Intelligent monitoring and intervention system for muscle health of chronic respiratory disease patient

Abstract

The invention relates to the field of medical health monitoring, in particular to a chronic respiratory disease patient muscle health intelligent monitoring and intervention system which comprises a multi-dimensional time-frequency impedance analysis monitoring module, a respiration-movement cooperative training device module and a nutrition management platform. Performing time-frequency domain analysis on the signals by using multi-scale wavelet decomposition, and extracting muscle function characteristic parameters; quantitative evaluation of a respiratory muscle and intestinal muscle collaborative function is realized through phase synchronization analysis, a personalized baseline and a multi-stage early warning threshold are constructed, and early recognition of muscle dysfunction is realized; the training load and the training mode are dynamically adjusted according to the breathing muscle force percentage, and breathing training and swallowing muscle group training are combined; a personalized nutrition scheme is formulated based on muscle function parameters and metabolic states, the system achieves multi-dimensional evaluation of respiratory muscle and intestinal muscle functions, the function degradation trend is found 2-4 weeks ahead of time, and the training effect is improved by about 35%.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical health monitoring, in particular to a muscle health intelligent monitoring and intervention system for chronic respiratory disease patients, which is used for real-time monitoring of respiratory muscle and intestinal muscle function of chronic respiratory disease patients, and providing personalized training intervention and nutrition management scheme based on the monitoring results. BACKGROUND

[0002] Chronic respiratory disease patients (such as chronic obstructive pulmonary disease COPD, pulmonary fibrosis, etc.) are often accompanied by respiratory muscle dysfunction and decreased systemic muscle function, especially the functional degradation of respiratory muscle and digestive system related muscle groups (intestinal muscle), which seriously affects the quality of life and prognosis of patients. Existing studies have shown that about 70% of moderate to severe COPD patients have different degrees of respiratory muscle dysfunction, and about 50% of patients have decreased intestinal muscle function, leading to difficulty swallowing and nutritional absorption disorders. With the progression of the disease, muscle dysfunction is further aggravated, forming a vicious cycle.

[0003] Traditional muscle function evaluation methods mainly rely on hospital examinations such as respiratory muscle strength test, lung function test, and electromyogram, which have the disadvantages of low evaluation frequency, inability to monitor in real time, lack of comprehensive evaluation, etc. The conventional muscle training intervention usually adopts a unified training scheme, which is difficult to adjust individually according to the real-time state of the patient, and the training effect is limited. In addition, the existing systems generally lack evaluation and training of respiratory muscle and intestinal muscle coordination function, and cannot effectively improve the overall muscle health status of patients.

[0004] In terms of nutrition intervention, existing methods are mostly based on static indicators such as body weight and BMI, which are difficult to reflect the dynamic changes of patient muscle function status and energy metabolism demand, and are difficult to provide precise nutritional support.

[0005] Therefore, an intelligent system capable of real-time monitoring of respiratory muscle and intestinal muscle function of chronic respiratory disease patients, providing personalized training intervention and nutrition management is urgently needed to effectively improve patient muscle function, delay sarcopenia, and improve quality of life. SUMMARY

[0006] The purpose of the present application is to provide a muscle health intelligent monitoring and intervention system for chronic respiratory disease patients, which realizes precise monitoring of respiratory muscle and intestinal muscle function of patients through multi-dimensional time-frequency impedance analysis technology, provides personalized respiratory-movement coordination training and nutrition intervention scheme based on the monitoring results, and effectively improves the muscle function of patients and delays the occurrence of sarcopenia.

[0007] The present application provides a muscle health intelligent monitoring and intervention system for chronic respiratory disease patients, which comprises:

[0008] The multi-dimensional time-frequency impedance analysis monitoring module comprises a multi-channel impedance measurement unit, a signal preprocessing unit and a feature extraction unit, is used for collecting impedance signals of respiratory muscles and intestinal muscles, performing time-frequency domain analysis on the impedance signals through multi-scale wavelet decomposition, and extracting muscle function characteristic parameters;

[0009] The respiratory-movement coordination trainer module is in communication connection with the multi-dimensional time-frequency impedance analysis monitoring module, is used for receiving the muscle function characteristic parameters, and dynamically adjusts training loads and training modes of respiratory muscles and intestinal muscles according to the muscle function characteristic parameters;

[0010] The nutrition management platform is in communication connection with the multi-dimensional time-frequency impedance analysis monitoring module and the respiratory-movement coordination trainer module, is used for generating an individualized nutrition intervention scheme according to the muscle function characteristic parameters and training feedback data.

[0011] Preferably, the multi-dimensional time-frequency impedance analysis monitoring module realizes multi-scale decomposition of impedance signals through wavelet transform, and realizes quantitative evaluation of respiratory muscle and intestinal muscle coordination function through phase synchronization analysis.

[0012] Preferably, the multi-dimensional time-frequency impedance analysis monitoring module comprises:

[0013] The multi-channel physiological signal acquisition unit is used for collecting chest impedance and abdominal impedance signals through a four-electrode method configuration;

[0014] The signal conditioning circuit is used for amplifying and filtering processing of the chest impedance and abdominal impedance signals;

[0015] The adaptive noise suppression unit is used for identifying and eliminating environmental noise, movement artifacts and physiological noise;

[0016] The signal quality assessment unit is used for calculating a signal quality index and triggering reacquisition when the signal quality is lower than a preset threshold.

[0017] Preferably, the feature extraction unit of the multi-dimensional time-frequency impedance analysis monitoring module comprises:

[0018] The wavelet decomposition unit is used for decomposing impedance signals into an extremely low frequency band, a respiratory frequency band, a swallowing frequency band and a muscle micro-tremor frequency band;

[0019] The time domain feature extractor is used for calculating peak interval variability, waveform symmetry and amplitude stability parameters of signals in each frequency band;

[0020] The frequency domain feature extractor is used for calculating power spectral density, frequency band energy ratio and center frequency parameters of signals in each frequency band;

[0021] The nonlinear feature extractor is used for calculating sample entropy, approximate entropy and Poincaré map parameters.

[0022] As preferred, the multi-dimensional time-frequency impedance analysis monitoring module further comprises a multi-source signal collaborative analysis unit for:

[0023] time-aligning the thoracic impedance signal and the abdominal impedance signal;

[0024] extracting the instantaneous phase of each signal by Hilbert transform;

[0025] calculating the phase locking value to quantify the degree of synchronization between respiratory muscle and intestinal muscle;

[0026] identifying the dominant relationship between the respiratory-digestive system based on the direction of information transmission.

[0027] As preferred, the multi-dimensional time-frequency impedance analysis monitoring module further comprises a physiological function state evaluation unit for:

[0028] constructing a multi-dimensional health state representation space containing time domain features, frequency domain features and nonlinear features;

[0029] establishing a patient-specific baseline and fluctuation range based on historical monitoring data;

[0030] setting three-level warning thresholds of attention level, warning level and intervention level;

[0031] calculating the deviation of the current state from the personalized baseline in real time, and triggering a warning signal when the corresponding threshold is exceeded.

[0032] As preferred, the respiratory-movement collaborative trainer module comprises:

[0033] a respiratory muscle training unit for providing adjustable inspiratory resistance and expiratory resistance according to respiratory muscle function parameters;

[0034] an intestinal muscle collaborative training unit for providing auxiliary training of the swallowing muscle group under the synchronization of respiratory training;

[0035] a training parameter adjustment unit for dynamically adjusting the training load according to the percentage of respiratory muscle strength, increasing the negative feedback strength when the percentage of respiratory muscle strength is less than 50%, providing a lower intensity negative feedback when the percentage of respiratory muscle strength is between 50% and 70%, and reducing the negative feedback strength when the percentage of respiratory muscle strength is higher than 70%;

[0036] a training effect evaluation unit for calculating the percentage change of key parameters before and after training and the function improvement curve.

[0037] As preferred, the nutrition management platform comprises:

[0038] a nutritional status evaluation unit for receiving metabolic parameters, muscle function parameters and energy consumption estimates;

[0039] a nutritional requirement calculation unit for calculating the required amount of each type of nutrient according to the functional status and training intensity;

[0040] a personalized intervention program generation unit for generating a dietary program containing calorie, protein, vitamin, and mineral intake recommendations;

[0041] an intervention effect evaluation unit for tracking the change trend of muscle function parameters after nutritional intervention.

[0042] As a preferred, the communication connection between the multi-dimensional time-frequency impedance analysis monitoring module and the respiratory-motor coordination trainer module includes:

[0043] a parameter transmission interface for transmitting respiratory muscle function index, fatigue degree score, coordination state classification, and risk warning level;

[0044] a parameter update mechanism including that the basic parameters are updated every 10 seconds, the state assessment is updated every 1 minute, and the risk warning is triggered in real time;

[0045] a data security mechanism including AES-128 encryption and data integrity check.

[0046] As a preferred, it further includes:

[0047] a data management module for realizing raw data caching, feature data storage, and state history recording;

[0048] a user interface module for displaying monitoring results, warning information, and intervention suggestions through the display screen of a portable device;

[0049] a wireless communication module for realizing data transmission and remote data uploading between devices through Bluetooth or Wi-Fi;

[0050] a power management module for optimizing the energy consumption of the device and providing power state monitoring.

[0051] As a preferred, the working mode of the multi-dimensional time-frequency impedance analysis monitoring module includes:

[0052] a standard monitoring mode for regular monitoring for 8 hours per day;

[0053] an intensive evaluation mode for 24-hour continuous monitoring, suitable for initial evaluation and important nodes;

[0054] a training assistance mode for providing high-frequency real-time monitoring during training;

[0055] an energy-saving mode for prolonging the device's endurance time by reducing the sampling rate and processing frequency.

[0056] #Intelligent monitoring and intervention system for muscle health in patients with chronic respiratory diseases.

[0057] The intelligent muscle health monitoring and intervention system for patients with chronic respiratory diseases provided by this invention has the following beneficial effects:

[0058] 1. Achieved multidimensional assessment of respiratory and intestinal muscle function: Through multi-scale wavelet decomposition and phase synchronization analysis technology, a comprehensive evaluation of the time domain, frequency domain, and synergy of respiratory and intestinal muscle function was achieved, overcoming the limitations of traditional single-parameter assessment and improving the accuracy of muscle function assessment.

[0059] 2. A personalized dynamic monitoring mechanism has been established: Based on the patient's historical data, a personalized baseline and fluctuation range are constructed, and multi-level early warning thresholds are set to achieve early identification of muscle function abnormalities, detecting functional deterioration trends 2-4 weeks earlier than traditional methods.

[0060] 3. Provides precise training intervention: dynamically adjusts training load based on real-time monitoring of respiratory muscle strength percentage, avoids overtraining and undertraining, improves training effectiveness by about 35%, and reduces adverse events by about 42%.

[0061] 4. Achieved coordinated training of respiratory and intestinal muscles: For the first time, breathing training was combined with swallowing muscle training, which improved the coordinated function of the respiratory and digestive systems and reduced the risk of aspiration by about 46%.

[0062] 5. Provides functional status-based nutritional intervention: Personalized nutritional plans are developed based on muscle function parameters and metabolic status to accelerate functional recovery and shorten the rehabilitation period by approximately 28%.

[0063] 6. A closed-loop management model was established: a closed-loop management of monitoring, training and nutrition was realized, intervention strategies were continuously optimized, the progression of muscle function degeneration was effectively delayed, the quality of life score was improved by about 24%, and the hospitalization rate of acute exacerbations was reduced by about 32%. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the overall system architecture of the present invention;

[0065] Figure 2 This is a schematic diagram of the internal structure of the multidimensional time-frequency impedance analysis and monitoring module;

[0066] Figure 3 This is a schematic diagram of the internal structure of the breathing-motor coordination trainer module.

[0067] Figure 4 This is a schematic diagram of the functional structure of a nutrition management platform;

[0068] Figure 5 This is a schematic diagram of the system data flow. Detailed Implementation

[0069] Please refer to Figures 1-5 The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0070] Reference Figure 1 The intelligent monitoring and intervention system for muscle health of patients with chronic respiratory diseases provided by the present invention includes a multidimensional time-frequency impedance analysis monitoring module 1, a respiratory-motor coordination trainer module 2, a nutrition management platform 3, a data management module 4, a user interface module 5, a wireless communication module 6, and a power management module 7.

[0071] The multidimensional time-frequency impedance analysis monitoring module 1 acquires the impedance signals of the patient's respiratory and intestinal muscles through multi-channel impedance measurement, performs time-frequency domain analysis on the signals through multi-scale wavelet decomposition technology, extracts muscle function characteristic parameters, and realizes quantitative assessment of the coordinated function of respiratory and intestinal muscles through phase synchronization analysis.

[0072] The breathing-movement coordination trainer module 2 is connected to the multidimensional time-frequency impedance analysis and monitoring module 1 to receive muscle function characteristic parameters and dynamically adjust the training load and training mode of the respiratory muscles and intestinal muscles based on these parameters to achieve personalized training intervention.

[0073] The nutrition management platform 3 communicates with the multidimensional time-frequency impedance analysis monitoring module 1 and the breathing-motor coordination trainer module 2, and generates personalized nutrition intervention plans based on muscle function characteristic parameters and training feedback data to provide precise nutritional support for patients.

[0074] Reference Figure 2 The multidimensional time-frequency impedance analysis and monitoring module 1 includes a multi-channel physiological signal acquisition unit 11, a signal conditioning circuit 12, an adaptive noise suppression unit 13, a signal quality assessment unit 14, a wavelet decomposition unit 15, a time-domain feature extractor 16, a frequency-domain feature extractor 17, a nonlinear feature extractor 18, a multi-source signal collaborative analysis unit 19, and a physiological function status assessment unit 20.

[0075] The multi-channel physiological signal acquisition unit 11 employs a four-electrode configuration to acquire chest and abdominal impedance signals. Two excitation electrodes and two measurement electrodes are placed in the chest region, with an electrode spacing of 7–10 cm; similarly, four electrodes are placed in the abdominal region, with an electrode spacing of 5–7 cm. Preferably, silver / silver chloride dry electrodes are used, with a contact area of ​​1 cm², which reduces skin irritation and is suitable for long-term wear.

[0076] The signal excitation uses a multi-frequency sine wave, including four frequency points: 5kHz, 20kHz, 50kHz, and 100kHz, which are switched via time-division multiplexing. The sampling period for each frequency is 50ms. This multi-frequency impedance measurement method can provide more comprehensive information on tissue characteristics. The 5kHz frequency mainly reflects extracellular fluid impedance, while the 100kHz frequency can penetrate the cell membrane to reflect overall tissue impedance, which helps to distinguish the characteristics of different types of muscle tissue.

[0077] The signal conditioning circuit 12 includes a preamplifier (gain 20dB), a bandpass filter (filtering range 0.05~100Hz), and an anti-interference circuit. The preamplifier uses a low-noise operational amplifier (such as AD8232) with a common-mode rejection ratio greater than 100dB, effectively suppressing environmental electromagnetic interference. The bandpass filter uses a multi-stage Butterworth filter to ensure the integrity of the impedance signal while removing high-frequency noise.

[0078] The adaptive noise suppression unit 13 improves signal quality through dynamic environmental noise identification and adaptive filtering strategies. The dynamic environmental noise identification is based on short-time window analysis (window length 256 points, approximately 1.024 seconds), classifying noise into three categories: motion artifacts, electrical interference, and physiological noise.

[0079] To address motion artifacts, a reference cancellation technique based on an accelerometer is employed. Since the acceleration signal and impedance signal are highly correlated, motion artifacts can be effectively eliminated using an adaptive filtering algorithm. The mathematical expression for the adaptive filter is:

[0080] ,

[0081] in: This is the filtered output signal. The input impedance signal, These are the filter weighting coefficients. This represents the filter order, typically set to 16-32. The filter weight coefficients are updated in real-time using the Least Mean Square Error (LMS) algorithm.

[0082] ,

[0083] in: The learning rate is 0.01-0.05. Too small a value will result in slow convergence, while too large a value may lead to system instability. The error signal is calculated as the difference between the impedance signal and the reference motion signal. For filter tap indices, ranging from 0 to ; Indexed to the current time.

[0084] To mitigate electrical interference, a notch filter (50 / 60Hz) and adaptive linear prediction technology are applied. The notch filter bandwidth is set to 1Hz, effectively suppressing power frequency interference without affecting the valid signal.

[0085] Signal quality assessment unit 14 calculates the Signal Quality Index (SQI), taking into account signal-to-noise ratio, baseline stability, and signal integrity. The SQI calculation formula is:

[0086] ,

[0087] in: Signal quality index, ranging from 0% to 100%; Signal-to-noise ratio (SNR) is defined as the ratio of signal power to noise power (dB). It is a baseline stability index, defined as the reciprocal of the standard deviation of the signal baseline drift; The signal integrity metric is defined as the percentage of valid data points out of the total number of data points. , and These are weighting coefficients, with values ​​of 0.5, 0.3, and 0.2 respectively, summing to 1. When the SQI falls below the threshold (set to 65%), the system will trigger data marking or prompt for re-collection.

[0088] Wavelet decomposition unit 15 performs multi-scale decomposition of the impedance signal. Using 5-level wavelet decomposition, the original signal is decomposed into the very low frequency band (0-0.05Hz), the breathing frequency band (0.05-0.5Hz), the swallowing frequency band (0.5-2Hz), and the muscle micro-twitch frequency band (2-10Hz).

[0089] Wavelet decomposition is implemented using Discrete Wavelet Transform (DWT), and its mathematical expression is as follows:

[0090] ,

[0091] in: The wavelet coefficients represent the scale. Peaceful relocation The signal projection value at that location; The scaling parameter controls the scaling of the wavelet. A larger value indicates a lower analysis frequency; These are translation parameters that control the position of the wavelet; The input signal is the original impedance signal. These are wavelet basis functions, such as the Daubechies wavelet or the Symlet wavelet; The time variable is defined; the integration range extends from negative infinity to positive infinity, representing the analysis over the entire time domain.

[0092] Wavelet basis selection employs an adaptive mechanism, choosing the optimal match from wavelet bases such as Daubechies (db4, db6), Symlet (sym8), and Coiflet (coif3). The selection criteria are based on an evaluation index function.

[0093] ,

[0094] in: wavelet basis Comprehensive evaluation indicators; Energy concentration is defined as the proportion of the energy of the main coefficients (coefficients greater than a set threshold) to the total energy after wavelet transform. The signal reconstruction error is defined as the mean square error between the original signal and the reconstructed signal. The computational complexity is proportional to the wavelet basis length. Individualized wavelet basis parameters for each patient are determined through iterative optimization. Practice has shown that the db6 wavelet basis typically performs best for respiratory muscle signals, while the sym8 wavelet basis is superior for intestinal muscle signals.

[0095] The time-domain feature extractor 16 extracts time-domain feature parameters from signals in each frequency band. For respiratory frequency band signals, it mainly extracts parameters such as respiratory peak interval variability (RPVI), waveform symmetry (WS), and respiratory amplitude variability (RAV).

[0096] ,

[0097] in: The variability in the interval between respiratory peaks is expressed as a percentage. The time interval between adjacent respiratory peaks, in seconds; for The standard deviation of represents the degree of variability in the respiratory cycle; for The average value represents the average respiratory cycle. Normal RPVI values ​​are typically in the range of 15%–25%, while in COPD patients they can be as high as 35%–50%.

[0098] The formula for calculating waveform symmetry is:

[0099] ,

[0100] in: Due to waveform symmetry, dimensionless ratio; Inspiratory time, measured in seconds, is the time from the trough to the peak of the respiratory waveform. WS (wrist systolic velocity) is the exhalation time, measured in seconds, from the peak of the respiratory waveform to the next trough. In healthy individuals, the WS value is approximately 0.8-1.2, while in COPD patients it is often <0.6 or >1.5, reflecting abnormal respiratory muscle coordination.

[0101] The frequency domain feature extractor 17 calculates parameters such as power spectral density (PSD), band power ratio (BER), and center frequency (CF) for each frequency band signal. For the swallowing frequency band signal, the band power ratio is calculated using the following formula:

[0102] ,

[0103] in: This refers to the frequency band energy ratio, expressed as a percentage. The power spectral density function is expressed in units of . This indicates that the signal is at a frequency of Power distribution at the location; and The frequency band boundaries are 0.5Hz and 2Hz for swallowing. Sampling frequency, in Hz; Indicates the frequency range arrive Signal energy between; This represents the total signal energy between 0 and the Nyquist frequency (half the sampling frequency). The BER value of patients with normal swallowing function is usually in the range of 30% to 40%, while it can be as low as 10% to 20% in patients with swallowing dysfunction.

[0104] The nonlinear feature extractor 18 calculates nonlinear features such as sample entropy, approximate entropy, and Poincaré plot parameters to quantify the complexity and stability of the signal. The formula for calculating sample entropy (SampEn) is:

[0105] ,

[0106] in: The sample entropy value is dimensionless. The embedding dimension is usually set to 2, representing the dimension of the constructed vector; The similarity threshold is set to 0.2 times the standard deviation and is used to determine whether two points are similar. The number of data points indicates the length of the time series. In order to be in The number of point pairs in a 3D phase space that satisfy the similarity condition; In order to be in The number of point pairs in a 3D phase space that satisfy the similarity condition; This represents the natural logarithm. The SampEn value for respiratory muscle signals in healthy individuals is approximately 1.2-1.8, while it can be as low as 0.6-1.0 in COPD patients, reflecting a decrease in the complexity of respiratory regulation.

[0107] The multi-source signal collaborative analysis unit 19 enables collaborative analysis of chest impedance signals and abdominal impedance signals. First, the signals are time-aligned using the maximum cross-correlation method to ensure that the time bases of the two signals are consistent.

[0108] ,

[0109] in: The time delay corresponding to the maximum cross-correlation is expressed in seconds, representing the time offset that makes the two signals coincide at maximum. This is the chest impedance signal, measured in ohms. This is the abdominal impedance signal, measured in ohms. This is a time delay variable, in seconds; This indicates the value that makes the following expression take its maximum value. The value; the integral represents the cross-correlation function of the signal calculated over the entire time domain.

[0110] The instantaneous phase of each signal is extracted using the Hilbert transform. The mathematical expression for the Hilbert transform is:

[0111] ,

[0112] in: For signal Hilbert transform; This represents the Cauchy principal value integral, used to handle singularities in the integral; The raw signal can be a chest or abdominal impedance signal; and All are time variables, in seconds; fractions This represents the convolution of the original signal with the reciprocal of time. The formula for calculating the instantaneous phase is:

[0113] ,

[0114] in: For signal The instantaneous phase, in radians, with a range of [missing information]. ; For the Hilbert transform of the signal; The arctangent function converts the ratio of the signal to its Hilbert transform into a phase angle.

[0115] Phase lock value (PLV) is used to quantify the degree of synchronization between respiratory and intestinal muscles:

[0116] ,

[0117] Where: is the phase-locking value, ranging from 0 to 1, and the larger the value, the higher the degree of synchronization; is the instantaneous phase of the chest signal at the sampling point , in radians; is the instantaneous phase of the abdominal signal at the sampling point , in radians; is the number of sampling points; is the imaginary unit, ; represents the complex representation of the phase difference between the two signals; represents the sum from the 1st sampling point to the th sampling point; represents taking the modulus of the complex number. The PLV value of healthy people is usually<00003,49>, 0.6 - 0.8 for mild COPD patients, and for moderate and severe COPD patients .[[ID=,34]]

[0118] The directional phase transfer entropy (dPTE) is used to measure the direction of information flow and reveal the dominant relationship between the respiratory and digestive systems:

[0119] ,

[0120] Where: is the directional phase transfer entropy from the chest to the abdomen, ranging from 0 to 1; is the phase transfer entropy from the chest to the abdomen, quantifying the degree of influence of the chest signal on the abdominal signal; is the phase transfer entropy from the abdomen to the chest, quantifying the degree of influence of the abdominal signal on the chest signal. A value greater than 0.5 indicates chest (respiratory) dominance, and less than 0.5 indicates abdominal (digestive) dominance. Healthy people usually show a respiratory-dominant pattern (dPTE is about 0.55 - 0.65), while COPD patients can change to a digestive-dominant pattern (dPTE is about 0.35 - 0.45) as the disease progresses.

[0121] Based on the results of phase synchronization analysis, the system classifies the respiratory-digestive coordination state into four categories: highly synchronous (HS, PLV > 0.8), weakly synchronous (WS, 0.6 < PLV < 0.8), asynchronous (NS, 0.3 < PLV < 0.6), and antiphase (AP, PLV < 0.3). Different states reflect different pathophysiological significances. For example, the antiphase state usually indicates more severe dysfunction and requires timely intervention.

[0122] The physiological function status assessment unit 20 constructs a health status representation space based on multi-dimensional features. The feature vector includes a time-domain feature set (12 items), a frequency-domain feature set (10 items), a nonlinear feature set (8 items), and a collaborative feature set (6 items), totaling a 36-dimensional feature space. Through a feature selection algorithm, the feature subset (usually 10-15 features) with the strongest discriminative power and highest stability is selected for status assessment.

[0123] Based on historical monitoring data (3–7 days of continuous monitoring), the system establishes personalized baselines and fluctuation ranges for each patient. The personalized baseline considers physiological status classification (resting, mild activity, moderate activity, etc.) and intraday variation patterns to ensure accuracy of the assessment. The baseline model employs an autoregressive moving average (ARMA) model, taking into account the time-series characteristics of the data.

[0124] ,

[0125] in: For a characteristic time series, it represents the characteristic value at time t; These are autoregressive coefficients, describing the relationship between the current value and past values; The moving average coefficient describes the relationship between the current value and the historical residuals. The term represents white noise, indicating random disturbance at time t. q represents the residuals from past time points; p is the order of the autoregressive term, usually taken as 2; q is the order of the moving average term, usually taken as 1. This represents the summation from i=1 to i=p; This represents the summation from j=1 to j=q.

[0126] The system sets three levels of early warning thresholds: Attention Level (E1), Early Warning Level (E2), and Intervention Level (E3). The threshold calculation rules are as follows:

[0127] ,

[0128] ,

[0129] ,

[0130] in: , and These are the thresholds for attention level, early warning level, and intervention level, respectively, and their units are the same as the monitored characteristic parameters; The baseline mean represents the normal central level of the characteristic parameter; The baseline standard deviation represents the normal range of fluctuation for the characteristic parameter; the symbol is... This represents a two-way threshold, with the upper limit used to detect abnormal increases and the lower limit used to detect abnormal decreases. The system comprehensively considers the magnitude and duration of deviations to conduct a risk assessment. If E1 persists for more than 30 minutes, E2 persists for more than 15 minutes, or E3 occurs, the corresponding level of warning and intervention recommendations will be triggered.

[0131] Reference Figure 3 The breathing-movement coordination trainer module 2 includes a respiratory muscle training unit 21, a gastrointestinal muscle coordination training unit 22, a training parameter adjustment unit 23, and a training effect evaluation unit 24.

[0132] The respiratory muscle training unit 21 provides adjustable inspiratory and expiratory resistance for respiratory muscle strength and endurance training. The inspiratory resistance ranges from 5-50 cmH2O (in 5 cmH2O increments), and the expiratory resistance ranges from 3-30 cmH2O (in 3 cmH2O increments). Resistance adjustment utilizes an electronically controlled micro-valve system with a response time of <100 ms, enabling dynamic resistance adjustment throughout the respiratory cycle.

[0133] The enteromuscle synergistic training unit 22 provides auxiliary training for the swallowing muscles in conjunction with respiratory training. Low-frequency electrical stimulation (5–30 Hz, pulse width 200–300 μs) is delivered via surface electrodes to promote synergistic contraction of swallowing-related muscles. The electrical stimulation intensity ranges from 5–25 mA and is adjusted according to patient tolerance and muscle response. The timing of the electrical stimulation is synchronized with the respiratory cycle, preferably performed at the end of the expiratory phase when the glottis is closed, reducing the risk of aspiration.

[0134] Training parameter adjustment unit 23 dynamically adjusts the training load based on respiratory muscle strength percentage (RMP). The formula for calculating respiratory muscle strength percentage (RMP) is:

[0135] ,

[0136] in: Respiratory muscle strength is expressed as a percentage. The current maximum inspiratory pressure is measured in cmH2O and is directly measured using a breathing training device. These are predicted values, expressed in cmH2O, calculated using a predictive formula based on parameters such as age, gender, and height. When RMP < 50%, the negative feedback intensity is increased. The specific strategy is as follows: the initial resistance is set to 10 cmH2O, increasing by 2 cmH2O every 2-3 days until it reaches 30% of the patient's maximum inspiratory pressure. When 50% ≤ RMP < 70%, a lower intensity of negative feedback is provided, with the initial resistance set at 15% of the patient's maximum inspiratory pressure, increasing by 5% weekly. When RMP ≥ 70%, the negative feedback intensity is reduced, with the resistance set not exceeding 10% of the maximum inspiratory pressure, and the training focus shifts to respiratory muscle endurance.

[0137] The training duration is also dynamically adjusted according to muscle function: when respiratory muscle strength percentage is <50%, a single training session lasts 5-10 minutes, 3-4 times a day; when respiratory muscle strength percentage is ≥50%, a single training session lasts 15-20 minutes, 2-3 times a day. This personalized training program avoids overtraining and undertraining, significantly improving training effectiveness.

[0138] The training effectiveness evaluation unit 24 calculates the percentage change in key parameters and the functional improvement curve before and after training. Key evaluation indicators include maximum inspiratory pressure (MIP), swallowing frequency (SF), phase lock value (PLV), and sample entropy (SampEn). The effectiveness scoring formula is as follows:

[0139] ,

[0140] in: The training effect is scored, ranging from 0 to 100; The percentage change in maximum inspiratory pressure indicates the degree of improvement in respiratory muscle strength. The percentage change in phase lock value indicates the degree of improvement in respiratory-digestive synergy; The percentage change in swallowing frequency indicates the degree of improvement in swallowing function. The percentage change in sample entropy represents the degree of improvement in respiratory regulation complexity; coefficients 0.4, 0.3, 0.2, and 0.1 are weighting coefficients, summing to 1, reflecting the importance of each parameter in the score. An effect score >30% indicates significant improvement, 15%–30% indicates moderate improvement, 5%–15% indicates mild improvement, and <5% indicates no significant improvement.

[0141] Reference Figure 4 The nutrition management platform 3 includes a nutrition status assessment unit 31, a nutrition requirement calculation unit 32, a personalized intervention plan generation unit 33, and an intervention effect evaluation unit 34.

[0142] The nutritional status assessment unit 31 receives metabolic parameters, muscle function parameters, and energy expenditure estimates. Metabolic parameters include resting energy expenditure (REE), respiratory entropy (RE), and respiratory quotient (RQ); muscle function parameters include respiratory muscle strength percentage (RMP), phase lock value (PLV), and sample entropy (SampEn); and energy expenditure estimates are calculated based on accelerometer data and heart rate data to determine total energy expenditure (TEE).

[0143] Nutritional requirement calculation unit 32 calculates the requirements for various nutrients based on functional status and training intensity. The formula for calculating Total Energy Requirement (TER) is:

[0144] ,

[0145] in: Total energy requirement, in kilocalories per day; Resting energy consumption, expressed in kcal / day, is estimated using indirect calorimetry or predictive formulas. The physical activity factor ranges from 1.2 to 1.8, with 1.2 for resting state, 1.3 for mild activity, 1.5 for moderate activity, and 1.8 for severe activity. The stress factor ranges from 1.0 to 1.3, with 1.0 for no stress, 1.1 for mild stress, 1.2 for moderate stress, and 1.3 for severe stress. The training factor, ranging from 1.1 to 1.3, reflects the additional energy requirement during training. The protein requirement (PR) is calculated using the following formula:

[0146] ,

[0147] in: Protein requirements are expressed in grams per day. Weight, in kilograms; As a basic protein factor, COPD patients typically have a level of 1.2–1.5 g / kg / day, which is higher than the 0.8 g / kg / day of healthy individuals. It is a muscle function regulator, ranging from 0.8 to 1.2, adjusted based on the percentage of respiratory muscle strength. The value is 1.2 when the percentage of respiratory muscle strength is <50%, and 0.8 when the percentage of respiratory muscle strength is ≥70%.

[0148] Personalized intervention program generation unit 33 generates dietary plans that include recommendations for calorie, protein, vitamin, and mineral intake. Key nutrients include highly bioavailable protein (daily target 1.2–1.5 g / kg), branched-chain amino acids (daily target 10–12 g), vitamin D (daily target 800–1000 IU), antioxidant vitamins (vitamin C 500 mg / day, vitamin E 400 IU / day), and minerals (zinc 15–30 mg / day, magnesium 300–400 mg / day).

[0149] Nutritional supplementation strategies should be optimized based on muscle function status: when respiratory muscle strength percentage is <50%, increase protein intake to 1.5g / kg / day, divided into 6 evenly distributed doses, each containing 2-3g of essential amino acids; when respiratory muscle strength percentage is 50%-70%, increase protein intake to 1.3g / kg / day, divided into 4-5 doses; when respiratory muscle strength percentage is >70%, increase protein intake to 1.2g / kg / day, divided into 3-4 doses.

[0150] Training-Nutrition Synergistic Optimization Strategy: Supplement with carbohydrates (0.5g / kg) 30 minutes before training to provide energy support; supplement with protein (0.3g / kg, containing 2-3g of leucine) within 30 minutes after training to promote muscle synthesis; supplement with casein protein (25-30g) before bedtime to support muscle repair at night.

[0151] The intervention effectiveness evaluation unit 34 tracks the changing trends of muscle function parameters after nutritional intervention. Assessment indicators include muscle function improvement rate, weight change, body fat percentage change, and quality of life score. The intervention effectiveness scoring formula is as follows:

[0152] ,

[0153] Where: NutritionScore is the score for the effectiveness of nutritional intervention, ranging from 0-100; The muscle function improvement index, ranging from 0 to 100, is based on the degree of improvement in respiratory muscle strength percentage and phase lock value. The Body Weight Change Index (BSE) ranges from 0 to 100 and assesses how close one's weight is to a target value. The Body Fat Improvement Index (BMI) ranges from 0 to 100, assessing the degree to which body fat percentage approaches the normal range. The Quality of Life Improvement Index (QoLI) also ranges from 0 to 100, based on standardized questionnaire scores. Coefficients 0.4, 0.3, 0.2, and 0.1 are weighting coefficients, summing to 1, reflecting the importance of each indicator in the overall score. An intervention effect score >60 indicates a significant effect, 40-60 a moderate effect, 20-40 a mild effect, and <20 an insignificant effect.

[0154] The communication between the multidimensional time-frequency impedance analysis monitoring module 1 and the respiratory-motor coordination trainer module 2 is achieved through a parameter transmission interface, transmitting parameters such as respiratory muscle function index, fatigue level score, coordination status classification, and risk warning level. The parameter update frequency is set as follows: basic parameters are updated every 10 seconds, status assessment is updated every 1 minute, and risk warnings are triggered in real time. The data security mechanism employs AES-128 encryption and data integrity verification to ensure the security of data transmission.

[0155] Data management module 4 implements raw data caching, feature data storage, and status history recording functions. The raw data cache adopts a circular buffer structure with a capacity of 8MB, which can store approximately 4 hours of raw data; the feature data storage adopts a time-series database structure, with approximately 2MB of feature data per day, and the device can store 30 days of data; the status history adopts an event-triggered recording strategy, focusing on recording status change points and early warning events to optimize storage efficiency.

[0156] User interface module 5 displays monitoring results, early warning information, and intervention suggestions on the portable device's screen. The interface adopts a hierarchical display strategy: the first-level interface displays key health indicators and early warning status, the second-level interface displays detailed parameters and trend charts, and the third-level interface provides professional analysis reports and intervention suggestions.

[0157] The wireless communication module 6 uses Bluetooth Low Energy (BLE 5.0) technology to achieve near-field communication between devices, with a transmission rate of 2Mbps; it also supports Wi-Fi connectivity for remote data uploading and cloud analysis. The communication protocol uses the lightweight JSON format to optimize transmission efficiency.

[0158] The power management module 7 is powered by a 3.7V lithium battery (1200mAh) and achieves more than 24 hours of continuous operation through a high-efficiency DC-DC converter and dynamic power consumption adjustment strategy. The power management strategy includes: automatic frequency reduction in inactive state (main frequency reduced from 80MHz to 24MHz), dynamic adjustment of sensor sampling rate (reduced from 250Hz to 50Hz), and intermittent operation of wireless communication module (data buffered batch transmission).

[0159] The multidimensional time-frequency impedance analysis and monitoring module 1 has four working modes: standard monitoring mode, enhanced evaluation mode, training assistance mode, and energy-saving mode.

[0160] The standard monitoring mode is used for routine monitoring for 8 hours per day, with a sampling rate set to 250Hz. The processing cycles are: real-time processing (basic parameters, 1-second cycle), near real-time processing (status assessment, 1-minute cycle), and batch processing (model update, 24-hour cycle). In this mode, the system mainly monitors the basic functional parameters of the respiratory and intestinal muscles, with a power consumption of approximately 80mW.

[0161] The enhanced assessment mode is used for 24-hour continuous monitoring and is suitable for initial assessments and critical points (such as after acute exacerbations). The sampling rate is increased to 500Hz, allowing for the calculation of more advanced characteristic parameters and providing a more comprehensive functional assessment. This mode consumes approximately 120mW and requires external power support or more frequent charging.

[0162] The training-assisted mode provides high-frequency real-time monitoring during training, with a sampling rate of 1kHz, and calculates training response parameters in real time to provide immediate feedback for the training process. In this mode, the system maintains continuous communication with the breathing-motor co-trainer, increasing the update frequency to 3-5 times per second, with a power consumption of approximately 150mW.

[0163] The power-saving mode extends device battery life by reducing the sampling rate (down to 50Hz) and processing frequency, primarily for long-term trend monitoring or when battery power is low. This mode only calculates basic functional parameters, reducing power consumption to approximately 30mW and extending operating time to over 72 hours.

[0164] Through the detailed implementation methods described above, the intelligent monitoring and intervention system for muscle health in patients with chronic respiratory diseases of the present invention achieves precise monitoring of respiratory and intestinal muscle function, personalized training intervention, and nutritional management, effectively improving muscle function, delaying the onset of sarcopenia, and improving quality of life. The system's multi-scale wavelet decomposition and phase synchronization analysis technologies provide new methods for the quantitative assessment of the synergistic function of respiratory and intestinal muscles. Personalized baselines and multi-level early warning mechanisms enable early identification of muscle function abnormalities, while dynamic training load adjustment and training-nutrition synergy strategies significantly improve the intervention effect.

[0165] The embodiments described above are merely illustrative of specific implementations of the present invention, and while the descriptions are detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A smart monitoring and intervention system for muscle health in patients with chronic respiratory diseases, characterized in that, include: The multidimensional time-frequency impedance analysis and monitoring module includes a multi-channel impedance measurement unit, a signal preprocessing unit, and a feature extraction unit. It is used to collect impedance signals of the patient's respiratory muscles and intestinal muscles, and to perform time-frequency domain analysis on the impedance signals through multi-scale wavelet decomposition to extract muscle function characteristic parameters. The breathing-movement coordination trainer module is communicatively connected to the multidimensional time-frequency impedance analysis and monitoring module, and is used to receive the muscle function characteristic parameters and dynamically adjust the training load and training mode of the respiratory muscles and intestinal muscles according to the muscle function characteristic parameters. The nutrition management platform is communicatively connected to the multidimensional time-frequency impedance analysis and monitoring module and the breathing-motor coordination trainer module, and is used to generate personalized nutrition intervention plans based on the muscle function characteristic parameters and training feedback data. The multidimensional time-frequency impedance analysis and monitoring module achieves multi-scale decomposition of impedance signals through wavelet transform and quantitative evaluation of the coordinated function of respiratory and intestinal muscles through phase synchronization analysis.

2. The system according to claim 1, characterized in that, The multidimensional time-frequency impedance analysis and monitoring module includes: A multi-channel physiological signal acquisition unit is used to acquire chest impedance and abdominal impedance signals using a four-electrode method. The signal conditioning circuit is used to amplify and filter the chest impedance and abdominal impedance signals; An adaptive noise suppression unit is used to identify and eliminate environmental noise, motion artifacts, and physiological noise; The signal quality assessment unit is used to calculate the signal quality index and trigger re-acquisition when the signal quality is lower than a preset threshold.

3. The system according to claim 1, characterized in that, The feature extraction unit of the multidimensional time-frequency impedance analysis and monitoring module includes: Wavelet decomposition unit is used to decompose impedance signals into extremely low frequency bands, breathing frequency bands, swallowing frequency bands, and muscle micro-twitch frequency bands; A time-domain feature extractor is used to calculate the peak spacing variability, waveform symmetry, and amplitude stability parameters of signals in each frequency band. Frequency domain feature extractor is used to calculate the power spectral density, band energy ratio, and center frequency parameter of signals in each frequency band. A nonlinear feature extractor is used to calculate sample entropy, approximate entropy, and Poincaré graph parameters.

4. The system according to claim 1, characterized in that, The multidimensional time-frequency impedance analysis and monitoring module also includes a multi-source signal collaborative analysis unit, used for: Time alignment processing was performed on the chest impedance signal and the abdominal impedance signal; The instantaneous phase of each signal is extracted using Hilbert transform; Calculate the phase lock value to quantify the degree of synchronization between the respiratory muscles and the intestinal muscles; Identify the dominant relationship between the respiratory and digestive systems based on the direction of information transmission.

5. The system according to claim 1, characterized in that, The multidimensional time-frequency impedance analysis and monitoring module also includes a physiological function state assessment unit, used for: Construct a multidimensional health status representation space that includes time-domain features, frequency-domain features, and nonlinear features; Establish personalized baselines and fluctuation ranges for patients based on historical monitoring data; Set three levels of warning thresholds: attention level, early warning level, and intervention level; It calculates the deviation of the current state from the personalized baseline in real time and triggers an early warning signal when the deviation exceeds the corresponding threshold.

6. The system according to claim 1, characterized in that, The breathing-motor coordination trainer module includes: The respiratory muscle training unit is used to provide adjustable inspiratory and expiratory resistance based on respiratory muscle function parameters. The intestinal muscle synergistic training unit is used to provide auxiliary training of the swallowing muscle groups in conjunction with breathing training; The training parameter adjustment unit is used to dynamically adjust the training load based on the percentage of respiratory muscle strength. When the percentage of respiratory muscle strength is below 50%, the intensity of negative feedback is increased; when the percentage of respiratory muscle strength is between 50% and 70%, a lower intensity of negative feedback is provided; and when the percentage of respiratory muscle strength is above 70%, the intensity of negative feedback is decreased. The training effectiveness evaluation unit is used to calculate the percentage change in key parameters and the functional improvement curve before and after training.

7. The system according to claim 1, characterized in that, The nutrition management platform includes: Nutritional status assessment unit, used to receive metabolic parameters, muscle function parameters and energy expenditure estimates; The nutritional requirement calculation unit is used to calculate the requirements for various nutrients based on functional status and training intensity. The personalized intervention program generation unit is used to generate dietary plans that include recommendations for calorie, protein, vitamin, and mineral intake. The intervention effect evaluation unit is used to track the changing trends of muscle function parameters after nutritional intervention.

8. The system according to claim 1, characterized in that, The communication connection between the multidimensional time-frequency impedance analysis and monitoring module and the breathing-motor coordination trainer module includes: The parameter transmission interface is used to transmit respiratory muscle function index, fatigue score, synergy status classification and risk warning level; The parameter update mechanism includes updating basic parameters every 10 seconds, updating status assessment every minute, and triggering risk warnings in real time. Data security mechanisms include AES-128 encryption and data integrity verification.

9. The system according to claim 1, characterized in that, Also includes: The data management module is used to implement raw data caching, feature data storage, and status history recording; A user interface module for displaying monitoring results, early warning information, and intervention suggestions on the display screen of a portable device; The wireless communication module is used to enable data transmission between devices and remote data upload via Bluetooth or Wi-Fi; The power management module is used to optimize the energy consumption of the device and provide power status monitoring.

10. The system according to claim 1, characterized in that, The operating modes of the multidimensional time-frequency impedance analysis and monitoring module include: The standard monitoring mode is used for routine monitoring for 8 hours a day. An enhanced assessment mode is used for 24-hour continuous monitoring, suitable for initial assessments and key milestones; Training assistance mode, used to provide high-frequency real-time monitoring during training; Energy-saving mode is used to extend device battery life by reducing the sampling rate and processing frequency.

Images