Method, device and electronic equipment for real-time monitoring of electroencephalogram status

Abstract

The present application provides a method, device and electronic device for real-time monitoring of EEG status, the method comprising: verifying the acquisition accuracy of the EEG acquisition device through a preset accuracy verification model; the verification process of the preset accuracy verification model comprises: multi-frequency point filtering, eigenvector calculation, statistic calculation, control limit determination and comparison of statistics and control limits based on the test data sequence of the EEG acquisition device and the corresponding standard data sequence; when the acquisition accuracy of the EEG acquisition device meets the requirements, the EEG signal data sequence is obtained through the EEG acquisition device; the EEG signal data sequence is identified by the multi-scale window EEG state analysis model to obtain the EEG signal identification index. The present application is a multi-scale window EEG state real-time monitoring method based on high-frequency rhythm, which assists in the assessment of anesthesia depth by real-time monitoring of EEG state.

Description

Technical Field

[0001] The present application relates to the field of electroencephalogram (EEG) signal technology, and in particular to a method, device and electronic equipment for real-time monitoring of EEG status. Background Art

[0002] In the process of anesthesia monitoring, EEG plays a vital role, and its role is reflected in the assessment of anesthesia depth, prevention of intraoperative awareness, and adjustment of anesthetic drug dosage. It not only helps to improve the safety and effectiveness of anesthesia, but also improves the patient's postoperative recovery and long-term prognosis. At present, the application of EEG in anesthesia depth is mainly through quantitative EEG signal indicators, such as the bispectral index and entropy index. These indicators are mainly based on the quantitative generation of EEG signals in low-frequency bands (below 50Hz), while higher-frequency bands have not been effectively used, mainly because the amplitude of the EEG signal in the band is small and is easily affected by interference, and the analysis is difficult, although the band is closely related to cognitive function and consciousness state.

[0003] A large number of studies have shown that under general anesthesia, oscillations and oscillations in the frontal lobe are a reliable sign of anesthetic unconsciousness. For example, under propofol or sevoflurane anesthesia, the energy of the bands and bands of the frontal EEG will increase significantly, while the higher-frequency bands and bands will be suppressed. However, in recent years, studies have shown that these common EEG characteristics cannot be observed for different patient groups. For example, for some elderly patients, the brain function of their frontal area gradually degenerates with age, which makes it difficult to observe the oscillation characteristics of their frontal lobe. Furthermore, the characteristics presented by the low-frequency bands of different EEG signals of anesthetic drugs on patients are very different, which makes it quite difficult to quantify the consistency of EEG indicators when evaluating the depth of anesthesia. Therefore, quantitative indicators relying on these low-frequency EEG bands may fail under certain specific conditions.

[0004] How to solve the anesthesia monitoring problem under the above-mentioned specific conditions becomes a difficult point. Summary of the invention

[0005] The purpose of the present application is to provide a method, device and electronic device for real-time monitoring of EEG status, which adopts a multi-scale window real-time monitoring method of EEG status based on high-frequency rhythm to assist in the assessment of anesthesia depth through real-time monitoring of EEG status.

[0006] In a first aspect, the present application provides a method for real-time monitoring of EEG status, the method comprising: verifying the acquisition accuracy of an EEG acquisition device through a preset accuracy verification model; the verification process of the preset accuracy verification model comprises: multi-frequency point filtering, eigenvector calculation, statistic calculation, control limit determination, and comparison of statistics and control limits based on a test data sequence of the EEG acquisition device and a corresponding standard data sequence; when the acquisition accuracy of the EEG acquisition device meets the requirements, acquiring an EEG signal data sequence through the EEG acquisition device; performing state recognition on the EEG signal data sequence using a multi-scale window EEG state analysis model to obtain an EEG signal recognition index; the state recognition process of the multi-scale window EEG state analysis model comprises: bandpass filtering, artifact detection, characterization processing, EEG statistic sequence determination, binarization sequence determination, and weighted processing based on the EEG signal data sequence; the EEG statistic sequence determination and the binarization sequence determination are completed based on an EEG training model sequence and an EEG control limit sequence determined based on an anesthesia state EEG data sequence.

[0007] Furthermore, the above-mentioned step of verifying the acquisition accuracy of the EEG acquisition device through a preset accuracy verification model includes: obtaining a test data sequence of the EEG acquisition device, and a standard data sequence corresponding to the test data sequence; using a bandpass filter to perform multi-frequency filtering on the standard data sequence and the test data sequence, respectively, to obtain a standard frequency signal sequence and a test frequency signal sequence; calculating the eigenvector of each frequency signal in the standard frequency signal sequence and the test frequency signal sequence, respectively, to obtain a standard eigenvector sequence set and a test eigenvector sequence set; using a local outlier factor algorithm to calculate the standard eigenvector sequence set and the test eigenvector sequence set, to obtain a standard statistic sequence and a test statistic sequence; using a kernel density method to determine a standard control limit sequence of the standard statistic sequence; comparing the test statistic sequence with the standard control limit sequence to determine whether the acquisition accuracy of the EEG acquisition device meets the requirements.

[0008] Furthermore, the above-mentioned step of using the local outlier factor algorithm to calculate the standard feature vector sequence set and the test feature vector sequence set to obtain the standard statistic sequence and the test statistic sequence includes: using the local outlier factor algorithm to calculate the standard feature vector sequence set to obtain the LOF statistic sequence and the standard training model sequence corresponding to the standard feature vector sequence set; using the LOF statistic sequence corresponding to the standard feature vector sequence set as the standard statistic sequence; inputting the test feature vector sequence set into the standard training model sequence to obtain the test statistic sequence.

[0009] Furthermore, the above step of comparing the test statistic sequence with the standard control limit sequence to determine whether the acquisition accuracy of the EEG acquisition device meets the requirements includes: if each statistic in the test statistic sequence is smaller than its corresponding control limit in the standard control limit sequence, it is determined that the acquisition accuracy of the EEG acquisition device meets the requirements.

[0010] Furthermore, the above-mentioned step of using the multi-scale window EEG state analysis model to perform state recognition on the EEG signal data sequence to obtain the EEG state recognition index includes: band-pass filtering the EEG signal data sequence to obtain a first signal sequence after filtering; performing artifact detection on the first signal sequence to obtain a second signal sequence after eliminating artifact interference; selecting time windows of multiple scales to perform characterization processing on the second signal sequence to obtain an EEG feature sequence set; inputting the EEG feature sequence set into the EEG training model sequence to determine the EEG statistic sequence at each scale; comparing the EEG statistic sequence at each scale with the corresponding EEG control limit sequence to obtain a binarized sequence; and weighting the binarized sequences at all scales to obtain an EEG state recognition index.

[0011] Furthermore, the process of determining the above-mentioned EEG training model sequence and EEG control limit sequence is as follows: obtain an EEG data sequence in anesthesia state; perform bandpass filtering on the EEG data sequence in anesthesia state to obtain a filtered third signal sequence; select time windows of multiple scales to characterize the third signal sequence to obtain an EEG feature sequence set in anesthesia state; use the local outlier factor algorithm to calculate the EEG feature sequence set in anesthesia state to obtain an EEG statistic sequence and an EEG training model sequence in anesthesia state at each scale; and use the kernel density method to determine the EEG control limit sequence of the EEG statistic sequence in anesthesia state.

[0012] Furthermore, the above-mentioned step of performing artifact detection on the first signal sequence to obtain a second signal sequence after eliminating artifact interference includes: taking the absolute value of the first signal sequence to obtain an amplitude signal sequence; determining a maximum value sequence corresponding to the amplitude signal sequence; eliminating strong interference artifacts from the maximum value sequence according to a set threshold to obtain an extreme value sequence; fitting the distribution curve of the extreme value sequence using kernel density estimation, and obtaining all minimum values ​​of the distribution curve, and determining the position corresponding to the first minimum value less than a specified value as the artifact threshold; eliminating artifact interference based on the artifact threshold to obtain a second signal sequence.

[0013] In the second aspect, the present application also provides a real-time monitoring device for EEG status, which includes: an equipment accuracy verification module, which is used to verify the acquisition accuracy of the EEG acquisition device through a preset accuracy verification model; the verification process of the preset accuracy verification model includes: multi-frequency point filtering, eigenvector calculation, statistic calculation, control limit determination and comparison of statistics and control limits based on the test data sequence of the EEG acquisition device and the corresponding standard data sequence; an EEG signal acquisition module, which is used to acquire the EEG signal data sequence through the EEG acquisition device when the acquisition accuracy of the EEG acquisition device meets the requirements; an EEG state recognition module, which is used to use a multi-scale window EEG state analysis model to perform state recognition on the EEG signal data sequence to obtain EEG signal recognition indicators; the state recognition process of the multi-scale window EEG state analysis model includes: bandpass filtering, artifact detection, characterization processing, EEG statistic sequence determination, binarization sequence determination and weighted processing based on the EEG signal data sequence; the EEG statistic sequence determination and binarization sequence determination are completed based on the EEG training model sequence and the EEG control limit sequence determined based on the anesthesia state EEG data sequence.

[0014] In a third aspect, the present application further provides an electronic device, comprising a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method described in the first aspect above.

[0015] In a fourth aspect, the present application further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the method described in the first aspect above.

[0016] In the real-time monitoring method, device and electronic device of EEG state provided by the present application, the acquisition accuracy of the EEG acquisition device is verified by a preset accuracy verification model; the verification process of the preset accuracy verification model includes: multi-frequency point filtering, feature vector calculation, statistic calculation, control limit determination and comparison of statistics and control limits based on the test data sequence of the EEG acquisition device and the corresponding standard data sequence; when the acquisition accuracy of the EEG acquisition device meets the requirements, the EEG signal data sequence is obtained through the EEG acquisition device; the EEG signal data sequence is state-identified using a multi-scale window EEG state analysis model to obtain an EEG signal identification index; the state identification process of the multi-scale window EEG state analysis model includes: bandpass filtering, artifact detection, characterization processing, EEG statistic sequence determination, binarization sequence determination and weighted processing based on the EEG signal data sequence; EEG statistic sequence determination and binarization sequence determination are completed based on the EEG training model sequence and the EEG control limit sequence determined by the EEG data sequence of the anesthesia state. The present application is a multi-scale window EEG state real-time monitoring method based on high-frequency rhythm, which assists in the assessment of anesthesia depth by real-time monitoring of EEG state. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] 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.

[0018] Figure 1 A flowchart of a method for real-time monitoring of EEG status provided in an embodiment of the present application;

[0019] Figure 2 A schematic diagram of a real-time monitoring process of EEG state provided in an embodiment of the present application;

[0020] Figure 3 A verification model flow chart provided for an embodiment of the present application;

[0021] Figure 4 A flow chart of an EEG state recognition model provided in an embodiment of the present application;

[0022] Figure 5 A structural block diagram of a device for real-time monitoring of EEG status provided in an embodiment of the present application;

[0023] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The technical solution of the present application will be described clearly and completely below in conjunction with the embodiments. 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 ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0025] The current application of EEG in anesthesia depth is mainly through quantitative EEG signal indicators, such as the bispectral index and entropy index. These indicators are mainly based on the quantitative generation of low-frequency band (below 50Hz) EEG signals, while higher-frequency bands have not been effectively used. EEG signals are a comprehensive reflection of the effects of anesthetic drugs on the cerebral cortex. These effects are not only reflected in the changes in the characteristics of the low-frequency band of EEG, but also in the high-frequency band of EEG. Ignoring the analysis of the high-frequency band may lose a lot of important information for the analysis of the brain state, thereby affecting the assessment of the entire anesthesia depth.

[0026] Based on this, the embodiments of the present application provide a method, device and electronic device for real-time monitoring of EEG status, which adopts a multi-scale window real-time monitoring method of EEG status based on high-frequency rhythm to assist in the assessment of anesthesia depth through real-time monitoring of EEG status.

[0027] To facilitate understanding of this embodiment, a method for real-time monitoring of EEG status disclosed in an embodiment of the present application is first introduced in detail.

[0028] Figure 1 A flowchart of a method for real-time monitoring of EEG status provided in an embodiment of the present application, Figure 2 A schematic diagram of a real-time monitoring process of EEG state provided by an embodiment of the present application is shown, and the method specifically comprises the following steps:

[0029] Step S102, verifying the acquisition accuracy of the EEG acquisition device through a preset accuracy verification model; the verification process of the preset accuracy verification model includes: multi-frequency point filtering based on the test data sequence of the EEG acquisition device and the corresponding standard data sequence, characteristic vector calculation, statistic calculation, control limit determination, and comparison between the statistic and the control limit;

[0030] Step S104, when the acquisition accuracy of the EEG acquisition device meets the requirement, the EEG signal data sequence is acquired through the EEG acquisition device;

[0031] Step S106, using the multi-scale window EEG state analysis model to perform state recognition on the EEG signal data sequence to obtain EEG signal recognition indicators; the state recognition process of the multi-scale window EEG state analysis model includes: bandpass filtering based on the EEG signal data sequence, artifact detection, characterization processing, EEG statistic sequence determination, binarization sequence determination and weighted processing; the EEG statistic sequence determination and binarization sequence determination are completed based on the EEG training model sequence and the EEG control limit sequence determined based on the anesthesia state EEG data sequence.

[0032] Considering a piece of collected EEG signal, the EEG signal can be decomposed according to the following model:

[0033] ;

[0034] in Represents the EEG signal, Indicates the background noise of the acquisition device. Indicates the artifact interference during the acquisition process. It can be found in the formula that the background noise of the device has a continuous impact on EEG acquisition, so the device for collecting EEG needs to have a stable background noise. In addition, due to the high frequency of EEG signals The amplitude of the band is small and its distribution range is wider than other frequency bands. Therefore, before collecting EEG, it is necessary to verify the background noise of the EEG collection device to ensure that stable EEG characteristics can be obtained. That is, the specific steps of verifying the collection accuracy of the EEG collection device through the preset accuracy verification model in step S102 are as follows, which can be combined with Figure 3 The verification model flowchart shown is understood:

[0035] (1) Obtaining the test data sequence of the EEG acquisition device , and the test data sequence The corresponding standard data series ;

[0036] Standard data series For a given standard signal, the test data sequence To collect the device self-test signal, such as the signal collected when the device sensor is in a short-circuited state.

[0037] (2) Use bandpass filters to filter the standard data sequences And the test data sequence Perform multi-frequency filtering to obtain a standard frequency signal sequence and a test frequency signal sequence;

[0038] For example, select frequency points, and use bandpass filters to filter the standard data sequence And the test data sequence Filter to obtain the standard frequency signal sequence And the test frequency signal sequence ;

[0039] (3) Calculate the standard frequency signal sequence respectively And the test frequency signal sequence The characteristic vector of each frequency point signal in is obtained to obtain a standard characteristic vector sequence set and a test characteristic vector sequence set;

[0040] Specifically, the feature vector sequence of each frequency point signal in the frequency point signal sequence is calculated respectively. Here, the peak-to-peak value and mean value of the signal are selected as the calculation features. The specific process is as follows:

[0041] Assume that the length of each frequency point sequence to be analyzed is , select window , whose length is , then the peak-to-peak value and mean feature vector sequence of each frequency point sequence is calculated , ,in represents the peak-to-peak value, Represents the mean.

[0042] (4) Using the local outlier factor algorithm, the standard feature vector sequence set and the test feature vector sequence set are calculated to obtain the standard statistic sequence and the test statistic sequence;

[0043] Specifically, the local outlier factor algorithm (LOF algorithm) is used to calculate the standard feature vector sequence set to obtain the LOF statistic sequence and the standard training model sequence corresponding to the standard feature vector sequence set; the LOF statistic sequence corresponding to the standard feature vector sequence set is used as the standard statistic sequence; the test feature vector sequence set is input into the standard training model sequence to obtain the test statistic sequence.

[0044] For example, the local outlier factor algorithm is used to calculate the LOF statistics of each frequency point signal feature vector sequence, and the statistical set is obtained respectively. and , the specific process is as follows:

[0045] Standard data series The accuracy meets the standard frequency signal sequence corresponding to the acquisition requirements The standard eigenvector sequence set of , then the feature vector sequence If there is no outlier data, then the local outlier factor algorithm is used to obtain The corresponding LOF statistic sequence and the standard training model sequence For the test data sequence The corresponding test frequency signal sequence The feature vector sequence in the test feature vector sequence set Input to the corresponding standard training model sequence In the above example, we can get the LOF statistic sequence .

[0046] The basic idea of ​​the local outlier factor algorithm is to first calculate a local reachable density for each data point based on the data density around the data point, and then further calculate an outlier factor for each data point through the local reachable density. The outlier factor identifies the degree of outlier of a data point. The larger the factor value, the higher the degree of outlier, and the smaller the factor value, the lower the degree of outlier.

[0047] (5) Determine the standard statistical sequence using the kernel density method The standard control limit sequence ;

[0048] Specifically, the kernel density estimation can be used to fit the LOF statistic sequence The distribution curve is used to compare with the standard data series The proportion of abnormal data in To determine the corresponding standard control limits .

[0049] Kernel density estimation is a nonparametric method in statistics used to estimate the probability density function of a random variable. This method uses kernel smoothing to estimate the probability density function by placing the kernel function on each data point and summing them.

[0050] (6) Compare the test statistic sequence with the standard control limit sequence to determine whether the acquisition accuracy of the EEG acquisition equipment meets the requirements.

[0051] If each statistic in the test statistic sequence is smaller than its corresponding control limit in the standard control limit sequence, it is determined that the acquisition accuracy of the EEG acquisition device meets the requirements, otherwise it does not meet the requirements.

[0052] For example, the test statistic series and standard control limit sequence For comparison, if The statistics in the If the device is smaller than 1, it means that the device meets the accuracy requirement at this frequency point. If the accuracy requirements are met at all frequency points, it means that the equipment accuracy verification has passed.

[0053] During general anesthesia, anesthetic drugs act on the brain, not only changing the characteristics of EEG signals below 50Hz, but also suppressing high-frequency EEG signals (70-100Hz). As shown in the figure, under the action of different drugs, the high-frequency rhythm of EEG signals (70-100Hz) is statistically significantly different, and this change is that the anesthesia state is suppressed, while the characteristics of EEG signals below 50Hz are not consistent under the action of different drugs. This makes it difficult to use a unified indicator to evaluate the brain state of EEG signals below 50Hz.

[0054] Therefore, the embodiment of the present application also proposes a brain state recognition model based on high-frequency rhythm (70-100Hz), and the process of performing EEG state recognition through the model, that is, the above step S106 uses a multi-scale window EEG state analysis model to perform state recognition on the EEG signal data sequence, and the step of obtaining the EEG signal recognition index specifically includes the following process, which can be referred to Figure 4 The flowchart shown is understood:

[0055] (1) performing bandpass filtering on the EEG signal data sequence to obtain a first signal sequence after filtering;

[0056] For example, the collection time is EEG signal data sequence Perform bandpass filtering to obtain the first signal sequence after filtering , where the filter cutoff frequency is ,and ;

[0057] (2) Performing artifact detection on the first signal sequence to obtain a second signal sequence after eliminating artifact interference; Perform artifact detection to obtain the second signal sequence after eliminating artifact interference Here, the artifact detection method first uses a fixed threshold to eliminate obvious strong interference artifacts, and then performs extreme value distribution fitting uncorrelated method on the obtained signal to eliminate suspected artifact interference. The specific steps are as follows:

[0058] (2.1) Take the absolute value of the first signal sequence to obtain the amplitude signal sequence;

[0059] Artifact-containing signal Take the absolute amplitude signal ;

[0060] (2.2) Determine the maximum value sequence corresponding to the amplitude signal sequence;

[0061] For amplitude signal Get all its maximum value sequences ;

[0062] (2.3) Eliminate strong interference artifacts from the maximum value sequence according to the set threshold to obtain the extreme value sequence;

[0063] For the maximum sequence Setting a fixed threshold , excluding strong interference artifacts, the extreme value sequence is recorded as ;

[0064] (2.4) Use kernel density estimation to fit the distribution curve of the extreme value sequence and obtain all the minimum values ​​of the distribution curve. The position corresponding to the first minimum value that is less than the specified value is determined as the artifact threshold;

[0065] Fitting a Series Using Kernel Density Estimation The distribution curve of , obtain all the minimum values ​​of the distribution curve S, and find the position corresponding to the first minimum value less than ε , which is the artifact threshold, and ε is an empirical value.

[0066] (2.5) Eliminate artifact interference based on the artifact threshold to obtain a second signal sequence.

[0067] (3) Selecting time windows of multiple scales to perform characterization processing on the second signal sequence to obtain an EEG feature sequence set;

[0068] Select multiple scale time windows to filter the second signal sequence Perform feature processing to obtain the EEG feature sequence set ,in Indicates the number of selected scale windows. Here, the peak-to-peak value and mean value of the signal are selected as the calculation features;

[0069] (4) Input the EEG feature sequence set into the EEG training model sequence to determine the EEG statistic sequence at each scale;

[0070] The EEG feature sequence set Input to EEG training model sequence And calculate the LOF statistic at each scale, that is, the EEG statistic sequence ;

[0071] (5) Compare the EEG statistic sequence at each scale with the corresponding EEG control limit sequence to obtain a binary sequence;

[0072] The EEG statistics sequence at each scale The corresponding EEG control limit sequence Compare the sequences Smaller than Assign a value of 0, otherwise 1, and get a binary sequence ; The binary sequence is a sequence consisting of 01.

[0073] (6) Perform weighted processing on the binarized sequences at all scales to obtain the EEG state recognition index.

[0074] The binarized sequences at all scales are weighted according to the following formula to obtain the EEG state recognition index: :

[0075] ;

[0076] in is the length of the binary sequence.

[0077] Furthermore, the above EEG training model sequence and EEG control limit sequence The determination process is as follows:

[0078] (1) Obtaining EEG data sequence in anesthesia state ;

[0079] (2) Bandpass filtering is performed on the EEG data sequence of the anesthesia state to obtain a filtered third signal sequence; Perform bandpass filtering to obtain the filtered third signal sequence , where the filter cutoff frequency is ,and ;

[0080] (3) Selecting multiple time windows to characterize the third signal sequence and obtaining an EEG feature sequence set of the anesthesia state;

[0081] Select multiple time windows to filter the third signal sequence Perform feature processing to obtain the EEG feature sequence set of anesthesia state ,in Indicates the number of selected scale windows. Here, the peak-to-peak value and mean value of the signal are selected as the calculation features;

[0082] (4) Using the local outlier factor algorithm, the EEG feature sequence set of the anesthesia state is calculated to obtain the EEG statistic sequence of the anesthesia state and the EEG training model sequence at each scale;

[0083] The local outlier factor algorithm is used to calculate the LOF statistic at each scale, that is, the EEG statistic sequence of anesthesia state and EEG training model sequence ;

[0084] (5) Determine the EEG control limit sequence of the EEG statistical sequence under anesthesia using the kernel density method. Determine the EEG statistical sequence of the EEG under anesthesia using the kernel density method EEG control limit sequence .

[0085] The real-time EEG state monitoring method provided in the embodiment of the present application first verifies the acquisition accuracy of the EEG acquisition device through the EEG acquisition equipment accuracy verification model; then acquires the EEG signal through the EEG acquisition device that meets the requirements, and uses the multi-scale window EEG state analysis model to perform state recognition, which can improve the recognition accuracy.

[0086] Based on the above method embodiment, the present application embodiment also provides a real-time monitoring device for EEG status, see Figure 5 As shown, the device includes: an equipment accuracy verification module 52, which is used to verify the acquisition accuracy of the EEG acquisition device through a preset accuracy verification model; the verification process of the preset accuracy verification model includes: multi-frequency point filtering, feature vector calculation, statistic calculation, control limit determination and comparison of statistics and control limits based on the test data sequence of the EEG acquisition device and the corresponding standard data sequence; an EEG signal acquisition module 54, which is used to obtain the EEG signal data sequence through the EEG acquisition device when the acquisition accuracy of the EEG acquisition device meets the requirements; an EEG state recognition module 56, which is used to use a multi-scale window EEG state analysis model to perform state recognition on the EEG signal data sequence to obtain an EEG signal recognition index; the state recognition process of the multi-scale window EEG state analysis model includes: bandpass filtering, artifact detection, characterization processing, EEG statistic sequence determination, binarization sequence determination and weighted processing based on the EEG signal data sequence; the EEG statistic sequence determination and binarization sequence determination are completed based on the EEG training model sequence and the EEG control limit sequence determined based on the anesthesia state EEG data sequence.

[0087] Furthermore, the above-mentioned device accuracy verification module 52 is used to obtain a test data sequence of the EEG acquisition device, and a standard data sequence corresponding to the test data sequence; use a bandpass filter to perform multi-frequency filtering on the standard data sequence and the test data sequence, respectively, to obtain a standard frequency signal sequence and a test frequency signal sequence; calculate the eigenvector of each frequency signal in the standard frequency signal sequence and the test frequency signal sequence, respectively, to obtain a standard eigenvector sequence set and a test eigenvector sequence set; use a local outlier factor algorithm to calculate the standard eigenvector sequence set and the test eigenvector sequence set, to obtain a standard statistic sequence and a test statistic sequence; use a kernel density method to determine the standard control limit sequence of the standard statistic sequence; compare the test statistic sequence with the standard control limit sequence to determine whether the acquisition accuracy of the EEG acquisition device meets the requirements.

[0088] Furthermore, the above-mentioned device accuracy verification module 52 is used to calculate the standard feature vector sequence set using the local outlier factor algorithm to obtain the LOF statistic sequence and the standard training model sequence corresponding to the standard feature vector sequence set; use the LOF statistic sequence corresponding to the standard feature vector sequence set as the standard statistic sequence; input the test feature vector sequence set into the standard training model sequence to obtain the test statistic sequence.

[0089] Furthermore, the device accuracy verification module 52 is used to determine that the acquisition accuracy of the EEG acquisition device meets the requirements if each statistic in the test statistic sequence is smaller than its corresponding control limit in the standard control limit sequence.

[0090] Furthermore, the above-mentioned EEG state recognition module 56 is used to perform bandpass filtering on the EEG signal data sequence to obtain a first signal sequence after filtering; perform artifact detection on the first signal sequence to obtain a second signal sequence after eliminating artifact interference; select time windows of multiple scales to perform characterization processing on the second signal sequence to obtain an EEG feature sequence set; input the EEG feature sequence set into the EEG training model sequence to determine the EEG statistic sequence at each scale; compare the EEG statistic sequence at each scale with the corresponding EEG control limit sequence to obtain a binarized sequence; and perform weighted processing on the binarized sequences at all scales to obtain an EEG state recognition index.

[0091] Furthermore, the above-mentioned EEG state recognition module 56 is used to execute the following process of determining the EEG training model sequence and the EEG control limit sequence: obtaining the EEG data sequence of the anesthesia state; performing bandpass filtering on the EEG data sequence of the anesthesia state to obtain a filtered third signal sequence; selecting time windows of multiple scales to perform characterization processing on the third signal sequence to obtain an EEG feature sequence set of the anesthesia state; using the local outlier factor algorithm to calculate the EEG feature sequence set of the anesthesia state to obtain the EEG statistic sequence and the EEG training model sequence of the anesthesia state at each scale; and using the kernel density method to determine the EEG control limit sequence of the EEG statistic sequence of the anesthesia state.

[0092] Furthermore, the above-mentioned EEG state recognition module 56 is used to take the absolute value of the first signal sequence to obtain an amplitude signal sequence; determine the maximum value sequence corresponding to the amplitude signal sequence; eliminate strong interference artifacts from the maximum value sequence according to the set threshold to obtain an extreme value sequence; use kernel density estimation to fit the distribution curve of the extreme value sequence, and obtain all minimum values ​​of the distribution curve, and determine the position corresponding to the first minimum value less than the specified value as the artifact threshold; eliminate artifact interference based on the artifact threshold to obtain a second signal sequence.

[0093] The device provided in the embodiment of the present application has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the embodiment of the device, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0094] The present application also provides an electronic device, such as Figure 6 As shown, it is a schematic diagram of the structure of the electronic device, wherein the electronic device includes a processor 61 and a memory 60, the memory 60 stores computer executable instructions that can be executed by the processor 61, and the processor 61 executes the computer executable instructions to implement the above method.

[0095] exist Figure 6 In the illustrated embodiment, the electronic device further includes a bus 62 and a communication interface 63 , wherein the processor 61 , the communication interface 63 and the memory 60 are connected via the bus 62 .

[0096] Among them, the memory 60 may include a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 63 (which can be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. can be used. The bus 62 can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 62 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0097] The processor 61 may be a serial circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by a hardware serial logic circuit in the processor 61 or an instruction in the form of software. The above processor 61 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor 61 reads the information in the memory and completes the steps of the method of the above embodiment in combination with its hardware.

[0098] An embodiment of the present application also provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the above method. The specific implementation can be found in the aforementioned method embodiment, which will not be repeated here.

[0099] The computer program products of the methods, devices, and electronic devices provided in the embodiments of the present application include a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be found in the method embodiments, which will not be repeated here.

[0100] Unless otherwise specifically stated, the relative steps, numerical expressions and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present application.

[0101] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0102] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0103] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method for real-time monitoring of EEG status, characterized in that: The method comprises: Verify the acquisition accuracy of the EEG acquisition device through a preset accuracy verification model; the verification process of the preset accuracy verification model includes: multi-frequency point filtering based on the test data sequence of the EEG acquisition device and the corresponding standard data sequence, characteristic vector calculation, statistic calculation, control limit determination and comparison between the statistic and the control limit; When the acquisition accuracy of the EEG acquisition device meets the requirement, acquiring an EEG signal data sequence through the EEG acquisition device; The state of the EEG signal data sequence is identified by using a multi-scale window EEG state analysis model to obtain an EEG signal identification index, including: band-pass filtering the EEG signal data sequence to obtain a first signal sequence after filtering; performing artifact detection on the first signal sequence to obtain a second signal sequence after eliminating artifact interference; selecting time windows of multiple scales to perform characterization processing on the second signal sequence to obtain an EEG feature sequence set; inputting the EEG feature sequence set into an EEG training model sequence to determine an EEG statistic sequence at each scale; comparing the EEG statistic sequence at each scale with the corresponding EEG control limit sequence to obtain a binarized sequence; weighting the binarized sequences at all scales to obtain an EEG state identification index; The process of determining the EEG training model sequence and the EEG control limit sequence is as follows: Acquire an EEG data sequence of anesthesia state; perform bandpass filtering on the EEG data sequence of anesthesia state to obtain a filtered third signal sequence; select time windows of multiple scales to perform characterization processing on the third signal sequence to obtain an EEG feature sequence set of anesthesia state; calculate the EEG feature sequence set of anesthesia state using a local outlier factor algorithm to obtain an EEG statistic sequence of anesthesia state and an EEG training model sequence at each scale; and determine the EEG control limit sequence of the EEG statistic sequence of anesthesia state using a kernel density method.

2. The method according to claim 1, characterized in that The steps of verifying the acquisition accuracy of the EEG acquisition device through the preset accuracy verification model include: Acquire a test data sequence of an EEG acquisition device and a standard data sequence corresponding to the test data sequence; Using a bandpass filter to perform multi-frequency point filtering on the standard data sequence and the test data sequence respectively to obtain a standard frequency point signal sequence and a test frequency point signal sequence; Calculating the characteristic vector of each frequency point signal in the standard frequency point signal sequence and the test frequency point signal sequence respectively to obtain a standard characteristic vector sequence set and a test characteristic vector sequence set; The standard feature vector sequence set and the test feature vector sequence set are calculated using a local outlier factor algorithm to obtain a standard statistic sequence and a test statistic sequence; Determining a standard control limit sequence of the standard statistic sequence using a kernel density method; The test statistic sequence is compared with the standard control limit sequence to determine whether the acquisition accuracy of the EEG acquisition device meets the requirements.

3. The method according to claim 2, characterized in that The step of calculating the standard feature vector sequence set and the test feature vector sequence set by using a local outlier factor algorithm to obtain a standard statistic sequence and a test statistic sequence comprises: Calculating the standard feature vector sequence set using a local outlier factor algorithm to obtain a LOF statistic sequence and a standard training model sequence corresponding to the standard feature vector sequence set; Using the LOF statistic sequence corresponding to the standard feature vector sequence set as the standard statistic sequence; The test feature vector sequence set is input into the standard training model sequence to obtain a test statistic sequence.

4. The method according to claim 2, characterized in that: The step of comparing the test statistic sequence with the standard control limit sequence to determine whether the acquisition accuracy of the EEG acquisition device meets the requirements includes: If each statistic in the test statistic sequence is smaller than its corresponding control limit in the standard control limit sequence, it is determined that the acquisition accuracy of the EEG acquisition device meets the requirements.

5. The method according to claim 1, characterized in that The step of performing artifact detection on the first signal sequence to obtain a second signal sequence after eliminating artifact interference includes: Taking the absolute value of the first signal sequence to obtain an amplitude signal sequence; Determine a maximum value sequence corresponding to the amplitude signal sequence; Eliminate strong interference artifacts from the maximum value sequence according to a set threshold to obtain an extreme value sequence; Fitting the distribution curve of the extreme value sequence by using kernel density estimation, obtaining all the minimum values ​​of the distribution curve, and determining the position corresponding to the first minimum value less than the specified value as the artifact threshold; Artifact interference is eliminated based on the artifact threshold to obtain a second signal sequence.

6. A device for real-time monitoring of EEG status, characterized in that: The device comprises: The device accuracy verification module is used to verify the acquisition accuracy of the EEG acquisition device through a preset accuracy verification model; the verification process of the preset accuracy verification model includes: multi-frequency point filtering based on the test data sequence of the EEG acquisition device and the corresponding standard data sequence, feature vector calculation, statistic calculation, control limit determination and comparison between statistics and control limits; An EEG signal acquisition module, used to acquire an EEG signal data sequence through the EEG acquisition device when the acquisition accuracy of the EEG acquisition device meets the requirements; The EEG state recognition module is used to perform state recognition on the EEG signal data sequence using a multi-scale window EEG state analysis model to obtain an EEG signal recognition index, including: bandpass filtering the EEG signal data sequence to obtain a first signal sequence after filtering; performing artifact detection on the first signal sequence to obtain a second signal sequence after eliminating artifact interference; selecting time windows of multiple scales to perform characterization processing on the second signal sequence to obtain an EEG feature sequence set; inputting the EEG feature sequence set into an EEG training model sequence to determine an EEG statistic sequence at each scale; comparing the EEG statistic sequence at each scale with the corresponding EEG control limit sequence to obtain a binary sequence; The binarized sequences at all scales are weighted to obtain an EEG state recognition index; the EEG training model sequence and the EEG control limit sequence are determined as follows: an EEG data sequence of anesthesia state is obtained; the EEG data sequence of anesthesia state is band-pass filtered to obtain a filtered third signal sequence; time windows of multiple scales are selected to characterize the third signal sequence to obtain an EEG feature sequence set of anesthesia state; the EEG feature sequence set of anesthesia state is calculated using a local outlier factor algorithm to obtain an EEG statistic sequence of anesthesia state and an EEG training model sequence at each scale; and the EEG control limit sequence of the EEG statistic sequence of anesthesia state is determined using a kernel density method.

7. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the method according to any one of claims 1 to 5.

Images