Brain metabolism state classification device and method, storage medium and program product

Abstract

The embodiment of the invention provides a brain metabolism state classification device and method, a storage medium and a program product. The method comprises the following steps that on the basis of sample data, the incidence relation between a brain metabolism index and a plurality of physiological indexes and the incidence relation between the brain metabolism index and a plurality of artery indexes are determined, and the artery indexes are related to the carotid artery and the vertebral artery; according to an association relationship, screening from the plurality of physiological indexes and the plurality of artery indexes to obtain a target physiological index and a target artery index; constructing and training a brain metabolism state classification model based on the target artery index and the target physiological index; wherein the brain metabolism state classification model is used for classifying the brain metabolism state of the user based on the index value of the target artery index of the user and the index value of the target physiological index of the user. According to the technical scheme provided by the embodiment of the invention, the classification accuracy of the brain metabolism state classification model can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a brain metabolic state classification device, method, storage medium and program product. Background Art

[0002] Currently, in order to evaluate the state of brain metabolism, it is necessary to use magnetic resonance imaging (MRI) equipment to obtain brain magnetic resonance data and evaluate the brain metabolic state based on the brain magnetic resonance data.

[0003] However, MRI equipment is large in size and complex to operate, making it difficult to adapt to some special scenarios, such as aerospace scenarios, outdoor emergency scenarios, etc. In other words, in special scenarios, it is difficult to obtain MRI data, and it is impossible to evaluate the brain metabolic state. Summary of the invention

[0004] In view of the above problems, the present application is proposed to provide a brain metabolic state classification device, method, storage medium and program product that solve the above problems or at least partially solve the above problems.

[0005] In a first aspect of the present application, a method for constructing a brain metabolic state classification model is provided, comprising:

[0006] Based on the sample data, determining the correlation between the brain metabolism index and a plurality of physiological indexes and the correlation between the brain metabolism index and a plurality of arterial indexes, wherein the plurality of arterial indexes are related to the carotid artery and the vertebral artery;

[0007] According to the correlation between the brain metabolic index and the multiple physiological indexes and the correlation between the brain metabolic index and the multiple arterial indexes, a target physiological index and a target arterial index are screened from the multiple physiological indexes and the multiple arterial indexes;

[0008] Based on the target arterial index and the target physiological index, construct and train a brain metabolic state classification model;

[0009] The brain metabolic state classification model is used to classify the user's brain metabolic state based on the index value of the user's target arterial index and the index value of the user's target physiological index.

[0010] In a second aspect of the present application, a device for classifying brain metabolic states is provided, comprising:

[0011] A transceiver module, used for receiving ultrasonic data sent by an ultrasonic data acquisition device, wherein the ultrasonic data includes arterial data acquired by the ultrasonic acquisition device at the user's carotid artery and vertebral artery;

[0012] A processor, configured to: extract an index value of a target artery index of the user from the ultrasound data; determine input information of a pre-trained brain metabolic state classification model according to the index value of the target artery index of the user and the index value of the target physiological index of the user; input the input information into the brain metabolic state classification model, so that the brain metabolic state classification model classifies the brain metabolic state of the user;

[0013] Wherein, the brain metabolic state classification model is constructed based on the target arterial index and the target physiological index.

[0014] According to a third aspect of the present application, a computer-readable storage medium storing a computer program is provided, wherein the computer program can implement any of the above-described methods when executed by a computer.

[0015] According to a fourth aspect of the present application, a computer program product is provided, comprising a computer program, which implements any of the above methods when executed by a processor.

[0016] In the technical solution provided in the embodiment of the present application, based on the sample data, the correlation between the brain metabolic index and the physiological index and the correlation between the brain metabolic index and the arterial index are determined, and based on the correlation, relatively important target physiological indexes and target arterial indexes are screened out from a large number of physiological indexes and arterial indexes, and a brain metabolic state classification model is constructed based on the screened target physiological indexes and target arterial indexes, which helps to improve the classification accuracy of the brain metabolic state classification model. Moreover, when classifying the user's brain metabolic state, the constructed brain metabolic state classification model does not rely on brain magnetic resonance data, but relies on some arterial index data related to the carotid artery and vertebral artery and some index data of physiological indicators, which are relatively easy to obtain. It can be seen that the technical solution provided in the embodiment of the present application is suitable for some special scenarios, such as: aerospace scenarios, outdoor first aid scenarios.

[0017] In the technical solution provided in the embodiment of the present application, the brain metabolic state classification device can classify the user's brain metabolic state by using the data collected by the ultrasonic data acquisition device at the user's carotid artery and vertebral artery and some basic physiological indicators of the user. It can be seen that the brain metabolic state classification device provided in the embodiment of the present application can evaluate the user's brain metabolic state in some special scenarios, such as: aerospace scenarios, outdoor first aid scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 A schematic diagram of a process for constructing a brain metabolic state classification model provided in one embodiment of the present application;

[0020] Figure 2 A structural block diagram of a brain metabolic state classification device provided in one embodiment of the present application;

[0021] Figure 3 A schematic diagram of a brain metabolic state classification system provided in one embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below according to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0023] In addition, some of the processes described in the specification, claims and the above-mentioned figures of the present application include multiple operations that appear in a specific order, and these operations may not be executed in the order in which they appear in this article or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0024] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0025] At present, the assessment of brain metabolic status mainly relies on brain magnetic resonance imaging data. However, brain magnetic resonance equipment is complicated to operate, needs to be controlled by specialized technicians, and is often large in size and usually installed in fixed places such as hospitals. Therefore, the assessment of brain metabolic status is difficult to adapt to some special scenarios. For example, in aerospace scenarios, due to changes in gravity in the space environment and limited space in the space capsule, it is impossible to use large-scale inspection equipment in related technologies to examine the astronauts' brains, resulting in the inability to assess the astronauts' brain metabolic status in the space environment. For another example, in outdoor first aid scenarios, the accident site is usually inconvenient for transportation (remote or congested nearby), and the injured are often difficult to transport to hospitals with inspection equipment in a timely manner. Therefore, first aid personnel are often unable to promptly learn the injured's brain metabolic status, affecting the treatment of the injured.

[0026] In order to solve or partially solve the above technical problems, the embodiment of the present application proposes a brain metabolic state classification device, wherein the brain metabolic state classification device uses the data collected by the ultrasonic data acquisition device at the user's carotid artery and vertebral artery and some basic physiological indicators of the user to classify the user's brain metabolic state. It can be seen that the brain metabolic state classification device provided by the embodiment of the present application can evaluate the user's brain metabolic state in some special scenarios, such as: aerospace scenarios, outdoor first aid scenarios.

[0027] The brain metabolic state classification device classifies the user's brain metabolic state based on a brain metabolic state classification model. A method for constructing a brain metabolic state classification model is described below. Figure 1 A flowchart of a method for constructing a brain metabolic state classification model provided in an embodiment of the present application. The execution subject of the method may include a terminal device and / or a server, and the present embodiment of the application does not specifically limit this.

[0028] Among them, terminal equipment: also known as terminal. It can be a mobile terminal, a fixed terminal or a portable terminal, such as a mobile phone, a station, a unit, a device, a multimedia computer, a multimedia tablet, an Internet node, a communicator, a desktop computer, a laptop computer, a notebook computer, a netbook computer, a tablet computer, a personal communication system device, a personal navigation device, a personal digital assistant, an audio / video player, a digital camera / camcorder, a positioning device, a television receiver, a radio broadcast receiver, an electronic book device, a game device or any combination thereof, including accessories, peripherals or any combination thereof of these devices.

[0029] Server: A server can be one or more servers. A server can also be a physical server or a virtual server. A server can be an independent physical server, a server cluster or a distributed system consisting of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, and big data and artificial intelligence platforms.

[0030] like Figure 1 As shown, the method comprises:

[0031] 101. Based on the sample data, determine the correlation between brain metabolic indicators and multiple physiological indicators, as well as the correlation between brain metabolic indicators and multiple arterial indicators.

[0032] Wherein, the cerebral metabolic index includes one or more of cerebral blood flow (CBF), oxygen extraction fraction (OEF) and cerebral oxygen metabolic rate (CMRO2). Wherein, cerebral blood flow, oxygen extraction fraction and cerebral oxygen metabolic rate are parameters for evaluating the cerebral metabolic state.

[0033] Among them, cerebral blood flow refers to the amount of blood passing through the brain tissue per unit time (usually per minute). The two indicators of oxygen uptake fraction and cerebral oxygen metabolic rate can be used to evaluate the state of brain oxygen metabolism. Together, they describe how the brain uses oxygen and the balance of oxygen supply and demand. The oxygen uptake fraction refers to the ratio of oxygen extracted from the blood into the brain tissue, which reflects the demand and utilization efficiency of brain tissue for oxygen. An increase in the oxygen uptake fraction may indicate an increased demand for oxygen in the brain tissue, while a decrease in the oxygen uptake fraction may mean a decrease in the demand for oxygen in the brain tissue or an excess supply of oxygen. Changes in cerebral blood flow can reflect the oxygen and nutrient supply of brain tissue. An increase in cerebral blood flow may be related to increased brain activity or pathological conditions (such as inflammation), while a decrease in cerebral blood flow may be related to cerebral ischemia or hypoxia. The cerebral oxygen metabolic rate refers to the rate at which brain tissue consumes oxygen, usually expressed in micromoles / 100 grams of brain tissue / minute (μmol / 100g / min). Changes in the cerebral oxygen metabolic rate directly reflect the metabolic activity of brain tissue. An increase in brain oxygen metabolic rate may be associated with increased brain activity or pathological states, while a decrease in brain oxygen metabolic rate may be associated with brain dysfunction or inhibition. Together, these three indicators form a framework for assessing brain oxygen metabolic status, and they can be used alone or in combination to provide more comprehensive brain oxygen metabolic information.

[0034] The index values ​​of brain metabolic indices can be obtained by analyzing brain magnetic resonance data, wherein the brain magnetic resonance data may include but are not limited to: quantitative blood oxygen level dependence (Quantitative BOLD, qBOLD), multiple delayed arterial spin labeling (Arterial Spin Labeling, ASL), multi-echo quantitative magnetic susceptibility imaging (Quantitative Susceptibility Mapping, QSM), T1 structural image and 4Dflow.

[0035] Among them, multi-delay ASL can obtain 5-7 parameters in a test lasting about 3-5 minutes, and obtain more accurate cerebral blood flow through arterial transit time correction, rather than assuming that arterial transit time is equal to marker delay time, thus achieving a technical upgrade. It not only does not require the use of contrast agents, it can avoid radiation damage, and provides multi-parameter information, improving the measurement accuracy based on single-delay ASL.

[0036] The multiple physiological indicators may include, but are not limited to: age, gender, body mass index (BMI), heart rate, systolic blood pressure, diastolic blood pressure, the ratio of systolic blood pressure to diastolic blood pressure, and electrocardiogram indicators obtained by analyzing electrocardiogram data.

[0037] These physiological indicators are important indicators for assessing individual health status and monitoring disease progression. Due to the interaction between the heart and the brain, multiple ECG indicators are obtained by performing various cardiac function analyses such as heart rate variability (HRV) analysis, atrial fibrillation analysis, QRS complex (Q wave-R wave-S wave complex) analysis, and pacing analysis on ECG data. Using these physiological indicators is helpful for the assessment of brain oxygen metabolism.

[0038] The multiple arterial indices are related to the carotid artery and the vertebral artery. The carotid artery supplies blood to the front 2 / 3 (anterior circulation) of the cerebral hemisphere, and the vertebral artery supplies blood to the back 1 / 3 (posterior circulation) of the cerebral hemisphere. By analyzing the ultrasound data of the carotid artery and vertebral artery, more detailed arterial indices can be obtained, which is conducive to exploring the corresponding relationship between the neck and the brain.

[0039] The multiple arterial indices may include but are not limited to: vascular inner diameter, vascular elasticity, and hemodynamic parameters, wherein the vascular inner diameter may be divided into the carotid artery vascular inner diameter and the vertebral artery vascular inner diameter, the vascular elasticity may be divided into the carotid artery vascular elasticity and the vertebral artery vascular elasticity, and the hemodynamic parameters may be divided into the carotid artery hemodynamic parameters and the vertebral artery hemodynamic parameters. Among them, the hemodynamic parameters may include: systolic peak blood flow velocity, end-diastolic blood flow velocity, mean blood flow velocity, resistance index, and the like.

[0040] In some embodiments, the sample data includes an index value of a brain metabolism index of each sample user among a plurality of sample users, an index value of a plurality of physiological indexes, and an index value of a plurality of arterial indexes.

[0041] In some embodiments, the sample data may be from a public biomedical database, such as the UK Biobank (UKB), and / or a locally constructed biomedical database.

[0042] The sample data can be statistically analyzed by statistical methods to determine the correlation between the brain metabolic index and multiple physiological indexes and the correlation between the brain metabolic index and multiple arterial indexes. For example, the correlation analysis can be performed by drawing a scatter plot between the indexes and then performing correlation coefficient analysis, partial correlation or multiple correlation analysis, variance analysis, regression analysis, and receiver operating characteristic (ROC) curve analysis and other statistical or machine learning methods.

[0043] Among them, the correlation relationship is used to characterize the degree of correlation between indicators.

[0044] According to the correlation relationship, a correlation matrix between brain metabolic indicators and multiple physiological indicators and a correlation matrix between brain metabolic indicators and multiple arterial indicators can be constructed. The values ​​in the correlation matrix are related to the specific statistical method used. For example, if correlation analysis is performed, the values ​​in the matrix represent the degree of correlation (or correlation coefficient) and credibility between the indicators. Among them, credibility is used to characterize the credibility of the degree of correlation, and credibility can be determined based on the probability of significance test (p). The specific implementation of correlation analysis can be referred to the prior art, and the embodiments of the present application do not specifically limit this.

[0045] 102. According to the correlation between the brain metabolic index and the multiple physiological indexes and the correlation between the brain metabolic index and the multiple arterial indexes, select a target physiological index and a target arterial index from the multiple physiological indexes and the multiple arterial indexes.

[0046] In some embodiments, target correlations between the brain metabolic index and multiple physiological indexes and target correlations between the brain metabolic index and multiple arterial indexes are determined based on correlations between the brain metabolic index and multiple physiological indexes and correlations between the brain metabolic index and multiple arterial indexes; target physiological indexes and target arterial indexes are screened from the multiple physiological indexes and the multiple arterial indexes based on the target correlations between the brain metabolic index and multiple physiological indexes and the target correlations between the brain metabolic index and multiple arterial indexes.

[0047] Among them, the target correlation degree between indicators is determined based on the correlation degree and credibility between indicators.

[0048] Exemplarily, the correlation degree between indicators can be determined as the target correlation degree between indicators, or, when the credibility between indicators is greater than or equal to a preset credibility threshold, the correlation degree between indicators can be determined as the target correlation degree between indicators, otherwise, the target correlation degree between indicators is set to 0.

[0049] In practical applications, importance ranking or threshold-based judgment can be used for screening.

[0050] Exemplarily, according to the target correlation degree between the brain metabolic index and the multiple physiological indexes and the target correlation degree between the brain metabolic index and the multiple arterial indexes, the multiple physiological indexes and the multiple arterial indexes are ranked in importance, and according to the result of the importance ranking, the target physiological index and the target arterial index are screened from the multiple physiological indexes and the multiple arterial indexes. For example, when the importance ranking is in descending order, the top N indexes are selected as the target physiological index and the target arterial index. Wherein, N is an integer greater than or equal to 1.

[0051] Exemplarily, the physiological index and the arterial index whose target correlation degree with the brain metabolic index is greater than a preset correlation threshold are determined as the target physiological index and the target arterial index.

[0052] As an option, when the brain metabolic index includes multiple of cerebral blood flow, oxygen uptake fraction and brain oxygen metabolic rate, since each brain metabolic index is very important for evaluating the brain metabolic state, the target physiological index and target arterial index related to each brain metabolic index can be screened out in the above manner for each brain metabolic index. In other words, the index set finally screened includes the target physiological index and target arterial index related to each brain metabolic index. That is, the above step 102 can be implemented by the following steps:

[0053] 1021. According to the correlation between the cerebral blood flow and the multiple physiological indicators and the correlation between the cerebral blood flow and the multiple arterial indicators, a first target physiological indicator related to the cerebral blood flow and a first target arterial indicator related to the cerebral blood flow are screened from the multiple physiological indicators and the multiple arterial indicators.

[0054] 1022. According to the correlation between the oxygen uptake fraction and the multiple physiological indices and the correlation between the oxygen uptake fraction and the multiple arterial indices, select a second target physiological indices related to the oxygen uptake fraction and a second target arterial indices related to the oxygen uptake fraction from the multiple physiological indices and the multiple arterial indices.

[0055] 1023. According to the correlation between the brain oxygen metabolic rate and the multiple physiological indicators and the correlation between the brain oxygen metabolic rate and the multiple arterial indicators, a third target physiological indicator related to the brain oxygen metabolic rate and a third target arterial indicator related to the brain oxygen metabolic rate are screened from the multiple physiological indicators and the multiple arterial indicators.

[0056] The union of a first target physiological index related to the cerebral blood flow, a first target arterial index related to the cerebral blood flow, a second target physiological index related to the oxygen uptake fraction, a second target arterial index related to the oxygen uptake fraction, a third target physiological index related to the brain oxygen metabolic rate, and a third target arterial index related to the brain oxygen metabolic rate can be determined as the final screened index set.

[0057] 103. Based on the target arterial index and the target physiological index, a brain metabolic state classification model is constructed and trained.

[0058] The target arterial index and the target physiological index are used as input parameters of the brain metabolic state classification model, and the brain metabolic state classification model is trained using training samples. The training samples can be generated based on sample data in a locally constructed biomedical database. The training samples may include: index values ​​of target physiological indicators of sample users, index values ​​of target arterial indicators, and real brain metabolic state categories. The index values ​​of target physiological indicators and index values ​​of target arterial indicators of sample users can be input into the brain metabolic state classification model to obtain a predicted brain metabolic state category; according to the real brain metabolic state category and the predicted brain metabolic state category, the model is optimized for parameters. The specific parameter optimization method can be implemented using existing optimization methods, and the embodiments of the present application do not specifically limit this.

[0059] The brain metabolic state classification model is used to classify the user's brain metabolic state based on the index value of the user's target arterial index and the index value of the user's target physiological index.

[0060] It should be noted that the classification mentioned in this application can be understood as a level division.

[0061] In the technical solution provided in the embodiment of the present application, based on the sample data, the correlation between the brain metabolic index and the physiological index and the correlation between the brain metabolic index and the arterial index are determined, and based on the correlation, relatively important target physiological indexes and target arterial indexes are screened out from a large number of physiological indexes and arterial indexes, and a brain metabolic state classification model is constructed based on the screened target physiological indexes and target arterial indexes, which helps to improve the classification accuracy of the brain metabolic state classification model. Moreover, when classifying the user's brain metabolic state, the constructed brain metabolic state classification model does not rely on brain magnetic resonance data, but relies on some arterial index data related to the carotid artery and vertebral artery and some index data of physiological indicators, which are relatively easy to obtain. It can be seen that the technical solution provided in the embodiment of the present application is suitable for some special scenarios, such as: aerospace scenarios, outdoor first aid scenarios.

[0062] In an optional implementation, the step of “determining, based on the sample data, the correlation between the brain metabolic index and the plurality of physiological indexes and the correlation between the brain metabolic index and the plurality of arterial indexes” in 101 above may include:

[0063] 1011. Based on sample data in a public biomedical database and / or based on sample data in a locally constructed medical database, determine the correlation between the brain metabolic index and multiple physiological indexes and the correlation between the brain metabolic index and multiple arterial indexes.

[0064] Currently, public biomedical databases, such as the UKB database, have the advantages of large sample data volume and low acquisition cost, but oxygen uptake fraction, brain oxygen metabolic rate, and ultrasound spectrum data of the carotid and vertebral arteries are not collected. In order to make up for this defect, a local medical database is constructed, which includes oxygen uptake fraction, brain oxygen metabolic rate, and ultrasound spectrum data of the carotid and vertebral arteries.

[0065] Therefore, in an optional manner, the correlation between brain metabolic indicators and multiple physiological indicators and the correlation between brain metabolic indicators and multiple arterial indicators can be determined based on sample data in a public biomedical database and sample data in a locally constructed medical database.

[0066] Currently, the UKB database includes multiple physiological indicators of sample users, the inner diameters of the carotid and vertebral arteries, the elasticity of the carotid and conical arteries, and the respective index values ​​of cerebral blood flow, but does not include the oxygen uptake fraction, cerebral oxygen metabolic rate, and hemodynamic parameters of the carotid and vertebral arteries of the sample users.

[0067] Exemplarily, the multiple arterial indices include vascular inner diameter and vascular elasticity; the above 1011 may include:

[0068] S11. Based on sample data in a public biomedical database, statistically analyze the correlation between multiple physiological indicators and cerebral blood flow, the correlation between the inner diameter of the blood vessel and the cerebral blood flow, and the correlation between the elasticity of the blood vessel and the cerebral blood flow.

[0069] Exemplarily, the arterial index includes: hemodynamic parameters; the above 1011 may include:

[0070] S12. Based on the sample data in the locally constructed medical database, statistically calculate the correlation between multiple physiological indicators and the oxygen uptake fraction, the correlation between multiple physiological indicators and the brain oxygen metabolic rate, the correlation between the hemodynamic parameters and the oxygen uptake fraction, and the correlation between the hemodynamic parameters and the brain oxygen metabolic rate.

[0071] The statistical method of the association relationship can refer to the corresponding content in the above embodiment, and is not specifically limited here.

[0072] Of course, if the public biomedical database adds oxygen uptake fraction, brain oxygen metabolic rate, and ultrasound spectrum data of the carotid artery and vertebral artery in the future, the correlation between brain metabolic indicators and multiple physiological indicators and the correlation between brain metabolic indicators and multiple arterial indicators can be determined entirely based on the sample data in the public biomedical database.

[0073] If the sample size in the locally constructed biomedical database is large enough, the correlation between brain metabolic indicators and multiple physiological indicators and the correlation between brain metabolic indicators and multiple arterial indicators can also be determined entirely based on the sample data in the locally constructed biomedical database.

[0074] In practical applications, a first correlation matrix between brain metabolic indicators and multiple physiological indicators and a second correlation matrix between brain metabolic indicators and multiple arterial indicators can be constructed based on the correlation obtained based on statistics of public biomedical databases; a first correlation network is constructed based on the first correlation matrix and the second correlation matrix.

[0075] According to the correlation obtained by statistics based on the locally constructed biomedical database, a third correlation matrix between brain metabolic indicators and multiple physiological indicators and a fourth correlation matrix between brain metabolic indicators and multiple arterial indicators can be constructed; based on the third correlation matrix and the fourth correlation matrix, a second correlation network is constructed.

[0076] A target association relationship network is constructed according to the first association relationship network and the second association relationship network.

[0077] It should be noted that in order to simplify the association relationship network, in the process of generating the association relationship network, when the correlation between two indicators is greater than or equal to the first set threshold and the credibility is greater than or equal to the second set threshold, the association relationship between the two indicators is constructed in the network, otherwise the association relationship between the two indicators is not constructed.

[0078] In this way, the indicator screening process in the above step 102 can be performed based on the target association relationship network.

[0079] In some embodiments, the brain metabolic state classification model is used to predict the index value of the user's brain oxygen metabolism level index based on the index value of the user's target arterial index and the index value of the user's target physiological index, and compare the index value of the brain oxygen metabolism level index with the index value ranges corresponding to each of multiple preset brain metabolic state categories to determine the user's brain metabolic state category.

[0080] The above-mentioned brain oxygen metabolism level indicators are used to characterize the brain oxygen metabolism level.

[0081] Exemplarily, when the index value of the user's brain oxygen metabolism level index is within the index value range corresponding to the first brain metabolism state category, the user's brain metabolism state category is determined to be the first brain metabolism state category. The first brain metabolism state category is one of a plurality of preset brain metabolism state categories.

[0082] In practical applications, multiple preset brain metabolic state categories can also be regarded as multiple preset brain metabolic level levels. Multiple brain metabolic state categories can be divided according to actual needs. The number of brain metabolic state categories can be set according to actual needs, and the embodiments of the present application do not specifically limit this. Exemplarily, the number of brain metabolic state categories is two, namely normal and abnormal. Another exemplary example is that the number of brain metabolic state categories is three, namely normal, mildly abnormal and severely abnormal. The index value range corresponding to each brain metabolic state category can be determined by statistical sample data. The following introduces a method for statistically analyzing the index value range, which includes the following steps:

[0083] 104. Obtain from the sample data the cerebral blood flow value, oxygen uptake fraction value, and cerebral oxygen metabolic rate value of the first sample user and the cerebral blood flow value, oxygen uptake fraction value, and cerebral oxygen metabolic rate value of the second sample user.

[0084] The first sample users are from the case group, and the second sample users are from the control group.

[0085] There may be multiple first sample users, and there may be multiple second sample users.

[0086] In practical applications, the above sample data can be obtained from a locally constructed biomedical database.

[0087] The cerebral blood flow value, oxygen uptake fraction value and cerebral oxygen metabolic rate value of the first sample user and the second sample user are data of the gray matter area of ​​the brain.

[0088] 105. Determine an index value of a brain oxygen metabolism level index of the first sample user according to the cerebral blood flow value, the oxygen uptake score value, and the brain oxygen metabolic rate value of the first sample user.

[0089] In an optional implementation, a preset formula may be used to calculate the index value of the brain oxygen metabolism level index of the first sample user.

[0090] For example, the following formula (1) can be used to calculate the index value of the brain oxygen metabolism level index Z:

[0091] Z=ω1·CBF+ω2·OEF+ω3·CMRO2(1)

[0092] Among them, CBF is cerebral blood flow, OEF is oxygen uptake fraction, CMRO2 is brain oxygen metabolic rate, ω1 is the weight corresponding to cerebral blood flow, ω2 is the weight corresponding to oxygen uptake fraction, and ω3 is the weight corresponding to brain oxygen metabolic rate. The size of the above multiple weights can be determined according to the importance of each indicator in the assessment of brain metabolic status in different environments.

[0093] 106. Determine an index value of a brain oxygen metabolism level index of the second sample user according to the cerebral blood flow value, the oxygen uptake score value, and the brain oxygen metabolic rate value of the second sample user.

[0094] In some embodiments, the index values ​​of the brain oxygen metabolism level indexes of the first sample user and the second sample user may be calculated in the same calculation method.

[0095] 107. Determine the index value range of each of the multiple preset brain metabolic state categories according to the distribution of the index value of the brain oxygen metabolism level index of the first sample user and the index value of the brain oxygen metabolism level index of the second sample user.

[0096] In actual applications, the distribution of index values ​​of brain oxygen metabolism level indicators of multiple first sample users (hereinafter referred to as the first distribution) and the distribution of index values ​​of brain oxygen metabolism level indicators of multiple second sample users (hereinafter referred to as the second distribution) can be statistically analyzed, and the index value range of each of the multiple preset brain metabolic state categories is determined based on the distribution.

[0097] Exemplarily, the first distribution includes: the minimum index value A and the maximum index value B among the index values ​​of the brain oxygen metabolism level index of multiple first sample users. The second distribution includes the minimum index value C and the maximum index value D among the index values ​​of the brain oxygen metabolism level index of multiple second sample users. Since the first sample users are from the case group and the second sample users are from the control group, B is generally smaller than C. Exemplarily, when the brain metabolic state category distribution is normal and abnormal, the intermediate value E between B and C can be taken, and the normal corresponding index value range is: greater than or equal to E, and the abnormal corresponding index value range is: less than E. Among them, E can be regarded as a threshold. Exemplarily, when the brain metabolic state category distribution is normal, mildly abnormal and severely abnormal, the intermediate value F between A and B can be taken, the normal corresponding index value range is: greater than or equal to E, the mildly abnormal corresponding index value range is: greater than or equal to F and less than E, and the severe abnormal corresponding index range is: less than F.

[0098] Among them, the brain metabolic state classification model may include but is not limited to models such as naive Bayes, support vector machine, K nearest neighbor, logistic regression, decision tree and random forest.

[0099] Exemplarily, the final screened indicator set may include: peak systolic blood flow velocity, end-diastolic blood flow velocity, mean blood flow velocity, resistance index, pulsatility index, systolic / diastolic ratio, vascular inner diameter, body temperature, blood pressure, and heart rate. These indicators are the input features, that is, the input of the classification model.

[0100] In other embodiments, the brain metabolic state classification model includes: a first sub-model for classifying the user's cerebral blood flow, a second sub-model for classifying the user's oxygen uptake fraction, a second sub-model for classifying the user's brain oxygen metabolic rate, and a comprehensive module; the comprehensive module is used to classify the user's brain metabolic state according to the classification results of the first sub-model, the classification results of the second sub-model, and the classification results of the third sub-model.

[0101] In an optional embodiment, the integration module may classify the user's brain metabolic state according to a combination of the classification result of the first sub-model, the classification result of the second sub-model, and the classification result of the third sub-model.

[0102] Exemplarily, the output of each sub-model is a level. Assume that each sub-model corresponds to three levels, such as the three levels corresponding to the first sub-model are A1, A2 and A3, the three levels corresponding to the second sub-model are B1, B2, B3, and the three levels corresponding to the third sub-model are C1, C2 and C3. Among them, the first level (A1, B1 and C1) is used to characterize how much the change is above, the second level (A2, B2 and C2) is used to characterize whether the change is within the normal range, and the third level (A3, B3 and C3) is used to characterize how much below. Except for the output result of the second level (normal), the output results of the other two levels are considered abnormal, that is, the combination of output results of A2, B2 and C2 is normal, and the combination of other output results indicates abnormal changes, which need to be paid special attention and dealt with according to the situation.

[0103] Since the functions of different sub-models are different, in order to improve their predictive ability, these three sub-models can be constructed based on different indicators. That is, the above 103 "based on the target arterial indicator and the target physiological indicator, construct and train a brain metabolic state classification model" can be implemented by the following steps:

[0104] 1031. Based on the first target physiological index and the first target arterial index, construct and train the first sub-model.

[0105] The first target arterial index and the first target physiological index are used as input parameters of the first sub-model, and the first sub-model is trained using training samples. The training samples can be generated based on sample data in a locally constructed biomedical database. The training samples may include: the index value of the first target physiological index of the sample user, the index value of the first target arterial index, and the actual cerebral blood flow category. The index value of the first target physiological index of the sample user and the index value of the first target arterial index can be input into the first sub-model to obtain a predicted cerebral blood flow category; according to the actual cerebral blood flow category and the predicted cerebral blood flow category, the parameters of the first sub-model are optimized. The specific parameter optimization method can be implemented using an existing optimization method, and the embodiments of the present application do not specifically limit this.

[0106] The first sub-model is used to classify the user's cerebral blood flow based on the index value of the user's first target arterial index and the index value of the user's first target physiological index to obtain a cerebral blood flow category.

[0107] 1032. Construct and train the second sub-model based on the second target physiological index and the second target arterial index.

[0108] The second target arterial index and the second target physiological index are used as input parameters of the second sub-model, and the second sub-model is trained using training samples. The training samples can be generated based on sample data in a locally constructed biomedical database. The training samples may include: the index value of the second target physiological index of the sample user, the index value of the second target arterial index, and the actual brain oxygen uptake score category. The index value of the second target physiological index of the sample user and the index value of the second target arterial index can be input into the second sub-model to obtain a predicted brain oxygen uptake score category; according to the actual cerebral blood flow category and the predicted brain oxygen uptake score category, the second sub-model is optimized for parameters. The specific parameter optimization method can be implemented using an existing optimization method, and the embodiments of the present application do not specifically limit this.

[0109] The second sub-model is used to classify the user's brain oxygen uptake score based on the index value of the user's second target arterial index and the index value of the user's second target physiological index to obtain a brain oxygen uptake score category.

[0110] 1033. Based on the third target physiological index and the third target arterial index, construct and train the third sub-model.

[0111] The third target arterial index and the third target physiological index are used as input parameters of the third sub-model, and the third sub-model is trained using training samples. The training samples can be generated based on sample data in a locally constructed biomedical database. The training samples may include: the index value of the third target physiological index of the sample user, the index value of the third target arterial index, and the real brain oxygen metabolic rate category. The index value of the third target physiological index of the sample user and the index value of the third target arterial index can be input into the third sub-model to obtain a predicted brain oxygen metabolic rate category; according to the real cerebral blood flow category and the predicted brain oxygen metabolic rate category, the third sub-model is optimized for parameters. The specific parameter optimization method can be implemented using an existing optimization method, and the embodiments of the present application do not specifically limit this.

[0112] The third sub-model is used to classify the user's brain oxygen metabolic rate based on the index value of the user's third target arterial index and the index value of the user's third target physiological index to obtain a brain oxygen metabolic rate category.

[0113] Among them, the first target physiological index is an index related to the cerebral blood flow among the target physiological indexes, the second target physiological index is an index related to the oxygen uptake fraction among the target physiological indexes, and the third target physiological index is an index related to the brain oxygen metabolic rate among the target physiological indexes; the first target arterial index is an index related to the cerebral blood flow among the target arterial indexes, the second target arterial index is an index related to the oxygen uptake fraction among the target arterial indexes, and the third target arterial index is an index related to the brain oxygen metabolic rate among the target arterial indexes.

[0114] In the embodiments of the present application, each sub-model may include but is not limited to: Naive Bayes, support vector machine, K nearest neighbor, logistic regression, decision tree, random forest and other models.

[0115] It should be noted that, in order to describe the brain state more precisely, the above-mentioned brain metabolic state classification model can classify the brain metabolic state for each brain region. In other words, the required index value can be collected for each brain region, and then the index value can be input into the model to obtain the classification result of the brain region.

[0116] Figure 2 This is a schematic diagram of the structure of the brain metabolic state classification device provided in the embodiment of the present application. Figure 2 As shown, the device comprises:

[0117] The transceiver module 201 is used to receive the ultrasonic data sent by the ultrasonic data acquisition device, wherein the ultrasonic data includes arterial data acquired by the ultrasonic acquisition device at the carotid artery and vertebral artery of the user;

[0118] The processor 202 is configured to: extract an index value of a target artery index of a user from the ultrasound data; determine input information of a pre-trained brain metabolic state classification model according to the index value of the target artery index and the index value of the target physiological index of the user; input the input information into the brain metabolic state classification model, so that the brain metabolic state classification model classifies the brain metabolic state of the user;

[0119] Wherein, the brain metabolic state classification model is constructed based on the target arterial index and the target physiological index.

[0120] Among them, the brain metabolic state classification device may include but is not limited to: a device integrated in any terminal device such as a smart phone, a tablet computer, a PDA (Personal Digital Assistant), a smart TV, a laptop computer, a desktop computer, a smart wearable device, etc. The device includes a transceiver module for receiving data to be processed (such as the neck blood flow data described below), and a processor for processing the data to be processed. The processor of the device can be mounted in the above-mentioned terminal device. The processor of the device can be integrated with the transceiver module in the same device, or can be integrated in different devices respectively, which is not limited in the embodiments of the present application. Optionally, the device also includes a display module for displaying the processing results of the device (such as classification results), such as a screen in a terminal device.

[0121] In practical applications, the transceiver module of the device can be connected to a communication with an inspection device (or ultrasonic data acquisition device) integrated with an ultrasonic sensor, and the inspection device is arranged on the target evaluation object (e.g., user) side. The inspection device is implemented as a neck inspection device integrated with an ultrasonic sensor, and the neck inspection device is connected to a device integrated with a transceiver module. Of course, in order to adapt to a variety of application scenarios, the connection method between the neck inspection device and the device integrated with a processor can be a wired connection or a wireless connection, such as WiFi, 5G, 4G, Bluetooth, etc.

[0122] In another embodiment, the transceiver module and the processor can be integrated into the same device. After acquiring the data to be processed from the neck examination device, the device analyzes the data to be processed and displays the processing results, such as issuing a voice message for early warning, or displaying the classification results of brain metabolic status of each brain region.

[0123] The ultrasound data may include ultrasound spectrum data, and the ultrasound spectrum data may be analyzed to obtain index values ​​of multiple arterial indexes, and then the index value of the target arterial index may be extracted from the index values ​​of the multiple arterial indexes.

[0124] In addition, the index value of the user's target physiological index can be manually input by the user through the screen, or the transceiver module receives the collected data sent by the physiological data acquisition device (for example: electrocardiogram monitor, blood pressure monitor, etc.), and then the processor extracts the index value of the user's target physiological index from the collected data sent by the physiological data acquisition device.

[0125] The target physiological index and target arterial index to be extracted may be stored in advance in the memory of the classification device. In this way, the subsequent processor may obtain the target physiological index and target arterial index to be extracted from the memory of the classification device, thereby completing the extraction of the index value of the target index.

[0126] In one embodiment, when the brain metabolic state classification model is a complete large model, the index value of the user's target arterial index and the index value of the user's target physiological index can be used as input information of the pre-trained brain metabolic state classification model.

[0127] In another embodiment, when the brain metabolic state classification model is composed of multiple sub-models, the input information of each sub-model can be determined, and then the input information of each sub-model is input into the sub-model to obtain the output result of the sub-model. Finally, the comprehensive module combines the output results of multiple sub-models to determine the final classification result.

[0128] The input information of the first sub-model may include: the index values ​​of the first target physiological index and the first target arterial index of the user. The input information of the second sub-model may include: the index values ​​of the second target physiological index and the second target arterial index of the user. The input information of the second sub-model may include: the index values ​​of the second target physiological index and the second target arterial index of the user.

[0129] It should be noted here that: for the contents of each step processed by the device provided in the embodiment of the present application that are not fully described in detail, please refer to the corresponding contents in the above embodiments, which will not be repeated here. In addition, in addition to executing the above steps, the device provided in the embodiment of the present application can also execute other parts or all of the steps in the above embodiments, which can be specifically referred to the corresponding contents in the above embodiments, which will not be repeated here.

[0130] Figure 3 This is a schematic diagram of the brain metabolic state classification system provided in the embodiment of the present application. Figure 3 As shown, the system includes: an ultrasonic data acquisition device 31 and a brain metabolic state classification device 32, and the ultrasonic data acquisition device 31 and the brain metabolic state classification device 32 are communicatively connected.

[0131] The brain metabolic state classification device 32 may include a transceiver module 321 and a processor 322. Optionally, the device 32 may also include a display module 323.

[0132] It should be noted here that the specific implementation and interaction of each device in the system provided in the embodiments of the present application can be found in the corresponding contents of the above embodiments, and will not be repeated here.

[0133] An embodiment of the present application also provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a computer, it can implement the steps or functions of the methods provided in the above-mentioned method embodiments.

[0134] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, can implement the steps or functions of the methods provided in the above-mentioned method embodiments.

[0135] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0136] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM (Read Only Memory) / RAM (Random Access Memory), a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0137] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for constructing a brain metabolic state classification model, characterized in that: include: Based on the sample data, determining the correlation between the brain metabolism index and a plurality of physiological indexes and the correlation between the brain metabolism index and a plurality of arterial indexes, wherein the plurality of arterial indexes are related to the carotid artery and the vertebral artery; According to the correlation between the brain metabolic index and the multiple physiological indexes and the correlation between the brain metabolic index and the multiple arterial indexes, a target physiological index and a target arterial index are screened from the multiple physiological indexes and the multiple arterial indexes; Based on the target arterial index and the target physiological index, construct and train a brain metabolic state classification model; The brain metabolic state classification model is used to classify the user's brain metabolic state based on the index value of the user's target arterial index and the index value of the user's target physiological index.

2. The method according to claim 1, characterized in that The brain metabolism index includes one or more of cerebral blood flow, oxygen uptake fraction and brain oxygen metabolic rate.

3. The method according to claim 2, characterized in that The determining, based on the sample data, the correlation between the brain metabolism index and the multiple physiological indexes and the correlation between the brain metabolism index and the multiple arterial indexes includes: Based on sample data in a public biomedical database and / or based on sample data in a locally constructed medical database, the correlation between the brain metabolic index and multiple physiological indexes and the correlation between the brain metabolic index and multiple arterial indexes are determined.

4. The method according to claim 3, characterized in that The multiple arterial indices include vascular inner diameter and vascular elasticity; Based on sample data from public biomedical databases, the correlation between brain metabolic indicators and multiple physiological indicators and the correlation between brain metabolic indicators and multiple arterial indicators are determined, including: Based on sample data in a public biomedical database, the correlation between multiple physiological indicators and cerebral blood flow, the correlation between the inner diameter of the blood vessel and the cerebral blood flow, and the correlation between the vascular elasticity and the cerebral blood flow are statistically analyzed.

5. The method according to claim 3, characterized in that: The arterial indices include: hemodynamic parameters; Based on the sample data in the locally constructed medical database, the correlation between brain metabolic indicators and multiple physiological indicators and the correlation between brain metabolic indicators and multiple arterial indicators are determined, including: Based on sample data in a locally constructed medical database, the correlation between multiple physiological indicators and the oxygen uptake fraction, the correlation between multiple physiological indicators and the brain oxygen metabolic rate, the correlation between the hemodynamic parameters and the oxygen uptake fraction, and the correlation between the hemodynamic parameters and the brain oxygen metabolic rate are statistically analyzed.

6. The method according to any one of claims 2 to 5, characterized in that According to the correlation between the brain metabolism index and the multiple physiological indexes and the correlation between the brain metabolism index and the multiple arterial indexes, a target physiological index and a target arterial index are screened from the multiple physiological indexes and the multiple arterial indexes, including: According to the correlation between the cerebral blood flow and the multiple physiological indices and the correlation between the cerebral blood flow and the multiple arterial indices, a first target physiological indices related to the cerebral blood flow and a first target arterial indices related to the cerebral blood flow are screened from the multiple physiological indices and the multiple arterial indices; According to the correlation between the oxygen uptake fraction and the multiple physiological indices and the correlation between the oxygen uptake fraction and the multiple arterial indices, a second target physiological indices related to the oxygen uptake fraction and a second target arterial indices related to the oxygen uptake fraction are screened from the multiple physiological indices and the multiple arterial indices; According to the correlation between the brain oxygen metabolic rate and the multiple physiological indicators and the correlation between the brain oxygen metabolic rate and the multiple arterial indicators, a third target physiological indicator related to the brain oxygen metabolic rate and a third target arterial indicator related to the brain oxygen metabolic rate are screened from the multiple physiological indicators and the multiple arterial indicators.

7. The method according to claim 6, characterized in that The association relationship includes: the degree of correlation and the credibility of the degree of correlation.

8. The method according to any one of claims 2 to 5, characterized in that in, The brain metabolic state classification model is used to predict the user's brain oxygen metabolism level index based on the index value of the user's target arterial index and the index value of the user's target physiological index, and compare the brain oxygen metabolism level index with the index value ranges of multiple preset brain metabolic state categories to determine the user's brain metabolic state category.

9. The method according to claim 8, characterized in that Also includes: Obtaining from the sample data a cerebral blood flow value, an oxygen uptake score value, and a cerebral oxygen metabolic rate value of a first sample user and a cerebral blood flow value, an oxygen uptake score value, and a cerebral oxygen metabolic rate value of a second sample user, wherein the first sample user is from a case group and the second sample user is from a control group; Determining an index value of a brain oxygen metabolism level index of the first sample user according to the cerebral blood flow value, the oxygen uptake score value, and the brain oxygen metabolism rate value of the first sample user; Determining an index value of a brain oxygen metabolism level index of the second sample user according to the cerebral blood flow value, the oxygen uptake score value, and the brain oxygen metabolic rate value of the second sample user; According to the distribution of the index value of the brain oxygen metabolism level index of the first sample user and the index value of the brain oxygen metabolism level index of the second sample user, the index value range of each of the multiple preset brain metabolic state categories is determined.

10. The method according to any one of claims 2 to 5, characterized in that The brain metabolic state classification model includes: a first sub-model for classifying the user's cerebral blood flow, a second sub-model for classifying the user's oxygen uptake fraction, a third sub-model for classifying the user's brain oxygen metabolic rate, and a comprehensive module; The comprehensive module is used to classify the brain metabolic state of the user according to the classification result of the first sub-model, the classification result of the second sub-model and the classification result of the third sub-model.

11. The method according to any one of claims 2 to 5, characterized in that Based on the target arterial index and the target physiological index, a brain metabolic state classification model is constructed and trained, including: Based on the first target physiological index and the first target arterial index, construct and train the first sub-model; constructing and training the second sub-model based on the second target physiological index and the second target arterial index; constructing and training the third sub-model based on the third target physiological index and the third target arterial index; Among them, the first target physiological index is an index related to the cerebral blood flow among the target physiological indexes, the second target physiological index is an index related to the oxygen uptake fraction among the target physiological indexes, and the third target physiological index is an index related to the brain oxygen metabolic rate among the target physiological indexes; the first target arterial index is an index related to the cerebral blood flow among the target arterial indexes, the second target arterial index is an index related to the oxygen uptake fraction among the target arterial indexes, and the third target arterial index is an index related to the brain oxygen metabolic rate among the target arterial indexes.

12. A brain metabolic state classification device, characterized in that: include: A transceiver module, used for receiving ultrasonic data sent by an ultrasonic data acquisition device, wherein the ultrasonic data includes arterial data acquired by the ultrasonic acquisition device at the user's carotid artery and vertebral artery; A processor, configured to: extract an index value of a target artery index of the user from the ultrasound data; determine input information of a pre-trained brain metabolic state classification model according to the index value of the target artery index of the user and the index value of the target physiological index of the user; input the input information into the brain metabolic state classification model, so that the brain metabolic state classification model classifies the brain metabolic state of the user; Wherein, the brain metabolic state classification model is constructed based on the target arterial index and the target physiological index.

13. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a computer, the method according to any one of claims 1 to 11 can be implemented.

14. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 11 is implemented.

Images