Respiratory effort evaluation system based on respiratory flow autopower spectrum

Abstract

The invention belongs to the technical field of breathing machines, and relates to a respiratory effort evaluation system based on a respiratory flow autopower spectrum. The system comprises a signal acquisition module used for acquiring a continuous breathing flow signal acquired by a flow sensor of the breathing machine in real time; the self-power spectrum calculation module is used for carrying out preprocessing and time-frequency conversion on the respiratory flow signals and calculating to obtain a self-power spectrum; the extraction module is used for extracting energy of a preset high-frequency band as a high-frequency energy component in the autopower spectrum, and the high-frequency band is higher than a main frequency band corresponding to the basic respiratory frequency of the human body; the statistical analysis module is used for establishing a baseline value for evaluation based on statistical data of high-frequency energy components in the historical respiratory flow data; the evaluation output module is used for comparing the high-frequency energy component at the current moment with the baseline value, when the high-frequency energy component is smaller than the baseline value, the respiratory effort state is evaluated to be abnormal, and otherwise, the respiratory effort state is normal.

Description

Technical Field

[0001] This invention belongs to the field of ventilator technology, and particularly relates to a respiratory effort assessment system based on respiratory flow autopower spectrum. Background Technology

[0002] Current ventilators, following specified parameters such as treatment pressure, respiratory cycle, and inspiratory-to-expiratory ratio, cyclically inflate the lungs, transition from inspiration to expiration, expel alveolar gas, and transition from expiration to inspiration to assist human respiration and address abnormal events during the respiratory process. This requires the ventilator to predict the human respiratory rhythm, including respiratory rate and inspiratory-to-expiratory ratio, to assist the respiratory rhythm controlled by the central nervous system with minimal conflict. However, current ventilator prediction technology is not mature enough to incorporate various states of the human respiratory rhythm into control. Furthermore, there are abrupt changes in respiratory rhythm and airway pressure caused by factors such as body position, disease, or others. These abrupt changes can lead to a mismatch between ventilator output and respiratory effort, and there are also numerous cases where the appropriate settings have not been titrated. Due to the influence of these external factors, the ventilator's pressure changes are inconsistent with the user's breathing actions, resulting in a phenomenon known as patient-ventilator asynchrony.

[0003] Regarding how to determine whether current ventilator control parameters match the patient's respiratory rhythm, current technologies mainly rely on subjective assessments based on patient-completed questionnaires. This lacks objective, quantifiable criteria for judging respiratory effort, and the assessment results are easily influenced by subjective factors. Methods for monitoring esophageal pressure and diaphragmatic electrical activity are overly stringent and unsuitable for routine treatment. Traditional methods, through questionnaire assessments and post-treatment data analysis, cannot achieve real-time monitoring of respiratory effort during ventilator use, making it difficult to promptly identify and adjust ventilator parameters to improve the patient's condition. While some methods using frequency domain analysis exist, which can identify patient-ventilator asynchrony events (such as invalid triggering or double triggering) by converting the frequency domain and examining the matching ratio of respiratory flow and pressure waveforms, these methods simply analyze synchrony through frequency domain analysis. Changes in respiratory synchrony may also be caused by lag in ventilator control response or occasional respiratory obstruction. This method cannot detect subtle tachypnea, which is often a precursor to respiratory events. Furthermore, this method lacks the ability to quantitatively assess respiratory rhythm coordination. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and to propose a respiratory effort assessment system based on the respiratory flow self-power spectrum.

[0005] In view of this, Embodiment 1 of the present invention provides a respiratory effort assessment system based on respiratory flow self-power spectrum, comprising: The signal acquisition module is used to acquire the continuous respiratory flow signal collected by the ventilator's flow sensor in real time; The self-power spectrum calculation module is used to preprocess the respiratory flow signal, perform time-frequency conversion, and calculate the self-power spectrum. The extraction module is used to extract the energy of a preset high-frequency band from the power spectrum as a high-frequency energy component, wherein the high-frequency band is higher than the main frequency band corresponding to the basic human respiratory rate. The statistical analysis module is used to establish a baseline value for assessment based on statistical data of high-frequency energy components in historical respiratory flow data; and The evaluation output module is used to compare the high-frequency energy component at the current moment with the baseline value. If the high-frequency energy component is less than the baseline value, the respiratory effort status is evaluated as abnormal; otherwise, it is normal.

[0006] Preferably, the preprocessing of the self-power spectrum calculation module includes: dividing the continuous respiratory flow signal into multiple data segments of fixed length, with overlap between adjacent data segments, and applying a window function to each data segment to reduce spectral leakage.

[0007] Preferably, the self-power spectrum of the self-power spectrum calculation module is calculated using the Welch method.

[0008] Preferably, in the extraction module, the preset high-frequency band is 0.5 Hz to 1.2 Hz.

[0009] Preferably, the processing procedure of the statistical analysis module includes: Identify and exclude abnormal respiratory event segments from the respiratory flow signal; the abnormal respiratory event segments include apnea, obstruction, and Cheyne-Stokes respiration. The average value is calculated using at least 20 auto-power spectrum data points during the steady breathing phase, and then multiplied by a preset correction factor to obtain the baseline value for evaluation; the correction factor ranges from 2.5 to 3.5.

[0010] Embodiment 2 of the present invention provides another respiratory effort assessment system based on respiratory flow self-power spectrum, comprising: The signal acquisition module is used to acquire the continuous respiratory flow signal collected by the ventilator's flow sensor in real time; The self-power spectrum calculation module is used to preprocess the respiratory flow signal, perform time-frequency conversion, and calculate the self-power spectrum. The frequency band marking module is used to obtain the current respiratory rate based on the continuous respiratory flow signal, and to mark the frequency bands within the set range of the respiratory rate point as the main frequency bands and the rest as other frequency bands. The indicator value calculation module is used to calculate the ratio based on the power spectrum of the current main frequency band and other frequency bands to obtain an indicator value that characterizes the degree of matching between respiratory effort and ventilator output; and The assessment output module is used to assess respiratory effort status based on indicated values.

[0011] Preferably, in the frequency band marking module, the frequency band within the respiratory frequency point setting range is R±v, where v is 50% of the current respiratory frequency R.

[0012] Preferably, the indication value obtained by the indication value calculation module is... for: in, The total power of the main frequency band, This represents the total power of other frequency bands.

[0013] Preferably, the processing procedure of the evaluation output module includes: The respiratory rate is between 3 and 60 breaths per minute, and the indicated value is... If the respiratory effort level exceeds the power threshold, the respiratory effort state is considered abnormal; otherwise, it is considered normal.

[0014] Preferably, the system further includes an adjustment instruction generation module for generating adjustment instructions when the assessed respiratory effort status is abnormal, to automatically or prompt the adjustment of the ventilator's treatment parameters, the treatment parameters including at least one of support pressure, trigger sensitivity, or respiratory rate.

[0015] Compared with the prior art, the advantages of the present invention are: 1. This invention is based on a ventilator flow sensor, which eliminates the need for additional detection devices at the patient's mouth and nose, thus reducing costs; 2. This invention focuses on the physiological significance of the power spectrum characteristics of different frequency bands, which can detect the slight respiratory tachypnea in the early stage of the mismatch between ventilator output and respiratory rhythm, and capture the precursor of such abnormal events earlier and more sensitively. 3. Power spectrum-based assessment makes it easier to quantify the resulting physiological changes without having to reassess the waveform of the work of breathing. Attached Figure Description

[0016] Figure 1 Here are flowcharts of respiratory flow signal preprocessing and power spectrum estimation for Examples 1 and 2; Figure 2 This is a schematic diagram of the screening and baseline establishment of normal and stable breathing data in Example 1. Detailed Implementation

[0017] This invention aims to systematically address the problem of ambiguous indicators in the monitoring and assessment of respiratory effort using existing non-invasive ventilators. Firstly, current technologies heavily rely on patients' subjective questionnaires and scale scores. This assessment method is not only lagging and discrete but also heavily influenced by individual subjective differences, failing to establish objective, unified, and quantifiable comfort measurement standards, resulting in insufficient consistency and reliability in clinical assessments. Secondly, traditional respiratory effort matching analysis is mostly based on post-treatment data retrospection, lacking the ability to monitor respiratory status in real-time and continuously during treatment. It cannot provide dynamic early warnings or immediate adjustments to ventilator parameters when respiratory effort mismatch occurs, missing the optimal intervention window. Alternatively, simple assessments of patient-ventilator synchronicity remain at the level of identifying time-domain waveforms of obvious abnormal events (such as invalid triggers or double triggers) or only perform basic frequency-domain synchronicity judgments, failing to quantify in depth from the perspective of respiratory rhythm coordination and subtle changes in neural control. Other algorithms assess respiratory effort through calculation and changes in the magnitude of respiratory effort. This method has extremely high requirements for the detection of relevant parameters, requiring additional detection equipment and specific detection procedures. Furthermore, existing algorithms are insufficient in mining the physiological information contained in respiratory flow signals. When patients are about to experience or are in a state of mild tachypnea due to insufficient ventilation or mismatch between ventilator pressure and respiratory rhythm—which is often a key early sign of declining respiratory effort matching—existing methods struggle to sensitively capture this weak but crucial rhythmic misalignment and respiratory drive changes from the high-frequency or specific spectral domain features of the signal. This results in insensitivity to the identification of "hidden" human-machine asynchrony and the resulting mismatch state, failing to achieve the leap from "post-event identification of obvious asynchrony" to "pre-event warning of potential mismatch."

[0018] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0019] Example 1 Embodiment 1 of the present invention provides a respiratory effort assessment system based on respiratory flow self-power spectrum, the system comprising: The signal acquisition module is used to acquire the continuous respiratory flow signal collected by the ventilator's flow sensor in real time; The self-power spectrum calculation module is used to preprocess the respiratory flow signal, perform time-frequency conversion, and calculate the self-power spectrum. The extraction module is used to extract the energy of a preset high-frequency band from the power spectrum as a high-frequency energy component, wherein the high-frequency band is higher than the main frequency band corresponding to the basic human respiratory rate. The statistical analysis module is used to establish a baseline value for evaluation based on statistical data of high-frequency energy components in historical respiratory flow data; The evaluation output module is used to compare the high-frequency energy component at the current moment with the baseline value. If the high-frequency energy component is less than the baseline value, the respiratory effort status is evaluated as abnormal; otherwise, it is normal.

[0020] The focus of work-effort analysis for respiration lies in the energy distribution in the frequency domain. The core of this system is to monitor and analyze the fluctuations in the high-frequency component of the respiratory flow signal power spectrum (typically referring to the portion above 0.5 Hz) in real time, thereby objectively and quantitatively assessing the degree of respiratory effort matching. The basic principle is that under stable conditions, the respiratory rhythm is robust, with dominant energy concentrated at the baseline respiratory rate (approximately 0.2-0.33 Hz for adults, i.e., 12-20 breaths / minute). However, when discomfort occurs (such as hypoventilation, patient-ventilator asynchrony, or early tachypnea), compensatory regulation by the respiratory center causes the respiratory rhythm to become unstable and the rate to increase. This manifests in the frequency domain as a significant increase in the energy of the high-frequency components, while the power spectrum of the dominant frequency component (0.2-0.5 Hz portion) corresponding to the normal respiratory rate remains relatively stable. Compared to directly calculating the change in total work, this method based on frequency domain energy ratios or frequency domain parameters has stronger noise immunity to respiratory flow, requires no additional observation equipment, and can be calculated solely from the ventilator-side flow data. This significantly reduces the impact of air leakage on the measured power. Furthermore, under normal conditions, the power of the main frequency component is stable, and the respiratory status can be determined by observing changes in the power spectrum data of the high-frequency component. The parameters can be controlled by observing the high-frequency component of the ventilator flow data.

[0021] By analyzing the gas flow characteristics of the ventilator system in actual application, the dominant frequency component and the high frequency component are distinguished. The ratio of the power spectrum of the dominant frequency component to the power spectrum of the high frequency component is found to be lower when the ventilator's work and breathing effort are matched, while the high frequency component accounts for a very high proportion when the patient is uncomfortable. This ratio is updated in real time during operation to achieve the indexation of comfort assessment. By monitoring this index, comfort can be identified to adjust ventilator parameters and optimize the patient experience.

[0022] Power spectral density (PSD) is a key tool for characterizing the distribution of signal power in the frequency domain. It is defined based on the square of the signal's spectral amplitude after time normalization. For a discretely sampled respiratory flow signal x[n], its power spectrum can be estimated using the periodogram method, expressed as:

[0023] Where N is the signal length and f is the frequency. It is the discrete-time Fourier transform (DTFT) of a discrete signal sequence of length N. This formula quantitatively characterizes the energy intensity of respiratory flow waveform data at each frequency component. As can be seen from the above formula, the energy intensity of respiratory flow in the high-frequency part can be estimated by the power spectrum.

[0024] At the implementation level, the specific implementation of this system is as follows: 1. Signal preprocessing To convert the time-domain respiratory flow signal to the frequency domain for analysis, the self-power spectrum calculation module uses the Welch method to estimate the power spectral density. Preprocessing includes: dividing the continuous signal provided by the ventilator flow sensor into multiple shorter data segments (each segment length L is typically 256 or 512 points), setting a 50% overlap between adjacent segments to increase the number of segments and improve the statistical reliability after averaging; and applying a window function (such as the Hanning window) to each data segment to reduce spectral leakage. Figure 1 As shown.

[0025] 2. Filtering based on normal and stable data The statistical analysis module's flow-based strategy is susceptible to numerous respiratory events, including apnea, obstruction, and Cheyne-Stokes respiration, all of which affect the work of breathing. When the ventilator identifies an event that may affect the flow-power ratio, the flow auto-power spectrum of that event must be excluded from the baseline calculation. The calculated auto-power spectrum range must represent a stable and normal respiratory waveform to ensure that abnormalities in flow-power do not lead to a large or very small portion being classified as abnormal. Figure 2 As shown.

[0026] 3. Baseline-based computation strategy After preprocessing and calculating the auto-power spectrum of the segmented flow data, the resulting auto-power spectrum is saved. The power spectrum integral value of the target high-frequency band (0.5–1.2 Hz) is calculated. The normal respiratory rate range for adults is 12–20 breaths / minute. To detect abnormal breathing, the data requirement is relaxed; a rate below 30 BPM is considered normal, corresponding to 0.5 Hz. Since the dominant frequency component of the respiratory flow is relatively stable, the statistical analysis module mainly compares changes in the high-frequency component, using a correction factor as the baseline.

[0027] The parameters of the baseline power spectrum Pbase are evaluated, where k represents the correction coefficient. Based on the self-power spectrum analysis of each respiratory waveform, prominent outliers are selected and compared with the average value. A coefficient of 3 is considered relatively safe based on the analysis of the actual respiratory waveform. However, there are some individuals whose respiratory waveforms show obvious abnormalities, but whose power spectrum shows relatively small abrupt changes. Therefore, the upper limit is set to 3.5, and the lower limit is set to 2.5. That is, the calculation range is 2.5-3.5. In this embodiment, the reference value PHigh must include at least 20 data points.

[0028] 4. Respiratory Effort State Judgment Strategy The evaluation output module compares the current respiratory power spectrum Pnow with Pbase:

[0029] Among them, positive and negative assessments of respiratory effort status A current power spectrum value greater than the baseline is identified as an abnormal respiratory effort state, while a value less than the baseline is identified as normal. Example 2 Embodiment 2 of the present invention provides a respiratory effort assessment system based on the power spectrum of respiratory flow. The system assesses respiratory effort by evaluating the energy distribution of the respiratory airflow. Based on the current human respiratory frequency R, frequency bands within a set range of this frequency are designated as the principal component, and all other frequency bands are designated as other frequency bands. The respiratory effort status is assessed by examining the power spectrum ratio of the principal component to other frequency bands. The system includes: The signal acquisition module is used to acquire the continuous respiratory flow signal collected by the ventilator's flow sensor in real time; The self-power spectrum calculation module is used to preprocess the respiratory flow signal, perform time-frequency conversion, and calculate the self-power spectrum; it also calculates the self-power spectrum of the respiratory airflow. :

[0030] in It is the autocorrelation function of respiratory airflow.

[0031] The two modules described above are implemented in the same way as in Example 1.

[0032] The frequency band marking module is used to obtain the current respiratory rate based on the continuous respiratory flow signal, and to mark the frequency bands within the set range of the respiratory rate point as the main frequency bands and the rest as other frequency bands. The indicator value calculation module is used to calculate the ratio based on the power spectrum of the current main frequency band and other frequency bands to obtain an indicator value that characterizes the degree of matching between breathing effort and ventilator output; the specific implementation is as follows: The observation interval R±v is set using the current respiratory rate R, where v is taken as 50% of the current respiratory rate. The actual value can be adjusted according to engineering applications. The total power of the observation frequency band is calculated. :

[0033] Total power of other sections :

[0034] Calculate the power ratio of the two sections. :

[0035] The assessment output module is used to assess respiratory effort status based on indicated values.

[0036] Using power ratio Positive and negative assessments of respiratory effort status. Here, a positive result indicates discomfort, and a negative result indicates no discomfort was detected. Based on practical engineering considerations, the respiratory rate is limited to 3 breaths per minute and 60 breaths per minute, with a power ratio... A respiratory effort level greater than a set threshold is considered positive; otherwise, it is considered negative. The corresponding piecewise function is as follows:

[0037] Where R is the respiratory rate. The power judgment threshold is generally set to 0.7, but can be adjusted according to the actual engineering application environment.

[0038] It is worth noting that in the embodiments of the above system, the modules included are divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional module are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0039] Invention point: 1. The role of the high-frequency respiratory flow band based on 0.5-1.2Hz in evaluating the matching degree between respiratory effort and ventilator output in Example 1; 2. Application of the high-frequency respiratory flow band based on 0.5-1.2Hz in ventilator parameter adjustment in Example 1; 3. The role of the correction coefficient range in determining the self-power spectrum of the high-frequency respiratory flow band based on 0.5-1.2Hz in Example 1; 4. The power spectrum ratio of the frequency band near the current respiratory rate and other frequency bands in Example 2, based on the observation interval R±v set according to the current respiratory rate R, plays a role in evaluating the degree of matching between respiratory effort and ventilator output; 5. The role of the power spectrum ratio of the frequency band near the current respiratory frequency and other frequency bands in adjusting ventilator parameters in Example 2, based on the observation interval R±v set according to the current respiratory frequency R.

[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the embodiments, those skilled in the art should understand that modifications or equivalent substitutions to the technical solutions of the present invention do not depart from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A respiratory effort assessment system based on respiratory flow self-power spectrum, characterized in that, include: The signal acquisition module is used to acquire the continuous respiratory flow signal collected by the ventilator's flow sensor in real time; The self-power spectrum calculation module is used to preprocess the respiratory flow signal, perform time-frequency conversion, and calculate the self-power spectrum. The extraction module is used to extract the energy of a preset high-frequency band from the power spectrum as a high-frequency energy component, wherein the high-frequency band is higher than the main frequency band corresponding to the basic human respiratory rate. The statistical analysis module is used to establish a baseline value for evaluation based on statistical data of high-frequency energy components in historical respiratory flow data; and The evaluation output module is used to compare the high-frequency energy component at the current moment with the baseline value. If the high-frequency energy component is less than the baseline value, the respiratory effort status is evaluated as abnormal; otherwise, it is normal.

2. The respiratory effort assessment system based on respiratory flow self-power spectrum according to claim 1, characterized in that, The preprocessing of the self-power spectrum calculation module includes: dividing the continuous respiratory flow signal into multiple fixed-length data segments with overlap between adjacent data segments, and applying a window function to each data segment to reduce spectral leakage.

3. The respiratory effort assessment system based on respiratory flow self-power spectrum according to claim 1, characterized in that, The self-power spectrum of the self-power spectrum calculation module is calculated using the Welch method.

4. The respiratory effort assessment system based on respiratory flow self-power spectrum according to claim 1, characterized in that, In the extraction module, the preset high-frequency band is 0.5 Hz to 1.2 Hz.

5. The respiratory effort assessment system based on respiratory flow self-power spectrum according to claim 1, characterized in that, The processing steps of the statistical analysis module include: Identify and exclude abnormal respiratory event segments from the respiratory flow signal; the abnormal respiratory event segments include apnea, obstruction, and Cheyne-Stokes respiration. The average value is calculated using at least 20 auto-power spectrum data points during the steady breathing phase, and then multiplied by a preset correction factor to obtain the baseline value for evaluation; the correction factor ranges from 2.5 to 3.

5.

6. A respiratory effort assessment system based on respiratory flow self-power spectrum, characterized in that, include: The signal acquisition module is used to acquire the continuous respiratory flow signal collected by the ventilator's flow sensor in real time; The self-power spectrum calculation module is used to preprocess the respiratory flow signal, perform time-frequency conversion, and calculate the self-power spectrum. The frequency band marking module is used to obtain the current respiratory rate based on the continuous respiratory flow signal, and to mark the frequency bands within the set range of the respiratory rate point as the main frequency bands and the rest as other frequency bands. The indicator value calculation module is used to calculate the ratio based on the power spectrum of the current main frequency band and other frequency bands to obtain an indicator value that characterizes the degree of matching between breathing effort and ventilator output. and The assessment output module is used to assess respiratory effort status based on indicated values.

7. The respiratory effort assessment system based on respiratory flow self-power spectrum according to claim 6, characterized in that, In the frequency band marking module, the frequency band within the respiratory frequency point setting range is R±v, where v is 50% of the current respiratory frequency R.

8. The respiratory effort assessment system based on respiratory flow self-power spectrum according to claim 6, characterized in that, The indicator value obtained by the indicator value calculation module for: ; in, The total power of the main frequency band, This represents the total power of other frequency bands.

9. The method for assessing respiratory effort based on respiratory flow self-power spectrum according to claim 8, characterized in that, The processing procedure of the evaluation output module includes: The respiratory rate is between 3 and 60 breaths per minute, and the indicated value is... If the respiratory effort level exceeds the power threshold, the respiratory effort state is considered abnormal; otherwise, it is considered normal.

10. The respiratory effort assessment system based on respiratory flow self-power spectrum according to claim 1 or claim 6, characterized in that, The system also includes an adjustment instruction generation module, which generates adjustment instructions when the assessed respiratory effort status is abnormal, to automatically or prompt the adjustment of the ventilator's treatment parameters, including at least one of support pressure, trigger sensitivity, or respiratory rate.

Images