A neuromuscular function assessment system

Abstract

The application discloses a neuromuscular function evaluation system and relates to the field of biomedical engineering.The system comprises a lower limb isometric muscle strength measuring device, an electromyographic signal acquisition device, a synchronous control module and an upper computer system, the upper computer system realizes real-time synchronous acquisition of muscle strength and electromyographic signals through the synchronous control module, and accurately captures neuromuscular activation and response processes.After the collected data is preprocessed, muscle strength peak value, force development rate, neuromuscular activation delay and electromyographic time-frequency domain parameters and various characteristics are extracted.A machine learning model is constructed by using the extracted characteristics, and is used for evaluation of the neuromuscular function state.The application improves the accuracy and automation degree of neuromuscular function evaluation through an objective evaluation mode, and is particularly suitable for function evaluation and rehabilitation monitoring of multiple types of people, such as the elderly, people with neuromuscular diseases, rehabilitation personnel after nerve injury or sports injury, etc.

Description

Technical Field

[0001] This invention relates to the field of biomedical engineering, and in particular to a neuromuscular function assessment system based on the simultaneous measurement of muscle strength and electromyographic signals. Background Technology

[0002] Neuromuscular function is crucial for maintaining human movement, daily life, and overall health. Its core indicators include muscle strength, endurance, coordination, and neuromuscular regulation. When neuromuscular function declines, individuals often exhibit decreased muscle strength, slowed reaction time, and weakened reserve functions, thereby increasing the risk of falls, disability, and death. Frailty, a common geriatric syndrome, stems from the decline of multiple system functions, with decreased muscle strength and reduced neuromuscular regulation being important pathological mechanisms. Studies show that frail individuals often exhibit lower limb muscle weakness and abnormal electromyographic signals; neuromuscular dysfunction can identify and predict frailty. Therefore, objective and accurate assessment of neuromuscular function is of great significance for early screening of frailty in the elderly. Besides frailty, neuromuscular dysfunction is also widely present in other populations. For example, patients with neuromuscular diseases (such as amyotrophic lateral sclerosis, myopathy, peripheral neuropathy, etc.) may exhibit persistent muscle weakness and nerve conduction disorders; people recovering from trauma or surgery often experience slow or incomplete functional recovery due to nerve or muscle damage; and patients with sports injuries may experience muscle weakness, delayed muscle activation, or asymmetry in specific muscle groups. These individuals also require quantitative assessment of neuromuscular function using objective methods to guide rehabilitation programs and monitor treatment effectiveness.

[0003] Currently, commonly used functional assessment methods still rely mainly on scales and clinical tests. However, these methods are highly subjective, have poor repeatability, and are difficult to capture subtle changes in neuromuscular function. Most existing measurement devices can only collect muscle strength or electromyographic signals individually, lacking the ability to collect data synchronously in real time and perform multidimensional analysis. This makes it difficult to reveal the temporal relationship between neuromuscular activation and muscle strength generation, and also makes it difficult to perform detailed analysis of different stages of the muscle strength curve.

[0004] Therefore, there is an urgent need for a neuromuscular function assessment system capable of simultaneously measuring muscle strength and electromyographic signals, and achieving comprehensive, objective, and automated feature extraction and analysis. This system could not only improve the accuracy of frailty screening in the elderly, but also provide a scientific basis for the diagnosis of neuromuscular diseases, functional monitoring of rehabilitation populations, and assessment of sports injuries, meeting the application needs of multiple scenarios such as clinical practice, rehabilitation, and community health management. Summary of the Invention

[0005] To address the issues of asynchronous muscle strength and electromyography signal measurements and limited feature extraction in existing technologies, this invention provides a neuromuscular function assessment system.

[0006] The technical solution of this invention is: a neuromuscular function assessment system, comprising:

[0007] Lower limb isometric muscle strength measuring device, used to measure muscle force signals generated when subjects perform isometric contractions of lower limb joints;

[0008] An electromyography (EMG) signal acquisition device is used to measure multi-channel EMG signals generated by subjects during isometric contractions of the lower limb joints.

[0009] The synchronization control module is used to receive control commands from the host computer system, generate synchronization trigger signals, and control the real-time synchronous acquisition of data from the lower limb isometric muscle strength measuring device and the electromyography signal acquisition device.

[0010] The signal processing and feature extraction module is used to preprocess the acquired muscle strength and electromyography (EMG) signals, and extract and calculate the basic parameters of muscle strength and the characteristics of each stage of muscle strength, the activation start time of EMG signals, the neuromuscular activation delay, the time domain features of EMG, and the frequency domain features of EMG.

[0011] The basic parameters of muscle strength extracted include: extracting peak muscle strength and corresponding time, time to reach 80% of maximum force, and calculating muscle strength symmetry and active antagonism ratio of the knee joint.

[0012] The features extracted from each stage of muscle strength include: dividing the muscle strength curve into stages, calculating the force development rate, maximum rate and corresponding time and stress value in the force development stage, calculating the difference before and after the force maintenance stage, attenuation angle and maintenance ability parameters, and calculating the attenuation rate, maximum attenuation rate and corresponding time and stress value in the force decay stage.

[0013] The machine learning analysis module constructs a machine learning model based on the feature parameters extracted by the signal processing and feature extraction modules, and outputs an objective score of neuromuscular functional status.

[0014] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0015] 1. This invention achieves real-time synchronous measurement of muscle strength and electromyographic signals through a synchronous control module, effectively capturing the dynamic changes in neuromuscular function;

[0016] 2. This invention comprehensively extracts multiple characteristic parameters of muscle strength and electromyography signals, including peak muscle strength, different parameters of the three stages of the force curve, electromyography frequency domain characteristics, and neuromuscular activation delay, to achieve an objective and detailed quantitative assessment of neuromuscular function.

[0017] 3. This invention uses a machine learning model for comprehensive analysis, which can effectively improve the accuracy and efficiency of objective assessment of neuromuscular dysfunction and meet the needs of clinical and community health management. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the evaluation system;

[0019] Figure 2 This is a flowchart of the evaluation method;

[0020] Figure 3 This is a schematic diagram of the measurement location;

[0021] Figure 4 This is a schematic diagram of signal characteristics.

[0022] In the diagram: 1-Lower limb isometric muscle strength measurement device; 2-Electromyography signal acquisition device; 3-Synchronization control module; 4-Host computer system. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0024] According to embodiments of the present invention, such as Figure 1 , Figure 2 , Figure 3 , Figure 4 As shown, a neuromuscular function assessment system.

[0025] Example 1:

[0026] like Figure 1 As shown, the neuromuscular function assessment system includes a lower limb isometric muscle strength measurement device 1, an electromyography signal acquisition device 2, a synchronization control module 3, and a host computer system 4. The host computer system 4 is responsible for sending commands to control the synchronization control module 3, thereby realizing the synchronous acquisition of data from the lower limb isometric muscle strength measurement device 1 and the electromyography signal acquisition device 2. The host computer system 4 includes a signal processing and feature extraction module and a machine learning analysis module.

[0027] The lower limb isometric muscle strength measuring device 1 is used to measure the muscle strength signal generated when the subject performs isometric contraction of the lower limb joints;

[0028] The electromyography signal acquisition device 2 is used to measure the multi-channel electromyography signals generated by the subject during isometric contraction of the lower limb joints;

[0029] The synchronization control module 3 receives instructions from the host computer system 4 and generates a synchronization trigger signal to achieve synchronous acquisition of muscle strength and electromyographic signals.

[0030] The signal processing and feature extraction module is used to preprocess and extract features from the acquired raw muscle strength and electromyography signals.

[0031] The raw muscle strength signal was preprocessed using a fourth-order Butterworth low-pass filter with a cutoff frequency set to 10 Hz, and phase distortion-free filtering was achieved through zero-phase filtering. The raw electromyography (EMG) signal was also preprocessed, including: removing the mean to eliminate the DC bias of the sEMG signal; applying a third-order Butterworth bandpass filter for 10–500 Hz; applying a notch filter to the EMG signal at 50 Hz and its harmonics; and smoothing using a wavelet thresholding denoising method based on the sym8 wavelet basis.

[0032] Extracting and calculating a series of basic parametric features of muscle force signals, including peak value. and corresponding time 80% of maximum force arrival time Muscle strength symmetry Knee joint active antagonism ratio ,in, This represents the maximum value of the signal on the left. This represents the maximum value of the signal on the right. This represents the maximum muscle strength of the relevant muscles during knee extension. This represents the maximum muscle strength of the relevant muscles when the knee joint is flexed.

[0033] The signal processing and feature extraction module, such as Figure 4 As shown, the muscle strength curve is divided into three stages: strength development, strength maintenance, and strength decay. By calculating the first derivative of the preprocessed muscle strength signal and combining it with threshold and stability judgment methods, the muscle strength initiation point A, the initiation point of the strength maintenance stage, the end point of the strength maintenance stage, and the end point of the strength decay stage are automatically identified in the entire muscle strength curve, thus realizing the three-stage division of the muscle strength curve. Among them, the muscle strength initiation point A is defined as the earliest moment when the signal exceeds 3 times the baseline mean and the first derivative is greater than 0.5 kg / s for 1 second; the strength maintenance stage initiation point is defined as the earliest moment after the rising stage, that is, the first derivative is mainly positive in the previous period and the absolute value of the first derivative is less than 0.5 kg / s for 1 second after this point; the strength maintenance stage end point is defined as the earliest moment after the stable stage, that is, the first derivative changes from near zero to a continuous negative trend before and after this point and the derivative is less than -0.5 kg / s for 1 second; the strength decay stage end point is defined as the moment when the signal drops to the baseline mean ± 3 times the standard deviation. The entire division process uses the full curve as a reference to ensure that the three stages of force development, maintenance and decay are continuous and do not overlap.

[0034] Calculate the characteristics of the three stages of the muscle strength curve separately, and calculate the rate of force development during the force development stage. Average strength development rate Maximum speed and corresponding time and stress value ,in, It is a time-varying force value function of the force development stage of the maximum force curve. To Differentiate, It is the force value at the end of the force development stage. It is the time when the development stage of power ends. It is the time corresponding to the muscle force initiation point A.

[0035] The difference before and after the calculation force holding phase attenuation angle The standard deviation between the actual force value and the force fitting curve ,in, It is the force value at the end of the force holding phase. It is the time when the force-holding phase ends. It is a time-varying force function of the force value during the force holding phase of the maximum force curve. It is a fitted curve function of the force value during the force holding phase. ,in , ,in, It is the total number of force values ​​during the force holding phase. It is each force value during the force holding phase. It is the time corresponding to each force value during the force holding phase.

[0036] decay rate during the computational power decay phase Average force decay rate Maximum decay rate and corresponding time and stress value ,in, It is a time-varying force function of the force decay phase of the maximum force curve. To Differentiate, It is the time when the force decay phase ends.

[0037] The activation start time of the electromyographic signal is defined as the electromyographic signal activation point B. By calculating the root mean square (RMS) envelope of the electromyographic signal, the point where it first exceeds "baseline mean + 8 times the standard deviation" is identified as the electromyographic signal activation point B.

[0038] Electromyographic time-domain and frequency-domain characteristics include: calculating the waveform length of electromyographic parameters related to motor unit activation. EMG peak amplitude Root mean square value Average absolute value EMG Points Zero cross Absolute power of CWT coefficients in continuous wavelet transform The change in the sign of the slope of electromyographic parameters related to the calculation and control of coordination ability. Wavelet entropy of CWT coefficients in continuous wavelet transform ,in, It is the number of sampling points. It is the total number of sampling points of the signal. It is a numerical function of the electromyographic signal that varies with the number of sampling points. It is a time variable. It is a scaling parameter used to control the scaling ratio of the mother wavelet. These are translation parameters used to control the translation position of the mother wavelet on the time axis. This is the center angular frequency of the mother wavelet, used to determine the dominant oscillation frequency of the wavelet. It is used to calculate the neuromuscular activation delay. ,in, It is the time corresponding to the muscle force initiation point A. It is the time corresponding to the activation point B of the electromyographic signal.

[0039] The machine learning analysis module is used to construct a machine learning model based on the aforementioned feature parameters, outputting an objective assessment of neuromuscular dysfunction. Specifically, it uses a series of basic parameter features extracted from the calculated muscle force signal, features of the three phases of the muscle force curve, electromyographic parameters related to motor unit activation, and electromyographic parameters related to control and coordination ability as the input feature vector, denoted as... The model employs supervised learning methods, including but not limited to support vector machines, random forests, logistic regression, gradient boosting trees, and artificial neural networks. During the training phase, historical datasets with clinical diagnostic labels or rehabilitation grading labels are utilized. The model is trained to establish a mapping relationship between feature parameters and neuromuscular functional states. During the evaluation phase, new subject features are incorporated. The input model yields a neuromuscular function status score. When a logistic regression model is used, its output probability is... ,in For the weight vector, This is the bias term. Based on the model output probability. The mapping method converts the classification probability values ​​output by the model into numerical scores. By setting threshold ranges, subjects were divided into normal and normal groups. Mild abnormality Moderate abnormality and severe abnormalities The above scoring methods can be used individually or in combination; this invention is not limited thereto.

[0040] Specifically, the lower limb isometric muscle strength measuring device 1 includes a force sensor for collecting muscle strength data generated by the subject when performing isometric contraction tasks of knee flexion and extension.

[0041] The electromyography (EMG) signal acquisition device 2 includes a wireless surface EMG sensor, which is placed on the surface of the target muscle of the subject to acquire surface EMG signals of the main force-generating muscles of the tested joint. Figure 3 As shown, when measuring knee extension, knee flexion, and ankle dorsiflexion on both sides, the electromyography (EMG) acquisition device 2 was positioned at the standard location corresponding to the target muscle group to accurately acquire EMG signals. Figure 3 When measuring left knee extension in A, the electromyography signal acquisition device 2 was placed on the surface of the vastus lateralis, rectus femoris, and vastus medialis muscles of the left quadriceps femoris. Figure 3 When measuring right ankle dorsiflexion in B, electromyography signal acquisition device 2 is placed on the surface of the tibialis anterior muscle of the right leg; Figure 3 When measuring right knee flexion in C, the electromyography signal acquisition device 2 is placed on the surface of the right hamstring muscle.

[0042] The synchronization control module 3 receives instructions from the host computer system 4 and generates a synchronization trigger signal to achieve real-time synchronization between the lower limb isometric muscle strength measuring device 1 and the electromyography signal acquisition device 2, ensuring precise alignment of the time axis of muscle strength data and electromyography data.

[0043] like Figure 2 The neuromuscular function assessment method includes the following steps: First, the host computer system 4 sends a control command to the synchronization control module 3 to activate the lower limb isometric muscle strength measurement device 1 and the electromyography signal acquisition device 2, thereby achieving synchronous acquisition of muscle strength and electromyography signals. The acquired raw signals are then uploaded to the host computer system 4.

[0044] The signal processing and feature extraction module preprocesses the electromyographic (EMG) signals, automatically identifies EMG activation points, and calculates EMG parameters related to motor unit activation and control coordination. It then combines the muscle force initiation point and EMG activation points to calculate neuromuscular activation delay. This segmented analysis method enables more precise quantification of neuromuscular functional characteristics at different stages.

[0045] Example 2:

[0046] Using the system described in Example 1, neuromuscular function assessment experiments were conducted on elderly individuals, patients with neuromuscular diseases, and rehabilitation populations whose neuromuscular function has declined due to trauma or sports injuries. Subjects completed isometric contraction tasks at different joints of the lower limbs, and muscle strength and electromyographic signals were simultaneously acquired by the lower limb isometric muscle strength measurement device 1 and the electromyographic signal acquisition device 2. The acquired signals were processed and feature extracted by the signal processing and feature extraction module, and then input into a machine learning analysis module to analyze the machine learning model. The system automatically outputs a neuromuscular function score, providing objective reference for clinical diagnosis and rehabilitation training.

[0047] In summary, this invention innovatively designs a neuromuscular function assessment system. Through synchronous control, it achieves real-time synchronous acquisition of isometric muscle strength and multi-channel electromyographic signals in the lower limbs. Combined with automatic feature extraction and machine learning analysis, it realizes objective, precise, and intelligent assessment of neuromuscular function. It is particularly suitable for the objective quantitative assessment of neuromuscular dysfunction. Application scenarios include early screening and auxiliary diagnosis of frailty risk in the elderly, functional assessment of patients with neuromuscular diseases, and functional monitoring of rehabilitation populations. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the spirit of this invention. All equivalent changes made to the structure, shape, and principle of this invention should be covered within the protection scope of this invention.

Claims

1. A neuromuscular function assessment system, characterized in that, include: Lower limb isometric muscle strength measuring device, used to measure muscle force signals generated when subjects perform isometric contractions of lower limb joints; An electromyography (EMG) signal acquisition device is used to measure multi-channel EMG signals generated by subjects during isometric contractions of the lower limb joints. The synchronization control module is used to receive control commands from the host computer system, generate synchronization trigger signals, and control the real-time synchronous acquisition of data from the lower limb isometric muscle strength measuring device and the electromyography signal acquisition device. The signal processing and feature extraction module is used to preprocess the acquired muscle strength and electromyography signals, and extract and calculate the basic parameters of muscle strength and the characteristics of each stage of muscle strength, the activation start time of electromyography signals, the neuromuscular activation delay, the time domain characteristics of electromyography, and the frequency domain characteristics of electromyography. The basic parameters of muscle strength extracted include: extracting peak muscle strength and corresponding time, time to reach 80% of maximum force, and calculating muscle strength symmetry and active antagonism ratio of the knee joint. The features extracted from each stage of muscle strength include: dividing the muscle strength curve into stages, calculating the force development rate, maximum rate and corresponding time and stress value in the force development stage, calculating the difference before and after the force maintenance stage, attenuation angle and maintenance ability parameters, and calculating the attenuation rate, maximum attenuation rate and corresponding time and stress value in the force decay stage. The machine learning analysis module constructs a machine learning model based on the feature parameters extracted by the signal processing and feature extraction modules, and outputs an objective score of neuromuscular functional status.

2. The neuromuscular function assessment system according to claim 1, characterized in that, The electromyography (EMG) signal acquisition device uses a wireless surface EMG acquisition module.

3. The neuromuscular function assessment system according to claim 1, characterized in that, The synchronization control module includes a microcontroller-based synchronization trigger circuit to achieve synchronous acquisition of muscle force signals and electromyographic signals.

4. The neuromuscular function assessment system according to claim 1, characterized in that, The muscle strength curve is divided into three stages: strength development, strength maintenance, and strength decay. The first derivative of the preprocessed muscle strength signal is calculated, and threshold and stability are considered for determination. The muscle strength initiation point A, the initiation point of the strength maintenance stage, the end point of the strength maintenance stage, and the end point of the strength decay stage are automatically identified throughout the entire muscle strength curve, thus achieving the three-stage division of the muscle strength curve. Specifically, the muscle strength initiation point A is defined as the earliest moment when the signal exceeds three times the baseline mean and the first derivative is greater than 0.5 kg / s for one second. The strength maintenance stage initiation point is defined as the earliest moment when the first derivative is positive in the preceding period and the absolute value of the first derivative is less than 0.5 kg / s for one second after that point. The strength maintenance stage end point is defined as the earliest moment when the first derivative changes from near zero to a continuously negative trend around that point, and the derivative is less than -0.5 kg / s for one second. The strength decay stage end point is defined as the moment when the signal drops to within ± three times the baseline mean.

5. The neuromuscular function assessment system according to claim 1, characterized in that, The machine learning analysis module employs support vector machine and random forest algorithms.

6. The neuromuscular function assessment system according to claim 1, characterized in that, The muscle force signal was preprocessed by using a fourth-order Butterworth low-pass filter with a cutoff frequency of 10 Hz, and phase distortion-free filtering was achieved through zero-phase filtering.

7. The neuromuscular function assessment system according to claim 1, characterized in that, The electromyography (EMG) signals were preprocessed, including: removing the mean to eliminate the DC bias of the EMG signals, using a third-order Butterworth filter for bandpass filtering from 10 to 500 Hz, using a notch filter to perform notch filtering of the EMG signals at 50 Hz and its harmonics, and using a wavelet threshold denoising method based on the sym8 wavelet basis for smoothing.

8. The neuromuscular function assessment system according to claim 1, characterized in that, The formula for calculating muscle strength symmetry is: The formula for calculating the active antagonistic ratio of the knee joint is: ,in, This represents the maximum value of the signal on the left. This represents the maximum value of the signal on the right. This represents the maximum muscle strength of the relevant muscles during knee extension. This represents the maximum muscle strength of the relevant muscles when the knee joint is flexed.

9. The neuromuscular function assessment system according to claim 4, characterized in that, Neuromuscular activation delay ,in, It is the time corresponding to the muscle force initiation point A. The time corresponding to the activation point B of the electromyographic signal is determined by calculating the root mean square envelope of the electromyographic signal. The point where the signal first exceeds the baseline mean plus 8 times the standard deviation of the electromyographic signal is identified as the activation point B of the electromyographic signal.

10. The neuromuscular function assessment system according to claim 1, characterized in that, The feature parameters extracted by the signal processing and feature extraction modules are used to construct the input feature vector. When inputting into a machine learning model, the output probability is when a logistic regression model is used. ,in For the weight vector, Bias term; rating By setting threshold ranges, subjects were divided into normal and normal groups. Mild abnormality Moderate abnormality and severe abnormalities .

Images