High-precision physiological parameter monitoring method for transcranial electrical stimulation process

Abstract

The invention relates to the technical field of transcranial electrical stimulation, in particular to a high-precision physiological parameter monitoring method for a transcranial electrical stimulation process. By acquiring comprehensive data of historical patients, including time sequence data of medical history, transcranial electrical stimulation parameters and physiological parameters, the group characteristics of the patients can be accurately identified. And clustering the historical samples according to the difference of the historical electrical stimulation parameters by using a clustering technology to ensure the consistency of the electrical stimulation parameters in the cluster. And further analyzing fluctuation similarity and medical history similarity of physiological parameter time sequence data in the cluster, determining confidence of the electrical stimulation parameter, and reflecting consistency of the physiological parameter to electrical stimulation response. According to the confidence coefficient and the current medical history data of the to-be-treated patient, classifying the to-be-treated patient into target clusters of similar illness states. Finally, the electrical stimulation parameters of the patient to be treated are set based on the historical data of the target cluster, the integrating degree of the parameters and the illness state is improved, and therefore the credibility of monitoring the physiological parameters in real time is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of transcranial electrical stimulation, and in particular to a high-precision physiological parameter monitoring method for a transcranial electrical stimulation process. Background Art

[0002] In modern life, due to the accelerated pace of work and increased competitive pressure, depression is a common mental illness that has a serious impact on the patient's life and mental health; transcranial pulse electrical stimulation technology, as a non-invasive neuromodulation technology, has been widely used. Transcranial pulse electrical stimulation technology applies current pulses to specific areas of the brain to regulate neural activity, thereby achieving the purpose of treatment. In order to monitor the patient's physiological condition in real time, artificial intelligence is usually combined with transcranial pulse electrical stimulation technology in the modern medical field. The patient's physiological parameters are monitored during the transcranial pulse electrical stimulation process to improve the therapeutic effect and safety of transcranial pulse electrical stimulation technology.

[0003] The therapeutic effect and safety of transcranial pulse electrical stimulation technology are highly dependent on the precise setting of electrical stimulation parameters, which affect the changes in the patient's physiological parameters. When monitoring the patient's physiological parameter data, the existing technology often uses the recommended parameters of clinical trials at the beginning of transcranial pulse electrical stimulation, and then adjusts it according to the user's feelings during the process. However, the recommended parameters lack sufficient consideration of individual differences among patients, which may lead to their inability to match the patient's own situation. This mismatch may increase the patient's physiological burden, thereby affecting the credibility of the monitoring results of the patient's physiological parameters. Summary of the invention

[0004] In order to solve the technical problem that the recommended parameters cannot match the patient's own conditions due to lack of sufficient consideration of individual differences of patients, and this mismatch may increase the patient's physiological burden, thereby affecting the credibility of the monitoring results of the patient's physiological parameters, the purpose of the present invention is to provide a high-precision physiological parameter monitoring method for transcranial electrical stimulation process, and the technical scheme adopted is as follows:

[0005] Acquire historical sample data of historical patients, and each historical sample data includes historical medical history data, various historical electrical stimulation parameter values ​​during transcranial pulse electrical stimulation, and various historical physiological parameter time series data;

[0006] Comprehensively analyzing the differences between various historical electrical stimulation parameter values ​​of historical sample data, performing cluster analysis on all historical sample data, and obtaining initial clusters;

[0007] In each initial cluster, the fluctuation similarity of various historical physiological parameter time series data of the historical sample data is analyzed, and the similarity of the historical medical history data between different historical sample data is quantified, so as to determine the confidence of the historical electrical stimulation parameter value in each initial cluster;

[0008] Analyze the confidence levels corresponding to all initial clusters, and determine the target cluster corresponding to the current medical history data based on the similarities between the current medical history data and the historical medical history data of the patient to be treated; based on the numerical characteristics of the historical sample data in the target cluster corresponding to the current medical history data, determine the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process and monitor the real-time physiological parameter data.

[0009] Furthermore, the method for obtaining the initial clustering clusters includes:

[0010] Comprehensively analyzing the differences between various historical electrical stimulation parameter values ​​of the historical sample data, and determining similarity values ​​of the electrical stimulation parameters between any two historical sample data;

[0011] The optimal K value of all historical sample data is obtained based on the silhouette coefficient method, the similarity value of the electrical stimulation parameters between any two historical sample data is used as the distance metric, and all historical sample data are clustered based on the K-means clustering algorithm and the optimal K value to obtain an initial cluster cluster.

[0012] Furthermore, the method for obtaining the similarity values ​​of the electrical stimulation parameters includes:

[0013] Analyze the difference in consistency between various historical electrical stimulation parameter values ​​and preset electrical stimulation parameter values ​​in any two historical sample data, and determine the consistency deviation factor between any two historical sample data;

[0014] Arrange various historical electrical stimulation parameter values ​​in each historical sample data in the same arrangement mode to obtain an electrical stimulation parameter value vector corresponding to each historical sample data, and calculate the cosine similarity of the electrical stimulation parameter value vectors of any two historical sample data;

[0015] According to the consistency deviation factor and cosine similarity between any two historical sample data, the electrical stimulation parameter similarity value between any two historical sample data is obtained, and the consistency deviation factor is negatively correlated with the electrical stimulation parameter similarity value, and the cosine similarity is positively correlated with the electrical stimulation parameter similarity value.

[0016] Furthermore, the method for obtaining the consistency deviation factor includes:

[0017] Compare each historical electrical stimulation parameter value in each historical sample data with the corresponding preset electrical stimulation parameter value, and count the number of historical electrical stimulation parameter values ​​that are the same as the corresponding preset electrical stimulation parameter value as the consistency parameter of each historical sample data;

[0018] The ratio of the consistency parameter to the number of types of historical electrical stimulation parameter values ​​is used as the electrical stimulation parameter consistency score corresponding to each historical sample data;

[0019] The absolute value of the difference between the electrical stimulation parameter consistency scores of any two historical sample data is used as the consistency deviation factor between any two historical sample data.

[0020] Furthermore, the method for obtaining the confidence level includes:

[0021] In each initial cluster, all historical sample data are combined in pairs to obtain all non-repeated sample data combinations;

[0022] Analyze the fluctuation similarity of various historical physiological parameter time series data of two historical sample data in each sample data combination, and determine the fluctuation similarity index corresponding to each sample data combination;

[0023] Based on the Jieba word segmentation tool, the historical medical history data of all historical sample data are segmented, and the word segmentation array corresponding to each historical sample data is obtained; the Jaccard coefficient between the word segmentation arrays of two historical sample data in each sample data combination is calculated as the similarity coefficient;

[0024] In each initial cluster, the fluctuation similarity index corresponding to the sample data combination is weighted by using the similarity coefficient between the word arrays of two historical sample data in the sample data combination, so as to obtain the confidence of the historical electrical stimulation parameter value in each initial cluster.

[0025] Furthermore, the method for obtaining the fluctuation similarity index includes:

[0026] For any sample data combination, calculate the DTW value between the same historical physiological parameter time series data of two historical sample data;

[0027] Analyze the similarity of the fluctuation range of the same historical physiological time series data in the two historical sample data in the sample data combination, and determine the distance factor of the two historical sample data under the same historical physiological time series data;

[0028] The product of the DTW value and the distance factor of the two historical sample data in the sample data combination under the same historical physiological parameter time series data is negatively correlated and normalized to obtain the fluctuation similarity factor of the two historical sample data under the same historical physiological parameter time series data;

[0029] The value after normalizing the mean value of the fluctuation similarity factor of the two historical sample data in the sample data combination under all types of historical physiological time series data is used as the fluctuation similarity index corresponding to the sample data combination.

[0030] Furthermore, the method for obtaining the distance factor includes:

[0031] For any sample data combination, the maximum value and the minimum value of each type of historical physiological time series data in each historical sample data in the sample data combination are combined into a maximum value array;

[0032] The Euclidean distance between the maximum value arrays of two historical sample data under the same historical physiological parameter time series data is used as the distance factor of the two historical sample data under the same historical physiological time series data in the sample data combination.

[0033] Furthermore, the analysis of the confidences corresponding to all initial clusters and the determination of the target cluster corresponding to the current medical history data in combination with the similarity between the current medical history data and the historical medical history data of the patient to be treated include:

[0034] The initial clusters with confidence greater than the preset confidence threshold are taken as candidate clusters;

[0035] The current medical history data of the patients to be treated are segmented based on the Jieba word segmentation tool to obtain the word segmentation array corresponding to the current medical history data;

[0036] Calculate the Jaccard coefficient between the word group corresponding to the current medical history data and the word array corresponding to each historical sample data in all candidate clusters as the similarity parameter;

[0037] The candidate cluster to which the historical sample data corresponding to the maximum value of all similar parameters of the current medical history data belongs is taken as the target cluster corresponding to the current medical history data.

[0038] Furthermore, the method of determining the electrical stimulation parameter value of the patient to be treated during the transcranial pulse electrical stimulation process and monitoring the real-time physiological parameter data based on the numerical characteristics of the historical sample data in the target cluster corresponding to the current medical history data includes:

[0039] In the target cluster, the mean of the same historical electrical stimulation parameter values ​​of all historical sample data is calculated as the standard value of each electrical stimulation parameter in the target cluster;

[0040] When the patient to be treated is subjected to transcranial pulse electrical stimulation, each electrical stimulation parameter is set to a standard value in the target cluster, and various real-time physiological parameter data of the patient to be treated during the transcranial pulse electrical stimulation process are obtained;

[0041] Monitor various real-time physiological parameter data and obtain monitoring results.

[0042] Furthermore, the monitoring of various real-time physiological parameter data to obtain monitoring results includes:

[0043] In the target cluster, the maximum and minimum values ​​of the same historical physiological parameter time series data of all historical sample data are obtained to form the fluctuation range of each physiological parameter;

[0044] For any physiological parameter, if the real-time physiological parameter data of the patient to be treated under the physiological parameter does not exist in the corresponding fluctuation range, it is considered that the physiological parameter of the patient to be treated is abnormal and an early warning is required; otherwise, it is considered normal and no early warning is required.

[0045] The present invention has the following beneficial effects:

[0046] Obtain historical sample data of historical patients, which includes historical medical history data, various historical electrical stimulation parameter values ​​during transcranial electrical stimulation, and various historical physiological parameter time series data; Obtaining comprehensive data of historical patients can more accurately identify the characteristics of different patient groups in the subsequent process, so as to accurately regulate the electrical stimulation parameters during the transcranial pulse electrical stimulation of the patients to be treated and achieve more accurate physiological parameter monitoring. With the help of clustering technology, cluster analysis is performed on the historical sample data based on the differences between various historical electrical stimulation parameter values ​​of the historical sample data, so as to form an initial cluster cluster. At this time, the historical electrical stimulation parameter values ​​used by the historical sample data in each initial cluster cluster have strong consistency. Then, in each initial cluster cluster, analyze the fluctuation similarity of various historical physiological parameter time series data of the historical sample data, as well as the similarity of the historical medical history data, and determine the confidence of the historical electrical stimulation parameter value in each initial cluster cluster. The confidence characterizes the consistency of the response of the physiological parameter changes of the historical sample data in each initial cluster cluster to the electrical stimulation parameter value, implying the effectiveness and reliability of the electrical stimulation parameter value. Then, according to the confidence corresponding to the initial cluster, combined with the similarity between the current medical history data of the patient to be treated and the historical medical history data, the current medical history data can be classified to determine the target cluster to which it belongs, that is, the initial cluster with a condition similar to that of the patient to be treated. Finally, based on the numerical characteristics of the historical sample data in the target cluster, the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process are set. The electrical stimulation parameter values ​​set in this way are more consistent with the current medical history data. Therefore, when the real-time physiological parameter data of the patient to be treated is monitored on this basis, the credibility of the monitoring results obtained will also be higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A method flow chart of a method for high-precision physiological parameter monitoring during transcranial electrical stimulation provided by one embodiment of the present invention;

[0049] Figure 2 A method flow chart of a method for obtaining an initial cluster provided by an embodiment of the present invention;

[0050] Figure 3 A flow chart of a method for obtaining confidence provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of a high-precision physiological parameter monitoring method for transcranial electrical stimulation process proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0052] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0053] The following is a detailed description of a method for high-precision physiological parameter monitoring during transcranial electrical stimulation provided by the present invention in conjunction with the accompanying drawings.

[0054] See also Figure 1 , which shows a method flow chart of a method for high-precision physiological parameter monitoring during transcranial electrical stimulation provided by an embodiment of the present invention, the method comprising the following steps:

[0055] Step S1: Acquire historical sample data of historical patients, and each historical sample data includes historical medical history data, various historical electrical stimulation parameter values ​​during the transcranial pulse electrical stimulation process, and various historical physiological parameter time series data.

[0056] Transcranial Pulsed Electrical Stimulation (TPES) is a non-invasive neuromodulation technology that has received extensive attention and application in the fields of neuroscience, mental health, and rehabilitation medicine in recent years. This technology modulates neural activity by applying current pulses to specific areas of the brain. However, the effectiveness and safety of TPES are highly dependent on the precise setting of electrical stimulation parameters.

[0057] In actual applications, due to individual differences, different patients will need different electrical stimulation parameter values ​​when undergoing transcranial pulse electrical stimulation. Usually, initially, the electrical stimulation parameter values ​​will often use the recommended parameters of clinical trials (preset electrical stimulation parameter values). When the patient feels unwell, the recommended parameters need to be adjusted to find electrical stimulation parameters that are more suitable for the patient's own situation. Inappropriate electrical stimulation parameters will increase the patient's physiological burden and have a greater impact on the credibility of the patient's physiological parameter monitoring results. Therefore, in an embodiment of the present invention, in order to ensure the credibility of the physiological parameter monitoring results, when the patient undergoes transcranial pulse electrical stimulation, it is necessary to set the electrical stimulation parameter value that best matches his or her own situation.

[0058] The embodiment of the present invention is based on artificial intelligence technology to analyze the historical sample data of historical patients in the hospital database to determine the electrical stimulation parameters that best match the patient's own condition to be treated. Each historical sample data should include various historical medical history data, various historical electrical stimulation parameter values ​​during transcranial pulse electrical stimulation, and various historical physiological parameter time series data. Specifically, in the embodiment of the present invention, the types of electrical stimulation parameters are set to stimulation intensity, stimulation frequency, pulse width, etc., the types of physiological parameters are set to blood pressure, blood oxygen, body temperature, etc., and the medical history data includes information such as the patient's medication and condition.

[0059] It should be noted that the types of electrical stimulation parameters, the types of physiological parameters, and the types of information contained in the medical history data can be adjusted according to the implementation scenario, and are not limited here. Since the initial recommended parameters may increase the patient's physiological burden when the patient receives transcranial pulse electrical stimulation, that is, the physiological parameters are abnormal, the doctor will adjust the initial recommended parameters. Therefore, the time period of various historical physiological parameter time series data obtained in the embodiment of the present invention is selected from the time period when the patient receives transcranial pulse electrical stimulation. The various historical physiological parameters are within the normal value range, and the historical electrical stimulation parameter value selects the electrical stimulation parameter value used in the historical physiological parameter time series data period, where the normal value range can be set according to the normal physiological parameters of the human body, for example, the normal value range corresponding to body temperature is 36°C-37.5°C, the normal value range corresponding to blood oxygen is 95%-100%, and the normal value range corresponding to blood pressure is systolic pressure: 90mmHg-139mmHg, diastolic pressure: 60mmHg-89mmHg.

[0060] Step S2: Comprehensively analyze the differences between various historical electrical stimulation parameter values ​​of the historical sample data, perform cluster analysis on all the historical sample data, and obtain initial clusters.

[0061] By comprehensively analyzing the historical sample data, especially focusing on the differences between various historical electrical stimulation parameter values, we can find the impact patterns of different electrical stimulation parameter values ​​on the patient's physiological response. Cluster analysis can further present these patterns in the form of initial clusters, thereby revealing the response characteristics of different patient groups to different electrical stimulation parameter values, so that historical sample data with similar historical electrical stimulation parameter values ​​are clustered into one category, providing a reference for the setting of electrical stimulation parameters for subsequent patients to be treated.

[0062] Preferably, in one embodiment of the present invention, the method for obtaining the initial clustering clusters includes:

[0063] See also Figure 2 , which shows a method flow chart of a method for obtaining an initial cluster in one embodiment of the present invention, the method comprising the following steps:

[0064] Step S201: comprehensively analyzing the differences between various historical electrical stimulation parameter values ​​of historical sample data, and determining similarity values ​​of electrical stimulation parameters between any two historical sample data.

[0065] Since the preset electrical stimulation parameter values ​​(the recommended parameters described in step S1) may cause a burden on the patient's physiological condition during the transcranial pulse electrical stimulation process, the doctor will adjust the preset electrical stimulation parameter values. Therefore, when analyzing the differences between the historical electrical stimulation parameter values ​​of the historical sample data, the differences in consistency between the various historical electrical stimulation parameter values ​​and the preset electrical stimulation parameter values ​​in any two historical sample data can be used as a measurement standard to determine the consistency deviation factor between any two historical sample data.

[0066] Specifically, each historical electrical stimulation parameter value in each historical sample data can be compared with the preset electrical stimulation parameter value corresponding to the same electrical stimulation parameter, and the number of historical electrical stimulation parameter values ​​that are identical to the corresponding preset electrical stimulation parameter values ​​can be counted as the consistency parameter of each historical sample data. Then, the ratio of the consistency parameter to the number of types of historical electrical stimulation parameter values ​​can be used as the electrical stimulation parameter consistency score corresponding to each historical sample data. The larger the consistency parameter, the larger the electrical stimulation parameter consistency score, indicating that a larger number of historical electrical stimulation parameter values ​​in the historical sample data have the same values ​​as the preset electrical stimulation parameter values.

[0067] It should be noted that the settings of the preset electrical stimulation parameter values ​​can be adjusted according to the implementation scenario. Specifically, for example, the stimulation intensity can be set to 0.6 mA, the stimulation frequency can be set to 50 Hz, the pulse width can be set to 80 microseconds, etc.

[0068] Analyze the difference in the consistency scores of the electrical stimulation parameters of any two historical sample data: take the absolute value of the difference in the consistency scores of the electrical stimulation parameters of any two historical sample data as the consistency deviation factor between any two historical sample data. The larger the value of the consistency deviation factor, the lower the similarity between the two historical sample data in terms of whether the historical electrical stimulation parameter values ​​have been adjusted.

[0069] In view of the fact that the historical electrical stimulation parameter values ​​between two historical sample data can more intuitively reflect similar situations, the various historical electrical stimulation parameter values ​​in each historical sample data are arranged in the same type arrangement method to obtain the electrical stimulation parameter value vector corresponding to each historical sample data, and the cosine similarity of the electrical stimulation parameter value vectors of any two historical sample data is calculated. At this time, the greater the cosine similarity, the higher the similarity of the electrical stimulation parameter values ​​of the two historical sample data.

[0070] Based on the above analysis, it can be seen that the larger the value of the consistency deviation factor, the lower the similarity of the electrical stimulation parameter values ​​of the two historical sample data; the larger the cosine similarity, the higher the similarity of the electrical stimulation parameter values ​​of the two historical sample data. Therefore, the electrical stimulation parameter similarity value between any two historical sample data is negatively correlated with the consistency deviation factor, and the electrical stimulation parameter similarity value is positively correlated with the cosine similarity. The formula model for the electrical stimulation parameter similarity value between any two historical sample data includes:

[0071]

[0072] Among them, DX represents the similarity value of the electrical stimulation parameters between any two historical sample data; θ represents the cosine similarity corresponding to the historical electrical stimulation parameter values ​​of any two historical sample data; YC represents the consistency deviation factor corresponding to any two historical sample data; γ1 represents the preset first parameter; norm() represents the normalization function.

[0073] In the formula model of the similarity value of the electrical stimulation parameters, since the value range of the cosine similarity is [-1, 1], the closer it is to 1, the more similar the two vectors are. In order to facilitate subsequent calculations, 1+ is used here to adjust the cosine similarity corresponding to the historical electrical stimulation parameter values ​​of any two historical sample data, so that the value range of θ+1 is [0, 2], and the closer it is to 2, the higher the similarity of the historical electrical stimulation parameter values ​​of the two historical sample data; the larger the value of the consistency deviation factor, the lower the similarity of the electrical stimulation parameter values ​​of the two historical sample data, so the consistency deviation factor is used as the denominator, θ+1 is used as the numerator, and the comparison value is normalized to obtain the similarity value of the electrical stimulation parameters between any two historical sample data. At this time, the larger the similarity value of the electrical stimulation parameters, the higher the similarity of the historical electrical stimulation parameter values ​​between the two historical sample data.

[0074] It should be noted that the purpose of presetting the first parameter γ1 is to prevent the denominator from being 0, and the value may be 0.001. The specific value may be adjusted according to the implementation scenario and is not limited here.

[0075] Step S202: performing cluster analysis on all historical sample data based on similarity values ​​of electrical stimulation parameters between historical sample data, thereby obtaining initial clusters.

[0076] Based on the silhouette coefficient method, the optimal K value of all historical sample data is obtained, and then the similarity value of the electrical stimulation parameters between any two historical sample data is used as the distance metric. Based on the K-means clustering algorithm and the optimal K value, all the historical sample data are clustered and analyzed to obtain the initial clustering clusters. At this time, the historical sample data in each initial clustering cluster have relatively consistent historical electrical stimulation parameter values.

[0077] It should be noted that the silhouette coefficient method and the K-means clustering algorithm are both well-known technologies and will not be described in detail here.

[0078] At this point, all historical sample data can be clustered based on historical electrical stimulation parameter values ​​to obtain all initial clusters.

[0079] Step S3: In each initial cluster, analyze the fluctuation similarity of various historical physiological parameter time series data of the historical sample data, and quantify the similarity of historical medical history data between different historical sample data, so as to determine the confidence of the historical electrical stimulation parameter value in each initial cluster.

[0080] Based on the process in step S2, an initial cluster consisting of historical sample data with similar historical electrical stimulation parameter values ​​can be obtained. Since various physiological parameters of the patient fluctuate in real time during the transcranial pulse electrical stimulation process, for each initial cluster, the fluctuation similarity of the historical physiological parameter time series data of the historical sample data is analyzed to determine whether the response of the physiological parameter change to the electrical stimulation parameter value is consistent, which can imply the effectiveness and reliability of the historical electrical stimulation parameters corresponding to the historical sample data in the initial cluster; at the same time, due to individual differences, if the similarity of the disease characteristics reflected by the historical sample data in the initial cluster is higher, it means that the historical electrical stimulation parameters corresponding to the historical sample data in the initial cluster perform better for a certain type of medical history data, which also means that the better the confidence, the more effective reference can be provided for the subsequent determination of the electrical stimulation parameters of the patient to be treated. Therefore, in each initial cluster, the similarity of the fluctuation of the historical physiological parameter time series data between the historical sample data and the similarity of the historical medical history data are analyzed to determine the confidence of the historical electrical stimulation parameter values ​​in each initial cluster.

[0081] Preferably, in one embodiment of the present invention, the method for obtaining the confidence level includes:

[0082] See also Figure 3 , which shows a method flow chart of a method for obtaining confidence in one embodiment of the present invention, the method comprising the following steps:

[0083] Step S301: In each initial cluster, all historical sample data are combined in pairs to obtain all non-repetitive sample data combinations.

[0084] Based on the above analysis, it can be seen that the confidence corresponding to each initial cluster is analyzed based on the similarity of fluctuations in historical physiological parameter time series data between historical sample data and the similarity of historical medical history data. Therefore, we can first combine all historical sample data in pairs in each initial cluster to obtain all non-repetitive sample data combinations, which is convenient for analyzing the similarities between sample data in subsequent processes.

[0085] The process of obtaining sample data combinations is illustrated by an example: for example, there are 4 historical sample data in an initial cluster, which are recorded as historical sample data 1-4 respectively. Then the sample data combinations in the initial cluster are (historical sample data 1, historical sample data 2), (historical sample data 1, historical sample data 3), (historical sample data 1, historical sample data 4), (historical sample data 2, historical sample data 3), (historical sample data 2, historical sample data 4), and (historical sample data 3, historical sample data 4).

[0086] Step S302: Analyze the fluctuation similarity of various historical physiological parameter time series data of two historical sample data in each sample data combination, and determine the fluctuation similarity index corresponding to each sample data combination.

[0087] Dynamic Time Warping (DTW) is an algorithm used to measure the similarity between two time series. It can handle the expansion and contraction of time series on the time axis, so as to more accurately evaluate the fluctuation similarity of two historical sample data in physiological parameter time series data. Therefore, for any sample data combination, the DTW value between the same historical physiological parameter time series data of two historical sample data is calculated. The smaller the DTW value, the higher the similarity between the two historical sample data under the same historical physiological parameter time series data.

[0088] The maximum value in a group of data represents the fluctuation range of a group of data. When the maximum values ​​of the historical physiological parameter time series data of two historical sample data are also relatively close, it can also be said that there is a high degree of similarity between the two. Therefore, the similarity of the fluctuation range of the same historical physiological time series data in the two historical sample data in the sample data combination is analyzed to determine the distance factor of the two historical sample data under the same historical physiological time series data: for any sample data combination, the maximum value array is composed of the maximum value and the minimum value of each historical physiological time series data in each historical sample data in the sample data combination. At this time, each historical sample data in the sample data combination has a maximum value array under each historical physiological time series data. The Euclidean distance between the maximum value arrays of the two historical sample data under the same historical physiological parameter time series data is used as the distance factor of the two historical sample data under the same historical physiological time series data in the sample data combination. The larger the distance factor, the higher the similarity of the fluctuation range of the two historical sample data under the same historical physiological parameter time series data.

[0089] Based on the above analysis, it can be known that for any sample data combination, the DTW value and distance factor of two historical sample data under the same historical physiological parameter time series data are negatively correlated with the similarity between the two historical sample data. Therefore, the product of the DTW value and the distance factor of the two historical sample data in the sample data combination under the same historical physiological parameter time series data can be negatively correlated and normalized to achieve logical relationship correction, thereby obtaining the fluctuation similarity factor of the two historical sample data under the same historical physiological parameter time series data. The larger the fluctuation similarity factor, the higher the fluctuation similarity of the two historical sample data under the same historical physiological parameter time series data.

[0090] It should be noted that the negative correlation mapping and normalization processing here can use exp(-x), where exp() represents an exponential function with the natural constant e as the base, and x represents the independent variable; the method for obtaining the DTW value is a well-known technology, and the process is not repeated here.

[0091] Finally, the fluctuation similarity factors of the two historical sample data under all types of historical physiological parameter time series data are combined: the mean value of the fluctuation similarity factors of the two historical sample data in the sample data combination under all types of historical physiological time series data is normalized and used as the fluctuation similarity index corresponding to the sample data combination. At this time, the larger the fluctuation similarity index is, the higher the degree of fluctuation similarity of the historical physiological parameter time series data of the two historical sample data in the sample data combination is, which can be regarded as the more consistent degree of influence of the historical electrical stimulation parameters corresponding to the initial clustering cluster to which they belong on their historical physiological parameter time series data. Normalization is a technical means well known to technicians in this field, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0092] Step S303: Based on the similarity between the historical medical history data of two historical sample data in each sample data combination, determine the similarity coefficient corresponding to each sample data combination.

[0093] Due to individual differences, the historical medical data of different historical sample data will be different. Therefore, during the transcranial pulse electrical stimulation process, the fluctuations of the historical physiological parameter time series data of the historical sample data with large differences in the historical medical data will be different. Therefore, based on the fluctuation similarity index calculated in step S302, it is also necessary to analyze the similarity of the historical medical data of the historical sample data, so that in each initial clustering cluster, the similarity of the historical physiological parameter time series data of the historical sample data with relatively similar historical medical data can be evaluated.

[0094] Jieba word segmentation tool is a Chinese text word segmentation tool that can segment historical medical history data into meaningful vocabulary units. Through analysis, it is more convenient to calculate the similarity between the historical medical history data of two historical sample data: based on Jieba word segmentation tool, the historical medical history data of all historical sample data are segmented, and the word segmentation array corresponding to each historical sample data is obtained; then the Jaccard coefficient between the word segmentation arrays of the two historical sample data in each sample data combination is calculated as the similarity coefficient. The larger the similarity coefficient, the higher the similarity between the word segmentation arrays of the two historical sample data.

[0095] It should be noted that the application process of the Jieba word segmentation tool and the calculation process of the Jaccard coefficient are well-known technologies and will not be described in detail here.

[0096] Step S304: In each initial cluster, the similarity coefficients and fluctuation similarity indexes between two historical sample data in all sample data combinations are comprehensively considered to determine the confidence level of the historical electrical stimulation parameter values ​​in each initial cluster.

[0097] The similarity coefficient between two historical sample data represents the similarity between their historical medical history data, that is, the similarity in the patients' physical health; the fluctuation similarity index between two historical sample data represents the similarity in the changes between their historical physiological parameter time series data, that is, the consistency in response to similar historical electrical stimulation parameters; combining the two can more accurately evaluate the influence of the electrical stimulation parameter values ​​in the initial clustering cluster on the historical physiological parameter time series data of the historical sample data with similar historical medical history data, thereby quantifying the confidence of the historical electrical stimulation parameter values ​​in each initial clustering cluster.

[0098] In each initial cluster, the fluctuation similarity index corresponding to the sample data combination is weighted by using the similarity coefficient between the word arrays of the two historical sample data in the sample data combination, so as to obtain the confidence of the historical electrical stimulation parameter value in each initial cluster. The formula model of confidence includes:

[0099]

[0100] Where XC represents the confidence of the historical electrical stimulation parameter values ​​in each initial cluster; n represents the total number of sample data combinations in each initial cluster; BX i XS represents the fluctuation similarity index corresponding to the i-th sample data combination in each initial cluster; i It represents the similarity coefficient corresponding to the i-th sample data combination in each initial cluster; norm() represents the normalization function.

[0101] In the confidence formula model, the fluctuation similarity index corresponding to each sample data combination is multiplied by the similarity coefficient, which can increase the weight of the sample data combination with high similarity of historical medical history data. The obtained product (BX i ×XS i ) is larger, indicating that the historical medical records of the historical sample data in the sample data combination are similar, and the feedback on the historical electrical stimulation parameter values ​​is also similar, then the confidence of the historical electrical stimulation parameter values ​​in the initial cluster is higher, that is, the reference value in the subsequent process is greater; so finally, all the sample data combinations in each initial cluster are combined with (BX i ×XS i ) is normalized as the confidence of the historical electrical stimulation parameter value in each initial cluster. The greater the confidence, the higher the similarity of the historical sample data in the initial cluster in terms of medical history data and physiological response, which implies that the effectiveness of the setting of the historical electrical stimulation parameter value in the initial cluster is better.

[0102] At this point, the confidence level of the historical electrical stimulation parameter values ​​of the historical sample data in each initial cluster can be obtained.

[0103] Step S4: Analyze the confidence levels corresponding to all initial clusters, and determine the target cluster corresponding to the current medical history data based on the similarities between the current medical history data and the historical medical history data of the patient to be treated; based on the numerical characteristics of the historical sample data in the target cluster corresponding to the current medical history data, determine the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process and monitor the real-time physiological parameter data.

[0104] In the above steps, the confidence of the historical electrical stimulation parameter values ​​of the historical sample data in each initial clustering cluster has been determined. Therefore, when determining the electrical stimulation parameters for transcranial pulse electrical stimulation of the patient to be treated, in order to find the electrical stimulation parameters that are more suitable for the current medical history data of the patient to be treated, the confidence corresponding to all the initial clustering clusters can be first analyzed, and the initial clustering clusters with higher confidence can be selected therefrom. The setting of the historical electrical stimulation parameter values ​​of such initial clustering clusters is often more reliable; at the same time, based on the similarity between the current medical history data (medical history information of the patient to be treated) and the historical medical history data, the target clustering cluster to which the current medical history data belongs is determined. The historical electrical stimulation parameter values ​​in the target clustering cluster are the electrical stimulation parameters that are more consistent with the current medical history data. Therefore, finally, based on the numerical characteristics of the historical sample data in the target clustering cluster corresponding to the current medical history data, the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process can be determined and the real-time physiological parameter data can be monitored, which can effectively improve the credibility of the monitoring results.

[0105] Preferably, in one embodiment of the present invention, the confidences corresponding to all initial clusters are analyzed, and the target cluster corresponding to the current medical history data is determined in combination with the similarity between the current medical history data and the historical medical history data of the patient to be treated, including:

[0106] First, the initial clusters with confidence greater than the preset confidence threshold are selected as candidate clusters. The screening of candidate clusters can ensure that the target clusters determined later are more reliable and have higher reference value. It should be noted that the preset confidence threshold is 0.6, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0107] Then, the current medical history data of the patients to be treated are segmented based on the Jieba segmentation tool to obtain the segmentation array corresponding to the current medical history data. The Jaccard coefficient between the segmentation phrases corresponding to the current medical history data and the segmentation arrays corresponding to each historical sample data in all candidate clusters is calculated as a similarity parameter. The similarity parameter can reflect the degree of overlap between the current medical history data and the historical medical history data of the historical sample data in the candidate clusters, that is, the larger the value of the similarity parameter, the higher the degree of similarity.

[0108] It should be noted that the application process of the Jieba word segmentation tool and the calculation process of the Jaccard coefficient are well-known technologies and will not be described in detail here.

[0109] Finally, since the larger the value of the similarity parameter is, the higher the similarity of the medical history information between the current medical history data and the historical medical history data is, the candidate clustering cluster to which the historical sample data corresponding to the maximum value of all similarity parameters of the current medical history data belongs is used as the target clustering cluster corresponding to the current medical history data. At this time, the historical electrical stimulation parameter values ​​corresponding to the historical sample data in the target clustering cluster have a higher degree of fit with the current medical history data of the patient to be treated. Therefore, in the subsequent process, the accuracy of determining the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process based on the historical electrical stimulation parameter values ​​of the historical sample data in the target clustering cluster will also be higher.

[0110] Since the historical electrical stimulation parameter values ​​of the historical sample data in the target cluster are more consistent with the current medical history data, after determining the target cluster corresponding to the current medical history data, the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process can be determined based on the numerical characteristics of the historical sample data in the target cluster corresponding to the current medical history data, and the real-time physiological parameter data generated can be monitored.

[0111] Preferably, in one embodiment of the present invention, based on the numerical characteristics of the historical sample data in the target cluster corresponding to the current medical history data, the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process are determined and the real-time physiological parameter data are monitored, including:

[0112] In the target cluster, the mean of the same historical electrical stimulation parameter values ​​of all historical sample data is calculated as the standard value of each electrical stimulation parameter in the target cluster, and the standard value is used as the reference value for the initial electrical stimulation parameter setting of the patient to be treated.

[0113] Then, when the patient to be treated is subjected to transcranial pulse electrical stimulation, each electrical stimulation parameter is set to a standard value in the target cluster, and various real-time physiological parameter data of the patient to be treated during the transcranial pulse electrical stimulation process are obtained in real time.

[0114] Various real-time physiological parameter data are monitored to obtain monitoring results: in the target cluster, the maximum and minimum values ​​of the same historical physiological parameter time series data of all historical sample data are obtained, and the maximum and minimum values ​​are used as a boundary value to form the fluctuation range of each physiological parameter.

[0115] For any physiological parameter, if the real-time physiological parameter data of the patient to be treated under the physiological parameter does not exist in the corresponding fluctuation range, it is considered that the physiological parameter of the patient to be treated is abnormal and an early warning is required. The doctor can further adjust the electrical stimulation parameter value based on the patient's current medical history data and various physiological parameter data; otherwise, it is considered normal and no early warning is required.

[0116] In summary, the historical sample data of historical patients are obtained, and the historical sample data include historical medical history data, various historical electrical stimulation parameter values ​​during transcranial electrical stimulation, and various historical physiological parameter time series data; obtaining comprehensive data of historical patients can more accurately identify the characteristics of different patient groups in the subsequent process, so as to accurately regulate the electrical stimulation parameters during the transcranial pulse electrical stimulation of the patients to be treated and achieve more accurate physiological parameter monitoring. With the help of clustering technology, the historical sample data are clustered and analyzed based on the differences between various historical electrical stimulation parameter values ​​of the historical sample data, so as to form an initial cluster cluster. At this time, the historical electrical stimulation parameter values ​​used by the historical sample data in each initial cluster cluster have strong consistency. Then, in each initial cluster cluster, the fluctuation similarity of various historical physiological parameter time series data of the historical sample data and the similarity of the historical medical history data are analyzed to determine the confidence of the historical electrical stimulation parameter value in each initial cluster cluster. The confidence characterizes the consistency of the response of the physiological parameter changes of the historical sample data in each initial cluster cluster to the electrical stimulation parameter value, which implies the effectiveness and reliability of the electrical stimulation parameter value. Then, according to the confidence corresponding to the initial cluster, combined with the similarity between the current medical history data of the patient to be treated and the historical medical history data, the current medical history data can be classified to determine the target cluster to which it belongs, that is, the initial cluster with a condition similar to that of the patient to be treated. Finally, based on the numerical characteristics of the historical sample data in the target cluster, the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process are set. The electrical stimulation parameter values ​​set in this way are more consistent with the current medical history data. Therefore, when the real-time physiological parameter data of the patient to be treated is monitored on this basis, the credibility of the monitoring results obtained will also be higher.

[0117] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0118] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A high-precision physiological parameter monitoring method for transcranial electrical stimulation, characterized in that: The method comprises: Acquire historical sample data of historical patients, and each historical sample data includes historical medical history data, various historical electrical stimulation parameter values ​​during transcranial pulse electrical stimulation, and various historical physiological parameter time series data; Comprehensively analyzing the differences between various historical electrical stimulation parameter values ​​of historical sample data, performing cluster analysis on all historical sample data, and obtaining initial clusters; In each initial cluster, the fluctuation similarity of various historical physiological parameter time series data of the historical sample data is analyzed, and the similarity of the historical medical history data between different historical sample data is quantified, so as to determine the confidence of the historical electrical stimulation parameter value in each initial cluster; Analyze the confidence levels corresponding to all initial clusters, and determine the target cluster corresponding to the current medical history data based on the similarities between the current medical history data and the historical medical history data of the patient to be treated; based on the numerical characteristics of the historical sample data in the target cluster corresponding to the current medical history data, determine the electrical stimulation parameter values ​​of the patient to be treated during the transcranial pulse electrical stimulation process and monitor the real-time physiological parameter data.

2. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 1, characterized in that: The method for obtaining the initial clustering clusters includes: Comprehensively analyzing the differences between various historical electrical stimulation parameter values ​​of the historical sample data, and determining similarity values ​​of the electrical stimulation parameters between any two historical sample data; The optimal K value of all historical sample data is obtained based on the silhouette coefficient method, the similarity value of the electrical stimulation parameters between any two historical sample data is used as the distance metric, and all historical sample data are clustered based on the K-means clustering algorithm and the optimal K value to obtain an initial cluster cluster.

3. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 2, characterized in that: The method for obtaining the similarity value of the electrical stimulation parameters comprises: Analyze the difference in consistency between various historical electrical stimulation parameter values ​​and preset electrical stimulation parameter values ​​in any two historical sample data, and determine the consistency deviation factor between any two historical sample data; Arrange various historical electrical stimulation parameter values ​​in each historical sample data in the same arrangement mode to obtain an electrical stimulation parameter value vector corresponding to each historical sample data, and calculate the cosine similarity of the electrical stimulation parameter value vectors of any two historical sample data; According to the consistency deviation factor and cosine similarity between any two historical sample data, the electrical stimulation parameter similarity value between any two historical sample data is obtained, and the consistency deviation factor is negatively correlated with the electrical stimulation parameter similarity value, and the cosine similarity is positively correlated with the electrical stimulation parameter similarity value.

4. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 3, characterized in that: The method for obtaining the consistency deviation factor includes: Compare each historical electrical stimulation parameter value in each historical sample data with the corresponding preset electrical stimulation parameter value, and count the number of historical electrical stimulation parameter values ​​that are the same as the corresponding preset electrical stimulation parameter value as the consistency parameter of each historical sample data; The ratio of the consistency parameter to the number of types of historical electrical stimulation parameter values ​​is used as the electrical stimulation parameter consistency score corresponding to each historical sample data; The absolute value of the difference between the electrical stimulation parameter consistency scores of any two historical sample data is used as the consistency deviation factor between any two historical sample data.

5. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 1, characterized in that: The method for obtaining the confidence level includes: In each initial cluster, all historical sample data are combined in pairs to obtain all non-repeated sample data combinations; Analyze the fluctuation similarity of various historical physiological parameter time series data of two historical sample data in each sample data combination, and determine the fluctuation similarity index corresponding to each sample data combination; Based on the Jieba word segmentation tool, the historical medical history data of all historical sample data are segmented, and the word segmentation array corresponding to each historical sample data is obtained; the Jaccard coefficient between the word segmentation arrays of two historical sample data in each sample data combination is calculated as the similarity coefficient; In each initial cluster, the fluctuation similarity index corresponding to the sample data combination is weighted by using the similarity coefficient between the word arrays of two historical sample data in the sample data combination, so as to obtain the confidence of the historical electrical stimulation parameter value in each initial cluster.

6. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 5, characterized in that: The method for obtaining the fluctuation similarity index includes: For any sample data combination, calculate the DTW value between the same historical physiological parameter time series data of two historical sample data; Analyze the similarity of the fluctuation range of the same historical physiological time series data in the two historical sample data in the sample data combination, and determine the distance factor of the two historical sample data under the same historical physiological time series data; The product of the DTW value and the distance factor of the two historical sample data in the sample data combination under the same historical physiological parameter time series data is negatively correlated and normalized to obtain the fluctuation similarity factor of the two historical sample data under the same historical physiological parameter time series data; The value after normalizing the mean value of the fluctuation similarity factor of the two historical sample data in the sample data combination under all types of historical physiological time series data is used as the fluctuation similarity index corresponding to the sample data combination.

7. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 6, characterized in that: The method for obtaining the distance factor includes: For any sample data combination, the maximum value and the minimum value of each type of historical physiological time series data in each historical sample data in the sample data combination are combined into a maximum value array; The Euclidean distance between the maximum value arrays of two historical sample data under the same historical physiological parameter time series data is used as the distance factor of the two historical sample data under the same historical physiological time series data in the sample data combination.

8. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 5, characterized in that: The analysis of the confidences corresponding to all initial clusters and combining the similarities between the current medical history data and the historical medical history data of the patient to be treated to determine the target cluster corresponding to the current medical history data includes: The initial clusters with confidence greater than the preset confidence threshold are taken as candidate clusters; The current medical history data of the patients to be treated are segmented based on the Jieba word segmentation tool to obtain the word segmentation array corresponding to the current medical history data; Calculate the Jaccard coefficient between the word group corresponding to the current medical history data and the word array corresponding to each historical sample data in all candidate clusters as the similarity parameter; The candidate cluster to which the historical sample data corresponding to the maximum value of all similar parameters of the current medical history data belongs is taken as the target cluster corresponding to the current medical history data.

9. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 1, characterized in that: The method of determining the electrical stimulation parameter value of the patient to be treated during the transcranial pulse electrical stimulation process and monitoring the real-time physiological parameter data based on the numerical characteristics of the historical sample data in the target cluster corresponding to the current medical history data includes: In the target cluster, the mean of the same historical electrical stimulation parameter values ​​of all historical sample data is calculated as the standard value of each electrical stimulation parameter in the target cluster; When the patient to be treated is subjected to transcranial pulse electrical stimulation, each electrical stimulation parameter is set to a standard value in the target cluster, and various real-time physiological parameter data of the patient to be treated during the transcranial pulse electrical stimulation process are obtained; Monitor various real-time physiological parameter data and obtain monitoring results.

10. A high-precision physiological parameter monitoring method for transcranial electrical stimulation according to claim 9, characterized in that: The monitoring of various real-time physiological parameter data to obtain monitoring results includes: In the target cluster, the maximum and minimum values ​​of the same historical physiological parameter time series data of all historical sample data are obtained to form the fluctuation range of each physiological parameter; For any physiological parameter, if the real-time physiological parameter data of the patient to be treated under the physiological parameter does not exist in the corresponding fluctuation range, it is considered that the physiological parameter of the patient to be treated is abnormal and an early warning is required; otherwise, it is considered normal and no early warning is required.

Images