be used to determine the signal transfer relationship of the tested object. When it is necessary to determine the signal transfer relationship of another tested object, training can be performed based on the trained object information embedding layer, the initial test signal of the tested object, the position of the initial test signal, and the collected test signal.

[0090] Some embodiments of this specification can more quickly and conveniently determine the signal transfer relationship of the subject through a signal transfer model, thereby facilitating denoising of the collected EEG signals. In addition, some embodiments of this specification can also extract relevant features of the subject's head and determine the impact of differences in the subject's head-related information on the subject's signal transfer relationship, thereby making the denoising of the collected EEG signals more accurate.

[0091] Figure 7 FIG. 7 is an exemplary flow chart of determining a target noise signal according to some embodiments of the present disclosure. In some embodiments, process 700 may be executed by the second determining module 220 .

[0092] In some embodiments, the signal transfer relationship may include multiple simple signal transfer relationships. The simple signal transfer relationship represents the signal change relationship when the first simple noise signal propagates between various positions of the object under test. For more information about the simple signal transfer relationship, please refer to Figure 4 The first simple noise signal is obtained by transforming the initial noise signal.

[0093] When the signal transfer relationship includes multiple simple signal transfer relationships, the process 700 can be executed to determine the target noise signal. Figure 7 As shown, process 700 may include the following steps:

[0094] Step 710: transform the initial noise signal into a plurality of first simple noise signals of different frequencies and intensities.

[0095] The first simple noise signal can be a simple signal obtained by transforming the initial noise signal. For example, the first simple noise signal can include any one or a combination of a sine wave signal, a cosine wave signal, and a square wave signal. The initial noise signal can be processed using a Fourier transform, a Laplace transform, a discrete cosine transform, or the like to obtain a plurality of first simple noise signals of different frequencies and intensities.

[0096] Step 720 : For each first simple noise signal, process the first simple noise signal based on the simple signal transfer relationship corresponding to the first simple noise signal to determine a second simple noise signal corresponding to the first simple noise signal.

[0097] The second simple noise signal may refer to a noise signal generated when the first simple noise signal propagates to other locations of the object under test.

[0098] In some embodiments, for each first simple noise signal, a second simple noise signal corresponding to the first simple noise signal at another location can be determined based on the first simple noise signal, the input location of the first simple noise signal, and a simple signal transfer relationship. Based on the simple signal transfer relationship and the input location of the first simple noise signal, a conduction function from the input location to each other location can be determined. Based on each conduction function and the first simple noise signal, a second simple noise signal at each other location can be determined.

[0099] Step 730: synthesize the second simple noise signals to obtain a target noise signal.

[0100] In some embodiments, for each position of the measured object, the second simple noise signals at that position can be synthesized to obtain a target noise signal at that position. The synthesis method can be an inverse transformation of the aforementioned transformation method for processing the initial noise signal into the first simple noise signal to obtain the target noise signal. For example, when the initial noise signal is processed into the first simple noise signal based on a Fourier transform, the second simple noise signals can be synthesized based on an inverse Fourier transform to obtain the target noise signal.

[0101] Some embodiments of this specification utilize simple noise signals and simple signal transfer relationships to conveniently acquire target noise signals, thereby simplifying operations and improving computational efficiency.

[0102] It should be noted that the above descriptions of the various processes are for illustration and purpose only and do not limit the scope of application of this specification. Those skilled in the art may make various modifications and alterations to the various processes under the guidance of this specification. However, such modifications and alterations are still within the scope of this specification.

[0103] While the basic concepts have been described above, it will be apparent to those skilled in the art that the detailed disclosure is merely illustrative and does not limit this specification. Although not explicitly stated herein, various modifications, improvements, and revisions to this specification may be made by those skilled in the art. Such modifications, improvements, and revisions are suggested in this specification and remain within the spirit and scope of the exemplary embodiments of this specification.

[0104] This specification also uses specific terms to describe the embodiments of this specification. For example, "one embodiment," "an embodiment," and / or "some embodiments" refer to a feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "one embodiment," "an embodiment," or "an alternative embodiment" two or more times in different locations in this specification do not necessarily refer to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of this specification may be appropriately combined.

[0105] In addition, unless expressly stated in the claims, the order of the processing elements and sequences, the use of alphanumeric characters, or the use of other names described in this specification are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some of the invention embodiments currently considered useful through various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the spirit and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0106] Similarly, it should be noted that, in order to simplify the presentation of this specification and thus facilitate understanding of one or more embodiments of the invention, the foregoing descriptions of the embodiments of this specification sometimes combine multiple features into a single embodiment, figure, or description thereof. However, this disclosure method does not imply that the subject matter of this specification requires more features than those recited in the claims. In fact, an embodiment may have fewer features than all of the features of a single disclosed embodiment.

[0107] In some embodiments, numbers are used to describe the quantity of components and attributes. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise stated, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the description and claims are approximate values, which may change according to the required characteristics of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the settings of such numerical values ​​are as accurate as possible within the feasible range.

[0108] Each patent, patent application, patent application publication, and other materials, such as articles, books, specifications, publications, and documents, cited in this specification is hereby incorporated by reference in its entirety. This includes application history documents that are inconsistent with or conflict with the content of this specification, as well as documents (currently or subsequently attached to this specification) that limit the broadest scope of the claims of this specification. It should be noted that if the descriptions, definitions, and / or terminology used in the accompanying materials are inconsistent or conflicting with the content of this specification, the descriptions, definitions, and / or terminology used in this specification will control.

[0109] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly described and illustrated in this specification.

Claims

1. A method for collecting and denoising EEG signals, characterized in that: include: Determining a signal transfer relationship of the subject, wherein the signal transfer relationship represents a signal change relationship when an initial noise signal propagates between various locations of the subject, and the initial noise signal is a signal of the collected EEG signal collected from the subject before the noise signal propagates; Determining a target noise signal based on the initial noise signal and the signal transfer relationship, where the target noise signal is the noise signal in the collected EEG signal; The target noise signal is removed from the collected EEG signal to obtain a noise-free target EEG signal.

2. The method according to claim 1, wherein Determining the signal transfer relationship includes: For each position of the measured object, input an initial test signal from the position, wherein the order of magnitude of the initial test signal is greater than the order of magnitude of a human brain electrical signal; Collect signals from other positions of the object to obtain collected test signals; The signal transfer relationship is determined based on the initial test signal and the acquired test signal.

3. The method according to claim 2, wherein The signal transfer relationship includes a signal transfer model, wherein the signal transfer model is obtained by training a machine learning model using training samples, wherein the training samples include multiple sets of initial test signals of different frequencies and intensities, locations where the initial test signals are input, and collected test signals corresponding to the initial test signals; The determining of the target noise signal based on the initial noise signal and the signal transfer relationship includes: The initial noise signal is processed based on the signal transfer model to obtain the target noise signal.

4. The method according to claim 1, wherein The signal transfer relationship includes a plurality of simple signal transfer relationships, each of which represents a signal change relationship when a first simple noise signal propagates between various positions of the object under test, wherein the first simple noise signal is obtained by transforming the initial noise signal. The determining of the target noise signal based on the initial noise signal and the signal transfer relationship includes: Converting the initial noise signal into a plurality of first simple noise signals of different frequencies and intensities; For each of the first simple noise signals, processing the first simple noise signal based on the simple signal transfer relationship corresponding to the first simple noise signal to determine a second simple noise signal corresponding to the first simple noise signal; The second simple noise signals are synthesized to obtain the target noise signal.

5. A system for collecting and denoising EEG signals, characterized in that: The system comprises: A first determining module is configured to determine a signal transfer relationship of the subject, wherein the signal transfer relationship represents a signal change relationship when an initial noise signal propagates between various locations of the subject, and the initial noise signal is a signal of the collected EEG signal collected from the subject before the noise signal propagates; a second determining module, configured to determine a target noise signal based on the initial noise signal and the signal transfer relationship, wherein the target noise signal is a noise signal in the collected EEG signal; The signal processing module is used to remove the target noise signal from the collected EEG signal to obtain a noise-free target EEG signal.

6. The system according to claim 5, wherein: The first determining module is further configured to: For each position of the subject, input an initial test signal from the position, wherein the order of magnitude of the initial test signal is greater than the order of magnitude of the EEG signal of the subject; Collect signals from other positions of the object to obtain collected test signals; The signal transfer relationship is determined based on the initial test signal and the acquired test signal.

7. The system according to claim 6, wherein: The signal transfer relationship includes a signal transfer model, wherein the signal transfer model is obtained by training a machine learning model using training samples, wherein the training samples include multiple sets of initial test signals of different frequencies and intensities, locations where the initial test signals are input, and collected test signals corresponding to the initial test signals; The second determining module is further configured to: The initial noise signal is processed based on the signal transfer model to obtain the target noise signal.

8. The system according to claim 5, wherein: The signal transfer relationship includes a plurality of simple signal transfer relationships, and the second determining module is further configured to: Converting the initial noise signal into a plurality of first simple noise signals of different frequencies and intensities; For each of the first simple noise signals, processing the first simple noise signal based on the simple signal transfer relationship corresponding to the first simple noise signal to determine a second simple noise signal corresponding to the first simple noise signal; The second simple noise signals are synthesized to obtain the target noise signal.

9. A device for collecting and denoising brain electrical signals, comprising a processor, characterized in that: The processor is used to execute the method for denoising collected EEG signals according to any one of claims 1 to 4.

10. A computer-readable storage medium storing computer instructions, wherein when a computer reads the computer instructions in the storage medium, the computer executes the method for denoising collected EEG signals according to any one of claims 1 to 4.

Images