Skeletal muscle spasm-to-contracture evolution rule quantitative analysis method and system and application

Abstract

The invention belongs to the technical field of medical data mining and rehabilitation evaluation, and provides a skeletal muscle spasm-to-contracture evolution rule quantitative analysis method and system and application, and the method comprises the steps: preprocessing spasm contracture time sequence data, and obtaining one-dimensional feature vector data; k-means clustering is carried out, state labeling of spasm and contracture is carried out, and clustering labels are obtained; determining the change rate of the torque characteristic and the adjacent angular velocity characteristic, and generating an eight-bit pseudo-sequential sequence representing the evolution rule from the spasm state to the contracture state in combination with a state transition threshold value determined by an ROC curve; constructing a Markov chain model to calculate a transition probability matrix from a spasm state to a contracture state, and identifying torque as a key driving feature of state evolution through grouping risk ratio; and constructing a dynamic correlation model based on the key driving features to complete quantitative analysis of the evolution rule from skeletal muscle spasm to contracture. According to the method, objective division from skeletal muscle spasm to contracture state can be realized, and the transition probability and key driving factors between the skeletal muscle spasm and the contracture state can be excavated.

Description

Technical Field

[0001] This application relates to the fields of medical data mining and rehabilitation assessment technology, and in particular to methods, systems and applications for quantitative analysis of the evolution of skeletal muscle spasms to contractures. Background Technology

[0002] Skeletal muscle spasticity and contracture are two common types of motor dysfunction in the field of neurorehabilitation. Both manifest as increased resistance during passive muscle stretching, but the sources and mechanisms of this increased resistance differ, and they are related in their pathological evolution. Spasticity is a sensory and motor control disorder caused by damage to upper motor neurons, characterized by intermittent or continuous involuntary muscle activation, and characteristically, increased velocity-dependent stretch reflexes. The pathogenesis of spasticity involves the loss of descending cortical inhibition of the spinal cord stretch reflex circuit, leading to overexcitation of α-motor neurons. Contracture refers to persistent shortening of the muscle-tendon unit, i.e., structural changes in the muscle, reduced elasticity, resulting in decreased passive muscle stretch compliance, restricted joint movement, and no velocity-dependent characteristics. Contracture often follows spasticity, representing a later stage of upper motor neuron damage, such as in stroke, traumatic brain injury, and spinal cord injury. The mechanisms of the two are vastly different, and their treatment pathways also differ significantly. Only accurate assessment and differentiation can lead to effective and precise interventions. In clinical practice, it is often necessary to collect multidimensional physiological indicators from patients, such as general clinical information, musculoskeletal ultrasound data, biomechanical data, and surface electromyography data, for comprehensive evaluation. However, existing technologies have the following problems:

[0003] 1. Data structure limitations: Medical data is mostly cross-sectional data, making it difficult to directly analyze the dynamic evolution of muscle pathological changes;

[0004] 2. Lack of quantitative classification of states: Relying on subjective indicators such as the traditional spasticity MAS scale score, it is impossible to accurately distinguish between spasticity and contracture based on objective quantitative data;

[0005] 3. Difficulty in uncovering evolutionary patterns: The evolution from muscle spasm to contracture is a latent process, and there is a lack of effective mathematical models to quantify its potential pathological change pathways.

[0006] To address the aforementioned issues, recent research has begun exploring new approaches that combine multimodal data with artificial intelligence technologies. For example, patent application CN119833150A discloses a method and system for evaluating spasticity and contractures based on multi-scale data feature fusion. The method includes: collecting biomechanical, surface electromyographic, musculoskeletal ultrasound, and clinical information data from patients with spasticity and contractures after clinical stroke; preprocessing the data, including chain equation multiple imputation, label encoding, normalization, feature concatenation, data transformation, and data augmentation; constructing a deep learning model based on a combination of CNN networks, multi-scale feature fusion, and self-attention mechanisms, including first, second, and third feature extraction modules, a feature fusion module, and a classification module; inputting the preprocessed data into the model for training, optimizing the model parameters through multiple iterations to obtain a deep learning model for spasticity evaluation; inputting the data to be evaluated into the model, and outputting the evaluation results. Although this method innovatively utilizes deep learning models to process multimodal data, its interpretability is poor and it is overly dependent on the amount of data. If applied to the analysis of the evolution from skeletal muscle spasticity to contracture, it cannot provide a clear ranking of feature contributions and it is difficult to clearly show the dynamic evolution path and key nodes from spasticity to contracture.

[0007] For example, patent application CN119007275A discloses an eye-tracking behavior classification method based on Gaussian mixture model and hidden Markov model. The hardware system for implementing the eye-tracking behavior classification method includes a desktop eye tracker, a host computer, and a display screen. The host computer contains Gaussian mixture model and hidden Markov model programs. The eye-tracking behavior classification operation steps are as follows: (1) Collect the raw eye-tracking data of the subject; (2) Preprocess the raw eye-tracking data by grouping the raw eye-tracking data using a clustering method; (3) Establish the Gaussian mixture model and hidden Markov model. (3.1) Establish a Gaussian mixture model; (3.2) Establish a hidden Markov model; (3.2.2) Optimize the initial hidden Markov model using the three or more preprocessed eye movement data through an optimization algorithm to obtain an optimized hidden Markov model; (4) Eye movement behavior classification: Based on the Gaussian mixture model optimized in step (3.1.2) and the hidden Markov model optimized in step (3.2.2), classify the eye movement behavior of the three or more preprocessed eye movement data; Based on the eye movement data at each time point, infer which of the following eye movement behaviors is the current eye movement behavior: fixation, saccade, or tracking eye movement behavior. This method strongly relies on the assumption that the data conforms to a Gaussian distribution, and is based on a parameter-dense model. Its training requires iterative optimization, resulting in high computational costs. It also tends to verify known patterns rather than discover natural grouping and state transition patterns in multimodal data. Summary of the Invention

[0008] In view of the shortcomings of the prior art, the present invention provides a method, system and application for quantitative analysis of the evolution law from skeletal muscle spasm to contracture. It can realize the objective division of skeletal muscle spasm to contracture state, and mine the transition probability and key driving factors between the two through angular velocity pseudo-time series and Markov chain. It is also applicable to the quantitative analysis of the progressive process of muscle dysfunction in cross-sectional medical data.

[0009] To achieve the above and related objectives, the present invention adopts the following technical solution:

[0010] The first aspect of this invention provides a method for quantitative analysis of the evolution from skeletal muscle spasm to contracture, comprising the following steps:

[0011] Step S100: Preprocess the collected spastic contracture time series data to obtain one-dimensional feature vector data;

[0012] Step S200: Perform K-means clustering based on one-dimensional feature vector data, and label the state of spasticity and contracture based on clinical indicators to obtain cluster labels;

[0013] Step S300: Based on clustering labels, determine the torque characteristics and rate of change of adjacent angular velocities in biomechanics, and combine the state transition threshold determined by ROC curves to generate an eight-bit pseudo-time series characterizing the evolution from spasm to contracture.

[0014] Step S400: Based on the eight-bit pseudo-time series, a Markov chain model is constructed to calculate the transition probability matrix between the spasticity and contracture states, and the torque is identified as the key driving feature of state evolution by group risk ratio; a dynamic correlation model is constructed based on the key driving feature to complete the quantitative analysis of the evolution law from skeletal muscle spasticity to contracture.

[0015] Furthermore, in step S100, the preprocessing includes missing value handling, label encoding, dimensionless conversion, and feature reshaping.

[0016] Furthermore, in step S200, K-means clustering based on one-dimensional feature vector data includes: determining the optimal number of clusters based on one-dimensional feature vector data through elbow quantization and silhouette coefficient verification, and then performing K-means clustering.

[0017] Furthermore, in step S300, the torque feature is that the maximum mutual information coefficient between the torque and the cluster label is 0.71.

[0018] Further, in step S300, the generation of an eight-bit pseudo-time series sequence characterizing the evolution from spasm to contracture state by combining the state transition threshold determined by the ROC curve includes: based on clustering labels, using the rate of change of adjacent angular velocity features as an indicator, plotting an ROC curve to determine the optimal threshold; starting from the initial state, traversing each adjacent rate of change of angular velocity features and comparing it with the optimal threshold, determining whether the angular velocity interval is a state continuation or a state transition, and generating an eight-bit pseudo-time series sequence, wherein the continuous determination of two state transitions is used as a stable transition indicator, and the position that satisfies the stable transition indicator is the transition point.

[0019] Further, in step S400, identifying torque as a key driving feature of state evolution through group risk ratio includes: dividing the eight-bit pseudo-time series into a high-torque group and a low-torque group according to the mean torque, and calculating the risk ratio, including:

[0020] (Formula 1),

[0021] In Formula 1, RR represents the risk ratio; P(0→1|high torque) represents the probability that the spastic state will evolve into the contracture state under high torque conditions; P(0→1|low torque) represents the probability that the spastic state will evolve into the contracture state under low torque conditions.

[0022] Furthermore, in step S100, the spastic contracture time-series data includes musculoskeletal ultrasound data, biomechanical data, and surface electromyography data.

[0023] A second aspect of this invention provides a quantitative analysis system for the evolution of skeletal muscle spasms into contractures, comprising:

[0024] The acquisition and preprocessing module is used to preprocess the acquired spastic contracture time series data to obtain one-dimensional feature vector data;

[0025] The clustering module is used to perform K-means clustering based on one-dimensional feature vector data, and to label the state of spasticity and contracture based on clinical indicators to obtain cluster labels;

[0026] The pseudo-time series construction module is used to determine the torque characteristics and rate of change of adjacent angular velocities in biomechanics based on clustering labels, and generate an eight-bit pseudo-time series sequence that characterizes the evolution from spasm to contracture state by combining the state transition threshold determined by ROC curves.

[0027] The Markov chain analysis module is used to construct a Markov chain model based on an octet pseudo-time series to calculate the transition probability matrix between the spasticity and contracture states, and to identify the torque as the key driving feature of state evolution by grouping risk ratio. Based on the key driving feature, a dynamic association model is constructed to complete the quantitative analysis of the evolution law from skeletal muscle spasticity to contracture.

[0028] A third aspect of the present invention provides a computer-readable storage medium storing computer-readable instructions thereon, which, when executed by a computer processor, cause the computer to perform the above-described quantitative analysis method for the evolution of skeletal muscle spasm to contracture.

[0029] A fourth aspect of the present invention provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described quantitative analysis method for the evolution of skeletal muscle spasm to contracture.

[0030] The beneficial technical effects of this invention are as follows:

[0031] This invention collects musculoskeletal ultrasound data, biomechanical data, and surface electromyography data, which can effectively reduce the error in transition probability estimation compared to traditional single-modality data analysis. Through K-means clustering and state labeling, it can identify the intermediate state from spasticity to contracture, avoiding the bias of traditional subjective scoring, and the classification of spasticity and contracture is highly accurate. Through torque characteristics and group risk ratio analysis, it identifies torque changes as a key driving factor for state evolution, providing a quantitative basis for clinical intervention. Furthermore, by using pseudo-temporal construction, it solves the problem of lack of temporal information in cross-sectional data, providing effective input for Markov chain analysis.

[0032] This invention can provide clear diagnostic criteria to help doctors identify the transition from spasticity to contracture at an early stage, and is applicable to the quantitative analysis of the progressive process of muscle dysfunction in cross-sectional medical data.

[0033] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0034] The accompanying drawings, incorporated in and forming part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without inventive effort. In the drawings:

[0035] Figure 1 Flowchart of the quantitative analysis method for the evolution of skeletal muscle spasm to contracture in this application;

[0036] Figure 2 This is a framework diagram of the quantitative analysis system for the evolution of skeletal muscle spasm to contracture in this application;

[0037] Figure 3 A schematic diagram of the structure of a computer system suitable for an embodiment of this application is shown. Detailed Implementation

[0038] Unless otherwise defined, all technical and / or scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should be understood that certain features of the invention (described in the context of separate embodiments for clarity) may also be provided in a single embodiment. Conversely, multiple features of the invention (described in the context of a single embodiment for brevity) may also be provided separately or in any suitable combination or, where appropriate, in any other described embodiment of the invention. Certain features described in the context of various embodiments will not be considered essential features of those embodiments unless the embodiment is inoperable without those elements. The invention is further illustrated below by specific examples; however, it should be noted that the specific process conditions and results described in the embodiments of the invention are merely illustrative and should not be construed as limiting the scope of protection of the invention. All equivalent changes or modifications made in accordance with the spirit and essence of the invention should be covered within the scope of protection of the invention.

[0039] Please see Figure 1 The flowchart of the quantitative analysis method for the evolution of skeletal muscle spasm to contracture in this application is detailed below:

[0040] Step S100: Preprocess the collected spastic contracture time series data to obtain one-dimensional feature vector data.

[0041] Specifically, the data collected in this application comes from 51 cases of spastic contracture patients admitted to the Department of Rehabilitation Medicine of the First Bethune Hospital of Jilin University from 2024 to 2025. These cases cover patients with different causes, severity of illness, and rehabilitation stages, ensuring the diversity and representativeness of the data. All data collection was done with the informed consent of the patients and was reviewed by the hospital's ethics committee to ensure the legality and safety of the data.

[0042] More specifically, the time-series data on spasticity and contractures in this application includes numerical characteristic data, such as musculoskeletal ultrasound data, biomechanical data, and surface electromyography (EMG) data. Musculoskeletal ultrasound data may include muscle thickness, cross-sectional area, echo intensity, fascial sliding characteristics, and muscle fiber angle data; biomechanical data may include joint range of motion, muscle tension, passive resistance torque, muscle stiffness, and spasticity threshold velocity; surface EMG data may include time-domain and frequency-domain characteristic data. The time-series data on spasticity and contractures also includes general clinical information, such as demographic data, medical history, treatment history, accompanying symptoms, and functional assessment grading data; it also includes category labels, marking each sample with "spasticity" or "contraction," based on clinical diagnostic criteria evaluated by multidisciplinary experts.

[0043] More specifically, the preprocessing in this application includes missing value handling, label encoding, dimensionless standardization, and feature reshaping. Preprocessing improves data integrity, enhances feature interpretability, eliminates dimensional differences, and optimizes data structure, thereby improving the performance and feasibility of machine learning models. Specifically, missing value handling includes using mean imputation or multiple imputation methods (MICE) for assessment indicators with a small number of missing values. For example, for missing biomechanical test values, missing values ​​can be predicted based on the average strength value of patients in the same group or by establishing a regression model. Label encoding includes label encoding for categorical variables, such as gender, MAS, and MTS assessment indicators, making them effectively processed by machine learning models. Dimensionless standardization includes using Z-score standardization to eliminate dimensions for continuous variables, including:

[0044] (Formula 2),

[0045] In Formula 2, μ is the mean, σ is the standard deviation, x is the original data, and x' is the standardized value.

[0046] More specifically, the feature reshaping of this application includes that the biomechanical data and surface electromyography data of this application are various observations measured at 8 angular velocities (from slow to fast), with each 8 rows of data forming a sample. In order to facilitate model analysis, the long format is converted into a wide format, that is, there are multiple observations in one row, and each observation corresponds to one column. The advantage is that it enhances the flexibility and visualization of the data, and facilitates various statistical analysis methods.

[0047] Step S200: Perform K-means clustering based on one-dimensional feature vector data, and label the state of spasm and contracture based on clinical indicators to obtain cluster labels.

[0048] More specifically, in this application, the one-dimensional eigenvector data is a feature matrix, and this feature matrix is ​​used as input for elbow quantization. The distortion index DI is calculated when k=2~6, and the elbow point at the abrupt change in slope is selected, including:

[0049] (Formula 3),

[0050] In Formula 3, C i For the i-th cluster, μ i It serves as the cluster center.

[0051] Furthermore, this application verifies the profile coefficients and calculates the profile coefficients SC when K=2, including:

[0052] (Formula 4),

[0053] In Formula 4, a i b is the distance from the sample to the same cluster.i The distance from the sample to the nearest heterogeneous cluster.

[0054] Furthermore, this application iterates through all samples, repeats the above steps, calculates the silhouette coefficient for each sample, and then averages the silhouette coefficients of all samples to obtain the overall silhouette coefficient for the current number of clusters K. The value of this overall silhouette coefficient is the indicator for evaluating the current clustering effect (K value). The above steps are repeated to cluster multiple candidate K values ​​and calculate their corresponding overall silhouette coefficients. The overall silhouette coefficients corresponding to different K values ​​are compared. The value of SC ranges from -1 to 1. The closer SC is to 1, the better the clustering effect, meaning that samples within clusters are compact and samples between clusters have high separation. An SC of around 0 indicates cluster overlap and low discriminative power. A negative SC value (close to -1) indicates that samples are misassigned, with the distance to different clusters being closer than to samples within the same cluster. Therefore, the optimal number of clusters in this application is the value that maximizes or is closest to 1 for the overall silhouette coefficient SC.

[0055] More specifically, after running the K-means algorithm, this application assigns a cluster label to each sample, such as cluster 0 and cluster 1. It then examines the samples in clusters 0 and 1 respectively, and calculates whether their clinical indicators meet preset medical standards. For samples in cluster 0, the percentage of samples in this cluster that simultaneously meet both of the following conditions is calculated: Modified Ashworth Scale (MAS) rating ≤ 2 (representing moderate to low muscle tone) and passive range of motion (PROM) of the ankle joint ≥ 30° (representing mild limitation of joint mobility). The percentage of these samples in the total number of samples in cluster 0 is calculated. For samples in cluster 1, the percentage of samples in this cluster that simultaneously meet both of the following conditions is calculated: Modified Ashworth Scale (MAS) rating ≥ 3 (representing severe to high muscle tone) and passive range of motion (PROM) of the ankle joint ≤ 30° (representing severe limitation of joint mobility). The percentage of these samples in the total number of samples in cluster 1 is calculated.

[0056] More specifically, based on the above statistical results, this application performs state labeling: if in cluster 0, the proportion of samples that simultaneously meet the conditions of MAS ≤ 2 and PROM ≥ 30° is ≥ 85%, then the cluster is labeled as "spastic state (C0)"; if in cluster 1, the proportion of samples that simultaneously meet the conditions of MAS ≥ 3 and PROM ≤ 30° is ≥ 85%, then the cluster is labeled as "contraction state (C1)". After the above state labeling, each sample in this application not only has a cluster label (0 or 1), but also a clear clinical state label (spasticity or contracture). This labeled dataset can be used for subsequent model training or analysis.

[0057] Step S300: Based on clustering labels, determine the torque characteristics and rate of change of adjacent angular velocities in biomechanics, and combine the state transition threshold determined by the ROC curve to generate an eight-bit pseudo-time series characterizing the evolution from spasm to contracture.

[0058] Specifically, the torque feature of this application is that the maximum mutual information coefficient between the torque and the cluster label is 0.71.

[0059] Specifically, this application combines the state transition threshold determined by the ROC curve to generate an eight-bit pseudo-time series that characterizes the evolution from spasm to contracture state. This includes: based on clustering labels, using the rate of change of adjacent angular velocity characteristics as an indicator, plotting an ROC curve to determine the optimal threshold; starting from the initial state, traversing each adjacent rate of change of angular velocity characteristics and comparing it with the optimal threshold, determining whether the angular velocity interval is a state continuation or a state transition, and generating an eight-bit pseudo-time series. Here, the continuous determination of two state transitions is used as a stable transition indicator, and the position that satisfies the stable transition indicator is the transition point.

[0060] More specifically, this application calculates features and clustering results using the maximum mutual information coefficient (MIC). A biomechanically relevant feature, torque (MIC=0.71), is selected as the core feature for generating state sequences.

[0061] More specifically, the formula for calculating the characteristic rate of change of adjacent angular velocities in this application is as follows:

[0062] (Formula 5),

[0063] In formula 5, △M i M represents the characteristic rate of change of adjacent angular velocities. i Let ω be the eigenvalue corresponding to the i-th angular velocity. i This represents the angular velocity value.

[0064] More specifically, this application uses clustering labels and the rate of change of the seven adjacent angular velocities generated by the eight angular velocities in the sample as indicators to plot ROC curves, including: traversing threshold T = 0.1-1.0, step size 0.01, and calculating the true positive rate (TPR) and false positive rate (FPR) corresponding to each T, as shown in the following formula:

[0065] (Formula 6),

[0066] In Formula 6, TP represents ΔM in sample C1. i The number of samples ≥T, FN is the number of samples with ΔM in C1. i < The number of T, FP is the ΔM in the C0 sample i The number of samples ≥T, where TN is the number of samples with ΔM in sample C0. i < The number of T.

[0067] Furthermore, based on the aforementioned true positive rate (TPR) and false positive rate (FPR), the Youden index J is calculated. The maximum value of J is selected as the optimal threshold T.

[0068] Furthermore, this application uses the state S1 at the first angular velocity point (slowest speed) as the initial state, determined by the cluster label, and sequentially judges the characteristic change rate ΔM of each adjacent angular velocity starting from the first angular velocity interval. i Relationship with the optimal threshold T: If ΔM i If the value is less than T, then it is determined that no state transition has occurred within this angular velocity interval, i.e., S. i+1 =S i Mark the angular velocity interval as the state duration; if ΔM i If the velocity is ≥T, then a state transition has occurred within this angular velocity interval, i.e., S. i+1 =1-S i The angular velocity interval is marked as a state transition. Here, two consecutive ΔM... i ≥T serves as a stable transition marker, and the position satisfying this condition is the transition point. Following the above rules, starting from S1, S2, S3, S4..., S8 are generated step by step, ultimately forming a binary state sequence of length 8, i.e., an eight-bit pseudo-time sequence (S1-S8, S...). i (∈{0,1}).

[0069] Step S400: Based on the eight-bit pseudo-time series, a Markov chain model is constructed to calculate the transition probability matrix between the spasticity and contracture states, and the torque is identified as the key driving feature of state evolution by group risk ratio; a dynamic correlation model is constructed based on the key driving feature to complete the quantitative analysis of the evolution law from skeletal muscle spasticity to contracture.

[0070] Specifically, identifying torque as a key driving feature of state evolution through group risk ratio includes: dividing the eight-bit pseudo-time series into high-torque and low-torque groups based on the mean torque, and calculating the risk ratio, including:

[0071] (Formula 1),

[0072] In Formula 1, RR represents the risk ratio; P(0→1|high torque) represents the probability that the spastic state will evolve into the contracture state under high torque conditions; P(0→1|low torque) represents the probability that the spastic state will evolve into the contracture state under low torque conditions.

[0073] More specifically, this application takes the eight-bit pseudo-temporal sequence generated from all the above samples as input and defines two spatial states: state 0 (S=0), representing spasm; and state 1 (S=1), representing contracture. Therefore, the state space is represented as S={0 (spasm), 1 (contraction)}.

[0074] More specifically, this application constructs a Markov chain model and calculates the probability P(i→j) of transitioning from any state i to state j, including the following steps:

[0075] (1) Transition count: Traverse all eight-bit pseudo-time series sequences and count the transition count of adjacent states. The formula is as follows:

[0076] (Formula 7),

[0077] In formula 7, The function is an indicator function, which is 1 when the condition is true. N(i→j) is the total number of transitions from state i to state j.

[0078] (2) Calculation of transition probability:

[0079] (Formula 8).

[0080] (3) Output transition probability matrix M: where P(0→1) represents the probability of the spasm state evolving into the contracture state, and P(1→0) represents the probability of the contracture state reversing.

[0081] (Formula 9).

[0082] More specifically, the essence of spasticity is abnormally high muscle tone. Sustained abnormal muscle tone generates abnormal joint torque, forcing the joint into a specific position (e.g., upper limb flexion, lower limb extension). This continuous, unbalanced torque causes mechanical stress stimulation. Sustained high torque load generates abnormal mechanical stress on soft tissues such as muscles, tendons, and joint capsules, inducing adaptive changes. To adapt to this continuous stress, connective tissue proliferates and degenerates, and the muscles themselves undergo disuse atrophy and fibrosis. These structural changes are the physical basis of contractures. Therefore, identifying the torque is crucial in this application. After identifying the torque, this application constructs a dynamic correlation model to quantify the quantitative relationship between torque and other parameters, such as the changes in torque with time (duration, frequency of attacks), load (spasticity intensity), and velocity (stretching speed), and correlates it with other indicators (e.g., electromyographic signals, muscle tone values, joint range of motion), thereby completing a quantitative analysis of the evolutionary pattern from skeletal muscle spasticity to contracture.

[0083] The present invention will be described in detail below through specific examples and embodiments. It should also be understood that the following embodiments are only for specific illustration of the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-essential improvements and adjustments made by those skilled in the art based on the above description of the present invention are within the scope of protection of the present invention. The specific process parameters, etc., in the following examples are merely examples within a suitable range; that is, those skilled in the art can make appropriate selections within the appropriate range based on the description herein, and are not intended to be limited to the specific values ​​in the examples below.

[0084] Example 1

[0085] (1) Preprocess the collected spasticity and contracture time series data:

[0086] 1) Handling missing values: For a small number of missing assessment indicators, mean imputation or multiple imputation methods based on MICE can be used. For example, for missing biomechanical test values, the missing values ​​can be predicted based on the average strength value of patients in the same group or by establishing a regression model.

[0087] 2) Label and encode categorical variables, such as gender, MAS and MTS evaluation indicators.

[0088] 3) For continuous variables, such as torque, average work done, and shear wave velocity, Z-score standardization is used to eliminate dimensions.

[0089] 4) Integrate the processed features to form a unified feature matrix. Biomechanical data and surface electromyography data consist of observations measured at 8 angular velocities, with each sample consisting of 8 rows of data. To facilitate model analysis, the long format is converted to a wide format, meaning that multiple observations exist in one row, and each observation corresponds to one column. The advantages are: enhanced data flexibility and visualization, and ease of use for various statistical analysis methods.

[0090] (2) K-means clustering:

[0091] 1) Determine the optimal number of clusters using the elbow method: When k=2, DI=125.6; when k=3, DI=98.2 (the slope changes from -40 to -27, and the "elbow point" is k=2).

[0092] Silhouette coefficient: When k=2, SC=0.73 (>0.65), verifying the effectiveness of clustering.

[0093] 2) Label the clustering results with spasms and contractures to obtain cluster labels:

[0094] (2.1) C0 (spasticity group): MAS score 1.8±0.3;

[0095] (2.2) C1 (contracture group): MAS score 3.2±0.4;

[0096] (2.3) In C0 (85.4%), the diagnosis was “spasm”, and in C1 (88.6%), the diagnosis was “contraction”.

[0097] (3) Pseudo-time sequence construction:

[0098] 1) Core feature selection: Calculate the MIC of each feature and the clustering results: extensor torque (0.71) > RMS value (0.62) > MAS score (0.58), and select extensor torque as the core feature.

[0099] 2) Calculation of characteristic change rate: C0 is calculated for a single sample as follows:

[0100] M=[5.2,5.8,6.3,6.7,7.0,7.2,7.5,7.7] (corresponding to ω1-ω8)

[0101] △M i =[5×0.6,5×0.5,...,5×0.2]= [3.0,2.5,2.0,1.5,1.0,1.5,1.0].

[0102] 3) Determination of ROC curve and optimal threshold T:

[0103] Input: 7 × 92 = 644 ΔM values ​​from 92 samples, labeled C0→0 and C1→1;

[0104] Traverse T = 0.5 - 5.0 (step size 0.1) and calculate TPR and FPR:

[0105] When T=2.0, TPR=0.78, FPR=0.22, J=0.56;

[0106] When T=2.5, TPR=0.82, FPR=0.14, J=0.68 (maximum);

[0107] The optimal threshold T was determined to be 2.5 (unit: Nm / rad / s). The MIC of each feature and the clustering result was calculated: torque (0.71) > RMS value (0.62) > MAS score (0.58), and torque was selected as the core feature.

[0108] 4) Generation of 8-bit pseudo-time series sequences:

[0109] Rule: Two consecutive △M i When T ≥ T, transition to state 1 from the first position;

[0110] For sample C1, ΔM = [2.3, 2.6, 2.8, 3.0, 2.9, 2.7, 2.5]; ΔM2 = 2.6 ≥ 2.5, ΔM3 = 2.8 ≥ 2.5, meaning two consecutive ΔM values...i ≥T satisfies the condition, and the transition point is i=2; the eight-bit pseudo-time sequence is: [0,0,1,1,1,1,1,1].

[0111] (4) Implementation and results of the Markov chain model:

[0112] 1) Transition Count Statistics: Traverse the 8-bit pseudo-time series of 92 samples and count the number of transitions between adjacent states:

[0113] N(0→0) = 320 times, N(0→1) = 160 times;

[0114] N(1→0) = 20 times, N(1→1) = 420 times.

[0115] 2) Calculation of the transition probability matrix:

[0116] (Formula 10).

[0117] 3) Key driving feature identification:

[0118] High torque group (torque ≥ 8): P(0→1) = 0.51;

[0119] Low torque group (torque < 8): P(0→1) = 0.18;

[0120] RR=0.51 / 0.18≈2.83, and "extensor torque ≥8" was identified as the key feature.

[0121] Based on the above key features and combined with a dynamic correlation model, a quantitative analysis of the evolutionary pattern from skeletal muscle spasm to contracture was completed.

[0122] Please see Figure 2 The above is a framework diagram of the quantitative analysis system for the evolution of skeletal muscle spasm to contracture, as described in this application, including:

[0123] The acquisition and preprocessing module 210 is used to preprocess the acquired spastic contracture time series data to obtain one-dimensional feature vector data;

[0124] Clustering module 220 is used to perform K-means clustering based on one-dimensional feature vector data, and to label the state of spasticity and contracture based on clinical indicators to obtain cluster labels;

[0125] The pseudo-time series construction module 230 is used to determine the torque characteristics and rate of change of adjacent angular velocity characteristics of biomechanics based on clustering labels, and generate an eight-bit pseudo-time series sequence that characterizes the evolution law from spasm to contracture state by combining the state transition threshold determined by ROC curve.

[0126] The Markov chain analysis module 240 is used to construct a Markov chain model based on an eight-bit pseudo-time series to calculate the transition probability matrix between the spasticity and contracture states, and to identify the torque as the key driving feature of state evolution by group risk ratio; based on the key driving feature, a dynamic association model is constructed to complete the quantitative analysis of the evolution law from skeletal muscle spasticity to contracture.

[0127] It should be noted that the quantitative analysis system for the evolution of skeletal muscle spasm to contracture provided in the above embodiments and the quantitative analysis method for the evolution of skeletal muscle spasm to contracture provided in the above embodiments belong to the same concept. The specific operation methods of each module and unit have been described in detail in the method embodiments and will not be repeated here. In practical applications, the quantitative analysis system for the evolution of skeletal muscle spasm to contracture provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.

[0128] Embodiments of this application also provide a computer device, including: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the computer device enables the quantitative analysis method for the evolution of skeletal muscle spasm to contracture provided in the above embodiments.

[0129] Figure 3 A schematic diagram of the structure of a computer system suitable for an embodiment of this application is shown. It should be noted that... Figure 3 The computer system 300 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0130] like Figure 3As shown, the computer system 300 includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage section 308 into a random access memory (RAM) 303, such as performing the methods described in the above embodiments. Various programs and data required for system operation are also stored in the RAM 303. The CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304. The following components are connected to the I / O interface 305: an input section 306 including a keyboard, mouse, etc.; an output section 307 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (local area network) card, modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A driver 310 is also connected to the I / O interface 305 as needed. Removable media 311, such as disks, optical discs, magneto-optical discs, semiconductor memories, etc., are installed on drive 310 as needed so that computer programs read from them can be installed into storage section 308 as needed.

[0131] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer tool programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 309, and / or installed from removable medium 311. When the computer program is executed by central processing unit (CPU) 301, it performs various functions defined in the system of this application.

[0132] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, flash memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. Computer programs contained on computer-readable media can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0133] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0134] The units described in the embodiments of this application can be implemented by tools or by hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the unit itself.

[0135] Another aspect of this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a computer's processor, causes the computer to perform the quantitative analysis method for the evolution of skeletal muscle spasms to contractures as described above. This computer-readable storage medium may be included in the computer device described in the above embodiments, or it may exist independently and not assembled into the computer device.

[0136] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the quantitative analysis method for the evolution of skeletal muscle spasms to contractures provided in the various embodiments above.

[0137] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for quantitatively analyzing the evolution rule from skeletal muscle spasm to contracture, characterized in that, The method comprises the following steps: Step S100, pre-processing the collected spasm and spasticity timing data to obtain one-dimensional feature vector data; Step S200, K-means clustering based on the one-dimensional feature vector data, and state labeling of spasm and spasticity based on the clustering results and clinical indicators to obtain clustering labels; Step S300, based on the clustering labels, determining the moment of force feature and the adjacent angular velocity feature change rate of biomechanics, combining the state transition threshold determined by the ROC curve to generate an eight-bit pseudo-time sequence representing the evolution rule of spasm to spasticity state; Step S400, based on the eight-bit pseudo-time sequence, constructing a Markov chain model to calculate the transition probability matrix between spasm and spasticity states, and identifying the moment of force as the key driving feature of state evolution through the grouped risk ratio; Based on the key driving feature, a dynamic correlation model is constructed to complete the quantitative analysis of the evolution rule of skeletal muscle spasm to spasticity.

2. The method of claim 1, wherein, In the step S100, the pre-processing includes missing value processing, label encoding, dimensionless and feature remodeling.

3. The method of claim 2, wherein, In the step S200, K-means clustering based on the one-dimensional feature vector data includes: based on the one-dimensional feature vector data, determining the optimal clustering number through elbow method quantification and contour coefficient verification, and performing K-means clustering.

4. The method of claim 3, wherein, In the step S300, the moment of force feature is the maximum mutual information coefficient of the moment of force and the clustering label, which is 0.

71.

5. The method of claim 4, wherein, In the step S300, combining the state transition threshold determined by the ROC curve to generate an eight-bit pseudo-time sequence representing the evolution rule of spasm to spasticity state includes: based on the clustering labels, taking the adjacent angular velocity feature change rate as an indicator, drawing an ROC curve to determine the optimal threshold; Starting from the initial state, each adjacent angular velocity feature change rate is compared with the optimal threshold to determine whether the angular velocity interval is a state duration or a state transition, and the eight-bit pseudo-time sequence is generated, wherein two consecutive state transitions are taken as stable transition markers, and the position meeting the stable transition marker is a transition point.

6. The method of claim 5, wherein, In the step S400, identifying the moment of force as the key driving feature of state evolution through the grouped risk ratio includes: dividing the eight-bit pseudo-time sequence into a high moment group and a low moment group according to the average moment of force, and calculating the risk ratio, including: (Formula 1), In formula 1, RR represents the risk ratio; P(0→1|high moment) represents the probability of spasm state evolving to spasticity state under the condition of high moment; P(0→1|low moment) represents the probability of spasm state evolving to spasticity state under the condition of low moment.

7. The method of claim 1, wherein, In the step S100, the spasm and spasticity timing data includes musculoskeletal ultrasound data, biomechanical data and surface electromyography data.

8. A system for quantitatively analyzing the evolution of skeletal muscle spasm to contracture, characterized in that, It comprises: A collection and preprocessing module for pre-processing the collected spasm and spasticity timing data to obtain one-dimensional feature vector data; A clustering module for K-means clustering based on the one-dimensional feature vector data, and state labeling of spasm and spasticity based on the clustering results and clinical indicators to obtain clustering labels; A pseudo-time sequence construction module is configured to determine a moment characteristic and an adjacent angular velocity characteristic change rate of biomechanics based on the clustering label, combine a state transition threshold determined by a ROC curve, and generate an eight-bit pseudo-time sequence sequence representing an evolution rule from spasm to contracture state; A Markov chain analysis module is configured to construct a Markov chain model based on the eight-bit pseudo-time sequence sequence to calculate a transition probability matrix between spasm and contracture states, and identify a moment as a key driving characteristic of state evolution through a grouped risk ratio. A dynamic correlation model is constructed based on the key driving characteristic to complete quantitative analysis of the evolution rule from skeletal muscle spasm to contracture.

9. A computer-readable storage medium, characterized in that, A computer readable instruction is stored thereon, and when the computer readable instruction is executed by a processor of a computer, the computer executes the skeletal muscle spasm to contracture evolution rule quantitative analysis method of any one of claims 1 to 7.

10. A computer device comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the skeletal muscle spasm to contracture evolution rule quantitative analysis method of any one of claims 1 to 7 when executing the computer program.

Images