Method, device, system, signal processing device and storage medium for calculating an off-line assessment index

Abstract

This application discloses a method, computing device, computing system, signal processing device, and storage medium for calculating an offline assessment index, relating to the field of medical device technology. The method includes: receiving a target diaphragm electrical signal; performing signal purification processing on the target diaphragm electrical signal to obtain a purified diaphragm electrical signal; extracting multi-dimensional physiological features based on the purified diaphragm electrical signal, including time-domain features, frequency-domain features, and stability features; the time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity within each respiratory cycle; the frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal; and the stability features include the coefficient of variation of diaphragm electrical activity intensity within multiple consecutive respiratory cycles; inputting the multi-dimensional physiological features into a fusion assessment model to obtain an offline assessment index. This solution solves the problem of low accuracy in offline assessment indices calculated by existing methods, thus improving the accuracy of the offline assessment index.

Description

Technical Field

[0001] This application relates to the field of medical device technology, and in particular to a method for calculating an offline assessment index, a computing device, a computing system, a signal processing device, and a storage medium. Background Technology

[0002] Mechanical ventilation is a commonly used life support method in intensive care. During the recovery of a patient's respiratory function, it is necessary to assess the patient's ability to wean off mechanical ventilation to determine whether to reduce or discontinue ventilation support.

[0003] The rapid shallow breathing index (RSBI) is a widely used weaning assessment index, calculated as the ratio of respiratory rate to tidal volume. However, the RSBI is essentially an indirect estimation of respiratory mechanics parameters, performing a simple proportional calculation based solely on respiratory rate and tidal volume. Its accuracy is limited, and it carries a high risk of misjudgment; that is, weaning failure may still occur even with a normal RSBI, while patients may successfully wean even with an abnormal RSBI. Therefore, the weaning assessment index calculated using current methods suffers from low accuracy. Summary of the Invention

[0004] This application provides a method, computing device, computing system, signal processing device, and storage medium for calculating an offline evaluation index, in order to solve the problem that the offline evaluation index calculated by existing methods has low evaluation accuracy.

[0005] In a first aspect, embodiments of this application provide a method for calculating an offline evaluation index, including: Receive the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module; wherein, the target diaphragm electrical signal is the signal obtained by the wireless surface diaphragm electrical acquisition module after preprocessing the original diaphragm electrical signal, and the original diaphragm electrical signal is acquired by surface electrodes attached to the preset diaphragm area of ​​the patient. The target diaphragm electrical signal is subjected to signal purification processing to obtain a purified diaphragm electrical signal; wherein, the signal purification processing includes at least one of ECG artifact filtering, power line interference filtering, and motion artifact filtering. Based on the purified diaphragm electrical signal, multi-dimensional physiological features are extracted, including time-domain features, frequency-domain features, and stability features. The time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity in each respiratory cycle. The frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal. The stability features include the coefficient of variation of diaphragm electrical activity intensity in multiple consecutive respiratory cycles. The multidimensional physiological features are input into a pre-constructed fusion evaluation model, so that the fusion evaluation model performs weighted fusion calculation on the multidimensional physiological features and outputs an offline evaluation index.

[0006] Secondly, embodiments of this application provide a device for calculating an offline evaluation index, characterized in that it includes: The receiving module is used to receive the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module; wherein, the target diaphragm electrical signal is the signal obtained by the wireless surface diaphragm electrical acquisition module after preprocessing the original diaphragm electrical signal, and the original diaphragm electrical signal is acquired by surface electrodes attached to a preset diaphragm area on the patient's body surface. The purification module is used to perform signal purification processing on the target diaphragm electrical signal to obtain a purified diaphragm electrical signal; wherein the signal purification processing includes at least one of ECG artifact filtering, power frequency interference filtering, and motion artifact filtering. An extraction module is used to extract multi-dimensional physiological features based on the purified diaphragm electrical signal. The multi-dimensional physiological features include time-domain features, frequency-domain features, and stability features. The time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity in each respiratory cycle. The frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal. The stability features include the coefficient of variation of diaphragm electrical activity intensity in multiple consecutive respiratory cycles. The model processing module is used to input the multi-dimensional physiological features into a pre-constructed fusion evaluation model, so that the fusion evaluation model performs weighted fusion calculation on the multi-dimensional physiological features and outputs an offline evaluation index.

[0007] Thirdly, embodiments of this application provide an offline assessment index calculation system, including: a signal processing device, a wireless body surface diaphragm electroacupuncture module, and body surface electrodes; The wireless surface diaphragm electrical signal acquisition module is used to acquire the original diaphragm electrical signal of a preset diaphragm region on the patient's body surface through the surface electrodes; and to preprocess the original diaphragm electrical signal to obtain the target diaphragm electrical signal. The signal processing device is used to receive the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module; perform signal purification processing on the target diaphragm electrical signal to obtain a purified diaphragm electrical signal; wherein, the signal purification processing includes at least one of ECG artifact filtering, power frequency interference filtering, and motion artifact filtering; based on the purified diaphragm electrical signal, extract multi-dimensional physiological features, including time-domain features, frequency-domain features, and stability features; wherein, the time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity in each respiratory cycle, the frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal, and the stability features include the coefficient of variation of diaphragm electrical activity intensity in multiple consecutive respiratory cycles; input the multi-dimensional physiological features into a pre-constructed fusion evaluation model so that the fusion evaluation model performs weighted fusion calculation on the multi-dimensional physiological features and outputs an offline evaluation index.

[0008] Fourthly, embodiments of this application provide a signal processing apparatus, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the offline evaluation index calculation method described above.

[0009] Fifthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the offline evaluation index calculation method described above.

[0010] Based on this, one technical solution provided in this application involves receiving the target diaphragm electrical signal sent by a wireless surface diaphragm electrical acquisition module to obtain a signal reflecting the state of diaphragm nerve electrical activity, enabling the calculation of the offline assessment index to be directly based on diaphragm functional state information. Furthermore, the target diaphragm electrical signal undergoes signal purification processing to obtain a high-quality purified diaphragm electrical signal. From the purified diaphragm electrical signal, peak amplitude and area under the curve are extracted as time-domain features, the center frequency calculated based on power spectral density is extracted as a frequency-domain feature, and the intensity of diaphragm electrical activity over multiple consecutive respiratory cycles is extracted. The coefficient of variation, as a stability feature, comprehensively quantifies the contraction intensity, fatigue state, and driving stability of the diaphragm from three dimensions: time domain, frequency domain, and stability. This effectively overcomes the information limitations of existing RSBI, which relies solely on simple proportional calculations based on two mechanical dimensions: respiratory rate and tidal volume. Furthermore, the aforementioned multi-dimensional physiological characteristics are input into a fusion assessment model for weighted fusion calculation, outputting a comprehensive offline assessment index. This offline assessment index integrates more dimensions of physiological information, enabling it to more comprehensively reflect the neuroelectrical activity state of the diaphragm and respiratory driving function, thereby improving the assessment accuracy of the offline assessment index. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of an offline evaluation index calculation system in one embodiment of this application; Figure 2 This is a schematic diagram of a wireless diaphragm electrophysiological data acquisition module in one embodiment of this application; Figure 3 This is a schematic diagram of a signal processing device according to an embodiment of this application; Figure 4 This is a flowchart of a method for calculating the offline evaluation index in one embodiment of this application; Figure 5 This is another flowchart of the calculation method of the offline evaluation index in one embodiment of this application; Figure 6 This is another flowchart of the calculation method of the offline evaluation index in one embodiment of this application; Figure 7 This is another flowchart of the calculation method of the offline evaluation index in one embodiment of this application; Figure 8 This is another flowchart of the calculation method of the offline evaluation index in one embodiment of this application; Figure 9 This is a schematic diagram of a display interface of a mechanical ventilation device in one embodiment of this application; Figure 10 This is a schematic diagram of an offline evaluation index calculation device in one embodiment of this application; Figure 11 This is another schematic diagram of a signal processing device in one embodiment of this application. Detailed Implementation

[0013] To make the technical problems, technical solutions, and beneficial effects solved by this application clearer, the following detailed description is provided in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0014] Existing technology: Mechanical ventilation is a critical life support method for saving the lives of patients with respiratory failure in the intensive care unit (ICU). Once the patient's primary disease improves and respiratory function gradually recovers, clinicians need to make a crucial decision—determining whether the patient can be safely weaned off mechanical ventilation support, i.e., weaning. The precise timing of weaning directly affects the patient's prognosis: weaning too early, when the patient's spontaneous breathing ability is insufficient to maintain effective ventilation, may lead to respiratory muscle fatigue, extubation failure, or even the need for reintubation, significantly increasing the incidence of complications such as device-associated pneumonia and airway injury, as well as mortality; weaning too late, on the other hand, prolongs the duration of mechanical ventilation, increases exposure to sedative drugs and the risk of device-associated complications, resulting in unnecessary occupation of medical resources and prolonged hospital stays.

[0015] In clinical practice, the most widely used weaning assessment indicator is the rapid shallow breathing index (RSBI), which is calculated as the ratio of respiratory rate (f) to tidal volume (Vt), i.e., RSBI = f / Vt. This indicator assesses whether a patient has the physiological reserve to be weaned off mechanical ventilation by measuring the patient's respiratory rate and tidal volume and using the ratio of the two. The measurement method of RSBI relies on respiratory mechanics parameters obtained from mechanical ventilation equipment or pulmonary function testing equipment, and is an indirect estimation indicator.

[0016] However, numerous clinical studies have shown that the accuracy of RSBI assessment is limited and carries a high risk of misjudgment. Specifically, even with normal RSBI values, a significant proportion of patients still experience respiratory failure after weaning and require reintubation; conversely, some patients with abnormal RSBI values ​​can be successfully weaned. The fundamental reason for this limitation is that RSBI is essentially a simple proportional calculation of respiratory mechanics parameters, estimating only respiratory rate and tidal volume. It reflects the overall mechanical state of the respiratory system but does not incorporate physiological information that directly reflects the functional state of the respiratory drive—the diaphragm. The success of weaning hinges on the diaphragm's ability to independently bear the ventilatory load after being weaned from mechanical ventilation. RSBI cannot directly assess the diaphragm's contractile capacity, fatigue state, and drive stability, thus its accuracy is insufficient to meet clinical needs.

[0017] Besides RSBI, clinical methods for assessing diaphragmatic function mainly include diaphragmatic ultrasound and transesophageal diaphragmatic electrodiaphragmatic monitoring (EAdi). Diaphragmatic ultrasound measures diaphragmatic movement and diaphragmatic thickening fraction through imaging, thereby indirectly assessing the diaphragm's contractile function. However, this method relies on the operator manually selecting measurement sections and marking measurement points, and the measurement results are greatly affected by the operator's experience, resulting in insufficient consistency in assessment. Furthermore, diaphragmatic ultrasound is an intermittent examination method, unable to achieve continuous dynamic monitoring of diaphragmatic function. Transesophageal diaphragmatic electrodiaphragmatic monitoring (EAdi) involves inserting a special catheter through the nose or mouth into the esophagus to collect diaphragmatic electrical signals at the diaphragm level, directly reflecting the diaphragm's neuroelectrical activity. However, this technique is invasive, the catheter insertion process is complex, patient tolerance is poor, complications such as mucosal damage or infection are easily caused, and the monitoring cost is high, making it difficult to use continuously for extended periods in routine clinical practice.

[0018] In summary, existing weaning assessment methods all suffer from the following technical limitations: While RSBI is a non-invasive and convenient bedside assessment method, its accuracy is limited and cannot meet the needs of clinical weaning decisions; diaphragmatic ultrasound, although it can directly assess diaphragmatic function, is an intermittent examination that relies on the operator's experience and cannot provide continuous monitoring; transesophageal diaphragmatic electromonitoring (TEEM) can directly acquire diaphragmatic electrical signals, but it is an invasive procedure that limits its use in routine clinical practice. Therefore, current technology lacks a solution that can non-invasively acquire signals reflecting diaphragmatic function and generate a highly accurate weaning assessment index based on these signals.

[0019] Based on this, the present application provides a method for calculating an offline assessment index, comprising: receiving a target diaphragm electrical signal transmitted by a wireless surface diaphragm electrical acquisition module; wherein the target diaphragm electrical signal is a signal obtained by the wireless surface diaphragm electrical acquisition module after preprocessing the original diaphragm electrical signal, and the original diaphragm electrical signal is acquired by a surface electrode attached to a preset diaphragm region on the patient's body surface; performing signal purification processing on the target diaphragm electrical signal to obtain a purified diaphragm electrical signal; wherein the signal purification processing includes the following operations: ECG artifact filtering, power line interference filtering, and motion artifact filtering. One less; based on the purified diaphragm electrophysiological signal, multi-dimensional physiological features are extracted, including time-domain features, frequency-domain features, and stability features; among them, the time-domain features include the peak amplitude and area under the curve of diaphragm electrophysiological activity in each respiratory cycle, the frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrophysiological signal, and the stability features include the coefficient of variation of diaphragm electrophysiological activity intensity in multiple consecutive respiratory cycles; the multi-dimensional physiological features are input into a pre-constructed fusion evaluation model so that the fusion evaluation model performs weighted fusion calculation on the multi-dimensional physiological features and outputs an offline evaluation index.

[0020] Compared to the closest prior art, the following advantages are achieved: By receiving raw diaphragmatic electrical signals from a wireless surface-mounted diaphragmatic electrical acquisition module, signals reflecting the electrical activity of the diaphragm are acquired in a completely non-invasive manner. This eliminates the need for esophageal catheters inserted via the nose or mouth, avoiding issues such as mucosal damage, infection risks, and poor patient tolerance associated with invasive procedures. Furthermore, the wireless surface-mounted diaphragmatic electrical acquisition module is a lightweight, battery-powered package that can be attached to a ventilator or placed at the bedside, eliminating cable constraints and significantly improving patient comfort and clinical deployment flexibility. This allows the calculation of the weaning assessment index to be directly based on diaphragmatic functional status information, overcoming the fundamental deficiency of RSBI, which relies solely on indirect calculations using respiratory mechanics parameters and cannot directly reflect diaphragmatic functional status. Based on this, signal purification processing is performed on the target diaphragmatic electrical signal to obtain a high-quality purified diaphragmatic electrical signal. From the purified diaphragmatic electrical signal, peak amplitude and area under the curve are extracted as time-domain features, the center frequency calculated based on power spectral density is extracted as a frequency-domain feature, and the coefficient of variation of diaphragmatic electrical activity intensity over multiple consecutive respiratory cycles is extracted as a stability feature. This invention provides a comprehensive quantitative characterization of the diaphragm's contraction strength, fatigue state, and driving stability from three dimensions: time domain, frequency domain, and stability. This effectively overcomes the limitations of existing RSBI methods, which rely solely on simple proportional calculations of respiratory rate and tidal volume. Furthermore, compared to diaphragmatic ultrasound, which relies on manual selection of measurement sections and marking of measurement points, and whose results are heavily influenced by operator experience, this invention's feature extraction process can be automatically executed by a processor, eliminating the need for manual intervention and enabling objective and consistent quantitative assessment. The aforementioned multi-dimensional physiological features are then input into a fusion assessment model for weighted fusion calculation, outputting a comprehensive offline assessment index. This index integrates physiological information from the time domain, frequency domain, and stability dimensions. Compared to RSBI's output based solely on simple proportional calculations of respiratory rate and tidal volume, this invention's offline assessment index contains richer information on the diaphragm's functional state, more comprehensively reflecting the diaphragm's neural electrical activity and respiratory driving function. This improves the accuracy of the offline assessment index and effectively reduces the risk of misjudgment in offline decisions.

[0021] The method for calculating the offline evaluation index provided in this application embodiment can be applied to, for example, Figure 1 The application environment shown. Specifically, the method for calculating the offline evaluation index is applied to the offline evaluation index calculation system, which includes, as shown in the example below. Figure 1The diagram shows a surface electrode 101, a wireless surface diaphragmatic electrophysiological acquisition module 102, a signal processing device 103, and a mechanical ventilation device 104. The mechanical ventilation device 104 includes a ventilator. The surface electrode 101 and the wireless surface diaphragmatic electrophysiological acquisition module 102, the wireless surface diaphragmatic electrophysiological acquisition module 102 and the signal processing device 103, and the signal processing device 103 and the mechanical ventilation device 104 can all be connected via wired means (such as wires or network cables) or wireless means (such as Bluetooth or Wi-Fi networks). This application does not limit the connection in this regard. The surface electrode 101 can be a disposable surface electrode.

[0022] The method for calculating the offline evaluation index provided in this application embodiment can be applied to, for example, Figure 1 The application environment shown. Specifically, the method for calculating the offline evaluation index is applied to the offline evaluation index calculation system, which includes, as shown in the example below. Figure 1 The diagram shows a surface electrode 101, a wireless surface diaphragmatic electrophysiological acquisition module 102, a signal processing device 103, and a mechanical ventilation device 104. The mechanical ventilation device 104 includes a ventilator. The surface electrode 101 and the wireless surface diaphragmatic electrophysiological acquisition module 102, the wireless surface diaphragmatic electrophysiological acquisition module 102 and the signal processing device 103, and the signal processing device 103 and the mechanical ventilation device 104 can all be connected via wired means (such as wires or network cables) or wireless means (such as Bluetooth or Wi-Fi networks). This application does not limit the connection in this regard. The surface electrode 101 can be a disposable surface electrode.

[0023] like Figure 2 As shown, the wireless surface diaphragm electrophysiological data acquisition module 102 is a lightweight, battery-powered packaged device. Its housing may have at least three electrode interfaces for connecting to surface electrodes 101. The wireless surface diaphragm electrophysiological data acquisition module 102 may internally include a second power module 121, a preamplifier 122, a bandpass filter 123, an analog-to-digital converter 124, and a third communication chip 125. The second power module 121 is connected to the preamplifier 122, the bandpass filter 123, the analog-to-digital converter 124, and the third communication chip 125, respectively, to power the wireless surface diaphragm electrophysiological data acquisition module 102. The input terminal of the preamplifier 122 is connected to the surface electrodes 101 via the electrode interfaces; the output terminal of the preamplifier 122 is connected to the input terminal of the bandpass filter 123; the output terminal of the bandpass filter 123 is connected to the input terminal of the analog-to-digital converter 124; and the output terminal of the analog-to-digital converter 124 is connected to the input terminal of the third communication chip 125. Among them, the preamplifier 122 is a high input impedance, low noise preamplifier; the lower limit of the passband frequency of the bandpass filter 123 can be 20Hz and the upper limit of the passband frequency can be 450Hz; the analog-to-digital converter 124 can be a high-precision analog-to-digital converter; and the third communication chip 125 can be a low-power wireless communication chip.

[0024] like Figure 3 As shown, the signal processing device 103 may include a first wireless communication unit 131, a second wireless communication unit 132, a processor 133, a memory 134, and a first power module 135. The first power module 135 is connected to the first wireless communication unit 131, the second wireless communication unit 132, the processor 133, and the memory 134, respectively, and is used to supply power to the signal processing device 103; the processor 133 is connected to the first wireless communication unit 131, the second wireless communication unit 132, and the memory 133, respectively. The first wireless communication unit 131 may be a Bluetooth receiver module, the second wireless communication unit 132 may be a Wi-Fi module, and the processor 133 may be an embedded microprocessor or a digital signal processor. The above is merely an example and does not constitute a limitation of this application.

[0025] It should be noted that, for ease of description of the embodiments, the hardware described above will not be labeled when detailing the embodiments of the calculation method for the offline evaluation index provided in this application. Please refer to the following... Figures 4 to 9 The steps of the calculation method for the offline evaluation index provided in the embodiments of this application are described in detail.

[0026] Firstly, such as Figure 4 As shown in the illustration, this application provides a method for calculating an offline evaluation index, which is described using a signal processing device as an example. The method includes the following steps: S201: Receives the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module.

[0027] As an example, the signal processing device can receive the target diaphragm electrical signal transmitted by the wireless surface diaphragm electrical acquisition module 102 through its first wireless communication unit. The target diaphragm electrical signal is the signal obtained after preprocessing the original diaphragm electrical signal by the wireless surface diaphragm electrical acquisition module. The original diaphragm electrical signal is acquired by surface electrodes attached to a predetermined area of ​​the patient's diaphragm. Specifically, the placement of the surface electrodes can be optimized according to the anatomical projection area of ​​the diaphragm. As a specific example, the surface electrodes can be three disposable surface electrode pads, respectively attached to the patient's right chest wall, left subcostal margin, and right subcostal margin, to acquire the original surface diaphragm electrical analog signal (i.e., the original diaphragm electrical signal) using a bipolar lead method; this is only an example.

[0028] The preprocessing of the raw diaphragm electrical signal by the wireless surface diaphragm electrical acquisition module can include pre-amplification, bandpass filtering, and analog-to-digital conversion. Pre-amplification, performed by the preamplifier of the module, amplifies the weak raw analog diaphragm electrical signal with high gain. Bandpass filtering, performed by the module's bandpass filter (with a lower passband frequency limit of 20Hz and an upper passband frequency limit of 450Hz), preserves the main energy components of the diaphragm electrical signal and initially filters out noise outside the passband. Analog-to-digital conversion, performed by the module's analog-to-digital converter, converts the analog signal into a digital signal. The preprocessed target diaphragm electrical signal is a digitized signal containing physiological information reflecting the electrical activity of the diaphragm nerves, but also includes noise components such as ECG artifacts, power line interference, and motion artifacts. The third communication chip of the wireless surface diaphragm electrical acquisition module and the first wireless communication unit of the signal processing device can establish a data connection through the Bluetooth Low Energy wireless communication protocol to realize the real-time transmission of the target diaphragm electrical signal.

[0029] In a specific example, the wireless surface diaphragm electrophysiological acquisition module can acquire and preprocess the raw surface diaphragm electrophysiological analog signal at a sampling rate of 2000Hz, and then transmit the preprocessed target diaphragm electrophysiological signal to the signal processing device in real time via Bluetooth Low Energy wireless communication protocol through its third communication chip. The signal processing device continuously receives this data stream through its first wireless communication unit.

[0030] As can be seen, this step wirelessly receives the target diaphragm electrical signal through the first wireless communication unit of the signal processing device. The target diaphragm electrical signal is a digital signal obtained by pre-amplifying, band-pass filtering and analog-to-digital conversion of the wireless surface diaphragm electrical acquisition module. This enables signal transmission from the surface electrode to the signal processing device without the need for catheter insertion, effectively avoiding patient discomfort caused by invasive procedures and reducing the technical implementation difficulty and clinical deployment cost in the signal acquisition process.

[0031] S202: Perform signal purification processing on the target diaphragm electrical signal to obtain the purified diaphragm electrical signal.

[0032] As an example, the processor can perform signal purification processing on the target diaphragm electrical signal received by the first wireless communication unit. Signal purification processing refers to the operation of filtering or suppressing various noise components in the target diaphragm electrical signal using digital signal processing technology. Its purpose is to extract high-quality diaphragm electrical activity signals from the noisy target diaphragm electrical signal. Since the target diaphragm electrical signal contains various noise components such as ECG artifacts, power line interference, and motion artifacts, these noises overlap with the diaphragm electrical signal in the frequency band. If they are not filtered out, they will affect the accuracy of subsequent physiological feature extraction.

[0033] Specifically, signal purification processing can include at least one of the following: ECG artifact filtering, power line interference filtering, and motion artifact filtering. ECG artifact filtering removes ECG signal components from the target diaphragm electrical signal. ECG artifacts are one of the most significant sources of interference in surface diaphragm electrical signal acquisition; the periodic spikes generated by cardiac electrical activity on the surface electrodes overlap with the diaphragm electrical signal in the time domain, and their amplitude is much higher than that of the diaphragm electrical signal. Power line interference removal removes 50Hz or 60Hz AC electric field-induced noise from the target diaphragm electrical signal. Power line interference is a sinusoidal interference induced in the measurement circuit by the AC electric field radiated from the power line, and it is the most common and strongest narrowband interference source in surface electrical signal acquisition. Motion artifact filtering removes low-frequency baseline drift and amplitude jumps caused by patient movement or breathing from the target diaphragm electrical signal. Motion artifacts are signal distortions caused by changes in electrode-skin contact due to patient movement or breathing.

[0034] As a specific example, ECG artifact filtering can employ adaptive filtering technology. This adaptive filtering uses the ECG signal as a reference signal and subtracts the ECG component from the target diaphragm electrical signal. For instance, the processor of signal processing device 103 synchronously acquires one ECG signal via body surface electrodes 101 as a reference signal. The adaptive filter uses this ECG signal as a reference to estimate the ECG artifact component in the target diaphragm electrical signal and subtracts the estimated component from the target diaphragm electrical signal, thereby effectively suppressing ECG interference. Power frequency interference filtering can employ a notch filter with a notch center frequency of 50Hz or 60Hz and a notch bandwidth of ±0.5Hz to ±2Hz from the center frequency. The reason for using 50Hz / 60Hz is that this is the most common power grid frequency standard globally and also the strongest and most critical narrowband interference source to suppress in body surface electrical signal acquisition. For example, in a current medical environment where the power grid frequency is 50Hz, the notch filter center frequency is set to 50Hz, and the notch bandwidth is the center frequency ±1Hz (i.e., 49Hz-51Hz). This notch filter deeply attenuates signals within this frequency band, thereby effectively suppressing 50Hz power frequency interference. Motion artifact filtering can employ a high-pass filter with a cutoff frequency of 0.5Hz to 1Hz to filter out baseline drift. For example, the processor of signal processing device 103 sets the cutoff frequency of the high-pass filter to 0.5Hz. This high-pass filter allows signal frequency components higher than 0.5Hz to pass through while attenuating low-frequency components lower than 0.5Hz, thereby effectively removing low-frequency baseline drift caused by patient movement or breathing.

[0035] In one possible implementation, a signal purification algorithm can be run by a processor to perform signal purification processing on the target diaphragm electrical signal, resulting in a purified diaphragm electrical signal (i.e., a high-quality diaphragm electrical signal).

[0036] As can be seen, this step performs signal purification processing on the target diaphragm electrical signal through the processor, filtering out noise components such as ECG artifacts, power frequency interference, and motion artifacts, and obtaining a purified diaphragm electrical signal, which provides a reliable data foundation for subsequent feature extraction.

[0037] S203: Based on the purified diaphragm electrical signal, extract multi-dimensional physiological features, including time-domain features, frequency-domain features, and stability features. Among them, the time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity in each respiratory cycle, the frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal, and the stability features include the coefficient of variation of diaphragm electrical activity intensity in multiple consecutive respiratory cycles.

[0038] As an example, a processor can extract multi-dimensional physiological features based on the purified diaphragm electroencephalogram obtained in step S202. Multi-dimensional physiological features refer to a set of physiological parameters obtained by quantifying the purified diaphragm electroencephalogram from multiple analytical dimensions such as time domain, frequency domain, and stability. These can include time domain features, frequency domain features, and stability features, etc., which are not limited here.

[0039] Specifically, time-domain characteristics refer to parameters obtained by quantitatively describing the electrical activity of the diaphragm in each respiratory cycle in the time domain. These parameters may include the peak amplitude and area under the curve of the diaphragm electrical activity in each respiratory cycle. The peak amplitude refers to the maximum amplitude of the diaphragm electrical signal reached in a respiratory cycle, reflecting the maximum contraction intensity of the diaphragm during that breath. The area under the curve refers to the integral of the diaphragm electrical signal amplitude over time in a respiratory cycle, reflecting the overall electrical activity intensity of the diaphragm during that breath.

[0040] Frequency domain characteristics refer to parameters obtained by quantifying the diaphragm electromyography (EMG) signal in the frequency domain. These parameters can include the center frequency calculated based on the power spectral density of the EMG signal. The power spectral density is the power distribution density function of the EMG signal in the frequency domain, reflecting the distribution characteristics of signal energy as a function of frequency. The center frequency is the first statistical moment of the power spectral density, reflecting the position of the centroid of the diaphragm EMG signal spectrum. When the diaphragm is fatigued, the spectrum of its EMG signal shifts towards lower frequencies, resulting in a decrease in the center frequency. Therefore, the center frequency can be used to quantitatively assess the fatigue state of the diaphragm.

[0041] Stability characteristics refer to parameters that quantify the fluctuation of diaphragmatic electrical activity intensity over multiple consecutive respiratory cycles in the time domain. These parameters may include the coefficient of variation (COP) of diaphragmatic electrical activity intensity over these cycles. The COP is the ratio of the standard deviation to the mean of diaphragmatic electrical activity intensity over multiple respiratory cycles, measuring the relative fluctuation of diaphragmatic electrical activity intensity across these cycles. A smaller COP indicates smaller differences in diaphragmatic electrical activity intensity across respiratory cycles and more stable respiratory drive; a larger COP indicates larger differences in diaphragmatic electrical activity intensity across respiratory cycles and less stable respiratory drive.

[0042] As can be seen, this step extracts multi-dimensional physiological characteristics such as peak amplitude, area under the curve, center frequency, and coefficient of variation from three dimensions: time domain, frequency domain, and stability, to comprehensively quantify the contraction intensity, fatigue state, and driving stability of the diaphragm.

[0043] S204: Input multi-dimensional physiological characteristics into a pre-built fusion assessment model so that the fusion assessment model performs weighted fusion calculation on the multi-dimensional physiological characteristics and outputs an offline assessment index.

[0044] As an example, the processor can input the multi-dimensional physiological features extracted in step S203 into a pre-constructed fusion assessment model. The fusion assessment model is a pre-established mathematical model used to perform weighted fusion calculations on multi-dimensional physiological features to generate an offline assessment index. Based on the physiological and clinical correlation between each physiological feature parameter and the offline outcome, the fusion assessment model assigns corresponding weight coefficients to each feature parameter, and can integrate multiple dimensions of physiological features into a comprehensive quantitative index through weighted fusion calculations.

[0045] Specifically, the fusion evaluation model performs normalization and weighted fusion calculations on the input time-domain features (such as peak amplitude and area under the curve), frequency-domain features (such as center frequency), and stability features (such as coefficient of variation), respectively. Normalization maps each feature parameter to the same numerical range, eliminating the influence of differences in dimensions and magnitudes on the fusion result. Weighted fusion calculation multiplies each normalized feature parameter by its corresponding weight coefficient, sums the results, maps them to a preset numerical range, and outputs a normalized offline evaluation index (i.e., SDE-VW Index).

[0046] The weaning assessment index (HII) is a comprehensive indicator used to quantitatively assess a patient's ability to wean off mechanical ventilation support, with a value ranging from 0 to 100. A higher HII indicates better diaphragmatic reserve, more stable respiratory drive, and a higher success rate in weaning off mechanical ventilation support. Conversely, a lower HII or a progressively declining trend indicates potential respiratory muscle fatigue, reduced respiratory efficiency, and an increased risk of weaning failure.

[0047] As can be seen, this step inputs multi-dimensional physiological characteristics into the fusion assessment model for weighted fusion calculation, and outputs a comprehensive offline assessment index. This index integrates physiological information from three dimensions: time domain, frequency domain, and stability, and can more comprehensively reflect the functional state of the diaphragm, thereby improving the assessment accuracy of the offline assessment index.

[0048] In summary, the technical solution provided in this application embodiment receives the target diaphragm electrical signal sent by a wireless surface diaphragm electrical acquisition module to obtain a signal reflecting the state of diaphragm nerve electrical activity, enabling the calculation of the offline assessment index to be directly based on diaphragm functional state information. Based on this, signal purification processing is performed on the target diaphragm electrical signal to obtain a high-quality purified diaphragm electrical signal. From the purified diaphragm electrical signal, peak amplitude and area under the curve are extracted as time-domain features, the center frequency calculated based on power spectral density is extracted as frequency-domain features, and the intensity of diaphragm electrical activity over multiple consecutive respiratory cycles is extracted. The coefficient of variation, as a stability feature, comprehensively quantifies the contraction intensity, fatigue state, and driving stability of the diaphragm from three dimensions: time domain, frequency domain, and stability. This effectively overcomes the information limitations of existing RSBI, which relies solely on simple proportional calculations based on two mechanical dimensions: respiratory rate and tidal volume. Furthermore, the aforementioned multi-dimensional physiological characteristics are input into a fusion assessment model for weighted fusion calculation, outputting a comprehensive offline assessment index. This offline assessment index integrates more dimensions of physiological information, enabling it to more comprehensively reflect the neuroelectrical activity state of the diaphragm and respiratory driving function, thereby improving the assessment accuracy of the offline assessment index.

[0049] In one embodiment, such as Figure 5 As shown, the extraction of multi-dimensional physiological features based on the purified diaphragm electrical signals includes the following steps: S231A: Identifies the start and end points of each respiratory cycle in the purified diaphragm electrical signal; S232A: Based on the start and end points of each respiratory cycle, extract the signal segment corresponding to each respiratory cycle from the purified diaphragm electrical signal; S233A: Within each signal segment, search along the time axis for the point of maximum amplitude and take the amplitude of the point of maximum amplitude as the peak amplitude; S234A: Within each signal segment, the amplitude of the signal segment is integrated along the time axis to obtain the area under the curve.

[0050] As an example, a processor can identify the start and end points of each respiratory cycle in the purified diaphragmatic electrophysiological signal. A respiratory cycle is the time interval from the start of one breath to the start of the next. In the purified diaphragmatic electrophysiological signal, each respiratory cycle corresponds to a segment of diaphragmatic electrical activity signal. The processor determines the start and end points of each respiratory cycle by analyzing the amplitude variation characteristics of the purified diaphragmatic electrophysiological signal, where the start point is the moment when diaphragmatic electrical activity begins to rise, and the end point is the moment when diaphragmatic electrical activity returns to the baseline level.

[0051] As a specific example, the processor can perform envelope detection on the purified diaphragm electrical signal to obtain a signal envelope. The starting point is the moment when the rising edge of the signal envelope crosses a preset amplitude threshold, and the ending point is the moment when the falling edge of the signal envelope returns to the baseline level. For example, the preset amplitude threshold can be set to twice the standard deviation of the baseline amplitude. When the rising edge of the signal envelope first exceeds this threshold, the processor can mark it as the start point of the respiratory cycle; when the falling edge of the signal envelope returns to the baseline level, the processor can mark it as the end point of the respiratory cycle.

[0052] The processor can: extract signal segments corresponding to each respiratory cycle from the purified diaphragmatic electroencephalogram (EEG) based on the start and end points of each respiratory cycle; segment the continuous purified diaphragmatic EEG from the processor using the start and end points as dividing boundaries to separate independent signal segments corresponding to each respiratory cycle, with each signal segment corresponding to the diaphragmatic electrical activity of a complete respiratory cycle; search for the point of maximum amplitude along the time axis within each signal segment and use the amplitude of the point of maximum amplitude as the peak amplitude; specifically, compare the amplitude of each sampling point on the time axis within a signal segment, find the sampling point with the largest amplitude, and use the amplitude of that sampling point as the peak amplitude of that respiratory cycle; integrate the amplitude of the signal segment along the time axis within each signal segment to obtain the area under the curve; specifically, integrate the amplitude of each sampling point in a signal segment along the time axis within a signal segment, and use the integration result as the area under the curve of that respiratory cycle. Therefore, through the above methods, the processor can calculate a set of time-domain features for each respiratory cycle, including the peak amplitude and area under the curve of that respiratory cycle.

[0053] As can be seen, this embodiment extracts time-domain features by identifying the start and end points of the respiratory cycle, extracting signal segments, searching for the point of maximum amplitude, and performing integral calculations. This allows for the quantification of the diaphragm's contraction intensity and overall electrical activity level within each respiratory cycle, providing time-domain physiological information for subsequent fusion evaluation.

[0054] In one embodiment, such as Figure 6 As shown, the extraction of multi-dimensional physiological features based on the purified diaphragm electrical signals includes the following steps: S231B: The purified diaphragm electrical signal is processed by framing to obtain multiple time frames that are continuous in time and overlap with each other. S232B: Multiply each time frame by a window function to obtain the corresponding windowed frame data; S233B: Perform a Fourier transform on each windowed frame of data to obtain the corresponding complex spectrum; S234B: Square the amplitude of each complex spectrum to obtain the corresponding power spectrum; S235B: Smooths the power spectrum of multiple time frames in the time dimension to obtain the power spectral density; S236B: Calculate the first moment based on the power spectral density to obtain the center frequency.

[0055] As an example, the purified diaphragm electroencephalogram (EEG) signal can be processed by a processor to obtain multiple temporally continuous and overlapping time frames. Frame processing refers to dividing a continuous signal into multiple short-time signal segments, each of which constitutes a time frame. Since the diaphragm EEG signal is a non-stationary signal with statistical characteristics that change over time, frame processing allows the signal to be approximated as a stationary signal within a relatively short time window of each frame, thus enabling frequency domain analysis. Partial overlap between adjacent time frames ensures a smooth transition of signal features and avoids information loss at frame boundaries. The frame length for frame processing can range from 256 milliseconds to 512 milliseconds, and the overlap rate between adjacent frames can be 50%, which is not limited here.

[0056] As a specific example, the processor can segment the purified diaphragm electrical signal into frames with a frame length of 256 milliseconds, with adjacent frames overlapping by 128 milliseconds. Each 256-millisecond signal segment constitutes one frame, and the next frame begins 128 milliseconds backward from the start position of the previous frame, thus obtaining multiple temporally continuous and overlapping time frames. Specifically, the processor can perform the following processing: Multiplying each time frame by a window function yields the corresponding windowed frame data. A window function is a real function that takes a value greater than zero within a finite time interval and a value of zero outside that interval. Multiplying a time frame by a window function can suppress spectral leakage caused by signal truncation in the subsequent Fourier transform. The window function can be a Hanning window.

[0057] Perform a Fourier transform on each windowed frame of data to obtain the corresponding complex spectrum. The Fourier transform is a Fast Fourier Transform (FFT). As a specific example, the processor can perform a 512-point FFT on each windowed frame of data to obtain the corresponding complex spectrum.

[0058] The power spectrum is obtained by squaring the amplitude of each complex spectrum. The square of the amplitude of the complex spectrum reflects the energy intensity of the signal at the corresponding frequency component.

[0059] The power spectrum of multiple time frames is smoothed over time to obtain the power spectral density. Power spectral density refers to the power distribution density function of a signal in the frequency domain. By smoothing the power spectrum of multiple time frames over time, random fluctuations between frames can be suppressed, and a stable power spectral density estimate can be obtained.

[0060] The center frequency is obtained by calculating the first moment from the power spectral density. Specifically, the first statistical moment of the power spectral density in the frequency domain can be calculated using a processor, and the result can be used as the center frequency.

[0061] As can be seen, this embodiment extracts frequency domain features by performing steps such as frame segmentation, windowing, Fourier transform, amplitude squaring, time-dimensional smoothing, and first-order moment calculation. This enables the quantitative characterization of the spectral centroid position of the diaphragm electrical signal, providing frequency domain physiological information for assessing diaphragm fatigue. By smoothing the power spectrum of multiple time frames in the time dimension, inter-frame random fluctuations can be suppressed, resulting in a stable power spectral density estimate and thus a more reliable center frequency.

[0062] In one embodiment, such as Figure 7 As shown, the extraction of multi-dimensional physiological features based on the purified diaphragm electrical signals includes the following steps: S231C: Select N consecutive respiratory cycles, where N is an integer greater than or equal to 30; S232C: Obtain the area under the curve for each respiratory cycle, resulting in a dataset consisting of N area under the curve values; S233C: Calculate the mean and standard deviation of the dataset; S234C: Calculate the ratio of the standard deviation to the mean to obtain the coefficient of variation.

[0063] As an example, the processor can select N consecutive respiratory cycles, where N is an integer greater than or equal to 30 and less than or equal to 60. N consecutive respiratory cycles refer to N temporally adjacent complete respiratory cycles. The number of selected respiratory cycles N should be sufficiently large (e.g., not less than 30) to ensure the reliability of the statistical analysis; at the same time, the value of N should not be too large (e.g., not greater than 60) to avoid an excessively long calculation window that fails to reflect the dynamic changes in diaphragmatic function in a timely manner. As a specific example, with N set to 30, the processor can select the most recent 30 consecutive respiratory cycles for stability feature calculation.

[0064] The processor can obtain the area under the curve (AUC) for each respiratory cycle, resulting in a dataset consisting of N AUC values. These N AUC values ​​correspond to the AUC of each of the selected N consecutive respiratory cycles. The AUC is a parameter calculated in the time-domain feature extraction step; in this embodiment, "diaphragmatic electrical activity intensity" specifically refers to the AUC for each respiratory cycle.

[0065] As a specific example, the area under the curve (AUC) values ​​for the 30 respiratory cycles are: 2.3, 2.5, 2.4, 2.6, 2.2, 2.7, 2.4, 2.5, 2.3, 2.6, 2.4, 2.5, 2.7, 2.3, 2.4, 2.6, 2.5, 2.4, 2.3, 2.7, 2.5, 2.6, 2.4, 2.5, 2.3, 2.6, 2.4, 2.5, 2.7, 2.3 (units). ), which constitute the dataset.

[0066] The processor calculates the mean of the dataset. and standard deviation Among them, the mean The standard deviation is the arithmetic mean of the areas under the curves in the dataset, reflecting the average level of diaphragmatic electrical activity; It refers to the arithmetic square root of the average of the squares of the differences between the area under each curve and the mean in the dataset, reflecting the dispersion of the intensity of diaphragmatic electrical activity.

[0067] The processor calculates the ratio of the standard deviation to the mean, yielding the coefficient of variation (CV). The coefficient of variation (CV) is the ratio of the standard deviation to the mean, and its calculation formula is: Standard deviation is used to measure the relative volatility of a set of data. Specifically, it can be processed by a processor. Divide by the mean The calculated result is used as the coefficient of variation (CV). The coefficient of variation is a dimensionless index of relative variability, used to measure the relative variability in the intensity of diaphragmatic electrical activity across multiple consecutive respiratory cycles.

[0068] The coefficient of variation (CV) directly quantifies the successive stability of respiratory drive. A small CV value indicates small differences in diaphragmatic electrical activity intensity across respiratory cycles, suggesting stable respiratory drive; a large CV value indicates large differences in diaphragmatic electrical activity intensity across respiratory cycles, suggesting unstable respiratory drive. Unstable drive may originate from early diaphragmatic fatigue, abnormal central regulation, or respiratory decompensation, and is an independent early warning indicator of weaning failure. In the subsequent fusion assessment model, a high CV value will reduce the weaning assessment index.

[0069] As can be seen, this embodiment extracts stability features by selecting N consecutive respiratory cycles, obtaining the area under the curve for each respiratory cycle, calculating the mean and standard deviation of the dataset, and calculating the ratio of the standard deviation to the mean to obtain the coefficient of variation, which can quantitatively characterize the stability of respiratory center drive.

[0070] In one embodiment, such as Figure 8 As shown, this involves selecting from N consecutive respiratory cycles and including the following steps: S231C1: Update N consecutive respiratory cycles using a sliding window method; S231C2: Whenever a new respiratory cycle is identified, obtain the area under the curve corresponding to the new respiratory cycle; S231C3: Add the area under the curve (AUC) of the new respiratory cycle to the dataset and remove the earliest AUC in the dataset to update the dataset.

[0071] As an example, the processor can update N consecutive respiratory cycles using a sliding window approach. Here, "selecting N consecutive respiratory cycles" is not a one-time operation, but rather a dynamic dataset containing the most recent N complete respiratory cycles is continuously maintained through the sliding window mechanism. Whenever a new respiratory cycle is identified, the processor of the signal processing device 103 obtains the area under the curve (AUC) corresponding to that new respiratory cycle; then, it adds the new AUC to the dataset, while removing the oldest AUC in the dataset to update the dataset. Through the sliding window approach, the dataset always remains a set of AUC values ​​for the most recent N respiratory cycles, achieving continuous real-time updates of stability features.

[0072] As a specific example, the current dataset can contain the area under the curve (AUC) values ​​for respiratory cycles #1 to #30. When a new respiratory cycle #31 is identified, the processor can add the AUC value corresponding to respiratory cycle #31 to the dataset and remove the AUC value corresponding to respiratory cycle #1. The updated dataset contains the AUC values ​​for respiratory cycles #2 to #31, and the dataset always maintains 30 of the most recent AUC values.

[0073] As can be seen, this embodiment updates the dataset by using a sliding window method, so that the calculation of the coefficient of variation is always based on the data of the most recent N respiratory cycles, which can reflect the latest trend of diaphragmatic function in a timely manner and realize continuous real-time monitoring of stability characteristics.

[0074] In one embodiment, such as Figure 1 As shown, after outputting the offline evaluation index, the following steps are also included: S205: Send the offline assessment index to the mechanical ventilation equipment, so as to control the first display area of ​​the display interface of the mechanical ventilation equipment to display the value of the offline assessment index in the form of a digital panel, and control the second display area of ​​the display interface of the mechanical ventilation equipment to display the evolution trajectory of the offline assessment index over time, including the currently output offline assessment index, in the form of a trend curve.

[0075] When the offline assessment index is lower than the preset safety threshold, an alarm trigger signal is sent to the mechanical ventilation equipment. The alarm trigger signal is used to drive the mechanical ventilation equipment to perform an alarm action.

[0076] As an example, the signal processing device can transmit the offline evaluation index to the mechanical ventilation device via its second wireless communication unit. The second wireless communication unit and the mechanical ventilation device can be connected via a Wi-Fi wireless communication protocol. The signal processing device and the mechanical ventilation device follow a predefined network communication protocol to ensure reliable, real-time transmission of data packets containing timestamps, offline evaluation index values, alarm flags, etc.

[0077] Specifically, such as Figure 9 As shown, after receiving the weaning assessment index, the mechanical ventilation device displays the index on its screen. The display interface of the mechanical ventilation device has a first display area and a second display area. The first display area displays the value of the currently output weaning assessment index in the form of a digital panel. The second display area displays the evolution trajectory of the weaning assessment index over time, including the currently output index, at multiple historical moments in the form of a trend curve. The horizontal axis of the trend curve represents time, and the vertical axis represents the weaning assessment index value. The weaning assessment index and the mechanical ventilation parameter waveforms of the mechanical ventilation device (such as pressure-time waveforms and flow-time waveforms) are displayed side by side on the same display interface, allowing clinicians to simultaneously assess the patient's weaning potential on the same integrated interface without shifting their gaze while observing and adjusting mechanical ventilation parameters.

[0078] As a specific example, the display interface of the mechanical ventilation equipment shows pressure-time waveforms and flow-time waveforms on the left side, and a dedicated "weaning assessment" area on the right side. The upper part of this dedicated area displays the current weaning assessment index value as "72" in the form of a digital panel, and the lower part plots the trend curve of this index over the past few hours. The horizontal axis of the trend curve is time, and the vertical axis is the weaning assessment index value. Clinicians can intuitively observe the dynamic change trend of the weaning assessment index through this trend curve.

[0079] When the weaning assessment index falls below a preset safety threshold, the signal processing device can send an alarm trigger signal to the mechanical ventilation device via its second wireless communication unit. The alarm trigger signal drives the mechanical ventilation device to perform an alarm action. The mechanical ventilation device also includes an audible and visual alarm unit; the alarm trigger signal triggers this unit to output an audible and visual alarm signal. The safety threshold can be preset by the clinical operator through the human-machine interface of the mechanical ventilation device.

[0080] As a specific example, clinical operators set a safety threshold of 45 through the human-machine interface of the mechanical ventilation device. When the weaning assessment index output by the processor falls below 45, the signal processing device sends an alarm trigger signal to the mechanical ventilation device through its second wireless communication unit. Upon receiving the alarm trigger signal, the mechanical ventilation device activates its audible and visual alarm unit to emit a flashing red light and a buzzer sound to alert clinicians to the risk of patient weaning.

[0081] As another concrete example, the processor can determine whether the duration for which the offline assessment index remains below a safety threshold has reached a preset duration. If the duration reaches the preset duration, an alarm trigger signal is sent via its second wireless communication unit to avoid false alarms caused by momentary signal fluctuations. For example, if the preset duration is set to 30 seconds, the processor can detect that the offline assessment index is below 45 for 30 seconds before sending an alarm trigger signal via its second wireless communication unit. If the offline assessment index only briefly falls below 45 before recovering above the safety threshold, no alarm is triggered, effectively preventing false alarms caused by momentary signal fluctuations.

[0082] As can be seen, this embodiment achieves deep integration of assessment results with clinical workflow by sending the weaning assessment index to the mechanical ventilation equipment and displaying it on the interface; by displaying the trajectory of the weaning assessment index over time through trend curves, the dynamic evolution trend of the patient's diaphragmatic function can be observed intuitively; and by implementing a threshold alarm mechanism, proactive warning of the risk of weaning failure is achieved, further improving the safety and efficiency of clinical decision-making.

[0083] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0084] Secondly, embodiments of this application provide a device for calculating an offline evaluation index, which corresponds one-to-one with the calculation method for the offline evaluation index in the above embodiments. For example... Figure 10 As shown, the offline assessment index calculation device includes a receiving module 301, a purification module 302, an extraction module 303, and a model processing module 304. Detailed descriptions of each functional module are as follows: The receiving module 301 is used to receive the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module; wherein, the target diaphragm electrical signal is the signal obtained by the wireless surface diaphragm electrical acquisition module after preprocessing the original diaphragm electrical signal, and the original diaphragm electrical signal is acquired by surface electrodes attached to a preset diaphragm area on the patient's body surface. The purification processing module 302 is used to perform signal purification processing on the target diaphragm electrical signal to obtain a purified diaphragm electrical signal; wherein, the signal purification processing includes at least one of ECG artifact filtering, power frequency interference filtering, and motion artifact filtering. Extraction module 303 is used to extract multi-dimensional physiological features based on the purified diaphragm electrical signal. The multi-dimensional physiological features include time-domain features, frequency-domain features, and stability features. The time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity in each respiratory cycle. The frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal. The stability features include the coefficient of variation of diaphragm electrical activity intensity in multiple consecutive respiratory cycles. The model processing module 304 is used to input the multi-dimensional physiological features into a pre-constructed fusion evaluation model, so that the fusion evaluation model performs weighted fusion calculation on the multi-dimensional physiological features and outputs an offline evaluation index.

[0085] In one embodiment, the extraction module 303 is further configured to: Identify the start and end points of each respiratory cycle in the purified diaphragm electrical signal; Based on the start and end points of each respiratory cycle, a signal segment corresponding to each respiratory cycle is extracted from the purified diaphragm electrical signal. Within each signal segment, the point of maximum amplitude is searched along the time axis, and the amplitude of the point of maximum amplitude is taken as the peak amplitude. Within each signal segment, the amplitude of the signal segment is integrated along the time axis to obtain the area under the curve.

[0086] In one embodiment, the extraction module 303 is further configured to: The purified diaphragm electrical signal is processed by frame segmentation to obtain multiple time frames that are continuous in time and overlap with each other. Multiply each of the aforementioned time frames by a window function to obtain the corresponding windowed frame data; Perform a Fourier transform on each windowed frame of data to obtain the corresponding complex spectrum; Square the amplitude of each complex spectrum to obtain the corresponding power spectrum; The power spectra of multiple time frames are smoothed in the time dimension to obtain the power spectral density; The first moment is calculated based on the power spectral density to obtain the center frequency.

[0087] In one embodiment, the extraction module 303 is further configured to: Select N consecutive respiratory cycles, where N is an integer greater than or equal to 30; Obtain the area under the curve for each respiratory cycle to obtain a dataset consisting of N area under the curve values; Calculate the mean and standard deviation of the dataset; The coefficient of variation is obtained by calculating the ratio of the standard deviation to the mean.

[0088] In one embodiment, the extraction module 303 is further configured to: The N consecutive respiratory cycles are updated using a sliding window method; Whenever a new respiratory cycle is identified, the area under the curve corresponding to the new respiratory cycle is obtained; The area under the curve of the new respiratory cycle is added to the dataset, and the earliest added area under the curve in the dataset is removed to update the dataset.

[0089] In one embodiment, the offline evaluation index calculation device further includes a transmission module (not shown in the figure); The sending module is used for: The offline assessment index is sent to the mechanical ventilation equipment to control the first display area of ​​the mechanical ventilation equipment's display interface to display the value of the offline assessment index in the form of a digital panel, and to control the second display area of ​​the mechanical ventilation equipment's display interface to display the evolution trajectory of the offline assessment index over time, including the currently output offline assessment index, in the form of a trend curve.

[0090] Specific limitations regarding the calculation device for the offline evaluation index can be found in the limitations on the calculation method for the offline evaluation index described above, and will not be repeated here. Each module in the aforementioned calculation device for the offline evaluation index can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of the processor in the signal processing device, or stored in software in the memory of the signal processing device, so that the processor can call and execute the operations corresponding to each module.

[0091] Thirdly, embodiments of this application provide a system for calculating an offline evaluation index, which corresponds one-to-one with the calculation methods for the offline evaluation index in the above embodiments. For example... Figure 1 As shown, the calculation system for the offline assessment index includes: a signal processing device 101, a wireless surface diaphragm electroacupuncture module 102, and surface electrodes 103. The wireless surface diaphragm electrical acquisition module 102 is used to acquire the original diaphragm electrical signal of a preset diaphragm region on the patient's body surface through the surface electrode 103; and to preprocess the original diaphragm electrical signal to obtain the target diaphragm electrical signal. The signal processing device 101 is used to receive the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module 102; perform signal purification processing on the target diaphragm electrical signal to obtain a purified diaphragm electrical signal; wherein, the signal purification processing includes at least one of ECG artifact filtering, power frequency interference filtering, and motion artifact filtering; based on the purified diaphragm electrical signal, extract multi-dimensional physiological features, including time-domain features, frequency-domain features, and stability features; wherein, the time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity in each respiratory cycle, the frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal, and the stability features include the coefficient of variation of diaphragm electrical activity intensity in multiple consecutive respiratory cycles; input the multi-dimensional physiological features into a pre-constructed fusion evaluation model so that the fusion evaluation model performs weighted fusion calculation on the multi-dimensional physiological features and outputs an offline evaluation index.

[0092] In one embodiment, the signal processing device 101 is further configured to: Identify the start and end points of each respiratory cycle in the purified diaphragm electrical signal; Based on the start and end points of each respiratory cycle, a signal segment corresponding to each respiratory cycle is extracted from the purified diaphragm electrical signal. Within each signal segment, the point of maximum amplitude is searched along the time axis, and the amplitude of the point of maximum amplitude is taken as the peak amplitude. Within each signal segment, the amplitude of the signal segment is integrated along the time axis to obtain the area under the curve.

[0093] In one embodiment, the signal processing device 101 is further configured to: The purified diaphragm electrical signal is processed by frame segmentation to obtain multiple time frames that are continuous in time and overlap with each other. Multiply each of the aforementioned time frames by a window function to obtain the corresponding windowed frame data; Perform a Fourier transform on each windowed frame of data to obtain the corresponding complex spectrum; Square the amplitude of each complex spectrum to obtain the corresponding power spectrum; The power spectra of multiple time frames are smoothed in the time dimension to obtain the power spectral density; The first moment is calculated based on the power spectral density to obtain the center frequency.

[0094] In one embodiment, the signal processing device 101 is further configured to: Select N consecutive respiratory cycles, where N is an integer greater than or equal to 30; Obtain the area under the curve for each respiratory cycle to obtain a dataset consisting of N area under the curve values; Calculate the mean and standard deviation of the dataset; The coefficient of variation is obtained by calculating the ratio of the standard deviation to the mean.

[0095] In one embodiment, the signal processing device 101 is further configured to: The N consecutive respiratory cycles are updated using a sliding window method; Whenever a new respiratory cycle is identified, the area under the curve corresponding to the new respiratory cycle is obtained; The area under the curve of the new respiratory cycle is added to the dataset, and the earliest added area under the curve in the dataset is removed to update the dataset.

[0096] In one embodiment, such as Figure 1 As shown, the calculation system for the offline assessment index also includes a mechanical ventilation device 104; The signal processing device 101 is further used for: The offline assessment index is sent to the mechanical ventilation device 104 to control the first display area of ​​the display interface of the mechanical ventilation device 104 to display the value of the offline assessment index in the form of a digital panel, and to control the second display area of ​​the display interface of the mechanical ventilation device 104 to display the evolution trajectory of the offline assessment index over time, including the currently output offline assessment index, in the form of a trend curve.

[0097] For specific limitations on the calculation system of the offline assessment index, please refer to the limitations on the calculation method of the offline assessment index mentioned above, which will not be repeated here.

[0098] Fourthly, embodiments of this application provide a signal processing device, the internal structure of which can be shown in the following diagram. Figure 11 As shown. The signal processing device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory of the signal processing device includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores a computer program. The internal memory provides an environment for the execution of the computer program in the non-volatile storage media. The network interface of the signal processing device is used to communicate with external devices or equipment (such as mechanical ventilation equipment and wireless surface diaphragmatic electrophysiological acquisition modules). When the computer program is executed by the processor, it implements the steps of the offline evaluation index calculation method described above.

[0099] In one embodiment, this application provides a signal processing apparatus, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the offline evaluation index calculation method described above.

[0100] Fifthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the offline evaluation index calculation method described above.

[0101] It should be noted that the functions or steps that the signal processing device or computer-readable storage medium can achieve are described in the relevant descriptions of the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0102] It should be noted that any AI models, software tools, or components not belonging to this company appearing in the embodiments of this application are merely illustrative examples and do not represent actual use. All user personal information involved in the embodiments of this application has been authorized (with the knowledge and consent) by the relevant parties or has been fully authorized by all parties, and the executing entity may obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with relevant laws and regulations and do not violate public order and good morals.

[0103] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for calculating an offline evaluation index, characterized in that, include: Receive the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module; wherein, the target diaphragm electrical signal is the signal obtained by the wireless surface diaphragm electrical acquisition module after preprocessing the original diaphragm electrical signal, and the original diaphragm electrical signal is acquired by surface electrodes attached to the preset diaphragm area of ​​the patient. The target diaphragm electrical signal is subjected to signal purification processing to obtain a purified diaphragm electrical signal; wherein, the signal purification processing includes at least one of ECG artifact filtering, power line interference filtering, and motion artifact filtering. Based on the purified diaphragm electrical signal, multi-dimensional physiological features are extracted, including time-domain features, frequency-domain features, and stability features. The time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity in each respiratory cycle. The frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal. The stability features include the coefficient of variation of diaphragm electrical activity intensity in multiple consecutive respiratory cycles. The multidimensional physiological features are input into a pre-constructed fusion evaluation model, so that the fusion evaluation model performs weighted fusion calculation on the multidimensional physiological features and outputs an offline evaluation index.

2. The method according to claim 1, characterized in that, Based on the purified diaphragm electrical signal, multi-dimensional physiological features are extracted, including: Identify the start and end points of each respiratory cycle in the purified diaphragm electrical signal; Based on the start and end points of each respiratory cycle, a signal segment corresponding to each respiratory cycle is extracted from the purified diaphragm electrical signal. Within each signal segment, the point of maximum amplitude is searched along the time axis, and the amplitude of the point of maximum amplitude is taken as the peak amplitude. Within each signal segment, the amplitude of the signal segment is integrated along the time axis to obtain the area under the curve.

3. The method according to claim 1, characterized in that, Based on the purified diaphragm electrical signal, multi-dimensional physiological features are extracted, including: The purified diaphragm electrical signal is processed by frame segmentation to obtain multiple time frames that are continuous in time and overlap with each other. Multiply each of the aforementioned time frames by a window function to obtain the corresponding windowed frame data; Perform a Fourier transform on each windowed frame of data to obtain the corresponding complex spectrum; Square the amplitude of each complex spectrum to obtain the corresponding power spectrum; The power spectra of multiple time frames are smoothed in the time dimension to obtain the power spectral density; The first moment is calculated based on the power spectral density to obtain the center frequency.

4. The method according to any one of claims 1 to 3, characterized in that, Based on the purified diaphragm electrical signal, multi-dimensional physiological features are extracted, including: Select N consecutive respiratory cycles, where N is an integer greater than or equal to 30; Obtain the area under the curve for each respiratory cycle to obtain a dataset consisting of N area under the curve values; Calculate the mean and standard deviation of the dataset; The coefficient of variation is obtained by calculating the ratio of the standard deviation to the mean.

5. The method according to claim 4, characterized in that, The selection of N consecutive respiratory cycles includes: The N consecutive respiratory cycles are updated using a sliding window method; Whenever a new respiratory cycle is identified, the area under the curve corresponding to the new respiratory cycle is obtained; The area under the curve of the new respiratory cycle is added to the dataset, and the earliest added area under the curve in the dataset is removed to update the dataset.

6. The method according to claim 1, characterized in that, After outputting the offline assessment index, the following is also included: The offline assessment index is sent to the mechanical ventilation equipment to control the first display area of ​​the mechanical ventilation equipment's display interface to display the value of the offline assessment index in the form of a digital panel, and to control the second display area of ​​the mechanical ventilation equipment's display interface to display the evolution trajectory of the offline assessment index over time, including the currently output offline assessment index, in the form of a trend curve.

7. A device for calculating an offline evaluation index, characterized in that, include: The receiving module is used to receive the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module; wherein, the target diaphragm electrical signal is the signal obtained by the wireless surface diaphragm electrical acquisition module after preprocessing the original diaphragm electrical signal, and the original diaphragm electrical signal is acquired by surface electrodes attached to a preset diaphragm area on the patient's body surface. The purification module is used to perform signal purification processing on the target diaphragm electrical signal to obtain a purified diaphragm electrical signal; wherein the signal purification processing includes at least one of ECG artifact filtering, power frequency interference filtering, and motion artifact filtering. An extraction module is used to extract multi-dimensional physiological features based on the purified diaphragm electrical signal. The multi-dimensional physiological features include time-domain features, frequency-domain features, and stability features. The time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity in each respiratory cycle. The frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal. The stability features include the coefficient of variation of diaphragm electrical activity intensity in multiple consecutive respiratory cycles. The model processing module is used to input the multi-dimensional physiological features into a pre-constructed fusion evaluation model, so that the fusion evaluation model performs weighted fusion calculation on the multi-dimensional physiological features and outputs an offline evaluation index.

8. A system for calculating an offline evaluation index, characterized in that, include: Signal processing device, wireless surface diaphragm electroacupuncture module and surface electrodes; The wireless surface diaphragm electrical acquisition module is used to acquire the original diaphragm electrical signal of a preset diaphragm region on the patient's body surface through the surface electrodes. The original diaphragm electrical signal is preprocessed to obtain the target diaphragm electrical signal; The signal processing device is used to receive the target diaphragm electrical signal sent by the wireless surface diaphragm electrical acquisition module. The target diaphragm electrical signal is subjected to signal purification processing to obtain a purified diaphragm electrical signal. The signal purification processing includes at least one of ECG artifact filtering, power frequency interference filtering, and motion artifact filtering. Based on the purified diaphragm electrical signal, multi-dimensional physiological features are extracted, including time-domain features, frequency-domain features, and stability features. The time-domain features include the peak amplitude and area under the curve of diaphragm electrical activity within each respiratory cycle; the frequency-domain features include the center frequency calculated based on the power spectral density of the diaphragm electrical signal; and the stability features include the coefficient of variation of diaphragm electrical activity intensity over multiple consecutive respiratory cycles. The multi-dimensional physiological features are input into a pre-constructed fusion evaluation model, which performs weighted fusion calculations on the multi-dimensional physiological features and outputs an offline evaluation index.

9. A signal processing apparatus, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for calculating the offline evaluation index as described in any one of claims 1 to 6.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method for calculating the offline evaluation index as described in any one of claims 1 to 6.

Images