Method and system for evaluating weakness risk based on lower limb myoelectricity and activation time sequence

Abstract

The invention provides a weakness risk assessment method and system based on lower limb myoelectricity and an activation time sequence, and relates to the technical field of medical auxiliary detection and biological information processing. The method comprises the following steps: acquiring historical lower limb surface myoelectricity data of a subject under a preset functional action, and calculating and storing a normal interval of individual characteristic parameters; then, original surface electromyogram signals of a target muscle group of the subject in a preset functional action period are collected, and digital filtering processing is carried out; and then multi-dimensional feature extraction is carried out, individual baseline correction is executed, multi-dimensional features after individual baseline correction are input into a pre-constructed cardiovascular asthenia risk assessment model, an asthenia probability value is calculated, and a risk assessment result is output. According to the invention, passive, quantitative and interpretable cardiovascular asthenia risk assessment in a daily scene is realized, individual differences are effectively corrected, and the accuracy of cardiovascular disease group asthenia identification and the intelligent level of clinical monitoring are improved.

Description

Technical Field

[0001] This invention relates to the field of medical auxiliary detection and bioinformatics processing technology, and more specifically, to a method and system for assessing the risk of frailty based on lower limb electromyography and activation timing. Background Technology

[0002] With the increasing aging of the global population, frailty identification in elderly patients with cardiovascular disease has become an important issue in clinical monitoring and rehabilitation. Frailty, as a typical geriatric syndrome, is characterized by a decline in physiological reserves and impairment of multiple system functions, and is clinically highly correlated with falls, prolonged hospitalization, functional loss, and an increased risk of death. Common cardiovascular diseases such as heart failure and coronary artery disease significantly accelerate the onset of frailty, making early identification and risk assessment of this population crucial. Currently, clinical identification of frailty mainly relies on questionnaires or basic function tests, such as the Fried frailty phenotype score, the 10-meter gait test, and balance function assessment. While these methods are widely used in clinical practice, they often have limitations such as high subjectivity, dependence on subject cooperation, and a lack of objective quantitative physiological markers. In particular, they are difficult to passively and conveniently monitor in the daily functional movements of the elderly, such as sitting-to-standing transitions or starting movements.

[0003] Surface electromyography (sEMG), as a non-invasive method for recording muscle electrical activity, can effectively reflect muscle activation intensity, muscle fatigue, and neuromuscular control characteristics. However, existing sEMG research largely focuses on short-term tests in a laboratory setting, and feature extraction is mostly concentrated on single time-domain indicators such as root mean square (RMS) and integral EMG, neglecting the deeper discriminative value of frequency-domain features such as median frequency, average power frequency, and activation sequence between muscle groups in assessing neuromuscular coordination and cardiovascular disease-related weakness. While existing biosignal monitoring patents involve the fusion of multimodal data, they fail to define the independent predictive value of frequency-domain and temporal features of specific lower limb muscle groups in cardiovascular disease populations and to construct relevant models.

[0004] Furthermore, existing technologies face numerous challenges in clinical application. Firstly, there are limitations in the data acquisition scenarios; the lack of trigger-based acquisition schemes linked to daily functional movements makes it difficult for the system to seamlessly integrate into nursing homes or hospital wards. Secondly, the models lack specificity; cardiovascular disease patients exhibit unique cardiogenic metabolic changes and muscle perfusion abnormalities, resulting in significantly different muscle electrophysiological manifestations compared to the general elderly population, often leading to inaccurate assessments using generic models. Thirdly, existing electromyography (EMG) acquisition systems largely rely on disposable patch electrodes and complex manual annotation processes, resulting in poor clinical portability and cumbersome operation, hindering their widespread adoption in home follow-up or routine bedside monitoring. Finally, due to the lack of effective individual baseline correction and threshold self-adaptation mechanisms, existing assessment methods lack robustness in the face of significant individual differences.

[0005] In view of the above, this application is hereby submitted. Summary of the Invention

[0006] This invention aims to provide a method and system for assessing frailty risk based on lower limb electromyography and activation timing, in order to address the shortcomings of existing methods in identifying frailty in elderly cardiovascular disease patients, such as over-reliance on subjective questionnaires, lack of convenient and objective quantitative physiological markers in daily life, and lack of specific models for cardiovascular disease populations and insufficient correction for individual differences in existing surface electromyography monitoring schemes.

[0007] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0008] A frailty risk assessment method based on lower limb electromyography and activation timing includes: S1. Acquire historical lower limb surface electromyography data of the subject under preset functional movements, calculate and store the normal range of individual characteristic parameters; the preset functional movements include at least three repeated sitting-to-standing movements or starting-to-walk movements; the individual characteristic parameters include the median mean frequency and standard deviation of each target muscle group. S2, collect raw surface electromyographic signals of the target muscle groups of the subject within a preset functional movement cycle, and perform digital filtering processing; the target muscle groups include the right rectus femoris, right medial gastrocnemius, and right tibialis anterior. S3, Multidimensional feature extraction is performed on the raw surface electromyography signal after digital filtering; the multidimensional feature extraction includes frequency domain feature extraction, activation time sequence feature extraction and clinical feature fusion. S4, Combine the historical lower limb surface electromyography data to perform individual baseline correction on the extracted multidimensional features; S5. Input the multidimensional features after individual baseline correction into the pre-constructed cardiovascular frailty risk assessment model, calculate the frailty probability value, and output the risk assessment result according to the preset interval in which the frailty probability value is located; the cardiovascular frailty risk assessment model is constructed using a multivariate logistic regression equation.

[0009] Preferably, the digital filtering process obtains the signal envelope of the original surface electromyography signal through full-wave rectification and low-pass filtering algorithms, and identifies the muscle activation initiation time based on the envelope detection algorithm; The criteria for identifying the muscle activation onset time are: the signal amplitude first exceeds twice the standard deviation of the resting segment mean and the duration exceeds fifty milliseconds; Its expression is: ; ; ; in, The original surface electromyography signal; The signal is after full-wave rectification; The linear envelope after low-pass filtering; For low-pass filtering; The mean of the envelope in the resting state; The standard deviation of the resting envelope.

[0010] Preferably, the frequency domain feature extraction is performed by performing a fast Fourier transform or power spectrum estimation on the segmented signal to calculate the median frequency of the right rectus femoris and right medial gastrocnemius muscles. The activation sequence feature extraction is defined as the activation sequence index by calculating the time difference between the right tibialis anterior muscle and the initiation event of the movement. The clinical feature fusion was achieved by acquiring the subject's age data and clinical function scale scores; the clinical function scale scores were obtained using the balance and gait test scale.

[0011] Preferably, the individual baseline correction is achieved by subtracting the median frequency mean of the individual feature parameters of the historical lower limb surface electromyography data from the extracted multidimensional features and dividing by the standard deviation of the individual feature parameters of the historical lower limb surface electromyography data, in order to achieve feature normalization.

[0012] Preferably, when collecting raw surface electromyography (EMG) signals of specific lower limb muscle groups within a preset functional movement cycle, a scene-triggered data acquisition method is adopted, that is, the data is collected by a pressure sensor array embedded under the subject's rehabilitation chair seat cushion; when the pressure sensor array detects that the pressure value drops from a preset high threshold corresponding to the subject's weight to a preset low threshold, it is determined as the start of the sitting-to-standing transition movement, and the synchronous acquisition and recording of raw surface EMG signals is initiated.

[0013] Preferably, in the multivariate logistic regression equation of the cardiovascular frailty risk assessment model, the median frequency of the right rectus femoris and the median frequency of the right medial gastrocnemius muscle are used as positive predictors, and the activation sequence index of the activation time sequence feature of the right tibialis anterior muscle is used as a negative predictor. By increasing the weight of frequency domain features in the model, a specific fit to the cardiovascular disease population can be achieved.

[0014] Preferably, the expression for the multivariate logistic regression equation is: ; in, The attenuation probability value; The intercept; , , These are the regression coefficients; This represents the median frequency of the right rectus femoris muscle. This represents the median frequency of the right medial gastrocnemius muscle. The activation sequence index represents the activation timing characteristics of the right tibialis anterior muscle.

[0015] Preferably, it further includes: generating a periodic report containing characteristic change curves based on the risk assessment results, and monitoring the subject's risk of deterioration across time dimensions.

[0016] This invention also provides a frailty risk assessment system based on lower limb electromyography and activation timing, comprising: The fabric-type lower limb electromyography (EMG) acquisition brace uses an elastic fabric base and integrates multiple conductive fabric electrodes inside. It is used to attach to the target muscle group of the subject's lower limb to collect surface EMG signals. A short-term trigger recognition module, including a pressure sensing unit embedded in the seat or an inertial measurement unit worn on the subject, is used to monitor the subject's motion characteristics in real time and generate a trigger; The portable data acquisition module is connected to the fabric-type lower limb electromyography acquisition brace via a magnetic interface. It is used to receive trigger pulses and perform signal amplification, analog-to-digital conversion, and local data caching. The system includes a processing unit and a remote server, wherein the remote server stores a computer program that can be executed by the processing unit to implement a frailty risk assessment method based on lower limb electromyography and activation timing as described above. The user interface, deployed on mobile terminals or medical workstations, is used to display the risk level and historical characteristics evolution trend of the subjects.

[0017] The present invention also provides a computer-readable storage medium storing computer-readable instructions, which, when executed by a processor of the device on which the computer-readable storage medium is located, implement the aforementioned method for assessing the risk of weakness based on lower limb electromyography and activation timing.

[0018] In summary, compared with the prior art, the present invention has the following beneficial effects: First, it offers a high degree of objectivity and quantification. This invention extracts the frequency domain and activation sequence characteristics of surface electromyography (EMG) signals from specific target muscle groups in the lower limbs (including the right rectus femoris, right medial gastrocnemius, and right tibialis anterior), providing objective neuromuscular physiological markers for assessing cardiovascular debility. These indicators directly reflect the subject's neural control ability and muscle metabolic state, overcoming the shortcomings of traditional clinical scales, which are heavily influenced by subjective factors and have poor repeatability.

[0019] Secondly, it achieves a scenario-based and seamless monitoring mode. This invention uses fabric electrode protectors combined with a daily functional movement triggering acquisition strategy, seamlessly integrating the monitoring process into the subject's daily life. Subjects do not need to go to a special laboratory, nor do they need professional personnel to perform complicated manual patching operations. Passive and long-term risk screening can be achieved in elderly care, rehabilitation, and home environments, and it has extremely high clinical applicability.

[0020] Third, it possesses a robust individualized correction mechanism. By introducing an individual baseline correction module, the system can automatically adapt the assessment threshold based on the subjects' historical data. This self-learning and self-adaptation capability effectively corrects for physiological heterogeneity among different individuals, significantly improving the model's assessment accuracy and robustness in complex real-world populations.

[0021] Fourth, the model has extremely high clinical interpretability. The lightweight cardiovascular frailty risk assessment model, built based on the physiological characteristics of specific muscle groups, has predictive factors that are highly consistent with clinical medical theory. For example, shifts in median frequency reflect muscle fatigue and fiber type transformation, while delays in activation sequences reflect impaired balance control. This transparent assessment process facilitates clinicians' understanding of the results and allows them to develop precise drug treatment or rehabilitation training programs.

[0022] Fifth, the system achieves a balance between low power consumption and long-term monitoring. The short-time trigger acquisition strategy ensures that the system only operates when critical actions occur, effectively reducing the processor's computational load and battery consumption. This design extends the wearable device's battery life, supports long-term, continuous trend analysis of frailty risk, and provides important data support for the prognostic management of elderly patients with cardiovascular disease.

[0023] Sixth, the hardware system structure is scientifically and rationally designed. The combination of fabric protective gear and magnetic interface balances wearing comfort and ease of operation. The application of conductive fabric electrodes not only improves the signal-to-noise ratio of signal acquisition but also enhances the durability of the device. The multi-sensor fusion triggering mechanism ensures the integrity and accuracy of signal segment extraction, laying a solid foundation for subsequent feature calculation.

[0024] In summary, this invention, through the deep integration of technical means, constructs a complete chain from signal perception and feature extraction to risk prediction, providing an efficient, accurate, and easily scalable technical approach for the early identification of frailty in elderly patients with cardiovascular disease. This invention not only enhances the intelligence level of clinical monitoring but also opens up new pathways for home-based rehabilitation and health management for the elderly. The methods and systems described in this invention can achieve early detection of frailty risk through in-depth mining of the electrophysiological characteristics of specific target muscle groups in the lower limbs, thereby gaining valuable time for the implementation of intervention measures. Attached Figure Description

[0025] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0026] Figure 1 This is a schematic diagram of a frailty risk assessment method based on lower limb electromyography and activation timing, as provided in Example 1.

[0027] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0029] Example 1 Embodiment 1 of the present invention provides a frailty risk assessment method based on lower limb electromyography and activation timing, which can be implemented by a frailty risk assessment device based on lower limb electromyography and activation timing (hereinafter referred to as risk assessment device), specifically, executed by one or more processors within the risk assessment device.

[0030] In this embodiment, the risk assessment device may be an electronic device equipped with a processor, which carries a computer program for the frailty risk assessment method based on lower limb electromyography and activation timing, and the computer program can be executed, such as a computer, smartphone, smart tablet, workstation, etc., which are not limited here.

[0031] In this embodiment, addressing the problems of existing technologies where surface electromyography (EMG) acquisition is mostly limited to laboratory environments, has limited signal characteristics, and does not consider the unique differences in muscle perfusion and metabolism in cardiovascular disease patients, this invention constructs a feature space containing the median frequency and activation sequence index of specific muscle groups, combines it with clinical functional scales, and utilizes individual baseline correction mechanisms and lightweight fusion models to achieve accurate classification of frailty risk in elderly cardiovascular disease patients.

[0032] like Figure 1 As shown, a frailty risk assessment method based on lower limb electromyography and activation timing includes steps S1 to S5.

[0033] S1. Acquire historical lower limb surface electromyography data of the subject under preset functional movements, calculate and store the normal range of individual characteristic parameters; the preset functional movements include at least three repeated sitting-to-standing movements or starting-to-walk movements; the individual characteristic parameters include the median mean frequency and standard deviation of each target muscle group.

[0034] Due to significant physiological differences among elderly individuals, especially among patients with cardiovascular disease who exhibit considerable heterogeneity in muscle mass and nerve conduction velocity, a baseline calibration procedure was performed before formal monitoring. Subjects, under the guidance of professionals, wore a fabric lower limb electromyography (EMG) acquisition brace and completed at least three repeated experiments with preset functional movements.

[0035] The preset movements included standing up from a standard-height rehabilitation chair and taking a forward step. Surface electromyography (EMG) signals from these three movements were then recorded and analyzed in real time, calculating the median mean and standard deviation of the target muscle groups, including the right rectus femoris and right medial gastrocnemius. These parameters were stored as individual baseline correction values ​​in the subject's physiological benchmark. By establishing this individualized mathematical benchmark, subsequent risk assessments shift from absolute value judgments to relative trend judgments, filtering out assessment errors caused by individual differences in basic physical fitness.

[0036] S2, collect raw surface electromyographic signals of the target muscle groups of the subject within a preset functional movement cycle, and perform digital filtering processing; the target muscle groups include the right rectus femoris, right medial gastrocnemius, and right tibialis anterior.

[0037] To achieve seamless monitoring, the system does not employ a continuous 24 / 7 sampling mode. Instead, it utilizes a short-time trigger recognition module to capture key functional movements. When collecting raw surface electromyography (EMG) signals of specific lower limb muscle groups within a preset functional movement cycle, a scene-triggered data acquisition method is used. This involves collecting data through a pressure sensor array embedded under the subject's rehabilitation chair seat. When the pressure sensor array detects a drop in pressure value from a preset high threshold corresponding to the subject's weight to a preset low threshold, it is determined as the start of a sitting-to-standing transition, and the synchronous acquisition and recording of raw EMG signals is initiated.

[0038] When the subject sits on the rehabilitation chair embedded with pressure sensors, the system is in a low-power standby state. When the pressure sensor array detects a rapid drop in the seat cushion pressure value from a preset high threshold corresponding to the subject's weight, and simultaneously the inertial measurement unit detects that the angular velocity of the torso leaning forward exceeds a preset threshold, the short-time trigger recognition module determines that the subject is performing a sitting-to-standing transition and immediately sends a trigger pulse to the portable data acquisition module. Upon receiving the pulse, the portable data acquisition module activates its internal 16-bit analog-to-digital converter, simultaneously recording surface electromyography signals from multiple channels at a sampling frequency of 1000Hz for 3 to 5 seconds, covering the entire process from preparation to standing to steady-state standing. This triggering mechanism ensures that the acquired data contains high-value functional information while reducing the system's average power consumption and extending the wearable device's battery life.

[0039] The acquired raw electromyography (EMG) signals are digitally filtered. The digital filtering process uses a full-wave rectification and low-pass filtering algorithm to obtain the signal envelope of the raw surface EMG signals, and an envelope detection algorithm is used to identify the muscle activation initiation time.

[0040] For example, by using a digital filter bank, the signal is first passed through a fourth-order Butterworth bandpass filter with a cutoff frequency of 20Hz to 450Hz to filter out low-frequency artifacts caused by human movement and high-frequency electromagnetic noise from the environment. Then, a 50Hz notch filter is used to remove power line interference. The preprocessed signal is then rectified using full-wave rectification and its envelope is extracted using a low-pass filter with a cutoff frequency of 6Hz.

[0041] Based on the envelope, the system executes a muscle activation onset time detection algorithm. The criteria for identifying the muscle activation onset time are: the signal amplitude first exceeds twice the standard deviation of the resting segment mean and the duration exceeds fifty milliseconds.

[0042] Its expression is: ; ; ; in, The original surface electromyography signal; The signal is after full-wave rectification; The linear envelope after low-pass filtering; For low-pass filtering; The mean of the envelope in the resting state; The standard deviation of the resting envelope.

[0043] For example, within a preset time window (e.g., 500ms) before the pressure trigger signal is generated, a segment of electromyographic signal in a completely relaxed state can be extracted as the resting segment.

[0044] The system automatically marks the identified points as the activation initiation times of the muscle group. This precise time-point anchoring enables quantitative analysis of the collaborative relationships between different muscle groups.

[0045] S3, perform multidimensional feature extraction on the original surface electromyography signal after digital filtering; the multidimensional feature extraction includes frequency domain feature extraction, activation time sequence feature extraction and clinical feature fusion.

[0046] This step involves extracting key indicators reflecting the degree of weakness from the segmented raw surface electromyography signals.

[0047] The frequency domain feature extraction is performed by performing a fast Fourier transform or power spectrum estimation on the segmented signal to calculate the median frequency of the right rectus femoris and right medial gastrocnemius muscles.

[0048] In patients with cardiovascular disease, the shift in median frequency due to type II myofibril atrophy or metabolic changes caused by insufficient peripheral blood perfusion has significant pathological indicative significance.

[0049] The activation sequence feature extraction is defined as the activation sequence index, which is calculated by measuring the time difference between the right tibialis anterior muscle and the initiation event of the movement. The activation sequence index reflects the recruitment efficiency of the central nervous system for the distal muscles of the lower limbs.

[0050] In this embodiment, the action initiation event can refer to the pressure drop trigger point detected by the pressure sensor array.

[0051] The clinical feature fusion was achieved by acquiring the subject's age data and clinical function scale scores; the clinical function scale scores were obtained using the balance and gait test scale.

[0052] The system uses an interactive interface to input clinical characteristics such as the subject's age and Tinetti balance and gait test scores to construct a multimodal feature vector.

[0053] S4. Combine the historical lower limb surface electromyography data to perform individual baseline correction on the extracted multidimensional features.

[0054] The extracted feature parameters are fed into the individual baseline correction module. The individual baseline correction involves subtracting the median frequency mean of the individual feature parameters from the historical lower limb surface electromyography (EMG) data from the extracted multidimensional features and dividing by the standard deviation of the individual feature parameters from the historical lower limb EMG data to achieve feature normalization.

[0055] S5. Input the multidimensional features after individual baseline correction into the pre-constructed cardiovascular frailty risk assessment model, calculate the frailty probability value, and output the risk assessment result according to the preset interval in which the frailty probability value is located; the cardiovascular frailty risk assessment model is constructed using a multivariate logistic regression equation.

[0056] The system uses a pre-constructed multivariate logistic regression equation to calculate risk probability. The coefficients of the regression equation are preset based on specific research data on cardiovascular disease populations.

[0057] In the multivariate logistic regression equation of the cardiovascular frailty risk assessment model, the median frequency of the right rectus femoris and the median frequency of the right medial gastrocnemius muscle are used as positive predictors, i.e., a decrease in frequency indicates an increased risk of frailty. The activation sequence index of the activation time sequence feature of the right tibialis anterior muscle is used as a negative predictor, i.e., the longer the activation delay time, the worse the neuromuscular coordination and the higher the risk. By increasing the weight of frequency domain features in the model, a specific fit to the cardiovascular disease population can be achieved.

[0058] The expression for the multivariate logistic regression equation is as follows: ; In a preferred embodiment, the multivariate logistic regression equation can also incorporate clinical characteristics (such as subject age and clinical function scale scores), for example, the expression is set as: ; in, The attenuation probability value; The intercept; , , , , These are the regression coefficients; This represents the median frequency of the right rectus femoris muscle. This represents the median frequency of the right medial gastrocnemius muscle. The activation sequence index represents the activation timing characteristics of the right tibialis anterior muscle. The subject's age; For clinical function scale scoring.

[0059] The model outputs a frailty probability value P between 0 and 1. The system can determine the risk level based on the range of the P value. For example, P less than 0.3 indicates low risk, 0.3 to 0.7 indicates medium risk, and greater than 0.7 indicates high risk. Considering the tendency of cardiovascular disease patients to experience fatigue, the risk scoring logic unit increases the weighting coefficients of frequency domain features in the model to achieve an accurate fit of frailty risk for this specific group.

[0060] In a preferred embodiment, the method further includes: S6, generating a periodic report containing characteristic change curves based on the risk assessment results, and monitoring the subject's risk of deterioration across time dimensions.

[0061] The assessment results are synchronized in real time to the user interface and the remote server via a wireless communication link. The user interface displays the subject's frailty profile in the form of radar charts and trend curves. If the system detects a rapid increase in a subject's risk level from low to high risk within a short period, or if the median frequency index shows a continuous downward trend, it automatically triggers an early warning logic, sending an alert to the medical staff or family members. The remote server stores long-term historical data and generates weekly or monthly reports, providing objective evidence for doctors to adjust rehabilitation plans or drug interventions.

[0062] In summary, compared with the prior art, the present invention has the following beneficial effects: This invention constructs a complete risk assessment scheme for cardiovascular disease frailty in the elderly through fabric electrode sensing, multi-sensor fusion triggering, extraction of specific target muscle groups in the lower limbs, and individualized risk modeling. While ensuring high-precision assessment, the system also considers wearing comfort and ease of operation, providing strong technical support for long-term health management and functional maintenance in cardiovascular disease patients. The parameters, muscle group locations, and algorithm logic in this embodiment are all optimized selections based on extensive clinical data validation, possessing extremely high practical value and promising prospects for widespread application.

[0063] Example 2 The second embodiment of the present invention also provides a frailty risk assessment system based on lower limb electromyography and activation timing, comprising: The fabric-type lower limb electromyography (EMG) acquisition brace uses an elastic fabric base and integrates multiple conductive fabric electrodes inside. It is used to attach to the target muscle group of the subject's lower limb to collect surface EMG signals. A short-term trigger recognition module, including a pressure sensing unit embedded in the seat or an inertial measurement unit worn on the subject, is used to monitor the subject's motion characteristics in real time and generate a trigger; The portable data acquisition module is connected to the fabric-type lower limb electromyography acquisition brace via a magnetic interface. It is used to receive trigger pulses and perform signal amplification, analog-to-digital conversion, and local data caching. The processing unit and a remote server, wherein the remote server stores a computer program that can be executed by the processing unit to implement a method for assessing the risk of frailty based on lower limb electromyography and activation timing as described in any one of claims 1-8.

[0064] The user interface, deployed on mobile terminals or medical workstations, is used to display the risk level and historical characteristics evolution trend of the subjects.

[0065] Furthermore, the arrangement structure of the conductive fabric electrodes of the fabric-type lower limb electromyography acquisition brace is as follows: a pair of differential main electrodes are set at the rectus femoris, semitendinosus / semimembranosus band, tibialis anterior and medial gastrocnemius muscles, and a common reference electrode is set at the ankle or the outside of the knee joint. The conductive fabric electrode is made using a silver fiber weaving process and its surface is covered with a conductive hydrogel layer.

[0066] A pair of differential master electrodes were placed at the belly of the rectus femoris muscle to capture active electromyographic signals during knee extension; electrodes were placed at the semitendinosus muscle to monitor antagonist muscle activity; electrodes were placed at the tibialis anterior muscle to assess ankle dorsiflexion control during gait initiation; and electrodes were placed at the medial gastrocnemius muscle to analyze force characteristics during the stance phase.

[0067] In addition, a common reference electrode is placed in the electrically neutral region of the subject's ankle or the outside of the knee joint, which, together with the differential amplifier circuit inside the portable data acquisition module, suppresses common-mode interference and improves the signal-to-noise ratio.

[0068] The conductive fabric electrode surface is covered with a thin layer of conductive hydrogel. This design reduces the contact resistance between the skin and the electrode, ensuring a stable signal waveform even in dry environments.

[0069] To further illustrate the practical application effects of the present invention, two specific implementation scenarios are given below.

[0070] Specific Implementation Scenario 1: Inpatient monitoring in a hospital rehabilitation department. In this scenario, the subject was an elderly patient with chronic heart failure and a tendency towards frailty. Upon admission, the patient's baseline was established in step S1 with the assistance of a rehabilitation therapist. The patient wore a brace with integrated conductive fabric electrodes. During daily rehabilitation training or standing up in the ward, the recognition module was briefly triggered to automatically capture the movement. The portable data acquisition module transmitted the collected signals from the rectus femoris and tibialis anterior muscles to the processing unit. Due to significant cardiogenic muscle hypoperfusion in this patient, the system detected a 15% decrease in the median frequency of the right rectus femoris compared to the baseline level in step S4, and an 80ms increase in the activation delay of the right tibialis anterior muscle. The model calculation in step S5 yielded a frailty probability value P of 0.82, indicating a high risk. The system immediately displayed a red warning on the user interface of the medical workstation. Based on this, the rehabilitation team adjusted the training intensity and increased physical therapy targeting lower limb blood oxygen metabolism, effectively preventing adverse events such as falls during subsequent activities.

[0071] Specific Implementation Scenario Two: Home-based Passive Monitoring in Elderly Care Facilities. The subjects were several elderly residents of a nursing home suffering from coronary heart disease. The residents wore fabric-type lower limb electromyography (EMG) braces daily, which resembled ordinary knee braces and did not affect their daily activities. Pressure sensors were installed on public seats and seats in the residents' rooms. Whenever an elderly person performed a sitting-to-standing transition, the system automatically completed a risk assessment. The processing unit used an individual baseline correction mechanism to filter out interference caused by differences in baseline muscle strength among different residents. A remote server aggregated the data over a long period and found that one resident's activation sequence index (AS) showed a slow upward trend over two weeks, suggesting a decline in their neural control ability. Although this resident's current clinical scale scores were still within the normal range, the system described in this invention detected subtle changes in physiological indicators in advance. Based on this, the nursing home intervened early, arranging targeted balance function strengthening training for the resident, achieving early detection and preventative maintenance of frailty risk.

[0072] During the operation of the system of this invention, the scientific design of the hardware provides a solid guarantee for data quality. The fabric-type lower limb electromyography (EMG) acquisition brace utilizes the flexibility of conductive fabric to ensure that the electrodes remain in close contact with the muscle bellies of the rectus femoris, semitendinosus, tibialis anterior, and medial head of the gastrocnemius muscle during the subject's dynamic transition movements. Compared with traditional disposable Ag / AgCl patch electrodes, the conductive fabric electrodes avoid drastic impedance fluctuations caused by skin sweating, and the magnetic interface design allows the subject or caregiver to easily remove the portable data acquisition module for charging. The brace itself can be routinely washed, greatly improving the durability of the device and user compliance.

[0073] At the software algorithm level, the lightweight logistic regression algorithm executed by the processing unit ensures real-time computation. Since the muscle electrophysiological characteristics of cardiovascular disease patients are significantly affected by hemodynamics, this invention focuses not only on the amplitude in the time domain but also on the spectral shift in the frequency domain during feature extraction. The calculation of the median frequency (MF), by finding the median of the power spectral density, can sensitively reflect changes in motor unit recruitment patterns. The introduction of the activation sequence index (AS) fills the gap in neuromuscular coordination assessment from a temporal perspective. This fusion of multidimensional features allows the system's identification of weakness to go beyond a single measure of muscle strength, delving into the operational efficiency of the neuromuscular system.

[0074] The individualized calibration mechanism is another core advantage of this invention. The individual baseline correction memory in the processing unit records the subject's "physiological fingerprint" in a relatively healthy state. During the assessment, the system calculates the coefficient of variation (COP) of the current characteristic value deviating from the baseline parameter. Only when the COP exceeds a preset threshold is a clinically significant risk change considered to have occurred. This design effectively corrects for individual background differences in elderly individuals due to age, disease duration, and medication use, improving the robustness of the assessment results.

[0075] This invention also fully considers scalability in practical deployment. The portable data acquisition module supports multiple wireless transmission protocols such as BLE or Wi-Fi, allowing for easy integration with the local area network of nursing homes or the HIS system of hospitals. The user interface not only displays the risk level but also provides personalized rehabilitation suggestions based on feature analysis results. For example, if the system analysis shows that the risk is mainly caused by a decline in median frequency, it recommends strengthening resistance training; if the risk is mainly caused by activation timing delay, it recommends strengthening balance and coordination exercises. This closed-loop management model elevates frailty risk assessment from a simple detection tool to an intelligent platform for decision support.

[0076] Example 3 The third embodiment of the present invention also provides a computer-readable storage medium storing computer-readable instructions. When the computer-readable instructions are executed by the processor of the device where the computer-readable storage medium is located, they implement the above-described method for assessing the risk of weakness based on lower limb electromyography and activation timing.

[0077] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for assessing frailty risk based on lower limb electromyography and activation timing, characterized in that, include: S1. Acquire historical lower limb surface electromyography data of the subject under preset functional movements, calculate and store the normal range of individual characteristic parameters; the preset functional movements include at least three repeated sitting-to-standing movements or starting-to-walk movements; the individual characteristic parameters include the median mean frequency and standard deviation of each target muscle group. S2, collect raw surface electromyographic signals of the target muscle groups of the subject within a preset functional movement cycle, and perform digital filtering processing; the target muscle groups include the right rectus femoris, right medial gastrocnemius, and right tibialis anterior. S3, Multidimensional feature extraction is performed on the raw surface electromyography signal after digital filtering; The multidimensional feature extraction includes frequency domain feature extraction, activation time-series feature extraction, and clinical feature fusion. S4, Combine the historical lower limb surface electromyography data to perform individual baseline correction on the extracted multidimensional features; S5. Input the multidimensional features after individual baseline correction into the pre-constructed cardiovascular frailty risk assessment model, calculate the frailty probability value, and output the risk assessment result according to the preset interval in which the frailty probability value is located; the cardiovascular frailty risk assessment model is constructed using a multivariate logistic regression equation.

2. The method for assessing frailty risk based on lower limb electromyography and activation timing as described in claim 1, characterized in that... The digital filtering process obtains the signal envelope of the original surface electromyography signal through full-wave rectification and low-pass filtering algorithms, and identifies the muscle activation initiation time based on the envelope detection algorithm; The criteria for identifying the muscle activation onset time are: the signal amplitude first exceeds twice the standard deviation of the resting segment mean and the duration exceeds fifty milliseconds; Its expression is: ； ； ； in, The original surface electromyography signal; The signal is after full-wave rectification; The linear envelope after low-pass filtering; For low-pass filtering; The mean of the envelope in the resting state; The standard deviation of the resting envelope.

3. The method for assessing frailty risk based on lower limb electromyography and activation timing as described in claim 1, characterized in that... The frequency domain feature extraction is performed by performing a fast Fourier transform or power spectrum estimation on the segmented signal to calculate the median frequency of the right rectus femoris and right medial gastrocnemius muscles. The activation sequence feature extraction is defined as the activation sequence index by calculating the time difference between the right tibialis anterior muscle and the initiation event of the movement. The clinical feature fusion was achieved by acquiring the subject's age data and clinical function scale scores; the clinical function scale scores were obtained using the balance and gait test scale.

4. The method for assessing the risk of frailty based on lower limb electromyography and activation timing as described in claim 1, characterized in that... The individual baseline correction is achieved by subtracting the median frequency mean of the individual feature parameters of the historical lower limb surface electromyography data from the extracted multidimensional features and dividing by the standard deviation of the individual feature parameters of the historical lower limb surface electromyography data, in order to normalize the features.

5. The method for assessing frailty risk based on lower limb electromyography and activation timing as described in claim 1, characterized in that... When collecting raw surface electromyography (EMG) signals of specific lower limb muscle groups within a preset functional movement cycle, a scenario-triggered data acquisition method is adopted, which involves collecting data through a pressure sensor array embedded under the subject's rehabilitation chair seat. When the pressure sensor array detects that the pressure value drops from a preset high threshold corresponding to the subject's weight to a preset low threshold, it is determined as the start of the sitting-to-standing transition movement, and the synchronous acquisition and recording of raw surface EMG signals is initiated.

6. The method for assessing frailty risk based on lower limb electromyography and activation timing as described in claim 3, characterized in that... In the multivariate logistic regression equation of the cardiovascular frailty risk assessment model, the median frequency of the right rectus femoris and the median frequency of the right medial gastrocnemius muscle are used as positive predictors, and the activation sequence index of the activation time sequence feature of the right tibialis anterior muscle is used as a negative predictor. By increasing the weight of frequency domain features in the model, a specific fit to the cardiovascular disease population can be achieved.

7. The method for assessing frailty risk based on lower limb electromyography and activation timing as described in claim 6, characterized in that... The expression for the multivariate logistic regression equation is as follows: ； in, The attenuation probability value; The intercept; , , These are the regression coefficients; This represents the median frequency of the right rectus femoris muscle. This represents the median frequency of the right medial gastrocnemius muscle. The activation sequence index represents the activation timing characteristics of the right tibialis anterior muscle.

8. The method for assessing frailty risk based on lower limb electromyography and activation timing as described in claim 1, characterized in that... It also includes: generating periodic reports containing characteristic change curves based on the risk assessment results, and monitoring the subject's risk of deterioration across time dimensions.

9. A frailty risk assessment system based on lower limb electromyography and activation timing, characterized in that, include: The fabric-type lower limb electromyography (EMG) acquisition brace uses an elastic fabric base and integrates multiple conductive fabric electrodes inside. It is used to attach to the target muscle group of the subject's lower limb to collect surface EMG signals. A short-term trigger recognition module, including a pressure sensing unit embedded in the seat or an inertial measurement unit worn on the subject, is used to monitor the subject's motion characteristics in real time and generate a trigger; The portable data acquisition module is connected to the fabric-type lower limb electromyography acquisition brace via a magnetic interface. It is used to receive trigger pulses and perform signal amplification, analog-to-digital conversion, and local data caching. The processing unit and the remote server, wherein the remote server stores a computer program that can be executed by the processing unit to implement a method for assessing the risk of weakness based on lower limb electromyography and activation timing as described in any one of claims 1-8. The user interface, deployed on mobile terminals or medical workstations, is used to display the risk level and historical characteristics evolution trend of the subjects.

10. A frailty risk assessment system based on lower limb electromyography and activation timing as described in claim 9, characterized in that, The conductive fabric electrodes of the fabric-type lower limb electromyography acquisition brace are arranged as follows: a pair of differential main electrodes are set at the rectus femoris, semitendinosus / semimembranosus band, tibialis anterior and medial gastrocnemius muscles, and a common reference electrode is set at the ankle or the outside of the knee joint. The conductive fabric electrode is made using a silver fiber weaving process and its surface is covered with a conductive hydrogel layer.

Images