Detection method and system for determining spinal cord injury degree and injury part

Abstract

The invention provides a detection method and system for determining a spinal cord injury degree and an injury part, and relates to the technical field of spinal cord injury detection. User data can be collected in real time through the wearable sensor, real-time monitoring of the physical condition of the user is achieved, time sequence data are collected by setting the fixed interval time, the illness state change of the user can be tracked for a long time, the illness state development trend can be found in time, and a doctor can adjust a treatment scheme according to the illness state change. Besides, the basic framework model and the analysis framework model are trained and optimized by utilizing case information with known results, so that the models can learn a large amount of diagnosis knowledge and experience, interference of data volume and noise can be reduced by a hierarchical analysis processing mode of the basic framework model and the analysis framework model, and the diagnosis efficiency and accuracy of the models are improved.

Description

Technical Field

[0001] This invention relates to the field of spinal cord injury detection technology, specifically to a detection method and system for determining the degree and location of spinal cord injury. Background Technology

[0002] Spinal cord injury is a serious neurological injury that can greatly affect a patient's motor, sensory, and autonomic nervous system functions. Accurately determining the extent and location of the spinal cord injury is crucial for developing subsequent treatment plans, arranging rehabilitation programs, and assessing prognosis.

[0003] In the medical field, the detection of spinal cord injury traditionally relies on imaging techniques such as X-rays, CT scans, and MRI to obtain structural information about the spinal cord. X-rays can be used to initially view the general shape of the spine, CT scans can more clearly show the details of the bones and the structures around the spinal cord, and MRI has high sensitivity to lesions of soft tissues such as the spinal cord, which can help doctors observe the morphology and signal changes of the spinal cord.

[0004] However, current common injury site detection systems are unable to fully capture the dynamic changes in a patient's condition over time, resulting in untimely and inaccurate understanding of the patient's condition development trend. They also cannot comprehensively reflect the user's physical condition from multiple perspectives, affecting the comprehensiveness of the diagnosis and making it difficult to cope with complex and diverse clinical situations. The accuracy and reliability of the diagnosis are greatly reduced. In addition, existing technologies cannot effectively reduce the amount of data and noise interference, resulting in low model diagnostic efficiency and susceptibility to irrelevant data, leading to erroneous diagnoses.

[0005] Based on this, this application proposes a detection method and system for determining the degree and location of spinal cord injury. Summary of the Invention

[0006] (a) Technical problems to be solved In view of the shortcomings of the prior art, the purpose of the present invention is to provide a detection system for determining the degree and location of spinal cord injury, which at least solves one or more of the above-mentioned technical problems existing in the prior art.

[0007] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: Firstly, this application proposes a detection method for determining the degree and location of spinal cord injury, the method comprising: The system acquires the user's current status data, which includes user basic information data, clinical information data, user imaging information data, and physiological indicator time series data. The physiological indicator time series data includes patient muscle activity data and skin temperature change data that are dynamically collected at fixed time intervals. A trained basic framework model is obtained, and the user's current state data is analyzed using the trained basic framework model to filter out key features. The key features include: blood lipid correlation features, image spatiotemporal features, user basic information data, clinical information data, and key clinical abnormal data in physiological indicator time series data. A trained analysis framework model is obtained, and the user's current state data and key features are input into the trained analysis framework model to detect and analyze the degree and location of the user's spinal cord injury.

[0008] In one embodiment, the formula for calculating the fixed time interval is: in, This indicates the set interval time. As an age point, For real-time user age, M represents the age of different users, and M represents the number of different users.

[0009] In one embodiment, the underlying framework model includes: a Gaussian Naive Bayes model, a CNN-LSTM model, and an XGBoost model.

[0010] Preferably, the user's current state data is analyzed using a pre-trained basic framework model to filter out key features, including: Based on the GNB model, the user's terahertz spectrum was analyzed to obtain data on the correlation between terahertz spectrum and blood lipids; Based on user image information data, the spatiotemporal features of the user's image are obtained using a CNN-LSTM model; the spatiotemporal features of the user's image include the user's spatial feature information and temporal change feature information. Based on user basic information data, clinical information data, and time series data of physiological indicators, the XGBoost model is used to obtain key clinical abnormal data of users.

[0011] In one embodiment, the analytical framework model includes a logistic regression algorithm model.

[0012] In one embodiment, the method further includes: Before performing analysis using the aforementioned basic framework model and analytical framework model, the basic framework model and analytical framework model are trained and optimized based on case information with known results.

[0013] In one embodiment, the user image information data includes X-ray data, CT skeletal structure scan data, and MRI spinal soft tissue imaging data.

[0014] In one embodiment, acquiring time-series data of user physiological indicators includes: Wearable sensors that utilize the built-in TDKICM-20649 integrated sensor collect data on changes in the user's muscle state.

[0015] In one embodiment, the method further includes: After obtaining the results of the analysis of the user's spinal cord injury degree and location, the results are compared with the doctor's opinion. If the doctor's opinion is the same as the analysis result, the analysis result is entered into the database as case information for subsequent use. If the doctor's opinion is different from the analysis result, the analysis result is deleted.

[0016] Secondly, this application also proposes a detection system for determining the degree and location of spinal cord injury, the system comprising: The data acquisition module is configured to acquire the user's current status data, which includes user basic information data, clinical information data, user imaging information data, and physiological indicator time series data. The physiological indicator time series data includes patient muscle activity data and skin temperature change data dynamically collected at fixed time intervals. The key feature filtering module is configured to acquire a trained basic framework model, analyze the user's current state data using the trained basic framework model, and filter out key features. The key features include: blood lipid correlation features, image spatiotemporal features, user basic information data, clinical information data, and key clinical abnormal data in physiological indicator time series data. The detection and analysis module is configured to acquire a trained analysis framework model, input the user's current state data and the key features into the trained analysis framework model, and detect and analyze the degree and location of the user's spinal cord injury.

[0017] (III) Beneficial Effects This invention provides a method and system for determining the degree and location of spinal cord injury. Compared with existing technologies, it has the following advantages: This application proposes a detection method for determining the degree and location of spinal cord injury. First, it acquires user current state data, including basic user information, clinical information, imaging data, and time-series physiological indicators. Then, it analyzes this data using a basic framework model to identify key features. Finally, it inputs the user's current state data and key features into the analytical framework model for detailed detection and analysis of the degree and location of spinal cord injury. This technique utilizes a hierarchical analysis approach between the basic and analytical framework models to reduce data volume and noise interference, thereby improving the model's diagnostic efficiency and accuracy. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the detection method for determining the degree and location of spinal cord injury according to the present invention. Figure 2 This is a schematic diagram of the detection system used in Example 2 to determine the degree and location of spinal cord injury. Detailed Implementation To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are described clearly and completely. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] This application provides a detection method for determining the degree and location of spinal cord injury, which solves the problem of low diagnostic efficiency and accuracy of existing detection models for the degree and location of spinal cord injury due to data volume and noise interference.

[0021] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0022] Example 1: Firstly, this invention proposes a detection method for determining the degree and location of spinal cord injury, see [link to relevant documentation]. Figure 1 The method includes: S1. Obtain user's current status data, which includes user basic information data, clinical information data, user imaging information data, and physiological indicator time series data. The physiological indicator time series data includes patient muscle activity data and skin temperature change data dynamically collected at fixed time intervals. S2. Obtain the trained basic framework model, use the trained basic framework model to analyze the user's current state data, and filter out key features. The key features include: blood lipid correlation features, image spatiotemporal features, user basic information data, clinical information data, and key clinical abnormal data in physiological indicator time series data. S3. Obtain the trained analysis framework model, input the user's current state data and the key features into the trained analysis framework model, and detect and analyze the degree and location of the user's spinal cord injury.

[0023] The following is in conjunction with the appendix Figure 1 The following details the implementation process of an embodiment of the present invention, including explanations of the specific steps S1-S3.

[0024] In hospital emergency departments, patients often present with suspected spinal cord injuries due to falls from heights, car accidents, or other traumatic injuries. These patients frequently experience symptoms such as limited lower limb movement and numbness. If the extent and location of the spinal cord injury cannot be quickly and accurately determined, it may delay the optimal time for surgical intervention or rehabilitation treatment, potentially leading to poor neurological function recovery and increasing the risk of lifelong paralysis. In such cases, medical staff can utilize the method proposed in this embodiment to detect the degree and location of the spinal cord injury, assisting in their assessment.

[0025] The detection method proposed in this implementation for determining the degree and location of spinal cord injury includes the following steps: S1. Obtain the user's current status data, which includes user basic information data, clinical information data, user imaging information data, and physiological indicator time series data. The physiological indicator time series data includes patient muscle activity data and skin temperature change data dynamically collected at fixed time intervals.

[0026] The system collects basic information about the patient (i.e., the patient, hereinafter the same), such as age and type of injury, as well as clinical information data such as terahertz spectrum and blood lipids. Simultaneously, it collects user imaging data, including X-ray and CT scans of bone structure and MRI imaging of the spinal cord and soft tissues. X-rays, also known as X-ray beams, are electromagnetic waves with extremely short wavelengths and high energy, belonging to ionizing radiation. CT scans of bone structure use X-ray beams to scan the area being examined. Different tissues absorb X-rays differently as they pass through the body. Detectors receive the X-ray signals passing through the body and convert them into electrical signals. Computers use complex algorithms to process these electrical signals and reconstruct tomographic images of the examined area. MRI, short for Magnetic Resonance Imaging, is a technology that differs from X-rays and CT scans in that it does not emit ionizing radiation and has extremely high resolution for soft tissues (such as the spinal cord, nerves, muscles, and internal organs).

[0027] Simultaneously, wearable sensors are fitted to patients to collect real-time 24-hour data on muscle activity and skin temperature changes, which are then combined into time-series data. When collecting this data, a fixed interval can be set for data collection; for example, data can be collected every n minutes within 24 hours, and all this data can be combined into a time-series dataset. If no specific time interval is set, time-series data is typically collected at 10-minute intervals.

[0028] In one embodiment, to reduce errors caused by individual differences and to acquire user time-series data more scientifically and reliably, a fixed interval is set. This fixed interval is calculated using scientific methods to obtain the time-series data acquisition interval. Specifically: Age information of different users in the group is extracted, and the spinal cord status of different users at different ages is collected. The critical age of spinal cord atrophy is analyzed, and the average critical point of many critical ages is calculated to obtain the age node.

[0029] The formula for calculating age milestones is as follows: in, As an age point, M represents the age of different users, and M represents the number of different users. The formula for calculating the set interval time is as follows: in, Indicates the set interval time; The real-time user age refers to the age of the user currently undergoing spinal cord injury assessment to determine the degree and location of the injury.

[0030] In one embodiment, the wearable sensor contains an integrated sensor of model TDKICM-20649 and is provided with an elastic cloth that can be wrapped around the legs and hands. The TDKICM-20649 integrated sensor is arranged in an X-shape in the elastic cloth and then placed on the user's legs and hands. This allows the wearable sensor to collect data on changes in the muscle state of the user's legs and hands.

[0031] S2. Obtain the trained basic framework model, analyze the user's current state data using the trained basic framework model, and filter out key features. The key features include: blood lipid correlation features, image spatiotemporal features, user basic information data, clinical information data, and key clinical abnormal data in physiological indicator time series data.

[0032] In one embodiment, the underlying framework models include: a Gaussian Naive Bayes (GNB) model, a CNN-LSTM model, and an XGBoost model. Specifically: 1) Obtain the correlation data between terahertz spectroscopy and blood lipids based on the GNB model.

[0033] Terahertz spectra of users are acquired using a terahertz spectrometer, and then blood lipid data is extracted. The characteristic optical responses of blood lipid molecules are captured using terahertz spectroscopy, and the correlation between the two is quantified using data processing and modeling methods. In analyzing the correlation between terahertz spectroscopy and blood lipid data, clinically tested standard blood lipid data are used as labels, and the extracted optical parameters are used as input to train a GNB model. The GNB model establishes a quantitative mapping relationship between terahertz optical parameters and blood lipid concentration by calculating the posterior probability at different blood lipid concentrations, ultimately achieving direct retrieval of blood lipid data through spectroscopy.

[0034] In practice, the time-domain signals of the user's blood lipid samples (such as triglycerides and cholesterol) and blood samples are first collected using a terahertz spectrometer. After preprocessing, key feature data of the spectral absorption coefficient curve are obtained, along with corresponding real label data such as blood lipid components and concentrations from hospital tests. The matched "spectral features - blood lipid labels" data are then divided into training and validation sets. The GNB model is trained using the training set data. The model automatically calculates the mean and variance of each spectral feature under different blood lipid categories to determine the Gaussian distribution parameters corresponding to each category. After training, the terahertz spectral features of an unknown blood sample are input. The model calculates the posterior probability of the sample belonging to different blood lipid states using the Gaussian probability density formula, thereby quantifying the correlation between spectral features and blood lipid data. Finally, it outputs blood lipid-related analysis results, such as blood lipid category and concentration range.

[0035] 2) Based on user image information data, use the CNN-LSTM model to obtain the spatiotemporal features of user images.

[0036] The CNN-LSTM model inputs X-ray and CT scans of the user's skeletal structure, as well as quantitative MRI time-series images, to extract spatial and temporal feature information of the human body. The spatial feature information, in particular, refers to static feature data captured through medical imaging that relates to the position, shape, relative relationships, and structural integrity of the human anatomy. This is crucial for accurately determining the "specific location" and "range" of spinal cord injury, and is not generalized whole-body spatial information, but rather targeted structural data focused on diagnostic and treatment needs. The temporal feature information, in particular, refers to dynamic change data captured through continuous monitoring that reflects the changes in the body's physiological state over time. This is crucial for determining the progression trend, injury stage, and treatment response of spinal cord injury, and differs from static spatial features, focusing on the evolution of the state over time.

[0037] In practice, spatial features are extracted from each frame of X-ray, CT, and MRI images of bone structure scans using three convolutional layers and two pooling layers. These spatial features include the clarity of the gray and white matter boundaries of the spinal cord, the cross-sectional area of ​​the spinal cord, and the signal intensity of the damaged area. Preferably, the kernel sizes of the three convolutional layers are 3×3, 3×3, and 5×5, respectively, and the maximum pooling stride of the two pooling layers is 2×2.

[0038] 3) Based on user basic information data, clinical information data, and time series data of physiological indicators, the XGBoost model is used to obtain key clinical abnormal data of users.

[0039] The user basic information data, clinical information data, and physiological indicator time series data obtained above are integrated, and then key feature screening is performed. The key feature screening is the key clinical abnormal data in the user basic information data, clinical information data, and physiological indicator time series data.

[0040] It should be noted that before identifying abnormal data (i.e., key clinical abnormal data), existing case information will be used to determine the user's clinical information data and physiological indicator time series data under normal conditions, thereby facilitating the judgment of abnormal data.

[0041] In one embodiment, before using the aforementioned basic framework models (GNB model, CNN-LSTM model, and XGBoost model), it is necessary to obtain case information with known results and use the case information with known results to train and analyze the basic framework models to achieve the expected accuracy before putting them into use.

[0042] S3. Obtain the trained analysis framework model, input the user's current state data and the key features into the trained analysis framework model, and detect and analyze the degree and location of the user's spinal cord injury.

[0043] In this step, based on the current status data of the user to be tested, and combined with the key features extracted above, a detailed analysis is conducted on the degree and location of the user's spinal cord injury.

[0044] Specifically, the analysis process for the degree and location of spinal cord injury is as follows: The user's current status data (including user basic information data, clinical information data, user image information data and physiological indicator time series data) collected in advance in S1, as well as the blood lipid correlation features, image spatiotemporal features, and key clinical abnormal data obtained using the basic framework model in S2, are input into the pre-trained analysis framework model. After analysis, the detection results of the degree and location of spinal cord injury are obtained.

[0045] After obtaining the analysis results of the spinal cord injury degree and location from the analytical framework model, the results are presented to the doctor, and the doctor's opinion is obtained. When the doctor's opinion is the same as the result obtained from the analytical framework model, the result information obtained from the analytical framework model will be entered into the framework database and used as case information for subsequent use. When the doctor's opinion is different from the result obtained from the analytical framework model, the result information obtained from the analytical framework model will be deleted.

[0046] In one embodiment, the analytical framework model is a logistic regression algorithm model.

[0047] In one embodiment, the analytical framework model needs to be trained on case information with known results before use, thereby improving the accuracy of the analytical framework model in analyzing the degree and location of spinal cord injuries in users. Only after meeting the doctor's requirements can it be used. Preferably, during training, a training target value is set. The analytical framework model is considered to have completed training only when its accuracy in analyzing case information with known results exceeds the training target value. Users can edit and set the training target value according to their actual needs.

[0048] This completes the entire process of the detection method for determining the degree and location of spinal cord injury as described in this application.

[0049] Example 2: Secondly, the present invention also provides a detection system for determining the degree and location of spinal cord injury, see [link to relevant documentation]. Figure 2 The system includes: The data acquisition module is configured to acquire the user's current status data, which includes user basic information data, clinical information data, user imaging information data, and physiological indicator time series data. The physiological indicator time series data includes patient muscle activity data and skin temperature change data dynamically collected at fixed time intervals. The key feature filtering module is configured to acquire a trained basic framework model, analyze the user's current state data using the trained basic framework model, and filter out key features. The key features include: blood lipid correlation features, image spatiotemporal features, user basic information data, clinical information data, and key clinical abnormal data in physiological indicator time series data. The detection and analysis module is configured to acquire a trained analysis framework model, input the user's current state data and the key features into the trained analysis framework model, and detect and analyze the degree and location of the user's spinal cord injury.

[0050] In one embodiment, the data acquisition module is configured as follows: The user basic information and clinical information acquisition unit is used to collect user basic information data and clinical information data. For example, the user basic information and clinical information acquisition unit can be connected to the hospital information system (HIS), electronic medical record system (EMR), clinical laboratory system (LIS), order system (OES), and nursing information system (NIS) to obtain user basic information data and clinical information data.

[0051] The image acquisition unit is used to connect to relevant acquisition equipment such as X-ray, CT and MRI to acquire the user's image information.

[0052] The dynamic data acquisition unit is used to connect to wearable sensor devices to collect 24-hour muscle activity data and skin temperature change data of users, and combine them into time-series data. The time-series data can be collected at set intervals.

[0053] In one embodiment, the system further includes: The database module is used to store case information with known results.

[0054] The model training and optimization module is used to call case information with known results from the database module to train and optimize the basic framework model and the analysis framework model.

[0055] The results output module is used to compare the results of the detection and analysis of the user's spinal cord injury degree and injury location with the doctor's opinion after obtaining the results. If the doctor's opinion is the same as the analysis result, the analysis result information is entered into the database of the database module as case information for subsequent use, and the detection result is output synchronously; if the doctor's opinion is different from the analysis result, the analysis result is deleted.

[0056] It is understood that the detection system for determining the degree and location of spinal cord injury provided in this embodiment of the invention corresponds to the detection method for determining the degree and location of spinal cord injury described above. The explanations, examples, and beneficial effects of the relevant content can be referred to the corresponding content in the detection method for determining the degree and location of spinal cord injury, and will not be repeated here.

[0057] In summary, compared with existing technologies, it has the following beneficial effects: 1. This application uses wearable sensors to collect real-time data on user muscle activity and skin temperature changes, enabling real-time monitoring of the user's physical condition. Furthermore, by collecting time-series data at set intervals, it is possible to track changes in the user's condition over a long period, promptly identify and monitor the development trend of the condition, and help doctors adjust treatment plans in a timely manner based on changes in the condition.

[0058] 2. This application utilizes a basic framework model and an analytical framework model, and trains and analyzes the model using case information with known results. This allows the model to learn a large amount of diagnostic knowledge and experience. The hierarchical processing method can reduce the amount of data and noise interference, and improve the diagnostic efficiency and accuracy of the model.

[0059] 3. By considering users' age information and combining it with the analysis of the critical age of spinal cord atrophy, this application can set personalized time parameters for users of different ages, which are more in line with the users' own physiological conditions. This can improve the targeting of data collection and analysis. Using such group data to determine age nodes can provide a more reliable basis for subsequent time calculations, reduce errors caused by individual differences, and help to capture spinal cord-related physiological changes more timely and accurately.

[0060] 4. This application demonstrates significant advantages in terms of accuracy, data utilization efficiency, and clinical applicability in spinal cord injury diagnosis through a multi-model collaborative basic framework, a logistic regression analysis framework, and a closed-loop doctor feedback system. By integrating and analyzing the lipid-related data, image spatiotemporal features, and key clinical abnormal data output by the basic framework, the reliability and interpretability of diagnostic results can be improved. Introducing doctor opinions as a result verification standard forms a closed loop of model output → doctor review → database update / result deletion, thus solving the pain point of traditional AI models being detached from clinical practice. It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0061] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method of detecting the degree and site of spinal cord injury, characterized by, The method comprises: obtaining user current state data, the current state data comprising user basic information data, clinical information data, user image information data and physiological index time series data, wherein the physiological index time series data comprises patient muscle activity data and skin temperature change data dynamically collected at fixed time intervals; obtaining a trained basic framework model, analyzing the user current state data using the trained basic framework model, and screening key features, the key features comprising blood lipid correlation features, image space-time features, user basic information data, clinical information data, and clinical key abnormal data in the physiological index time series data; obtaining a trained analysis framework model, inputting the user current state data and the key features into the trained analysis framework model, and detecting and analyzing the user's spinal cord injury degree and injury site.

2. The method of claim 1, wherein the method is used to determine the degree and site of spinal cord injury. The fixed time interval calculation formula is: wherein, represents a set interval time, is an age node, is a real-time user age, is an age of different users, and M is a number of different users.

3. The method of claim 1, wherein the method is used to determine the degree and location of spinal cord injury. The basic framework model comprises a Gaussian naive Bayes model, a CNN-LSTM model, and an XGBoost model.

4. The method of claim 3, wherein the method is used to determine the degree and site of spinal cord injury. The pre-trained basic framework model is used to analyze the user current state data and screen key features, which comprises: based on the GNB model, analyze the user's terahertz spectrum to obtain terahertz spectrum and blood lipid correlation data; based on user image information data, use the CNN-LSTM model to obtain user image space-time features, which include user spatial feature information and time change feature information; based on user basic information data, clinical information data and physiological index time series data, use the XGBoost model to obtain user clinical key abnormal data.

5. The method of claim 1, wherein, The analysis framework model comprises a logistic regression algorithm model.

6. The method of claim 1, wherein the method is used to determine the degree and location of spinal cord injury. The method further comprises: Before analyzing using the basic framework model and the analysis framework model, train and optimize the basic framework model and the analysis framework model based on known case information.

7. The method of claim 4, wherein the method is used to determine the degree and site of spinal cord injury. The user image information data comprises X-ray data, CT bone structure scan data, and MRI spinal cord soft tissue imaging data.

8. The method of claim 1, wherein the method is used to determine the degree and location of spinal cord injury. When obtaining user physiological index time series data, it comprises: use a wearable sensor with a built-in TDKICM-20649 integrated sensor to collect user muscle state change data.

9. The method of claim 1, wherein the method is used to determine the degree and location of spinal cord injury. The method further comprises: After obtaining the user's spinal cord injury degree and injury site detection and analysis result, compare the result with the doctor's opinion, if the doctor's opinion is the same as the analysis result, record the analysis result information in the database as case information for subsequent use, if the doctor's opinion is different from the analysis result, delete the analysis result.

10. A detection system for determining the extent and location of a spinal cord injury, characterized by, The system comprises: a data acquisition module configured to obtain user current state data, the current state data comprising user basic information data, clinical information data, user image information data and physiological index time series data, wherein the physiological index time series data comprises patient muscle activity data and skin temperature change data dynamically collected at fixed time intervals; The key feature screening module is configured to obtain the trained basic framework model, analyze the user current state data by using the trained basic framework model, and screen out key features, including blood lipid related features, image space-time features, user basic information data, clinical information data, and clinical key abnormal data in physiological index time series data; The detection and analysis module is configured to obtain the trained analysis framework model, input the user current state data and the key features into the trained analysis framework model, and detect and analyze the spinal cord injury degree and injury site of the user.

Images