ICU acquired weakness patient muscle function dynamic monitoring system

Abstract

The application relates to the technical field of personal health risk assessment, in particular to an ICU acquired weakness patient muscle function dynamic monitoring system, which comprises a data acquisition module and a data processing module, the data acquisition module is used for acquiring bending angle change data and muscle thickness change data of a muscle to be measured in the process of joint bending and stretching; the data processing module is used for determining the angular velocity deviation feature and the angular acceleration deviation feature of each data point in the bending angle change data, obtaining the motion quality weight of each data point; determining the difference between the muscle thickness of each data point and the standard muscle thickness, obtaining the muscle thickness change deviation feature through the weighting of the motion quality weight; determining the muscle thickness difference of the same bending angle in the bending and stretching process, obtaining the muscle thickness asymmetric change feature through the weighting of the motion quality weight, fusing the two features to obtain an abnormal index of muscle function, and improving the acquisition accuracy of the abnormal index of muscle function.

Description

Technical Field

[0001] This invention relates to the field of personal health risk assessment technology, specifically to a dynamic monitoring system for muscle function in ICU-acquired frailty patients. Background Technology

[0002] ICU-acquired weakness is a decline in muscle strength caused by a combination of critical illness, prolonged immobilization, and neuromuscular dysfunction. Patients with ICU-acquired weakness may experience serious consequences such as difficulty in limb movement. Therefore, monitoring muscle function in ICU-acquired weakness patients is necessary. Current monitoring procedures typically involve analyzing abnormal muscle function indicators in ICU-acquired weakness patients. Physicians then use these abnormal indicators, combined with other relevant patient symptoms, to comprehensively analyze and assess muscle function, achieving dynamic monitoring of muscle function in ICU-acquired weakness patients. Current methods often determine abnormal muscle function indicators by monitoring the muscle state of the target muscle (e.g., the biceps brachii) during joint movement. Specifically, this involves measuring the actual change in muscle thickness during movement and directly determining the abnormal muscle function indicator based on the difference between the actual change and a standard change. A greater difference indicates a greater abnormality. However, this monitoring method is relatively one-sided and simplistic, leading to significant monitoring errors. Summary of the Invention

[0003] To address the technical problem of large monitoring errors in existing methods for monitoring muscle function in ICU-acquired frailty patients, the present invention aims to provide a dynamic monitoring system for muscle function in ICU-acquired frailty patients, the specific technical solution of which is as follows: This invention provides a dynamic monitoring system for muscle function in ICU-acquired frailty patients, comprising: The data acquisition module is used to acquire data on the changes in bending angle of the muscle under test during joint bending and extension, as well as data on the changes in muscle thickness with changes in bending angle. The data processing module, electrically connected to the data acquisition module, is configured as follows: Determine the angular velocity deviation characteristics and angular acceleration deviation characteristics of each data point in the bending angle change data, and thereby obtain the motion mass weight of each data point; The difference between the muscle thickness at each data point in the muscle thickness variation data and the standard muscle thickness is determined. Under the weighting of the exercise quality weight, the deviation characteristics of muscle thickness variation are obtained. The standard muscle thickness is obtained from the standard variation relationship between the bending angle and the muscle thickness. The muscle thickness difference at the same bending angle during bending and extension is determined, and the asymmetric change characteristics of muscle thickness are obtained under the weighting of the motion quality weights. By combining the deviation characteristics of muscle thickness changes and the asymmetric characteristics of muscle thickness changes, abnormal indicators of muscle function are obtained.

[0004] In an exemplary embodiment, the process of obtaining the angular velocity deviation feature includes: The angular velocity of the bending angle at each data point is determined, and the difference between the angular velocity and the preset reference angular velocity is used to obtain the angular velocity deviation characteristic.

[0005] In an exemplary embodiment, the process of obtaining the angular acceleration deviation feature includes: Determine the angular acceleration of the bending angle at each data point; The absolute value of the angular acceleration is compared with the magnitude of a preset reference angular acceleration. When the absolute value of the angular acceleration is less than or equal to the preset reference angular acceleration, the angular acceleration deviation feature is set to 0. When the absolute value of the angular acceleration is greater than the preset reference angular acceleration, the angular acceleration deviation feature is obtained based on the difference between the absolute value of the angular acceleration and the preset reference angular acceleration. The angular acceleration deviation feature is positively correlated with the difference in angular acceleration.

[0006] In an exemplary embodiment, the process of obtaining the preset reference angular velocity includes: determining the median of the angular velocities of each data point as the preset reference angular velocity; The process of obtaining the preset reference angular acceleration includes: determining a preset multiple of the median of the absolute values ​​of the angular accelerations of each data point as the preset reference angular acceleration.

[0007] In an exemplary embodiment, the process of obtaining the standard variation relationship between the bending angle and muscle thickness includes: Within a range where the bending angle is less than a preset low angle threshold, a first target bending angle and a first target muscle thickness are determined. The first target bending angle is obtained by the bending angle where the motion mass weight satisfies a first preset condition, and the first target muscle thickness is obtained by the first target bending angle. Within a range where the bending angle is greater than a preset high angle threshold, a second target bending angle and a second target muscle thickness are determined. The second target bending angle is obtained by the bending angle where the motion mass weight satisfies the second preset condition, and the second target muscle thickness is obtained by the second target bending angle. Linear fitting is performed based on the first target bending angle and the first target muscle thickness, as well as the second target bending angle and the second target muscle thickness, to obtain a fitted straight line, which represents the standard variation relationship between the bending angle and the muscle thickness.

[0008] In an exemplary embodiment, the motion quality weights satisfying a first preset condition include: selecting a first preset number of motion quality weights with the largest values ​​within a range where the bending angle is less than a preset low angle threshold. The process of obtaining the first target bending angle and the first target muscle thickness includes: calculating the average value of the bending angles corresponding to the first preset number of motion mass weights as the first target bending angle; and calculating the average value of the muscle thicknesses corresponding to the first preset number of motion mass weights as the first target muscle thickness. The motion quality weights satisfy the second preset condition, including: within the range where the bending angle is greater than a preset high angle threshold, selecting the second preset number of motion quality weights with the largest values; The process of obtaining the second target bending angle and the second target muscle thickness includes: calculating the average value of the bending angles corresponding to the second preset number of motion mass weights as the second target bending angle; and calculating the average value of the muscle thicknesses corresponding to the second preset number of motion mass weights as the second target muscle thickness.

[0009] In an exemplary embodiment, the process of obtaining the muscle thickness variation deviation feature includes: Based on the muscle thickness residuals and movement quality weights at each data point, a weighted sum of squared residuals of muscle thickness is obtained; the muscle thickness residuals represent the difference between the muscle thickness and the standard muscle thickness. The weighted root mean square error of muscle thickness is obtained based on the weighted sum of squared residuals. The ratio of the weighted root mean square error to the range of muscle thickness variation is calculated to obtain the deviation characteristics of the muscle thickness variation.

[0010] In an exemplary embodiment, the process of obtaining the asymmetric variation characteristics of muscle thickness includes: Based on the target motion quality weights, the differences in muscle thickness at the same bending angle during bending and extension are weighted and summed to obtain the degree of separation of muscle thickness changes during bending and extension; the target motion quality weights are obtained from the motion quality weights at the same bending angle during bending and extension. Determine the cumulative change in target muscle thickness during flexion and extension; The ratio of the degree of separation of the muscle thickness change to the cumulative amount of the target muscle thickness change is calculated to obtain the asymmetric change characteristics of the muscle thickness.

[0011] In an exemplary embodiment, the process of obtaining the cumulative change in target muscle thickness includes: Based on the target motion quality weight, the muscle thickness of each data point during the bending and extension processes is weighted and summed to obtain the first cumulative change in muscle thickness during the bending process and the second cumulative change in muscle thickness during the extension process. The minimum value between the first cumulative change in muscle thickness and the second cumulative change in muscle thickness is selected to obtain the target cumulative change in muscle thickness.

[0012] In one exemplary embodiment, the data processing module further performs the following steps: Obtain multiple abnormal indicators in the time series; Curve fitting is performed on multiple abnormal indicators over time to obtain and visualize the change curves of the abnormal indicators.

[0013] This invention offers the following advantages: It simultaneously collects angle changes and synchronous muscle thickness changes during joint movement, laying the foundation for subsequent analysis of motion smoothness and the dynamic response of muscle thickness with angle; it obtains motion quality weights based on the deviations in angular velocity and angular acceleration during joint angle changes, thereby quantifying different importance for different changes. Through weighted processing of motion quality weights, it ensures that the focus of attention for each data point corresponds to its importance, increasing the focus on key data, reducing the focus on noisy data, and improving the robustness of monitoring; it analyzes the differences between muscle thickness at each data point and standard muscle thickness, and in the context of motion quality... By weighting the magnitude of the movement, the deviation characteristics of muscle thickness changes are obtained, enabling the quantification of the difference between muscle thickness and the standard muscle thickness. Furthermore, the differences in muscle thickness at the same bending angle during joint flexion and extension are analyzed. By weighting the movement quality, the asymmetric change characteristics of muscle thickness are obtained, enabling the quantification of the differences in muscle thickness during different movement processes. Finally, the deviation characteristics of muscle thickness changes and the asymmetric change characteristics of muscle thickness are integrated, which not only retains the sensitivity to changes in muscle thickness relative to the standard, but also supplements the sensitivity to changes in muscle thickness between different movement states. The comprehensive analysis of the two can reduce the detection error caused by fluctuations of a single indicator, and ultimately improve the accuracy of abnormal indicators of muscle function. Attached Figure Description

[0014] Figure 1 This is a block diagram of a dynamic monitoring system for muscle function in ICU-acquired frail patients provided in an embodiment of the present invention; Figure 2 This is a data processing flowchart of the data processing module provided in an embodiment of the present invention. Detailed Implementation

[0015] This embodiment provides a dynamic monitoring system for muscle function in ICU-acquired frailty patients, hereinafter referred to as the system, which is used to acquire abnormal indicators of muscle function in ICU-acquired frailty patients and realize the monitoring of muscle function in ICU-acquired frailty patients.

[0016] like Figure 1 As shown, the system includes a data acquisition module and a data processing module. The data acquisition module and the data processing module are electrically connected. The data acquisition module outputs the acquired relevant data information to the data processing module. The data processing module processes the acquired data information to obtain abnormal indicators of muscle function in ICU-acquired weakness patients. This embodiment does not limit the specific implementation of the data acquisition module and the data processing module. Both can be software modules configured in a data processing chip, such as a CPU (Central Processing Unit), or they can be hardware modules.

[0017] In this embodiment, ICU-acquired weakness patients passively perform flexion and extension exercises on the muscles to be tested. The muscles to be tested, used to determine muscle function, can be selected based on actual needs, such as the biceps brachii or quadriceps femoris. This embodiment uses the biceps brachii, the most commonly used muscle to be tested, as an example. Passive flexion and extension exercises of the biceps brachii are typically performed using rehabilitation equipment. In an exemplary embodiment, the rehabilitation equipment is specifically a Continuous Passive Motion (CPM) device. By monitoring the changes in muscle thickness of the biceps brachii in ICU-acquired weakness patients during flexion and extension of the elbow joint, abnormal indicators of muscle function in these patients are assessed. This embodiment uses one complete flexion and extension exercise as one monitoring process to obtain abnormal indicators of muscle function during that monitoring process. One complete flexion and extension exercise includes one complete flexion and one complete extension of the biceps brachii. The order of these two movements is set according to actual needs; for example, flexion can be performed first, followed by extension. This embodiment determines the minimum and maximum flexion angles of the biceps brachii during joint flexion and extension. The minimum flexion angle is the joint angle of the biceps brachii when the joint is fully extended, typically set to 0°. The maximum flexion angle is the joint angle of the biceps brachii in its maximum flexion state, typically ranging from 120° to 140°; this embodiment uses 130° as an example. Therefore, the initial flexion angle of the biceps brachii flexion process is the minimum flexion angle, and the final flexion angle is the maximum flexion angle; similarly, the initial flexion angle of the biceps brachii extension process is the maximum flexion angle, and the final flexion angle is the minimum flexion angle.

[0018] The continuous passive motion device uses a motor to drive the joint in a preset, slow, painless flexion and extension movement, requiring the patient's muscles to be completely relaxed and without active force. This embodiment is equipped with a biceps brachii joint flexion angle detection device to collect the flexion angle of the biceps brachii during flexion and extension movements. Currently, flexion angle detection devices mainly include optical motion capture systems and inertial measurement units (IMUs), either of which can be selected. The optical motion capture system uses multiple high-speed cameras to capture markers attached to the joint, tracking their three-dimensional spatial positions to calculate the flexion angle. The IMU, worn on the limb, integrates an accelerometer, gyroscope, and magnetometer, and uses a sensor fusion algorithm (such as Kalman filtering) to calculate the orientation of the sensors in three-dimensional space, thus obtaining the flexion angle. This embodiment also includes a biceps brachii muscle thickness detection device to collect the muscle thickness of the biceps brachii during flexion and extension movements. Muscle thickness measurement devices typically come in two types, and either one can be chosen. The first type is a bedside ultrasound device, which uses high-frequency sound waves to generate real-time images of the muscle's cross-section. Image processing algorithms then identify and automatically calculate the vertical distance between the upper and lower interfaces of the muscle to obtain its thickness. The second type is a wearable sensor. This smart device, integrating fabric strain sensors, flexible force / position sensors, or multimodal sensing units, is placed close to the skin to sense changes in the muscle's circumference during contraction and converts this into an electrical signal, thus obtaining muscle thickness. It should be understood that muscle thickness always has a certain value during a complete bending or stretching motion. If the muscle thickness is zero, it indicates a serious anomaly in the muscle thickness data acquisition. In this case, further data processing is discontinued, and an abnormal data acquisition signal is output to facilitate timely investigation of the cause of the anomaly and the acquisition of normal data.

[0019] The data acquisition module is used to acquire data on the changes in flexion angle of the biceps brachii during flexion and extension, as well as data on the changes in muscle thickness as the flexion angle changes.

[0020] During passive biceps flexion and extension exercises in ICU-acquired weakness patients using a continuous passive motion device, the flexion angle and corresponding muscle thickness at each data point (i.e., each sampling moment) during the flexion and extension processes are acquired at a preset sampling frequency. The duration of a single flexion or extension process for ICU-acquired weakness patients is set according to the detection needs. In this embodiment, a longer duration can be set to improve detection accuracy; 10 seconds is used as an example. The flexion angle and muscle thickness are sampled simultaneously, and the sampling frequency is set according to actual needs, such as the duration of a single flexion or extension process. If the duration of a single flexion or extension process is 10 seconds, the sampling frequency can be set to 10 Hz. Therefore, for a complete bending process, the bending angles of each data point during the bending process are sorted chronologically to obtain bending angle change data, and the muscle thickness of each data point during the bending process is sorted chronologically to obtain muscle thickness change data as the bending angle changes. Similarly, for a complete stretching process, the bending angles of each data point during the stretching process are sorted chronologically to obtain bending angle change data, and the muscle thickness of each data point during the stretching process is sorted chronologically to obtain muscle thickness change data as the bending angle changes. It should be understood that the muscle thickness change data is composed of the muscle thickness of each data point; therefore, the muscle thickness change data is composed of multiple discrete muscle thickness values. This embodiment can fit the muscle thickness between every two adjacent data points, so that the muscle thickness change data is a continuous data curve. For any two adjacent data points, a linear fit is performed on the muscle thickness of these two data points to determine the continuous change in muscle thickness between these two data points, thereby obtaining the temporally continuous muscle thickness during the movement. The linear fitting formula in this embodiment is: Where y is the output, x is the input, k is the slope, and b is the intercept. It should be understood that linear fitting is essentially equivalent to linearly interpolating the muscle thickness between two adjacent data points.

[0021] It should be understood that each data point essentially represents a bending angle. Therefore, in this embodiment, a two-dimensional coordinate system can be constructed with the bending angle as the horizontal axis and muscle thickness as the vertical axis. The muscle thickness change data during the bending and extension processes, which vary with the bending angle, can be mapped onto the two-dimensional coordinate system to obtain two muscle thickness change curves. Although the changes in bending angle during the bending and extension processes are exactly opposite, after mapping to the two-dimensional coordinate system, both muscle thickness change curves represent the change in muscle thickness as the bending angle increases.

[0022] It should be noted that during actual data collection, if the minimum bending angle of the bending process is different from that of the extension process, the maximum value of these two minimum bending angles is used as the minimum bending angle for both the bending and extension processes. The bending angle and muscle thickness of the process in which the minimum bending angle is less than the maximum value are then deleted. If the maximum bending angle of the bending process is different from that of the extension process, the minimum value of these two maximum bending angles is used as the maximum bending angle for both the bending and extension processes. The bending angle and muscle thickness of the process in which the maximum bending angle is greater than the minimum value are then deleted. This ensures that the bending angle ranges for the bending and extension processes are the same, facilitating subsequent data processing. For example: if the minimum bending angle during the bending process is 0° and the minimum bending angle during the stretching process is 3° during the actual data collection process, then 3° is taken as the minimum bending angle for both the bending and stretching processes. Bending angles below 3° and their corresponding muscle thicknesses during the bending process are removed, so that the bending process also starts from 3°. If the maximum bending angle during the bending process is 125° and the maximum bending angle during the stretching process is 130° during the actual data collection process, then 125° is taken as the maximum bending angle for both the bending and stretching processes. Bending angles above 125° and their corresponding muscle thicknesses during the stretching process are removed, so that the stretching process ends at 125°.

[0023] The data processing flow of the data processing module is as follows: Figure 2 As shown, it includes the following steps: Step S100: Determine the angular velocity deviation characteristics and angular acceleration deviation characteristics of each data point in the bending angle change data, thereby obtaining the motion mass weight of each data point; Step S200: Determine the difference between the muscle thickness of each data point in the muscle thickness change data and the standard muscle thickness, and obtain the deviation characteristics of muscle thickness change under the weighting of exercise quality weight; Step S300: Determine the difference in muscle thickness at the same bending angle during bending and extension, and obtain the characteristics of asymmetric changes in muscle thickness under the weighting of movement quality. Step S400: Integrate the deviation features of muscle thickness changes and the asymmetric features of muscle thickness changes to obtain abnormal indicators of muscle function.

[0024] The following is a detailed explanation of each step.

[0025] Step S100: Determine the angular velocity deviation characteristics and angular acceleration deviation characteristics of each data point in the bending angle change data, thereby obtaining the motion mass weight of each data point.

[0026] During passive biceps flexion and extension exercises, ICU-acquired weakness patients often exhibit involuntary resistance in their biceps, affecting angular velocity and angular acceleration at various data points and causing deviations in these values. Consequently, the data quality varies across different data points.

[0027] For any process in the bending or stretching process, taking the bending process as an example, the angular velocity deviation feature and angular acceleration deviation feature of each data point in the bending angle change data during the bending process are obtained. The angular velocity deviation feature represents the deviation of angular velocity, and the angular acceleration deviation feature represents the deviation of angular acceleration. Based on these two deviation features, the motion mass weight of each data point is obtained. The motion mass weight represents the data quality of the data point and reflects the reliability of the data point.

[0028] For any data point, the angular velocity deviation characteristic is used to determine the angular velocity of that data point. The angular velocity can be the instantaneous angular velocity calculated through the central difference. For the i-th data point, the angular velocity of the i-th data point is calculated as follows: ; in, This represents the angular velocity of the i-th data point. This represents the bending angle of the (i+1)th data point. This represents the bending angle of the (i-1)th data point. This represents the unit time, i.e., the time length between two adjacent data points. It should be noted that the angular velocities of the first and last data points can be obtained using a forward or backward first-order difference calculation method. Specifically, the angular velocity of the first data point is calculated by subtracting the bending angle of the first data point from the bending angle of the second data point, and then dividing this by the unit time. Similarly, the angular velocity of the last data point is calculated by subtracting the bending angle of the second-to-last data point from the bending angle of the last data point, and then dividing this by the unit time.

[0029] As the bending angle gradually increases during the bending process, the angular velocity at each data point is positive. Conversely, the angular velocity at each data point is negative during the stretching process. The positive or negative value of the angular velocity only indicates its direction. For the angular velocity at each data point during the stretching process, the absolute value needs to be taken to ensure that the angular velocity is positive for ease of data processing.

[0030] This embodiment presupposes a reference angular velocity, which characterizes the overall angular velocity of each data point during the bending process. In an exemplary embodiment, the angular velocities of each data point are sorted from smallest to largest, and the median of these angular velocities is extracted and used as the reference angular velocity. It should be understood that during the passive flexion and extension of the biceps brachii using a continuous passive motion device, the device is in operation. Therefore, each data point has a certain angular velocity, meaning the angular velocity cannot be 0. Even if there are cases where the angular velocity is 0, it is only a very small number of data points. Therefore, the median of the obtained angular velocities cannot be 0. If the median angular velocity is 0, it indicates that the bending process of the ICU-acquired weakness patient is abnormal. The acquired data is discarded, and the ICU-acquired weakness patient undergoes passive flexion and extension again.

[0031] The difference between the angular velocity of each data point's bending angle and the reference angular velocity is determined. Based on this difference, an angular velocity deviation characteristic is obtained. The greater the difference between the angular velocity and the reference angular velocity, the stronger the angular velocity deviation characteristic; the two are positively correlated. In an exemplary embodiment, one method for quantifying the angular velocity deviation characteristic is given below: ; in, This represents the angular velocity deviation characteristic of the i-th data point. Indicates the reference angular velocity. This represents the difference between the angular velocity of the i-th data point and the reference angular velocity, that is, the degree of relative deviation of the angular velocity from the reference angular velocity; This represents a normalization function, such as the Softsign function. Additionally, regarding the relative deviation of the angular velocity from the reference angular velocity, this embodiment can preset an upper limit value for the relative deviation (e.g., 4). When the actual relative deviation exceeds this upper limit value, it indicates that the angular velocity is significantly higher than the reference angular velocity. Therefore, the actual relative deviation is limited to this upper limit value to avoid the subsequent motion mass weights being too small. This yields the angular velocity deviation characteristics for each data point.

[0032] In ICU-acquired weakness patients, both genuine muscle spasms and unexpected events such as equipment movement (i.e., non-physiological noise) can cause fluctuations in angular acceleration. Among these, angular acceleration changes caused by pathological resistance, which fall within the normal physiological frequency and amplitude range, are a key feature that needs to be quantified and preserved; while non-physiological noise produces extremely high-frequency, non-physiological abnormal angular accelerations. Therefore, it is necessary to differentiate between these two situations based on the degree of angular acceleration deviation to obtain angular acceleration deviation characteristics.

[0033] Calculate the absolute value of the angular acceleration at the bending angle of the i-th data point. This embodiment predetermines a reference angular acceleration, which characterizes the overall angular acceleration of each data point during bending. In an exemplary embodiment, the absolute values ​​of the angular accelerations of each data point are sorted from smallest to largest, the median is extracted, and then the product of the median and a preset multiple is calculated. This product is used as the reference angular acceleration. The preset multiple can be obtained through statistical experiments using a large number of historical samples. Based on these experiments, this embodiment provides a recommended value range (e.g., 2~3.5), with a value of 3 as an example. This preset reference angular acceleration serves as the boundary for distinguishing between real pathological resistance and non-physiological noise. It should be understood that in other embodiments, if the median value of the extracted absolute values ​​of angular acceleration is 0, the reference angular acceleration is directly set as the preset angular acceleration limit value, which serves as the boundary for distinguishing between real pathological resistance and non-physiological noise. The preset angular acceleration limit value can be obtained from historical data statistics: under the same bending process duration and sampling frequency, determine the angular acceleration under several real pathological resistance conditions in history, and take the largest angular acceleration as the preset angular acceleration limit value.

[0034] The absolute value of the angular acceleration of the bending angle at the i-th data point is compared with the magnitude of the reference angular acceleration. When the absolute value of the angular acceleration is less than or equal to the reference angular acceleration, it indicates that the angular acceleration fluctuation of the bending angle at the i-th data point originates from the patient's true pathological resistance characteristics. In this case, the angular acceleration deviation characteristic of the i-th data point is set to 0, indicating that the angular acceleration is within the normal range. When the absolute value of the angular acceleration is greater than the reference angular acceleration, it is determined that the bending action of the i-th data point exceeds the physiologically explainable range of intensity and is subject to interference from extremely high-frequency non-physiological noise. The angular acceleration deviation characteristic of the i-th data point is obtained based on the difference between the angular acceleration of the i-th data point and the reference angular acceleration. The larger the difference in angular acceleration, the higher the degree of interference from extremely high-frequency non-physiological noise, and the larger the value of the angular acceleration deviation characteristic. Therefore, the angular acceleration deviation characteristic is positively correlated with the angular acceleration difference. In an exemplary embodiment, the following is a method for quantifying the angular acceleration deviation characteristic: ; in, This represents the angular acceleration deviation characteristic of the i-th data point. This represents the absolute value of the angular acceleration at the i-th data point. Indicates the reference angular acceleration. This indicates the degree of deviation of the absolute value of the actual angular acceleration from the reference angular acceleration. Furthermore, this embodiment can preset an upper limit value (e.g., 4) for the relative deviation of the absolute value of the actual angular acceleration from the reference angular acceleration. When the actual relative deviation exceeds this upper limit value, it indicates that the angular acceleration is significantly higher than the reference angular velocity. In this case, the actual relative deviation is limited to this upper limit value to avoid the subsequent motion mass weight being too small. Thus, the angular acceleration deviation characteristics of each data point are obtained.

[0035] The angular velocity deviation feature and angular acceleration deviation feature of the i-th data point measure the motion quality of the i-th data point from two aspects: velocity deviation and drastic velocity fluctuation (i.e., acceleration deviation), respectively. As analyzed above, a larger value for both the angular velocity and angular acceleration deviation features indicates poorer motion quality. Therefore, based on the angular velocity and angular acceleration deviation features of the i-th data point, a motion quality weight is determined. This motion quality weight is negatively correlated with both the angular velocity and angular acceleration deviation features. Based on this logical analysis, one method for obtaining the motion quality weight is given below: ;in, This represents the motion mass weight of the i-th data point. Thus, the motion mass weights of each data point during the bending process are obtained. Similarly, the motion mass weights of each data point during the extension process are obtained.

[0036] Step S200: Determine the difference between the muscle thickness of each data point in the muscle thickness change data and the standard muscle thickness, and obtain the deviation characteristics of muscle thickness change under the weighting of exercise quality weight.

[0037] Whether in the bending or stretching process, muscle thickness increases with the increase of the bending angle, and this change is usually relatively uniform, close to a linear change. However, in ICU-acquired weakness patients, due to pathological changes such as fibrosis and edema, this uniformity of change is disrupted, causing the muscle thickness change data to deviate from the ideal shape. The degree of deviation reflects the uneven damage to the internal muscle structure.

[0038] Taking the bending process as an example, the standard muscle thickness at various data points in ICU-acquired weakness patients during bending is determined. The standard muscle thickness at each data point is obtained from the standard variation relationship between the bending angle and muscle thickness during the bending process. The standard variation relationship describes the standard change in muscle thickness as the bending angle changes during bending. The following is a specific process for obtaining the standard variation relationship between the bending angle and muscle thickness: First, a starting two-dimensional coordinate point is determined within a small range of bending angles. This starting two-dimensional coordinate point includes a bending angle and a muscle thickness, defined as the first target bending angle and the first target muscle thickness, respectively. In an exemplary embodiment, a low-angle threshold is preset, serving as the dividing point for determining a small bending angle. The range of small bending angles is defined as bending angles less than this low-angle threshold. The specific value of this low-angle threshold can be adaptively set based on the specific value of the minimum bending angle. Typically, the angle difference between the two is approximately 15°. Taking a minimum bending angle of 0° as an example, the low-angle threshold can be set to 15°.

[0039] Within a range where the bending angle is less than the low angle threshold, it is determined whether the motion quality weight of each data point meets a first preset condition. A motion quality weight meeting the first preset condition indicates a higher motion quality weight, thus selecting the bending angle with the higher motion quality weight. In an exemplary embodiment, the motion quality weights of each data point within the range where the bending angle is less than the low angle threshold are sorted from largest to smallest. A first preset number of motion quality weights with the largest values ​​are selected. The first preset number can be one or more; this embodiment uses three as an example. The three motion quality weights with the largest values ​​are then selected to obtain the bending angles corresponding to these three motion quality weights. The average value of these three bending angles is then calculated and used as the first target bending angle. Based on the first target bending angle, the first target muscle thickness is obtained. In an exemplary embodiment, the average value of the muscle thickness corresponding to these three motion quality weights is calculated and used as the first target muscle thickness. Alternatively, if the first preset number is one, the bending angle corresponding to that motion quality weight is directly used as the first target bending angle, and the muscle thickness corresponding to that motion quality weight is used as the first target muscle thickness. The first target bending angle and the first target muscle thickness constitute the starting two-dimensional coordinate point.

[0040] Within a range of values ​​indicating a large bending angle, a cutoff two-dimensional coordinate point is determined. This cutoff point includes a bending angle and a muscle thickness, defined as the second target bending angle and the second target muscle thickness, respectively. In an exemplary embodiment, a high-angle threshold is preset. This high-angle threshold serves as the dividing point for determining a large bending angle, and the range of values ​​indicating a large bending angle is the range where the bending angle exceeds this high-angle threshold. The specific value of this high-angle threshold can be adaptively set based on the specific value of the maximum bending angle. Typically, the angle difference between the two is approximately 15°. For example, if the maximum bending angle is 130°, the high-angle threshold can be set to 115°.

[0041] Within the range where the bending angle is greater than the high angle threshold, it is determined whether the motion quality weight of each data point meets the second preset condition. Meeting the second preset condition also indicates a higher motion quality weight, thus selecting the bending angle with the higher motion quality weight. In an exemplary embodiment, the motion quality weights of each data point within the range where the bending angle is greater than the high angle threshold are sorted from largest to smallest. A second preset number of motion quality weights with the largest values ​​are selected. The second preset number can be one or more; this embodiment uses three as an example. The three motion quality weights with the largest values ​​are then selected to obtain the bending angles corresponding to these three motion quality weights. The average value of these three bending angles is then calculated and used as the second target bending angle. The second target muscle thickness is obtained based on the second target bending angle. In an exemplary embodiment, the average value of the muscle thickness corresponding to these three motion quality weights is calculated and used as the second target muscle thickness. Alternatively, if the second preset number is one, the bending angle corresponding to that motion quality weight is directly used as the second target bending angle, and the muscle thickness corresponding to that motion quality weight is used as the second target muscle thickness. The second target bending angle and the second target muscle thickness constitute the cutoff two-dimensional coordinate point.

[0042] In a two-dimensional coordinate system, a linear fit is performed on the starting and ending two-dimensional coordinate points to obtain a fitted straight line. This fitted straight line represents the standard variation relationship between the bending angle and muscle thickness during the bending process. By inputting the bending angle of each data point into this fitted straight line, the muscle thickness of each data point in the fitted straight line can be obtained, which is defined as the standard muscle thickness of each data point.

[0043] The difference between the muscle thickness at each data point in the muscle thickness variation data and the standard muscle thickness at the same data point is determined. Taking the i-th data point as an example, the difference between the muscle thickness at the i-th data point and the standard muscle thickness at the i-th data point is defined as the muscle thickness residual at the i-th data point. In an exemplary embodiment, the absolute value of the difference between the muscle thickness at the i-th data point and the standard muscle thickness at the i-th data point is calculated as the muscle thickness residual. The larger the muscle thickness residual, the greater the deviation of the muscle thickness from the standard muscle thickness, and the more obvious the deviation feature of muscle thickness variation. Moreover, the larger the motion quality weight of the i-th data point, the more the influence of the muscle thickness residual needs to be positively amplified, so that the contribution of data points with high motion quality is large and the contribution of data points with low motion quality is small. Thus, the deviation feature of muscle thickness variation during bending process is obtained by weighting the motion quality weights of each data point.

[0044] Since the squared residual can avoid the cancellation of positive and negative values ​​and is more sensitive to large deviations, this embodiment provides the following calculation process for the deviation characteristics of muscle thickness changes: Based on the muscle thickness residuals at each data point and the corresponding motion quality weights, the weighted sum of squared residuals of muscle thickness is obtained: ; in, The weighted sum of squared residuals represents the muscle thickness, and I represents the number of data points during the bending process. This represents the muscle thickness at the i-th data point. This represents the standard muscle thickness at the i-th data point. This represents the muscle thickness residual at the i-th data point. This indicates taking the absolute value.

[0045] Then, based on the weighted sum of squared residuals, the weighted root mean square error of muscle thickness is obtained: ; Here, R represents the weighted root mean square error of muscle thickness. Dividing the sum of squared weighted residuals by the sum of all movement quality weights results in a result that reflects the average deviation of muscle thickness, thus eliminating the influence of the number of data points on the deviation. The resulting weighted average squared deviation is then squared to obtain the weighted average deviation magnitude, i.e., the weighted root mean square error. This value represents the actual deviation of muscle thickness from the standard after considering the differences in movement quality among the data points.

[0046] To determine the range of muscle thickness variation during the bending process, in one exemplary embodiment, the difference between the second target muscle thickness and the first target muscle thickness is used as the range of muscle thickness variation. It should be understood that, based on the relationship between the bending angle and muscle thickness, the range of muscle thickness variation is a positive value and cannot be negative or zero. If the range of muscle thickness variation is negative or zero, it indicates abnormal data acquisition or a severe abnormality in the biceps brachii of an ICU-acquired weakness patient. In such cases, no further processing is performed, and a severe abnormality signal is directly output to facilitate timely investigation of the cause of the abnormality.

[0047] Dividing the weighted root mean square error by the range of muscle thickness variation, i.e., calculating the ratio of the weighted root mean square error to the range of muscle thickness variation, yields the muscle thickness variation deviation characteristic. Dividing the weighted root mean square error, which characterizes the weighted average deviation amplitude of muscle thickness, by the total range of muscle thickness variation transforms the absolute average deviation amplitude into a relative distortion proportion relative to the total muscle thickness deformation. This eliminates the influence of the patient's own condition on the muscle thickness variation, resulting in a distortion characteristic representing the degree of non-linear distortion in the shape of the muscle thickness variation curve—the muscle thickness variation deviation characteristic.

[0048] Similarly, for the stretching process, the deviation characteristics of muscle thickness changes during the stretching process are obtained by following the same data processing procedure described above.

[0049] Step S300: Determine the difference in muscle thickness at the same bending angle during bending and extension, and obtain the characteristics of asymmetric changes in muscle thickness under the weighting of movement quality.

[0050] Healthy muscles exhibit minimal viscoelastic loss; the elastic potential energy stored during flexion is almost completely released during extension. Therefore, the changes in muscle thickness during flexion and extension are opposite processes, meaning their curves essentially overlap. However, in ICU-acquired weakness, due to pathological changes and abnormal viscoelasticity, the muscle thickness change curve during extension may not return along its original path, resulting in a significant delay. Consequently, a difference exists between the muscle thickness change curves during flexion and extension, creating an area difference. This area difference quantifies the increase in viscoelastic dissipation; a larger area difference indicates more abnormal components within the muscle and a more severe functional loss.

[0051] This embodiment transforms muscle thickness change data during flexion and extension into a continuous function sequence with the flexion angle as the independent variable. Interpolation and resampling are then performed at the same flexion angle step size (e.g., 2°) to obtain muscle thickness change data for both flexion and extension processes along the same flexion angle sequence dimension. It should be understood that each data point in the flexion and extension processes corresponds to a specific flexion angle. Furthermore, this embodiment can also employ the same spatial interpolation method to interpolate the motion mass weights of each data point along the flexion angle dimension.

[0052] To determine the difference in muscle thickness at the same bending angle during flexion and extension (specifically, the absolute value of the difference in muscle thickness), under normal circumstances, the difference in muscle thickness at the same bending angle is very small or even zero. A larger difference in muscle thickness indicates a more asymmetrical change in muscle thickness during flexion and extension, meaning a more pronounced asymmetrical change in muscle thickness. Furthermore, a higher weighting for motion quality in the data points necessitates amplifying the impact of muscle thickness differences, thus ensuring a larger contribution from data points with high motion quality and a smaller contribution from data points with low motion quality. Therefore, by weighting the data points according to motion quality, the asymmetrical change characteristics of muscle thickness during flexion and extension are obtained. The following is a specific process for obtaining these asymmetrical muscle thickness change characteristics: The j-th bending angle is set to any bending angle. Corresponding data points for the j-th bending angle during the bending process are determined, thereby determining the motion mass weight of the j-th bending angle during the bending process. Similarly, corresponding data points for the j-th bending angle during the extension process are determined, thereby determining the motion mass weight of the j-th bending angle during the extension process. Based on the motion mass weights of the j-th bending angle during the bending and extension processes, the target motion mass weight of the j-th bending angle is obtained. Since the motion mass weights of the j-th bending angle may differ during bending and extension, to ensure the reliability of data processing, this embodiment obtains the minimum value of the motion mass weights of the j-th bending angle during both processes and uses this minimum value as the target motion mass weight of the j-th bending angle.

[0053] Based on the target motion mass weights at each bending angle, the differences in muscle thickness at the same bending angle during bending and extension are weighted and summed to obtain the degree of separation of muscle thickness changes during bending and extension: ; in, This indicates the degree of separation in terms of muscle thickness changes during flexion and extension. This represents the target motion mass weight at the j-th bending angle. This represents the muscle thickness at the j-th bending angle during the bending process. The value represents the muscle thickness at the j-th flexion angle during extension, where J represents the number of flexion angles. The degree of separation between muscle thickness changes during flexion and extension is characterized by the weighted absolute area difference between the muscle thickness change curves during flexion and extension, reflecting the overall degree of separation between the flexion and extension processes.

[0054] Then, the cumulative change in target muscle thickness during the bending and extension processes is determined. Specifically: for the bending process, the area enclosed by the muscle thickness change curve and the horizontal axis is obtained as the first cumulative change in muscle thickness for the bending process. Furthermore, in this embodiment, the first cumulative change in muscle thickness for the bending process is obtained by weighted summation using the target motion mass weights. ; in, This represents the cumulative change in muscle thickness during the bending process.

[0055] For the stretching process, the area enclosed by the muscle thickness change curve and the horizontal axis is obtained as the second cumulative muscle thickness change for the stretching process. Furthermore, in this embodiment, the second cumulative muscle thickness change for the stretching process is obtained by weighted summation using the target motion mass weight. ; in, This indicates the cumulative change in the thickness of the second muscle during the stretching process.

[0056] The cumulative change in target muscle thickness during flexion and extension is used as a benchmark to determine the degree of asymmetry in muscle thickness change separation relative to this benchmark (i.e., asymmetric muscle thickness change characteristics). Therefore, to ensure the reliability of this asymmetry, this embodiment takes... and The minimum value in the range is used as the cumulative change in target muscle thickness during flexion and extension.

[0057] Calculate the ratio of the degree of separation of muscle thickness changes to the cumulative amount of target muscle thickness changes: The results are used as characteristics of asymmetric changes in muscle thickness, where min represents the minimum value function. Due to the degree of separation in muscle thickness changes... The weighted absolute area difference between the muscle thickness change curves representing the flexion and extension processes, and the cumulative change in target muscle thickness representing the total area formed by the muscle thickness change curves, are used to represent the ratio of the deformation hysteresis area of ​​muscle thickness. For muscles with normal elasticity, the muscle thickness change responses during extension and flexion tend to be consistent, resulting in a small or even zero asymmetric change characteristic in muscle thickness. Conversely, as viscous resistance increases, the separated area difference becomes larger, and the value of the asymmetric change characteristic in muscle thickness also increases accordingly.

[0058] Step S400: Integrate the deviation features of muscle thickness changes and the asymmetric features of muscle thickness changes to obtain abnormal indicators of muscle function.

[0059] The deviation characteristics of muscle thickness changes during flexion and extension reflect the degree to which muscle thickness deviates from the standard muscle thickness during these processes, corresponding to uneven damage to the internal muscle structure. This embodiment can calculate the average value of the muscle thickness change deviation characteristics during flexion and extension as the final muscle thickness change deviation characteristic.

[0060] The asymmetric variation in muscle thickness reflects the difference in muscle thickness during flexion and extension, corresponding to abnormal muscle viscoelasticity.

[0061] The deviation characteristics of final muscle thickness change and the asymmetric changes in muscle thickness are fused to obtain an abnormal index of muscle function, representing the overall degree of muscle function impairment. Specifically, the average value of the deviation characteristics of final muscle thickness change and the asymmetric changes in muscle thickness is calculated, and then this average value is normalized (e.g., using the Softsign function). The normalized result is used as the abnormal index of muscle function. The abnormal index of muscle function is output and visualized, for example, displayed on a screen, for viewing, enabling dynamic monitoring of muscle function. The higher the value of the abnormal index of muscle function, the more abnormal the muscle function.

[0062] Through the above process, abnormal indicators of muscle function are obtained for each monitoring process, thus obtaining multiple abnormal indicators of muscle function over time. Then, in this embodiment, a two-dimensional coordinate system is constructed with each monitoring process as the horizontal axis and the abnormal indicators of muscle function as the vertical axis. The multiple abnormal indicators of muscle function over time are mapped to this two-dimensional coordinate system and curve fitting is performed. Specifically, the curve fitting is performed using the least squares method to perform polynomial fitting (the order of the polynomial can be set to 3rd order), thereby obtaining the abnormal indicator change curve. The abnormal indicator change curve is then output and visualized, thus intuitively showing the changes in the abnormal indicators of muscle function of ICU-acquired weakness patients to date.

[0063] It should be noted that all data and information collected in this application have been obtained with full consent and authorization.

Claims

1. A dynamic monitoring system for muscle function in ICU-acquired frailty patients, characterized in that, include: The data acquisition module is used to acquire data on the changes in bending angle of the muscle under test during joint bending and extension, as well as data on the changes in muscle thickness with changes in bending angle. The data processing module, electrically connected to the data acquisition module, is configured as follows: Determine the angular velocity deviation characteristics and angular acceleration deviation characteristics of each data point in the bending angle change data, and thereby obtain the motion mass weight of each data point; The difference between the muscle thickness at each data point in the muscle thickness change data and the standard muscle thickness is determined, and the deviation characteristics of muscle thickness change are obtained under the weighting of the exercise quality weight. The standard muscle thickness is obtained from the standard variation relationship between the bending angle and the muscle thickness. The muscle thickness difference at the same bending angle during bending and extension is determined, and the asymmetric change characteristics of muscle thickness are obtained under the weighting of the motion quality weights. By combining the deviation characteristics of muscle thickness changes and the asymmetric characteristics of muscle thickness changes, abnormal indicators of muscle function are obtained.

2. The ICU-acquired frailty patient dynamic muscle function monitoring system as described in claim 1, characterized in that, The process of obtaining the angular velocity deviation feature includes: The angular velocity of the bending angle at each data point is determined, and the difference between the angular velocity and the preset reference angular velocity is used to obtain the angular velocity deviation characteristic.

3. The ICU-acquired frailty patient dynamic muscle function monitoring system as described in claim 2, characterized in that, The process of obtaining the angular acceleration deviation feature includes: Determine the angular acceleration of the bending angle at each data point; The absolute value of the angular acceleration is compared with the magnitude of a preset reference angular acceleration. When the absolute value of the angular acceleration is less than or equal to the preset reference angular acceleration, the angular acceleration deviation feature is set to 0. When the absolute value of the angular acceleration is greater than the preset reference angular acceleration, the angular acceleration deviation feature is obtained based on the difference between the absolute value of the angular acceleration and the preset reference angular acceleration. The angular acceleration deviation feature is positively correlated with the difference in angular acceleration.

4. The ICU-acquired frailty patient dynamic muscle function monitoring system as described in claim 3, characterized in that, The process of obtaining the preset reference angular velocity includes: determining the median of the angular velocities of each data point as the preset reference angular velocity; The process of obtaining the preset reference angular acceleration includes: determining a preset multiple of the median of the absolute values ​​of the angular accelerations of each data point as the preset reference angular acceleration.

5. The dynamic monitoring system for muscle function in ICU-acquired frailty patients as described in claim 1, characterized in that, The process of obtaining the standard variation relationship between the bending angle and muscle thickness includes: Within a range where the bending angle is less than a preset low angle threshold, a first target bending angle and a first target muscle thickness are determined. The first target bending angle is obtained by the bending angle where the motion mass weight satisfies a first preset condition, and the first target muscle thickness is obtained by the first target bending angle. Within a range where the bending angle is greater than a preset high angle threshold, a second target bending angle and a second target muscle thickness are determined. The second target bending angle is obtained by the bending angle where the motion mass weight satisfies the second preset condition, and the second target muscle thickness is obtained by the second target bending angle. Linear fitting is performed based on the first target bending angle and the first target muscle thickness, as well as the second target bending angle and the second target muscle thickness, to obtain a fitted straight line, which represents the standard variation relationship between the bending angle and the muscle thickness.

6. The ICU-acquired frailty patient dynamic muscle function monitoring system as described in claim 5, characterized in that, The motion quality weights satisfy the first preset condition, including: within the range where the bending angle is less than a preset low angle threshold, selecting the first preset number of motion quality weights with the largest values; The process of obtaining the first target bending angle and the first target muscle thickness includes: calculating the average value of the bending angles corresponding to the first preset number of motion mass weights as the first target bending angle; and calculating the average value of the muscle thicknesses corresponding to the first preset number of motion mass weights as the first target muscle thickness. The motion quality weights satisfy the second preset condition, including: within the range where the bending angle is greater than a preset high angle threshold, selecting the second preset number of motion quality weights with the largest values; The process of obtaining the second target bending angle and the second target muscle thickness includes: calculating the average value of the bending angles corresponding to the second preset number of motion mass weights as the second target bending angle; and calculating the average value of the muscle thicknesses corresponding to the second preset number of motion mass weights as the second target muscle thickness.

7. The ICU-acquired frailty patient dynamic muscle function monitoring system as described in claim 1, characterized in that, The process of obtaining the deviation feature of muscle thickness change includes: Based on the muscle thickness residuals and movement quality weights at each data point, a weighted sum of squared residuals of muscle thickness is obtained; the muscle thickness residuals represent the difference between the muscle thickness and the standard muscle thickness. The weighted root mean square error of muscle thickness is obtained based on the weighted sum of squared residuals. The ratio of the weighted root mean square error to the range of muscle thickness variation is calculated to obtain the deviation characteristics of the muscle thickness variation.

8. The ICU-acquired frailty patient dynamic muscle function monitoring system as described in claim 1, characterized in that, The process of obtaining the asymmetric variation characteristics of muscle thickness includes: Based on the target motion quality weights, the differences in muscle thickness at the same bending angle during bending and extension are weighted and summed to obtain the degree of separation of muscle thickness changes during bending and extension; the target motion quality weights are obtained from the motion quality weights at the same bending angle during bending and extension. Determine the cumulative change in target muscle thickness during flexion and extension; The ratio of the degree of separation of the muscle thickness change to the cumulative amount of the target muscle thickness change is calculated to obtain the asymmetric change characteristics of the muscle thickness.

9. The ICU-acquired frailty patient dynamic monitoring system as described in claim 8, characterized in that, The process of obtaining the cumulative change in the target muscle thickness includes: Based on the target motion quality weight, the muscle thickness of each data point during the bending and extension processes is weighted and summed to obtain the first cumulative change in muscle thickness during the bending process and the second cumulative change in muscle thickness during the extension process. The minimum value between the first cumulative change in muscle thickness and the second cumulative change in muscle thickness is selected to obtain the target cumulative change in muscle thickness.

10. The ICU-acquired frailty patient dynamic muscle function monitoring system as described in claim 1, characterized in that, The data processing module also performs the following steps: Obtain multiple abnormal indicators in the time series; Curve fitting is performed on multiple abnormal indicators over time to obtain and visualize the change curves of the abnormal indicators.

Images