Chronic pain electroencephalogram feature extraction method based on brain network

Abstract

The invention discloses a chronic pain electroencephalogram feature extraction method based on a brain network, and relates to the technical field of research and treatment of chronic pain, and the method comprises the following steps: 1, employing a 50Hz wave trap and independent component analysis to carry out the preprocessing of each channel electroencephalogram data collected from a chronic pain patient, removing power frequency interference, motion artifacts and eye movement noise in the electroencephalogram data; step 2, respectively extracting electroencephalogram data components of alpha, beta, theta and gamma frequency bands from the electroencephalogram data of each channel by using a zero-phase digital band-pass filter; according to the method, the cross-channel electroencephalogram network is constructed through an algorithm, cross-channel electroencephalogram feature extraction is achieved, cross-frequency electroencephalogram feature extraction is achieved by constructing the cross-frequency electroencephalogram network, and the chronic pain state is reflected more accurately by using the deep neural network and fusing multi-scale electroencephalogram information.

Description

Technical Field

[0001] The present invention relates to the technical field of research and treatment of chronic pain, and in particular to a method for extracting EEG features of chronic pain based on brain network. Background Art

[0002] Chronic pain refers to pain that lasts for more than 3 to 6 months. Unlike acute pain, chronic pain is an independent disease that lasts longer than the expected healing time. Chronic pain is one of the biggest causes of disability worldwide. Due to the lack of effective treatments for chronic pain, it may also make patients weak. Long-term pain has a devastating impact on patients' physical and mental health and quality of life. There are three key technical shortcomings in the method of extracting chronic pain EEG features based on brain networks. The current methods or studies only focus on a single EEG channel or brain region, and do not consider the EEG connection characteristics across channels or brain regions. It is impossible to comprehensively measure the entire brain as a network as a whole, and it is difficult to effectively discover the spatial features related to chronic pain. The current methods or studies only focus on a single EEG frequency range, and do not consider the association of brain functions reflected by different EEG frequency ranges, and it is difficult to effectively discover the frequency domain features related to chronic pain. The EEG features of current methods or research are very simple, and only the frequency scale is extracted. It is difficult to comprehensively consider different spatiotemporal characteristics to effectively explore the multi-scale EEG features related to chronic pain. Summary of the invention

[0003] The purpose of the present invention is to provide a method for extracting chronic pain EEG features based on brain network to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a method for extracting EEG features of chronic pain based on brain network, comprising the following steps: Step 1: Use a 50 Hz notch filter and independent component analysis to pre-process each channel of EEG data collected from chronic pain patients to remove power frequency interference, motion artifacts, and eye movement noise in the EEG data; Step 2: Use a zero-phase digital bandpass filter to extract the EEG data components of the alpha, beta, theta, and gamma frequency bands from each channel of the EEG data; Step 3: Calculate the adjacency matrix for each frequency band. This method uses the correlation of inter-channel signals to represent the distance of the brain network between channels, and performs feature selection on the values ​​in the adjacency matrix. Step 4: Calculate the adjacency matrix between each frequency band. This method uses the correlation between the channel signals between frequency bands to represent the distance between the brain networks of the frequency bands, and performs feature selection on the values ​​in the adjacency matrix. Step 5: Use a multi-layer perceptron with multiple hidden layers to construct a deep neural network, so as to regress the cross-space and cross-frequency brain network features obtained in steps 3 and 4, and obtain the chronic pain EEG index by fusing multi-scale information.

[0005] Preferably, during the preprocessing in step 1, when an abnormal situation is found, the abnormal data is recorded, and during the recording process, the data is classified and processed according to time periods.

[0006] Preferably, in step 2, the alpha frequency band represents 8-12 Hz, the beta frequency band represents 13-30 Hz, the theta frequency band represents 4-7 Hz, and the gamma frequency band represents 30-90 Hz.

[0007] Preferably, the correlation of the signals between the channels in step 3 represents the distance of the brain network between the channels, and the size of the adjacency matrix calculated is ,in is the number of EEG channels.

[0008] Preferably, in step 3, the too small correlation is regarded as the distance between the two channels being too far and resulting in no connection, so the first The frequency band The discrete EEG of the first channel The correlation distance of discrete EEG signals of channels is: ; in, Representative The frequency bands are used to filter the thresholds of channels that are too far away. In addition, because the adjacency matrix constructed by this correlation calculation is a diagonal matrix and all the elements on the diagonal are 1, only the upper half-angle matrix value of the adjacency matrix of each frequency band is taken as the feature.

[0009] Preferably, the size of the adjacency matrix calculated in step 4 is also , for the The frequency band and The adjacency matrix of the frequency bands, Line The column values ​​are passed through The frequency band Channel EEG and Frequency band The Pearson product-moment correlation coefficient of each channel EEG is calculated using the following formula: ; in, Representative The frequency band and The adjacency matrix of the frequency band Line The value of the column, and Respectively represent The frequency band The discrete EEG of the first channel The frequency band channel discrete EEG signal.

[0010] Preferably, in step 4, the too small correlation is regarded as the distance between the two channels being too far and resulting in no connection, so the first Frequency band The discrete EEG of the first channel The frequency band The correlation distance of discrete EEG signals of channels is: ; in, Representative The frequency band and The threshold used to filter channels that are too far away in the adjacency matrix of each frequency band.

[0011] Preferably, the chronic pain EEG index obtained by fusion in step five is analyzed by big data analysis technology during use, and the analyzed data is classified and collected after the analysis is completed, and a collection library is established by date for collection.

[0012] Compared with the prior art, the present invention has the following beneficial effects: The present invention constructs a cross-channel EEG network through an algorithm, thereby realizing the extraction of cross-channel EEG features. By constructing a cross-frequency EEG network, the extraction of cross-frequency EEG features is realized. By using a deep neural network and fusing multi-scale EEG information, the chronic pain state can be more accurately reflected. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 An overall structural diagram is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0014] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0015] See also Figure 1The present invention provides a technical solution: a method for extracting chronic pain EEG features based on brain network, comprising the following steps: Step 1: Use a 50 Hz notch filter and independent component analysis to pre-process each channel of EEG data collected from chronic pain patients to remove power frequency interference, motion artifacts, and eye movement noise in the EEG data; Step 2: Use a zero-phase digital bandpass filter to extract the EEG data components of the alpha, beta, theta, and gamma frequency bands from each channel of the EEG data; Step 3: Calculate the adjacency matrix for each frequency band. This method uses the correlation of inter-channel signals to represent the distance of the brain network between channels, and performs feature selection on the values ​​in the adjacency matrix. Step 4: Calculate the adjacency matrix between each frequency band. This method uses the correlation between the channel signals between frequency bands to represent the distance between the brain networks of the frequency bands, and performs feature selection on the values ​​in the adjacency matrix. Step 5: Use a multi-layer perceptron with multiple hidden layers to construct a deep neural network, so as to regress the cross-space and cross-frequency brain network features obtained in steps 3 and 4, and obtain the chronic pain EEG index by fusing multi-scale information.

[0016] During the preprocessing process in step 1, when an abnormal situation is found, the abnormal data is recorded, and during the recording process, it is classified and processed according to time periods.

[0017] In step 2, the alpha band represents 8~12Hz, the beta band represents 13~30Hz, the theta band represents 4~7Hz, and the gamma band represents 30~90Hz.

[0018] The correlation of the signals between channels in step 3 represents the distance of the brain network between channels. The size of the adjacency matrix calculated is ,in is the number of EEG channels. In the adjacency matrix, Line The column values ​​are passed through Channel EEG and The Pearson product-moment correlation coefficient of each channel EEG is calculated using the following formula: ; in, Representative The adjacency matrix of the frequency band Line The value of the column, and Respectively represent The frequency band The discrete EEG of the first channel The frequency band The channel discrete EEG signals, the matrix sizes of these discrete EEG signals are , is the number of sampling points.

[0019] In step 3, the small correlation is regarded as the distance between the two channels being too far and resulting in no connection. Therefore, the first The frequency band The discrete EEG of the first channel The correlation distance of discrete EEG signals of channels is: ; in, Representative The frequency bands are used to filter the thresholds of channels that are too far away. In addition, because the adjacency matrix constructed by this correlation calculation is a diagonal matrix and all the elements on the diagonal are 1, only the upper half-angle matrix value of the adjacency matrix of each frequency band is taken as the feature.

[0020] The size of the adjacency matrix calculated in step 4 is also , for the The frequency band and The adjacency matrix of the frequency bands, Line The column values ​​are passed through The frequency band Channel EEG and Frequency band The Pearson product-moment correlation coefficient of each channel EEG is calculated using the following formula: ; in, Representative The frequency band and The adjacency matrix of the frequency band Line The value of the column, and Respectively represent The frequency band The discrete EEG of the first channel The frequency band channel discrete EEG signal.

[0021] In step 4, the small correlation is regarded as the distance between the two channels being too far and resulting in no connection. Therefore, the first Frequency band The discrete EEG of the first channel The frequency band The correlation distance of discrete EEG signals of channels is: ; in, Representative The frequency band and The threshold used to filter channels that are too far away in the adjacency matrix of each frequency band.

[0022] The chronic pain EEG indicators obtained by fusion in step five are analyzed using big data analysis technology during use. After the analysis is completed, the analyzed data are classified and collected, and a collection library is established by date for collection.

[0023] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0024] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for extracting EEG features of chronic pain based on brain network, characterized by The following steps are involved: Step 1: Use 50 Hz notch filter and independent component analysis to preprocess each channel of EEG data collected from chronic pain patients; Step 2: Use a zero-phase digital bandpass filter; Step 3: Calculate the adjacency matrix for each frequency band. This method uses the correlation of inter-channel signals to represent the distance of the brain network between channels. Step 4: Calculate the adjacency matrix between each frequency band. This method uses the correlation between the channel signals between frequency bands to represent the distance between the brain networks of the frequency bands. Step 5: Build a deep neural network using a multi-layer perceptron with multiple hidden layers.

2. The method for extracting chronic pain EEG features based on brain network according to claim 1, characterized in that: During the preprocessing in step 1, when an abnormal situation is found, the abnormal data is recorded, and during the recording process, the data is classified and processed according to time periods.

3. The method for extracting chronic pain EEG features based on brain network according to claim 1, characterized in that: In step 2, the alpha frequency band represents 8-12 Hz, the beta frequency band represents 13-30 Hz, the theta frequency band represents 4-7 Hz, and the gamma frequency band represents 30-90 Hz.

4. The method for extracting chronic pain EEG features based on brain network according to claim 1, characterized in that: The correlation of the signals between channels in step 3 represents the distance of the brain network between channels. The size of the adjacency matrix calculated is ,in is the number of EEG channels.

5. The method for extracting chronic pain EEG features based on brain network according to claim 4, characterized in that: In the step 3, the too small correlation is regarded as the distance between the two channels being too far and resulting in no connection. Therefore, the first The frequency band The discrete EEG of the first channel The correlation distance of discrete EEG signals of channels is: ； in, Representative The frequency bands are used to filter the thresholds of channels that are too far away. In addition, because the adjacency matrix constructed by this correlation calculation is a diagonal matrix and all the elements on the diagonal are 1, only the upper half-angle matrix value of the adjacency matrix of each frequency band is taken as the feature.

6. The method for extracting chronic pain EEG features based on brain network according to claim 1, characterized in that: The size of the adjacency matrix calculated in step 4 is also , for the The frequency band and The adjacency matrix of the frequency bands.

7. The method for extracting chronic pain EEG features based on brain network according to claim 6, characterized in that: In the step 4, the too small correlation is regarded as the distance between the two channels being too far and resulting in no connection. Therefore, the first Frequency band The discrete EEG of the first channel The frequency band The correlation distance of discrete EEG signals of channels is: ； in, Representative The frequency band and The threshold used to filter channels that are too far away in the adjacency matrix of each frequency band.

8. The method for extracting chronic pain EEG features based on brain network according to claim 1, characterized in that: The chronic pain EEG index obtained by fusion in the step five is analyzed by big data analysis technology during use, and the analyzed data is classified and collected after the analysis is completed, and a collection library is established by date for collection.

Images