Method and device for modifying bifrequency index of electroencephalogram, equipment and storage medium

Abstract

This invention discloses a method, apparatus, device, and storage medium for correcting the bispectral index (BIS) of an electroencephalogram (EEG) signal. The method includes: acquiring feature data of an EEG signal from a target user during a preset first time period, the feature data including at least first feature values ​​of multiple sub-features of the EEG signal and a first bispectral index, the sub-features reflecting the degree of inhibition of the EEG signal; determining first algebraic values ​​corresponding to each sub-feature based on the first feature values ​​and a preset feature threshold, the algebraic values ​​reflecting whether the bispectral index is a valid value; determining a target correction coefficient using the first algebraic values, the first bispectral index, and a preset correction coefficient algorithm; and performing correction processing based on the target correction coefficient and the first bispectral index to obtain the target bispectral index. This method allows for the correction of the first bispectral index using a target correction coefficient, improving the accuracy of the BIS index determination and reducing errors caused by insufficient robustness.

Description

Technical Field

[0001] This invention relates to the field of electroencephalogram (EEG) signal processing technology, and in particular to a method, apparatus, device, and storage medium for correcting the dual-frequency index of EEG signals. Background Technology

[0002] During surgery, real-time and accurate monitoring of the patient's anesthesia status is crucial to prevent delayed intraoperative awareness and awakening. Electroencephalogram (EEG) signals reflect the neural activity of neurons in the cerebral cortex through electrical signals. Its advantages, including being non-invasive, providing large datasets, and allowing for continuous real-time monitoring, have led to its widespread application in brain-related medical diagnostics and monitoring. In clinical surgery, EEG signals are also used as an important indicator for assessing the depth of anesthesia. The bispectral index (BIS) and cerebral state index (CSI) are two of the most common monitoring indicators.

[0003] The Biological Intensity Index (BIS) uses four parameters—Burst Suppression Ratio (BSR), QUAZI suppression index, β ratio, and Synch Fast Slow (SFS)—of the EEG signal to determine the depth of anesthesia corresponding to the current EEG signal through a weighted summation. However, because the BIS index is modeled by fitting previously collected data, its robustness is somewhat lacking. Therefore, the accuracy of the BIS index determination still needs improvement. Summary of the Invention

[0004] The main objective of this invention is to provide a method, apparatus, device, and storage medium for correcting the bispectral index of electroencephalogram (EEG) signals, which can solve the problem of the lack of accuracy in determining the BIS index in the prior art.

[0005] To achieve the above objectives, the first aspect of the present invention provides a method for correcting the bifrequency index of electroencephalogram (EEG) signals, the method comprising:

[0006] Acquire feature data of the target user's electroencephalogram (EEG) signal during a preset first time period. The feature data includes at least the first feature value of multiple sub-features of the EEG signal and a first bi-frequency index. The sub-features are used to reflect the degree of inhibition of the EEG signal.

[0007] Based on the first feature value and the preset feature threshold, the first generation value corresponding to each sub-feature is determined. The generation value is used to reflect whether the dual-frequency index is a valid value.

[0008] The target correction coefficient is determined using the first generation numerical value, the first dual-frequency index, and a preset correction coefficient algorithm.

[0009] The target dual-frequency index is obtained by performing a correction process based on the target correction coefficient and the first dual-frequency index.

[0010] In one feasible implementation, determining the target correction coefficient using the first generation numerical value, the first dual-frequency index, and a preset correction coefficient algorithm includes:

[0011] The second dual-frequency index corresponding to the second time period before the first time period and the second generation value corresponding to various sub-features are obtained from the preset historical database. The historical database includes the correspondence between the target user's historical dual-frequency index, sub-features, generation values ​​and time.

[0012] The target correction coefficient is determined using the correction coefficient algorithm, the first dual-frequency index, the second dual-frequency index, various first-generation values, and second-generation values.

[0013] In one feasible implementation, the correction coefficient algorithm includes a weighted average algorithm corresponding one-to-one with each segment interval of the dual-frequency index. Then, determining the target correction coefficient using the correction coefficient algorithm, the first dual-frequency index, the second dual-frequency index, various first-generation values, and second-generation values ​​includes:

[0014] The mean of the dual-frequency index is determined by averaging the first dual-frequency index and the second dual-frequency index.

[0015] The algebraic mean of various sub-features is determined by calculating the mean of the first-generation and second-generation values.

[0016] Based on the dual-frequency exponential mean, a target weighted average algorithm is selected from the correction coefficient algorithm to identify the first segment interval where the dual-frequency exponential mean is located;

[0017] The algebraic mean is input into the target weighted average algorithm to determine the target correction coefficient.

[0018] In one feasible implementation, the step of performing correction processing based on the target correction coefficient and the first dual-frequency index to obtain the target dual-frequency index includes:

[0019] The dual-frequency index compensation value is determined based on the target correction coefficient, the first dual-frequency index, and the preset compensation algorithm.

[0020] The sum of the dual-frequency index compensation value and the first dual-frequency index is taken as the target dual-frequency index.

[0021] In one feasible implementation, the feature threshold includes the optimal feature thresholds for various sub-features within segmented intervals of different dual-frequency indices. The sub-features include at least burst suppression ratio, QUAZI suppression index, β ratio, and synchronization speed-to-slow ratio. Then, determining the first-generation values ​​corresponding to each sub-feature based on the first feature value and the preset feature thresholds includes:

[0022] Based on the first dual-frequency index, a target feature threshold for the second segment interval where the first dual-frequency index is located is determined from a preset feature threshold. The target feature threshold includes the optimal feature threshold for various sub-features in the second segment interval.

[0023] The first eigenvalues ​​of the burst suppression ratio, QUAZI suppression index, β ratio, and synchronization speed ratio are compared with their respective target feature thresholds to determine the comparison results of various sub-features;

[0024] If the comparison result is that the first feature value is greater than the target feature threshold, then the first generation value of the sub-feature is determined to be 1;

[0025] If the comparison result is that the first feature value is less than or equal to the target feature threshold, then the first generation value of the sub-feature is determined to be 0.

[0026] In one feasible implementation, the step of determining the first-generation values ​​corresponding to various sub-features based on the first feature value and a preset feature threshold further includes:

[0027] Obtain sample feature data for each segmented interval, wherein the sample feature data includes the correspondence between several sample dual-frequency indices and sample feature values ​​of various sub-features for each segmented interval;

[0028] The sample feature values ​​are input as breeding individuals into a preset hybrid rice model for breeding processing until the maximum number of iterations is reached, to obtain the best individual for the sample feature values ​​of various sub-features under each segment interval output by the hybrid rice model; and the best individual is used as the best feature threshold.

[0029] In one feasible implementation, the method further includes:

[0030] The target bifrequency index and the first-generation value are saved to a preset historical database, and the process returns to the step of obtaining the characteristic data of the target user's EEG signal in a preset first time period.

[0031] To achieve the above objectives, a second aspect of the present invention provides a device for correcting the bifrequency index of electroencephalogram (EEG) signals, the device comprising:

[0032] Data acquisition module: used to acquire feature data of the target user's electroencephalogram (EEG) signal during a preset first time period. The feature data includes at least the first feature value of multiple sub-features of the EEG signal and a first bi-frequency index. The sub-features are used to reflect the degree of inhibition of the EEG signal.

[0033] Algebraic determination module: used to determine the first algebraic values ​​corresponding to various sub-features based on the first feature value and the preset feature threshold, wherein the algebraic values ​​are used to reflect whether the dual-frequency index is a valid value;

[0034] Coefficient determination module: used to determine the target correction coefficient using the first generation numerical value, the first dual-frequency index, and a preset correction coefficient algorithm;

[0035] Index correction module: used to perform correction processing based on the target correction coefficient and the first dual-frequency index to obtain the target dual-frequency index.

[0036] To achieve the above objectives, a third aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps shown in the first aspect and any feasible implementation.

[0037] To achieve the above objectives, a fourth aspect of the present invention provides a computer device including a memory and a processor, the memory storing a computer program, which, when executed by the processor, causes the processor to perform the steps shown in the first aspect and any feasible implementation.

[0038] The embodiments of the present invention have the following beneficial effects:

[0039] This invention provides a method for correcting the bispectral index (BIS) of an electroencephalogram (EEG) signal. The method includes: acquiring feature data of an EEG signal from a target user during a preset first time period, wherein the feature data includes at least first feature values ​​of multiple sub-features of the EEG signal and a first bispectral index, the sub-features reflecting the degree of inhibition of the EEG signal; determining first algebraic values ​​corresponding to each sub-feature based on the first feature values ​​and a preset feature threshold, the algebraic values ​​reflecting whether the bispectral index is a valid value; determining a target correction coefficient using the first algebraic values, the first bispectral index, and a preset correction coefficient algorithm; and performing correction processing based on the target correction coefficient and the first bispectral index to obtain the target bispectral index. This method allows for the correction of the first bispectral index using a target correction coefficient, improving the accuracy of the BIS index determination and reducing errors caused by insufficient robustness. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] in:

[0042] Figure 1 This is a flowchart of a method for correcting the bi-frequency index of an electroencephalogram (EEG) signal according to an embodiment of the present invention.

[0043] Figure 2 This is another flowchart of a method for correcting the bi-frequency index of an electroencephalogram (EEG) signal according to an embodiment of the present invention;

[0044] Figure 3 This is a structural block diagram of a device for correcting the bi-frequency index of an electroencephalogram (EEG) signal according to an embodiment of the present invention.

[0045] Figure 4 This is a structural block diagram of a computer device in an embodiment of the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Please see Figure 1 , Figure 1 This is a flowchart of a method for correcting the bifrequency index of an electroencephalogram (EEG) signal according to an embodiment of the present invention, as shown below. Figure 1 The method shown can be applied to both terminals and servers. The terminal can be a desktop terminal or a mobile terminal; a mobile terminal can be at least one of a mobile phone, tablet, or laptop. The server can be a standalone server or a server cluster consisting of multiple servers. This embodiment uses a terminal application as an example. Figure 1 The method shown includes the following steps:

[0048] 101. Obtain feature data of the target user's electroencephalogram (EEG) signal during a preset first time period. The feature data includes at least the first feature value of multiple sub-features of the EEG signal and a first bi-frequency index. The sub-features are used to reflect the degree of inhibition of the EEG signal.

[0049] It should be noted that the method for correcting the bispectral index of EEG signals in this embodiment of the invention can also be applied to monitoring terminals and other devices capable of monitoring EEG signals, thereby enabling the acquisition of EEG signals from a target user and obtaining corresponding feature data. The target user is the user whose EEG signals are being collected, such as a patient under anesthesia during surgery. Furthermore, corresponding feature data can be extracted from the acquired EEG signals. This feature data includes, but is not limited to, the bispectral index (BIS index) of the EEG signal and feature values ​​of various sub-features. These sub-features include, but are not limited to, burst suppression ratio (BSR), QUAZI suppression index, β ratio, and Synch FastSlow ratio (SFS), etc., which can reflect the degree of inhibition of EEG signals. For example, feature values ​​of various sub-features can be extracted from the EEG signals, and the corresponding BIS index can be determined by weighted summation of these feature values. The BIS index can then reflect the depth of anesthesia corresponding to the EEG signal. In this embodiment, to improve the calculation and correction accuracy of the BIS index, the BIS index is calculated and corrected using data over a period of time. Specifically, feature data of the target user's EEG signal during a preset first time period is acquired. This feature data includes at least first feature values ​​of multiple sub-features of the EEG signal and a first bi-frequency index. The first feature values ​​are the feature values ​​of the multiple sub-features of the EEG signal during the preset first time period, and the first bi-frequency index is the bi-frequency index of the EEG signal during the preset first time period. For example, the multiple sub-features may be burst inhibition ratio, QUAZI inhibition index, β ratio, and synchronization speed ratio. Therefore, the first feature values ​​include the first feature value of the burst inhibition ratio, the first feature value of the QUAZI inhibition index, the first feature value of the β ratio, and the first feature value of the synchronization speed ratio. The preset first time period can be 2 seconds or other time lengths, which are not limited here. Taking a preset first time period of 2 seconds as an example, the corresponding feature data can be extracted each time a 2-second EEG signal is acquired.

[0050] 102. Based on the first feature value and the preset feature threshold, determine the first generation value corresponding to each sub-feature, wherein the generation value is used to reflect whether the dual-frequency index is a valid value;

[0051] After obtaining the first eigenvalue and the first dual-frequency index, the weighted sum of the eigenvalues ​​of various sub-features can be used to obtain the first dual-frequency index for subsequent correction processing to improve accuracy. Specifically, based on the first eigenvalue and a preset feature threshold, the first algebraic value corresponding to each sub-feature can be determined. This algebraic value reflects whether the dual-frequency index is a valid value. It should be noted that different sub-features have different feature thresholds. The feature threshold reflects the critical threshold of the sub-feature's eigenvalue, thus determining whether the sub-feature is valid. Meeting the feature threshold indicates the sub-feature is reliable; otherwise, it is unreliable. Since the first dual-frequency index is obtained by weighted summation of various sub-features, it can be used to reflect whether the dual-frequency index is a valid value. Therefore, the first algebraic value reflecting the validity of the dual-frequency index corresponding to each sub-feature can be determined using the first eigenvalue and the preset feature threshold. For example, comparing a sub-feature with its corresponding feature threshold will yield different algebraic values ​​depending on the comparison result, thus obtaining the algebraic value of that sub-feature.

[0052] 103. Determine the target correction coefficient using the first generation numerical value, the first dual-frequency index, and the preset correction coefficient algorithm;

[0053] Furthermore, after obtaining the algebraic value reflecting whether the dual-frequency index is effective, different correction coefficients can be obtained through the algebraic value. In order to improve the accuracy of the correction coefficient calculation, since the preset correction coefficient algorithm is not only related to the algebraic value, but also to the dual-frequency index obtained by weighted summation of various sub-features, the target correction coefficient is determined by using the first algebraic value, the first dual-frequency index, and the preset correction coefficient algorithm to improve the accuracy of the correction coefficient.

[0054] 104. Perform correction processing based on the target correction coefficient and the first dual-frequency index to obtain the target dual-frequency index.

[0055] After obtaining the target correction coefficient, the target correction coefficient and the first dual-frequency index can be used for correction to obtain the target dual-frequency index, where the target dual-frequency index is the index value after correction of the first dual-frequency index.

[0056] This invention provides a method for correcting the bispectral index (BIS) of an electroencephalogram (EEG) signal. The method includes: acquiring feature data of an EEG signal from a target user during a preset first time period, wherein the feature data includes at least first feature values ​​of multiple sub-features of the EEG signal and a first bispectral index, the sub-features reflecting the degree of inhibition of the EEG signal; determining first algebraic values ​​corresponding to each sub-feature based on the first feature values ​​and a preset feature threshold, the algebraic values ​​reflecting whether the bispectral index is a valid value; determining a target correction coefficient using the first algebraic values, the first bispectral index, and a preset correction coefficient algorithm; and performing correction processing based on the target correction coefficient and the first bispectral index to obtain the target bispectral index. This method allows for the correction of the first bispectral index using a target correction coefficient, improving the accuracy of the BIS index determination and reducing errors caused by insufficient robustness.

[0057] Please see Figure 2 , Figure 2 This is another flowchart of a method for correcting the bifrequency index of an electroencephalogram (EEG) signal according to an embodiment of the present invention, as shown below. Figure 2 The method shown includes the following steps:

[0058] 201. Obtain feature data of the target user's electroencephalogram (EEG) signal during a preset first time period. The feature data includes at least the first feature value of multiple sub-features of the EEG signal and a first bi-frequency index. The sub-features are used to reflect the degree of inhibition of the EEG signal.

[0059] 202. Based on the first feature value and the preset feature threshold, determine the first generation value corresponding to each sub-feature, wherein the generation value is used to reflect whether the dual-frequency index is a valid value;

[0060] It should be noted that steps 201 and 202 are related to... Figure 1 Steps 101 and 102 shown are similar and will not be repeated here to avoid repetition. Please refer to [link to relevant documentation] for details. Figure 1 The contents of steps 101 and 102 are shown.

[0061] In one feasible implementation, to further improve the accuracy of the algebraic value determination, the feature threshold will vary not only due to different sub-features but also due to different dual-frequency indices. Furthermore, the aforementioned feature thresholds include the optimal feature thresholds for various sub-features within different segmented intervals of dual-frequency indices. The aforementioned sub-features include at least the burst suppression ratio, QUAZI suppression index, β ratio, and synchronization speed ratio. Then, step 202 may specifically include the following steps A1 to A4:

[0062] A1. Based on the first dual-frequency index, determine the target feature threshold of the second segment interval where the first dual-frequency index is located from the preset feature thresholds. The target feature threshold includes the optimal feature threshold of various sub-features in the second segment interval.

[0063] The dual-frequency index is pre-divided into segments, such as [40, 60), [60, 70), and [70, 100). Since the same sub-feature exhibits different characteristics at different stages of anesthesia, and different sub-features at the same stage also exhibit different characteristics, the range of feature values ​​will vary, affecting the feature threshold. Therefore, to more accurately obtain the feature threshold for the first feature value, a feature threshold list can be pre-set. This list includes the optimal feature thresholds for different sub-features within each segment, thereby improving the accuracy of threshold determination. The optimal feature threshold can be obtained by training the sample data using a breeding algorithm such as the hybrid rice algorithm and stored in the terminal for direct retrieval during use.

[0064] A2. Compare the first feature values ​​of the burst suppression ratio, QUAZI suppression index, β ratio, and synchronization speed ratio with their respective target feature thresholds to determine the comparison results of various sub-features;

[0065] A3. If the comparison result is that the first feature value is greater than the target feature threshold, then the first generation value of the sub-feature is determined to be 1;

[0066] A4. If the comparison result is that the first feature value is less than or equal to the target feature threshold, then the first generation value of the sub-feature is determined to be 0.

[0067] Furthermore, different sub-features are compared with their corresponding target feature thresholds to determine the comparison result. Specifically, if the comparison result is that the first feature value is greater than the target feature threshold, the first generation value of the sub-feature is determined to be 1; otherwise, if the comparison result is that the first feature value is less than or equal to the target feature threshold, the first generation value of the sub-feature is determined to be 0.

[0068] For example, each segment interval is [40,60), [60,70), [70,100), and the sub-features include four types: burst suppression ratio, QUAZI suppression index, β ratio, and synchronization speed ratio. The preset feature thresholds include the four optimal feature thresholds for the four sub-features under the segment interval [40,60), the four optimal feature thresholds for the four sub-features under the segment interval [60,70), and the four optimal feature thresholds for the four sub-features under the segment interval [70,100). Furthermore, if the first dual-frequency index is 55, then its second segment interval is [40, 60). Then, the target feature threshold for the second segment interval [40, 60) is found from the preset feature thresholds. This target feature threshold is the four optimal feature thresholds for the four sub-features within the segment interval [40, 60). Then, the first feature value of the four sub-features at this time is compared with the four optimal feature thresholds for the four sub-features within the segment interval [40, 60). That is, the first feature value of the burst suppression ratio is compared with the optimal feature threshold of the burst suppression ratio, the first feature value of the QUAZI suppression index is compared with the optimal feature threshold of the QUAZI suppression index, the first feature value of the β ratio is compared with the optimal feature threshold of the β ratio, and the first feature value of the synchronization speed ratio is compared with the optimal feature threshold of the synchronization speed ratio. Four comparison results are obtained, and according to steps A3 and A4, corresponding first-generation values ​​are assigned to each sub-feature. The first-generation values ​​are 0 or 1. For example, if the first feature value of the burst suppression ratio is compared with the optimal feature threshold of the burst suppression ratio, and the result of the comparison is that the first feature value of the burst suppression ratio is greater than the target feature threshold, then the first generation value of the burst suppression ratio is determined to be 1. Conversely, if the result of the comparison is that the first feature value of the burst suppression ratio is less than or equal to the target feature threshold, then the first generation value of the burst suppression ratio is determined to be 0. The same applies to other types of sub-features, which will not be elaborated here.

[0069] In one feasible implementation, to further improve the accuracy of feature threshold determination, the hybrid rice algorithm is used to train the samples to obtain the aforementioned feature threshold. Therefore, steps B1 and B2 may be included before step 202:

[0070] B1. Obtain sample feature data under each segmented interval, wherein the sample feature data includes the correspondence between several sample dual-frequency indices and sample feature values ​​of various sub-features under the segmented interval;

[0071] It should be noted that, in order to obtain the optimal feature thresholds for various sub-features in different segment intervals, the sample data in this application is collected according to different segment intervals. That is, the sample feature data corresponds one-to-one with the segment intervals, thereby obtaining the sample feature data under each segment interval. Among them, the sample feature data includes the correspondence between several sample bifrequency indices under the segment intervals and the sample feature values ​​of various sub-features. For example, various sample EEG signals are collected. The sample EEG signals can be obtained from past clinical databases. The EEG signals and BIS index related data of patients during surgery are obtained from the clinical databases. The sample feature values ​​of various sub-features and the sample bifrequency indices are obtained from the sample EEG signals. The correspondence between the sample feature values ​​and the sample bifrequency indices is established. Then, the sample bifrequency indices in the corresponding segment intervals are clustered together according to the preset segment intervals, thereby obtaining several sample bifrequency indices under the segment intervals, and obtaining the correspondence between several sample bifrequency indices under each segment interval and the sample feature values ​​of various sub-features.

[0072] B2. Input the sample feature values ​​as breeding individuals into a preset hybrid rice model for breeding processing until the maximum number of iterations is reached, and obtain the best individual for the sample feature values ​​of various sub-features under each segment interval output by the hybrid rice model; and use the best individual as the best feature threshold.

[0073] Furthermore, the sample feature values ​​can be input as breeding individuals into a preset hybrid rice model for breeding processing until the maximum number of iterations is reached, obtaining the optimal individual for the sample feature values ​​of various sub-features in each segment interval of the hybrid rice model output; and the optimal individual is used as the optimal feature threshold. For example, several sample feature values ​​of any seed feature in different segment intervals are input into the hybrid rice model to obtain the optimal individual for any seed feature in different segment intervals. This optimal individual marks the optimal feature value of this sub-feature, so it can be used as the optimal feature threshold (referred to as the optimal threshold).

[0074] In one feasible implementation, the specific process for obtaining the optimal threshold using the hybrid rice algorithm is as follows: steps 1) to 8):

[0075] 1) First, initialize the model parameters of the hybrid rice algorithm, including population size, individual dimension, upper and lower bounds of encoding, maximum number of iterations, and maximum number of self-crosses;

[0076] 2) Initialize the population by setting thresholds for sub-feature parameters;

[0077] 3) Calculate the fitness value of each individual using the fitness function, and determine the quality of the individual based on the fitness value;

[0078] 4) Based on the results in 3), the population is divided into three subpopulations: maintenance line, sterile line, and restorer line, each accounting for 1 / 3. When the population size cannot be divided into three equal parts, the number of seeds of the maintenance line and the sterile line is the same, and the rest are classified as restorer lines.

[0079] 5) The maintainer line and the sterile line are crossed. The selection method for the cross is random selection, as shown in the following formula:

[0080]

[0081] in, It is the k-th gene of the i-th individual in the sterile line; It preserves the k-th gene of the a-th individual in the line. It is the k-th gene of the b-th individual in the sterile line, where r1 and r2 are random numbers between -1 and 1, and r1 + r2 ≠ 0. If the new individual after hybridization is superior, it is retained; otherwise, it is discarded.

[0082] The restorer line undergoes self-crossing, using the following formula:

[0083]

[0084] in, It is the i-th restorer individual New individuals produced through self-fertilization; X best It is the current optimal individual; It is the j-th individual in the currently randomly selected recovery line, and r3 is a random number between 0 and 1.

[0085] 6) Obtain the optimal individual;

[0086] 7) Return to steps 3) through 6) and repeat until the maximum number of iterations is reached;

[0087] 8) Output the optimal threshold for each sub-feature. That is, the optimal individual for each sub-feature.

[0088] Furthermore, repeat steps 1) to 8) according to different BIS index ranges until the optimal thresholds for the four sub-features are obtained for BIS indices of [40,60), [60,70), and [70,100), respectively. In other words, the final optimal thresholds will be obtained as 3 segment intervals * 4 sub-feature categories = 12.

[0089] To further improve the accuracy of the target correction coefficient, the target correction coefficient can be obtained based on steps 203 and 204, as follows.

[0090] 203. Obtain the second dual-frequency index corresponding to the second time period before the first time period and the second generation value corresponding to various sub-features from the preset historical database. The historical database includes the correspondence between the target user's historical dual-frequency index, sub-features, generation values ​​and time.

[0091] To avoid frequent and large fluctuations in the target correction coefficient, a weighted average of the results of the four sub-features can be calculated based on different anesthesia states to obtain the target correction coefficient. Specifically, the target correction coefficient can be calculated using historical data. The second bifrequency index and the second generation values ​​of various sub-features corresponding to the preset second time period before the preset first time period are obtained from the preset historical database. The historical database includes the correspondence between the target user's historical bifrequency index, sub-features, algebraic values ​​and time. In other words, the historical database is used to record the target user's historical EEG data, such as the characteristic data, algebraic values, and acquisition time of the EEG signal. For example, the historical database can include the correspondence between several past preset first time periods and the characteristic values ​​of the bifrequency index, algebraic values ​​and sub-features. For example, if the target user's 8-second EEG signal has been acquired before, and the preset first time period is 2 seconds, then the historical database will include 4 sets of data, that is, the bifrequency index, algebraic values ​​and characteristic values ​​of the sub-features corresponding to each 2-second period. In this application, the preset first time period and the preset second time period can be equal or unequal. Furthermore, since the algebraic values ​​are obtained by analyzing a segment of EEG signal within the preset first time period, the algebraic values ​​in the historical database are also obtained from historical time periods of the same length. Therefore, the preset second time period is N times the preset first time period, where N is a positive integer, thus obtaining an integer number of historical algebraic values ​​and bifrequency indices. For example, if the first time period is 2 seconds and N is 4, then the second time period can be 8 seconds. This allows for the acquisition of the second bifrequency index and the second algebraic value for these 8 seconds. The second bifrequency index can include the four historical bifrequency indices from these 8 seconds, and the second algebraic value includes 16 algebraic values ​​comprising four sub-features.

[0092] For example, after acquiring the raw EEG signal, a threshold method is used to make a judgment every 2 seconds. After recording for 10 seconds, the results of 5 judgments are counted. There are only two results: 1 and 0. 1 indicates that the sub-feature parameter is greater than the optimal threshold in this BIS index range. At this time, the bifrequency index is reliable. 0 indicates the opposite. At this time, the bifrequency index is not reliable. The number of 1s and 0s in 10 seconds can be further counted to comprehensively evaluate whether the bifrequency index is reliable or not.

[0093] 204. Using the aforementioned correction coefficient algorithm, the first dual-frequency index, the second dual-frequency index, various first-generation values, and second-generation values, determine the target correction coefficient;

[0094] Furthermore, the target correction coefficient can be determined using the correction coefficient algorithm, the first dual-frequency index, the second dual-frequency index, various first-generation values, and second-generation values, making the target correction coefficient more accurate. The correction coefficient algorithm includes a weighted average algorithm that corresponds one-to-one with each segment interval of the dual-frequency index.

[0095] Therefore, step 204 may include the following steps C1 to C4:

[0096] C1. Calculate the mean of the dual-frequency index using the first dual-frequency index and the second dual-frequency index;

[0097] C2. Calculate the mean value of the algebraic values ​​of various sub-features by using the first generation values ​​and the second generation values.

[0098] Continuing with the example above, if the first time period is 2 seconds and N is 4, then the second time period can be 8 seconds. This allows us to obtain the second dual-frequency index and the second-generation values ​​for these 8 seconds. The second dual-frequency index can include the four historical dual-frequency indices from these 8 seconds, and the second-generation values ​​include 16 algebraic values ​​for four sub-features. Adding the first dual-frequency index obtained this time, we get 5 dual-frequency indices and 20 algebraic values. Calculating the mean of the first and second dual-frequency indices is equivalent to calculating the mean of these 5 dual-frequency indices, resulting in the dual-frequency index mean. Calculating the mean of each first-generation value and second-generation value is equivalent to calculating the mean of the 5 algebraic values ​​for each sub-feature, resulting in the algebraic value mean for each sub-feature.

[0099] C3. Based on the mean of the dual-frequency index, select the target weighted average algorithm corresponding to the first segment interval where the mean of the dual-frequency index is located from the correction coefficient algorithm;

[0100] C4. Input the algebraic mean into the target weighted average algorithm to determine the target correction coefficient.

[0101] For example, the correction coefficient algorithm includes the following mathematical expression (1):

[0102]

[0103] In the formula, K is the target correction coefficient, BIS is the first dual-frequency index, and the coefficient a+b=1, always ensuring a>b. Different subscripts of coefficients a and b indicate different emphases on different sub-features within different BIS ranges. SFS, βRatio, BSR, and QUAZI represent, in order, the synchronization speed ratio, β rate, burst suppression ratio (also known as burst suppression ratio), and QUAZI suppression ratio of the BIS index, respectively. The subscript R represents the mean, i.e., SFS. R βRatioR BSR R And QUAZI R These represent the algebraic mean values ​​of various sub-features.

[0104] 205. Perform correction processing based on the target correction coefficient and the first dual-frequency index to obtain the target dual-frequency index.

[0105] It should be noted that step 205 and Figure 1 The content of step 104 shown is similar, and will not be repeated here to avoid repetition. Please refer to [link / reference needed] for details. Figure 1 The content of step 104 shown.

[0106] In one feasible implementation, to make the correction more accurate, step 205 may include the following steps D1 to D2:

[0107] D1. Determine the dual-frequency index compensation value based on the target correction coefficient, the first dual-frequency index, and the preset compensation algorithm;

[0108] For example, the compensation algorithm includes the following mathematical expression (2):

[0109] BIS b = (1-K)*log 10 (BIS 2 (2)

[0110] In the formula, BIS is the first dual-frequency index, K is the target correction coefficient, and BIS b The dual-frequency index compensation value is then obtained by inputting the target correction coefficient and the first dual-frequency index into the compensation algorithm. As shown in (2), the closer the K value is to 1, the closer the current BIS index value is to the judgment result of each sub-feature parameter; the closer it is to 0, the opposite is true. This produces different correction effects for different judgment results, avoiding excessive or insufficient correction.

[0111] D2. The sum of the dual-frequency index compensation value and the first dual-frequency index is taken as the target dual-frequency index.

[0112] For example, the formula for calculating the target dual-frequency index includes the following mathematical expression (3):

[0113] BIS f =BIS + (1-K)*log 10 (BIS 2 (3)

[0114] In the formula, BIS f The target dual-frequency index is (1-K)*log(BIS), where BIS is the first dual-frequency index. 10 (BIS2 Let be the dual-frequency index compensation value, and K be the target correction coefficient. Then, adding the dual-frequency index compensation value to the first dual-frequency index yields the target dual-frequency index.

[0115] Understandably, after completing the correction of the bifrequency index, the target bifrequency index and the first-generation value need to be saved to the preset historical database so that the data in the historical database can be updated. Furthermore, in order to achieve real-time correction of the bifrequency index, it is also necessary to return to the step of obtaining the characteristic data of the target user's EEG signal in the preset first time period and continue to perform correction processing.

[0116] This invention provides a method for correcting the bispectral index (BIS) of an electroencephalogram (EEG) signal. The method includes: acquiring feature data of an EEG signal from a target user during a preset first time period, the feature data including at least first feature values ​​of multiple sub-features of the EEG signal and a first bispectral index, the sub-features reflecting the degree of inhibition of the EEG signal; determining first-generation values ​​corresponding to various sub-features based on the first feature values ​​and preset feature thresholds, the algebraic values ​​reflecting whether the bispectral index is valid; acquiring a second bispectral index corresponding to a preset second time period prior to the first time period and second-generation values ​​corresponding to various sub-features from a preset historical database, the historical database including the correspondence between the target user's historical bispectral index, sub-features, algebraic values, and time; determining a target correction coefficient using a correction coefficient algorithm, the first bispectral index, the second bispectral index, various first-generation values, and second-generation values; and performing correction processing based on the target correction coefficient and the first bispectral index to obtain the target bispectral index. This method can use the target correction coefficient to correct the first bispectral index, improving the accuracy of determining the BIS index and reducing errors caused by insufficient robustness. By combining hybrid rice algorithms and BIS index sub-features, the robustness of the BIS index model can be improved by assisting the BIS index in monitoring the depth of anesthesia.

[0117] Please see Figure 3 , Figure 3 This is a structural block diagram of a dual-frequency index correction device for electroencephalogram (EEG) signals according to an embodiment of the present invention. Figure 3 The apparatus shown includes:

[0118] Data acquisition module 301: used to acquire feature data of the target user's electroencephalogram (EEG) signal during a preset first time period. The feature data includes at least the first feature value of multiple sub-features of the EEG signal and a first bi-frequency index. The sub-features are used to reflect the degree of inhibition of the EEG signal.

[0119] Algebraic determination module 302: is used to determine the first algebraic values ​​corresponding to various sub-features based on the first feature value and the preset feature threshold, wherein the algebraic values ​​are used to reflect whether the dual-frequency index is a valid value;

[0120] Coefficient determination module 303: used to determine the target correction coefficient using the first generation value, the first dual-frequency index and the preset correction coefficient algorithm;

[0121] Index correction module 304: used to perform correction processing based on the target correction coefficient and the first dual-frequency index to obtain the target dual-frequency index.

[0122] This invention provides a device for correcting the bispectral index (BIS) of an electroencephalogram (EEG) signal. The device includes: a data acquisition module for acquiring feature data of an EEG signal from a target user during a preset first time period. The feature data includes at least first feature values ​​of multiple sub-features of the EEG signal and a first bispectral index, where the sub-features reflect the degree of inhibition of the EEG signal; an algebraic determination module for determining first algebraic values ​​corresponding to various sub-features based on the first feature values ​​and a preset feature threshold, where the algebraic values ​​reflect whether the bispectral index is valid; a coefficient determination module for determining a target correction coefficient using the first algebraic values, the first bispectral index, and a preset correction coefficient algorithm; and an index correction module for performing correction processing based on the target correction coefficient and the first bispectral index to obtain the target bispectral index. This device can correct the first bispectral index using the target correction coefficient, improving the accuracy of the BIS index determination and reducing errors caused by insufficient robustness.

[0123] Figure 4 An internal structural diagram of a computer device in one embodiment is shown. This computer device can specifically be a terminal or a server. Figure 4 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program, which, when executed by the processor, causes the processor to perform the aforementioned methods. The internal memory may also store a computer program, which, when executed by the processor, causes the processor to perform the aforementioned methods. Those skilled in the art will understand that… Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0124] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform actions such as... Figure 1 or Figure 2The steps of the method shown.

[0125] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, causes the processor to perform the following actions: Figure 1 or Figure 2 The steps of the method shown.

[0126] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0127] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0128] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method of modifying a bispectral index of an electroencephalogram signal, characterized by, The method includes: Acquire feature data of the target user's electroencephalogram (EEG) signal during a preset first time period. The feature data includes at least the first feature value of multiple sub-features of the EEG signal and a first bi-frequency index. The sub-features are used to reflect the degree of inhibition of the EEG signal. Based on the first feature value and the preset feature threshold, the first generation value corresponding to each sub-feature is determined. The generation value is used to reflect whether the dual-frequency index is a valid value. The target correction coefficient is determined using the first generation numerical value, the first dual-frequency index, and a preset correction coefficient algorithm. The target dual-frequency index is obtained by performing a correction process based on the target correction coefficient and the first dual-frequency index. The feature thresholds include the optimal feature thresholds for various sub-features within different segmented intervals of dual-frequency indices. The sub-features include at least burst suppression ratio, QUAZI suppression index, β ratio, and synchronization speed-to-slow ratio. The step of determining the first-generation values ​​corresponding to various sub-features based on the first feature value and the preset feature thresholds includes: Based on the first dual-frequency index, a target feature threshold for the second segment interval where the first dual-frequency index is located is determined from a preset feature threshold. The target feature threshold includes the optimal feature threshold for various sub-features in the second segment interval. The first eigenvalues ​​of the burst suppression ratio, QUAZI suppression index, β ratio, and synchronization speed ratio are compared with their respective target feature thresholds to determine the comparison results of various sub-features. If the comparison result is that the first feature value is greater than the target feature threshold, then the first generation value of the sub-feature is determined to be 1; If the comparison result is that the first feature value is less than or equal to the target feature threshold, then the first generation value of the sub-feature is determined to be 0.

2. The method of claim 1, wherein, The step of determining the target correction coefficient using the first generation numerical value, the first dual-frequency index, and a preset correction coefficient algorithm includes: The second dual-frequency index corresponding to the second time period before the first time period and the second generation value corresponding to various sub-features are obtained from the preset historical database. The historical database includes the correspondence between the target user's historical dual-frequency index, sub-features, generation values ​​and time. The target correction coefficient is determined using the correction coefficient algorithm, the first dual-frequency index, the second dual-frequency index, various first-generation values, and second-generation values.

3. The method of claim 2, wherein, The correction coefficient algorithm includes a weighted average algorithm that corresponds one-to-one with each segment interval of the dual-frequency index. The step of determining the target correction coefficient using the correction coefficient algorithm, the first dual-frequency index, the second dual-frequency index, various first-generation values, and second-generation values ​​includes: The mean of the dual-frequency index is determined by averaging the first dual-frequency index and the second dual-frequency index. The algebraic mean of various sub-features is determined by calculating the mean of the first-generation and second-generation values. Based on the dual-frequency exponential mean, a target weighted average algorithm is selected from the correction coefficient algorithm to identify the first segment interval where the dual-frequency exponential mean is located; The algebraic mean is input into the target weighted average algorithm to determine the target correction coefficient.

4. The method of claim 1, wherein, The step of performing correction processing based on the target correction coefficient and the first dual-frequency index to obtain the target dual-frequency index includes: The dual-frequency index compensation value is determined based on the target correction coefficient, the first dual-frequency index, and the preset compensation algorithm. The sum of the dual-frequency index compensation value and the first dual-frequency index is taken as the target dual-frequency index.

5. The method of claim 1, wherein, The step of determining the first-generation values ​​corresponding to various sub-features based on the first feature value and a preset feature threshold also includes: Obtain sample feature data for each segmented interval, wherein the sample feature data includes the correspondence between several sample dual-frequency indices and sample feature values ​​of various sub-features for each segmented interval; The sample feature values ​​are input as breeding individuals into a preset hybrid rice model for breeding processing until the maximum number of iterations is reached, to obtain the best individual for the sample feature values ​​of various sub-features under each segment interval output by the hybrid rice model; and the best individual is used as the best feature threshold.

6. The method of claim 2, wherein, The method further includes: The target bifrequency index and the first-generation value are saved to a preset historical database, and the process returns to the step of obtaining the characteristic data of the target user's EEG signal in a preset first time period.

7. A device for modifying a bispectral index of an electroencephalogram signal, characterized in that The device includes: Data acquisition module: used to acquire feature data of the target user's electroencephalogram (EEG) signal during a preset first time period. The feature data includes at least the first feature value of multiple sub-features of the EEG signal and a first bi-frequency index. The sub-features are used to reflect the degree of inhibition of the EEG signal. Algebraic determination module: used to determine the first algebraic values ​​corresponding to various sub-features based on the first feature value and the preset feature threshold, wherein the algebraic values ​​are used to reflect whether the dual-frequency index is a valid value; Coefficient determination module: used to determine the target correction coefficient using the first generation numerical value, the first dual-frequency index, and a preset correction coefficient algorithm; Index correction module: used to perform correction processing based on the target correction coefficient and the first dual-frequency index to obtain the target dual-frequency index; The feature thresholds include optimal feature thresholds for various sub-features within different segmented intervals of dual-frequency indices. The sub-features include at least burst suppression ratio (PSR), QUAZI suppression index (QUIZI), β ratio, and synchronization speed-to-slow ratio (SSR). The algebraic determination module is specifically used for: determining a target feature threshold for the second segmented interval of the first dual-frequency index from preset feature thresholds, based on the first dual-frequency index. The target feature threshold includes optimal feature thresholds for various sub-features within the second segmented interval. The module compares the first feature values ​​of the burst suppression ratio, QUIZI suppression index, β ratio, and SSR with their respective target feature thresholds to determine the comparison results for each sub-feature. If the comparison result is that the first feature value is greater than the target feature threshold, the first algebraic value of the sub-feature is determined to be 1. If the comparison result is that the first feature value is less than or equal to the target feature threshold, the first algebraic value of the sub-feature is determined to be 0.

8. A computer readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it causes the processor to perform the steps of the method as described in any one of claims 1 to 6.

9. A computer device, comprising a memory and a processor, characterized in that, The memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of the method as described in any one of claims 1 to 6.

Images