A sleep apnea real-time monitoring and early warning system

Abstract

The present application relates to the technical field of medical identification, in particular to a sleep apnea real-time monitoring and early warning system, which comprises: an acquisition module, configured to acquire PSG data of a target user from a sleep time to a current time; a determination module, configured to determine all OSA events from the sleep time to the current time according to a spectrum of a respiratory effort wave and a spectrum of airflow of a window corresponding to each time in the PSG data; the determination module is further configured to determine a blood oxygen consumption period of a current OSA event; the determination module is further configured to determine a credible OSA event; and an early warning module, configured to construct an early warning model of the target user by using PSG data of the credible OSA event, and to perform sleep apnea early warning on the target user by using real-time PSG data of the target user collected in real time and the early warning model. The present application improves the accuracy of OSA diagnosis.

Description

Technical Field

[0001] This invention relates to the field of medical identification technology, specifically to a real-time monitoring and early warning system for sleep apnea. Background Technology

[0002] Obstructive sleep apnea (OSA), a prevalent sleep-related breathing disorder, is characterized by recurrent partial or complete obstruction of the upper airway during sleep, leading to periodic apnea and hypopnea. Currently, the core clinical diagnostic method for OSA is polysomnography (PSG) combined with the apnea-hypopnea index (AHI). PSG, as the standard for diagnosing OSA, requires long-term, synchronous, and continuous monitoring of multiple physiological indicators, including electroencephalography (EEG), electrooculography (EOG), electrocardiography (ECG), respiratory airflow, blood oxygen saturation, and chest and abdominal movements. Analysis of the monitoring data is then used to determine the presence and severity of OSA.

[0003] In some scenarios, existing PSG diagnostic technologies have significant limitations in clinical application. The core issue lies in the fact that the testing environment and methods are prone to triggering the "first night effect," leading to distorted monitoring data and affecting diagnostic accuracy. Specifically, PSG testing requires patients to complete the test in a specialized medical monitoring environment, where they wear multiple sensors and monitoring electrodes. The unfamiliar and unnatural sleep environment, along with the patient's subjective perception of being monitored, can induce anxiety, significantly increasing sleep alertness. In this state, patients frequently experience micro-awakenings during sleep, which directly activate the sympathetic nervous system, triggering physiological responses such as increased heart rate and unconscious bodily movements. It is important to note that the normal monitoring logic for OSA events is as follows: upper airway obstruction leads to apnea, resulting in a decrease in blood oxygen saturation and an increase in carbon dioxide concentration. These physiological changes trigger micro-awakenings, prompting the patient to unconsciously adjust their body posture to restore upper airway muscle tone, ultimately achieving airway reopening and respiratory recovery. However, the micro-awakenings caused by the "first night effect" are not due to the pathological event of OSA; they are non-pathological interference factors that generate a large amount of false monitoring data. These spurious monitoring data can confuse real OSA events with physiological microarousal responses, leading to misjudgments when analyzing the AHI index and other monitoring indicators. This makes it impossible to accurately identify the patient's true apnea and hypoventilation rate, thus affecting the accuracy of OSA diagnosis and potentially causing missed diagnoses, misdiagnoses, or misjudgments of the severity of the condition. Summary of the Invention

[0004] To address the technical problem of insufficient accuracy in OSA diagnosis, the present invention aims to provide a real-time monitoring and early warning system for sleep apnea.

[0005] To solve the above technical problems, the specific technical solution adopted is as follows:

[0006] This invention provides a real-time sleep apnea monitoring and early warning system, comprising: an acquisition module for acquiring PSG data of a target user from the time of sleep to the current time; a determination module for determining all OSA events from the time of sleep to the current time based on the spectrum of respiratory effort waves and the spectrum of airflow in the windows corresponding to each time moment in the PSG data, wherein the OSA events include awakening periods and apnea periods; the determination module is further configured to determine the blood oxygen consumption period of the current OSA event; the determination module is further configured to determine credible OSA events based on the blood oxygen saturation during the blood oxygen consumption period of the OSA event, the average amplitude of blood oxygen saturation and respiratory effort waves during the apnea period of the OSA event, the heart rate during the awakening period of the OSA event, and the heart rate during the apnea period of the OSA event; and an early warning module for constructing an early warning model for the target user using the PSG data of credible OSA events, and providing sleep apnea early warning to the target user using the real-time PSG data of the target user collected in real time and the early warning model.

[0007] Optionally, the determining module is further configured to: determine the degree of respiratory disturbance at each time point based on the spectrum of the respiratory effort wave corresponding to each time point in the PSG data, and determine the degree of airflow disturbance at each time point based on the spectrum of the airflow corresponding to each time point in the PSG data; determine the probability of micro-arousal at each time point based on the degree of respiratory disturbance and the degree of airflow disturbance; determine the target time point where the probability of micro-arousal is greater than or equal to the arousal judgment threshold; determine the window of the time point preceding the target time point as the arousal period; determine the difference coefficient of each time point in the window of the consecutive time points based on the amplitude and airflow of the respiratory effort wave at each time point in the window of the consecutive time points preceding the target time; determine the apnea period at the target time point based on the difference coefficient of each time point in the window of the consecutive time points; determine the arousal period and the apnea period as an OSA event, and determine all OSA events from the sleep time point to the current time point.

[0008] Optionally, the determining module is further configured to: calculate the absolute value of a first difference between the amplitude of each frequency in the spectrum of the respiratory effort wave of the window corresponding to the current time and the amplitude of each frequency in the spectrum of the respiratory effort wave of the window before the current time; determine the degree of respiratory disturbance at each time based on the absolute value of each first difference; calculate the absolute value of a second difference between the amplitude of each frequency in the spectrum of the airflow of the window corresponding to the current time and the amplitude of each frequency in the spectrum of the airflow of the window before the current time; and determine the degree of airflow disturbance at each time based on the absolute value of each second difference.

[0009] Optionally, the determination module is also used to: calculate the average difference coefficient of all times in a window of multiple consecutive times of the target time; determine the times when the difference coefficient is greater than the average value as pause times; and determine the interval formed by the closest pause time before the current pause time and the current pause time as the apnea period of the target time.

[0010] Optionally, the determining module is further configured to: determine the period between the first moment of the apnea period of the current OSA event and the last moment of the apnea period of the previous OSA event as the blood oxygen consumption period of the current OSA event.

[0011] Optionally, the determination module is further configured to: determine the dissolved oxygen demand coefficient of the current OSA event based on the blood oxygen saturation during the blood oxygen consumption period, the blood oxygen saturation during the apnea period, and the average amplitude of the respiratory effort wave; determine the respiratory interruption performance of the current OSA event based on the dissolved oxygen demand coefficient, the heart rate during the awakening period, and the heart rate during the apnea period; and screen for reliable OSA events with a respiratory interruption performance greater than the interruption threshold from all OSA events.

[0012] Optionally, the determining module is further configured to determine a first fitted slope of blood oxygen saturation during the blood oxygen consumption period of the current OSA event, a second fitted slope of blood oxygen saturation during the apnea period of the current OSA event, and the duration of the apnea period of the current OSA event; determine the maximum value of blood oxygen saturation during the blood oxygen consumption period of the current OSA event and the minimum value of blood oxygen saturation during the apnea period of the current OSA event; calculate a first ratio between the maximum and minimum values, and a second ratio between the first fitted slope and the second fitted slope; and determine the dissolved oxygen demand coefficient of the current OSA event based on the first ratio, the second ratio, the duration, and the average amplitude of the respiratory effort wave during the apnea period of the current OSA event.

[0013] Optionally, the determination module is also used to: calculate a third difference between the standard deviation of the heart rate during the awakening phase and the standard deviation of the heart rate during the apnea phase of the current OSA event; determine a third fitting slope of the heart rate during the apnea phase of the current OSA event; and determine the respiratory interruption performance of the current OSA event based on the third difference, the third fitting slope, and the dissolved oxygen demand coefficient.

[0014] Optionally, the early warning module is also used to: construct a coordinate system for all PSG data in all trusted OSA events with the same dimension, where the horizontal axis of the coordinate system is time and the vertical axis is the value of the dimension data; and use the least squares method to fit all data points in the coordinate system to obtain an early warning model for each dimension of the target user's PSG data.

[0015] Optional PSG data include: heart rate, respiratory effort wave, airflow, and blood oxygen saturation.

[0016] This invention offers the following advantages: By jointly analyzing respiratory effort wave spectrograms and airflow spectrograms, the awakening and apnea periods of OSA events are located, eliminating interference from non-OSA-related abnormal periods at the OSA event classification level. Furthermore, by utilizing multi-dimensional physiological characteristics such as blood oxygen saturation during blood oxygen consumption periods, blood oxygen saturation during apnea periods, average amplitude of respiratory effort waves, and heart rate differences between awakening and apnea periods, a credible OSA event judgment standard is constructed. Since the micro-awakening caused by the first-night effect lacks a corresponding pathological correlation chain of "apnea-blood oxygen change-heart rate adjustment," its related data is precisely filtered out, effectively avoiding the interference of pseudo-data on the AHI index and other core monitoring indicators, ensuring that all identified OSA events are genuine pathological events, and significantly improving the accuracy of OSA event identification. Furthermore, by employing a dual verification logic of accurately identifying genuine OSA events and screening credible OSA events, the reliability of the diagnostic basis is ensured from the data source: PSG data based on credible OSA events can accurately calculate core diagnostic indicators such as the AHI index, truly reflecting the severity of the patient's OSA condition; simultaneously, the joint judgment of multi-dimensional physiological characteristics can cover the pathological features of different types of OSA events, avoiding the limitations of single-indicator judgment and further improving the comprehensiveness and reliability of the diagnosis. Moreover, the personalized early warning model built based on credible OSA event data can accurately match the physiological characteristics of the target user with the OSA incidence pattern, improving the targeting of the early warning; by dynamically matching real-time collected PSG data with the early warning model, timely warnings can be issued when patients show signs of OSA events or when OSA events occur, facilitating rapid intervention by medical staff or patients' families and reducing the duration of OSA events and the resulting harm to the patient's body. Attached Figure Description

[0017] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of a real-time sleep apnea monitoring and early warning system according to an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of a real-time sleep apnea monitoring and early warning system provided in another embodiment of the present invention. Detailed Implementation

[0020] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a real-time sleep apnea monitoring and early warning system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0022] The specific solution of the real-time monitoring and early warning system for sleep apnea provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Example 1:

[0024] Please see Figure 1 This document illustrates a schematic diagram of a real-time sleep apnea monitoring and early warning system according to an embodiment of the present invention. The system includes: an acquisition module 101, used to acquire PSG data of a target user from the time of sleep to the current time; a determination module 102, used to determine all OSA events from the time of sleep to the current time based on the spectrum of respiratory effort waves and the spectrum of airflow corresponding to each time window in the PSG data, wherein OSA events include awakening periods and apnea periods; the determination module 102 is also used to determine the blood oxygen consumption period of the current OSA event; the determination module 102 is also used to determine credible OSA events based on the blood oxygen saturation during the blood oxygen consumption period of the OSA event, the average amplitude of blood oxygen saturation and respiratory effort waves during the apnea period of the OSA event, the heart rate during the awakening period of the OSA event, and the heart rate during the apnea period of the OSA event; and an early warning module 103, used to construct an early warning model for the target user using the PSG data of credible OSA events, and to provide sleep apnea early warning to the target user using the real-time collected PSG data of the target user and the early warning model.

[0025] Specifically, the PSG data in this embodiment of the invention includes: heart rate, respiratory effort wave, airflow, and blood oxygen saturation. This embodiment first places the target user (e.g., a patient) into the PSG monitoring room and performs corresponding data detection. The specific process of acquiring PSG data by module 101 is as follows: Heart rate data acquisition: Clean the skin at the electrode attachment site with a scrub. Attach the electrodes to the skin surface of the patient's right subclavian, left subclavian, and right lower abdomen to form a triangle. Each electrode is equipped with an electrocardiogram electrode sensor. Connect the electrocardiogram electrode sensor to an electrocardiogram device to acquire the heart rate at each moment. Respiratory effort wave and airflow acquisition: Use a respiratory sensing tape wrapped around the chest (usually at the level of the nipples) and abdomen (usually at the level of the navel). When breathing, changes in chest and abdominal dimensions cause changes in the inductance or resistance within the tape. The magnitude of the change is proportional to the respiratory effort, thus recording the change in resistance to obtain the respiratory effort wave. A thin catheter is placed in front of the nostrils, connected to a pressure sensor, to monitor the minute pressure changes generated during inhalation / exhalation, obtaining a pressure waveform, i.e., airflow. Blood oxygen saturation data: The probe (usually clip-on or adhesive) is clipped or attached to the fingertip (most commonly), earlobe, or toe. Blood oxygen saturation is collected using a pulse oximeter probe based on the photoplethysmography method. One side of the probe emits two wavelengths of light (red light and infrared light), and the other side receives it. Oxygenated hemoglobin and deoxygenated hemoglobin have different absorption rates for the two types of light. The receiver calculates the real-time blood oxygen saturation by analyzing the pulsatility changes of the projected and reflected light. At the same time, the pulsatility changes can also be used to obtain the pulse rate. If the pulse rate is used as the heart rate, heart rate collection is not necessary.

[0026] Furthermore, in this embodiment of the invention, the PSG data at the current moment is monitored, and the probability of micro-arousal at the current moment is assessed based on the respiratory effort wave and airflow turbulence. If the micro-arousal probability is met, the most recent OSA apnea period before the current moment is obtained based on the difference between the respiratory effort wave and airflow. The micro-arousal at the current moment and the obtained OSA apnea period are recorded as an OSA event at the current moment, and all OSA events from the start of monitoring to the current moment are obtained. Therefore, as an optional embodiment of the present invention, the determining module 102 is further configured to: determine the degree of respiratory disturbance at each time point based on the spectrum of the respiratory effort wave corresponding to each time point in the PSG data, and determine the degree of airflow disturbance at each time point based on the spectrum of the airflow corresponding to each time point in the PSG data; determine the micro-arousal probability at each time point based on the degree of respiratory disturbance and the degree of airflow disturbance; determine the target time point where the micro-arousal probability is greater than or equal to the arousal judgment threshold; determine the window of the time point adjacent to the target time point as the arousal period; determine the difference coefficient of each time point in the window of the consecutive time points based on the amplitude and airflow of the respiratory effort wave at each time point in the window of the consecutive time points before the target time; determine the apnea period at the target time point based on the difference coefficient of each time point in the window of the consecutive time points; determine the arousal period and the apnea period as an OSA event, and determine all OSA events from the sleep time point to the current time point.

[0027] Specifically, monitoring sleep apnea events primarily analyzes how, during sleep, relaxation or collapse of the pharyngeal muscles obstructs airflow, leading to apnea and decreased blood oxygen saturation. The brain detects this low oxygen saturation and initiates micro-arousals to adjust the patient's posture and sleeping position, restoring upper airway muscle tone and airway closure, thus resuming breathing. Therefore, the standard procedure for OSA event monitoring is: apnea causes a decrease in blood oxygen saturation and an increase in carbon dioxide, triggering micro-arousals in the brain. Unconscious activity adjusts body posture, restoring upper airway muscle tone and airway closure, thereby resuming breathing. A complete OSA event must include both apnea and the restoration of breathing through physical activity. Prolonged apnea can cause blood oxygen saturation to remain at a very low level during sleep, leading to poor sleep quality and daytime sleepiness. Therefore, timely warnings are crucial for patients with sleep apnea, and physicians use the frequency and recording of warnings to determine appropriate treatment methods.

[0028] Furthermore, because patients may snore, adopt postures unfavorable to breathing, or experience short periods of apnea during prolonged sleep, early warning only requires monitoring for prolonged apnea in conjunction with physical activity. However, monitoring for sleep apnea requires transferring the patient to a specialized medical environment. Since the patient's sleep is in a completely new environment, a "first night effect" occurs, significantly reducing sleep quality and depth, leading to prolonged periods of light sleep. This makes the patient more prone to unconscious activities, which are similar to micro-awakenings. This combination of short-term apnea and unconscious activity can be misdiagnosed as sleep apnea, resulting in an overestimation of the overall severity of OSA events (AHI index).

[0029] Therefore, this embodiment of the invention analyzes the multimodal data of patients when they have a combination of physical activity and apnea, and then monitors and warns of sleep apnea. Specifically, it filters the apnea-physical activity combination and then constructs an early warning model based on the filtered combination that only contains long-term apnea-physical activity.

[0030] Furthermore, when the patient is asleep, the patient's breathing is stable, meaning there is no disturbance in the respiratory effort wave and airflow. However, when the patient engages in physical activity, because the body is in a relaxed state for a long time during sleep, and the transition from a relaxed state to an active state requires the body to accumulate energy, this causes disturbance in the respiratory effort wave and airflow. Therefore, this application monitors the PSG data at the current moment and assesses the probability of micro-arousal at the current moment based on the disturbance in the respiratory effort wave and airflow.

[0031] Therefore, in this embodiment of the invention, a time window of 20 seconds is preset to obtain the spectrum of respiratory effort waves and airflow within each window, starting from the current moment and moving forward. Since respiratory effort waves and airflow are periodic data, changes in the amplitude of the chest cavity and airflow during activity will disrupt this spectrum. Since each window contains at least one respiratory cycle, by obtaining the spectrum and analyzing the differences in respiratory effort waves and airflow within it, the probability of micro-arousal at the current moment can be obtained. Based on this, as an optional embodiment of the invention, the determining module 102 is further configured to: calculate the absolute value of a first difference between the amplitude of each frequency in the spectrum of the respiratory effort wave corresponding to the current moment and the amplitude of each frequency in the spectrum of the respiratory effort wave of windows prior to the current moment; determine the degree of respiratory disturbance at each moment based on the absolute value of each first difference; calculate the absolute value of a second difference between the amplitude of each frequency in the spectrum of airflow corresponding to the current moment and the amplitude of each frequency in the spectrum of airflow of windows prior to the current moment; and determine the degree of airflow disturbance at each moment based on the absolute value of each second difference.

[0032] Specifically, the embodiments of the present invention use the following formula to calculate the degree of respiratory disturbance:

[0033]

[0034] In the above formula, The degree of respiratory disturbance at time i. Let x be the amplitude of the x-th frequency in the spectrum of the breathing effort wave at time i. Let x be the amplitude of the x-th frequency in the spectrum of the breathing effort wave in the m-th window before time i. This represents the inverse proportional normalization function.

[0035] Similarly, the degree of airflow turbulence at time i can be calculated using the same method as for calculating the degree of respiratory disturbance. The probability of micro-arousal at time i is calculated as follows: . This indicates taking the maximum value.

[0036] Furthermore, in this embodiment of the invention, the arousal judgment threshold can be determined according to the actual scenario, and in this embodiment, the value is 0.5. If the micro-arousal probability at time i is greater than or equal to 0.5, then the window of the adjacent time before time i is recorded as an arousal period.

[0037] Furthermore, since the amplitude of the respiratory effort wave generated by the patient's chest is directly proportional to the amplitude of the airflow change, this embodiment of the invention uses a pressure sensor to monitor the airflow. By monitoring the pressure change within the thin duct, i.e., during inhalation or exhalation, the airflow increases. According to Bernoulli's principle, the faster the flow rate, the lower the pressure. Therefore, the acquired airflow is obtained after inverse number post-processing. When apnea occurs, especially when obstructive apnea is caused by muscle collapse, the respiratory effort waveform of the chest and abdomen exists or even intensifies, but the airflow decreases significantly. Therefore, this embodiment of the invention analyzes the five windows before the i-th time, specifically using the following formula to calculate the difference coefficient of each time in the windows of multiple consecutive time periods before the target time:

[0038]

[0039] In the above formula, is the difference coefficient between respiration and airflow at time j within the 5 windows preceding time i. This represents the amplitude of the respiratory effort wave at time j within the 5 windows preceding time i. This represents the airflow magnitude at time j within the 5 windows preceding time i.

[0040] It should be noted that when calculating the difference coefficient, if the denominator of the calculation formula is 0, a zero-prevention parameter is added to the denominator before calculation. In this embodiment, the zero-prevention parameter is 0.001. In specific applications, implementers can set it according to specific circumstances.

[0041] Furthermore, as an optional embodiment of the present invention, the determining module 102 is also configured to: calculate the average value of the difference coefficients of all times in a window of multiple consecutive times of the target time; determine the times when the difference coefficient is greater than the average value as pause times; and determine the interval formed by the closest pause time before the current pause time and the current pause time as the apnea period of the target time.

[0042] Specifically, in this embodiment of the invention, the difference coefficients of all times within 5 windows are obtained, and the times when the difference coefficients of all times within 5 windows are greater than the average of the difference coefficients are recorded as pause times. The interval formed by the most recent pause time before the i-th time is recorded as an OSA apnea period, and the apnea period and the awakening period together constitute an OSA event.

[0043] Furthermore, since the assessment of OSA events is performed in real time, all OSA events of the patient from the onset of sleep to the current moment can be obtained.

[0044] Furthermore, among all the OSA historical events obtained, there were unconscious activities due to the first-night effect and unconscious activities due to prolonged apnea. However, the early warning model only considered unconscious activities caused by prolonged apnea. Therefore, it is necessary to screen all OSA events to find respiratory data indicating severe respiratory interruption in the patient. Due to differences in physical condition, especially sleep habits and body posture, the degree of airway collapse varies among patients. Traditional assessment of OSA severity mainly relies on physicians using a constant AHI threshold, which significantly affects the assessment of the patient's condition and early warning of apnea. Since blood oxygen saturation is crucial for maintaining normal metabolism, the value of blood oxygen saturation is relatively high. The duration of apnea and the respiratory effort wave during apnea, indicating the body's oxygen demand, are both affected by blood oxygen saturation. Therefore, this application obtains the dissolved oxygen demand coefficient for OSA events by analyzing the covariance between blood oxygen changes and respiration, and the relationship between the duration of respiratory interruption and the resulting drop in blood oxygen levels.

[0045] Therefore, as an optional embodiment of the present invention, the determining module 102 is further configured to: determine the period between the first moment of the apnea period of the current OSA event and the last moment of the apnea period of the previous OSA event as the blood oxygen consumption period of the current OSA event.

[0046] Specifically, in this embodiment of the invention, after acquiring each OSA event, each OSA event includes a sleep apnea period and a wakefulness period, the time period from the first moment of the sleep apnea period of each OSA event to the last moment of the preceding sleep apnea period is defined as the blood oxygen consumption period of each OSA event. It should be noted that for the first OSA event during the monitoring process, since there is no "previous OSA event," its blood oxygen consumption period is defined as the time from when the user enters sleep until the first moment of the sleep apnea period of that OSA event.

[0047] Furthermore, as an optional embodiment of the present invention, the determining module 102 is further configured to: determine the dissolved oxygen demand coefficient of the current OSA event based on the blood oxygen saturation during the blood oxygen consumption period of the current OSA event, the blood oxygen saturation during the apnea period of the current OSA event, and the average amplitude of the respiratory effort wave; determine the respiratory interruption performance of the current OSA event based on the dissolved oxygen demand coefficient, the heart rate during the awakening period of the current OSA event, and the heart rate during the apnea period; and screen credible OSA events from all OSA events whose respiratory interruption performance is greater than the interruption threshold.

[0048] Specifically, the longer the blood oxygen consumption period, the higher the patient's demand for oxygen. If the duration of the apnea period is longer, and the blood oxygen continues to drop during that period, and the amplitude of the respiratory effort wave during the apnea period is larger, it indicates that the body had taken in very low amounts of oxygen for a long period of time before the current OSA event, thus requiring a higher amount of oxygen. Therefore, the dissolved oxygen demand coefficient for the current OSA event can be obtained.

[0049] Furthermore, as an optional embodiment of the present invention, the determining module 102 is further configured to: determine a first fitting slope of blood oxygen saturation during the blood oxygen consumption period of the current OSA event, a second fitting slope of blood oxygen saturation during the apnea period of the current OSA event, and the duration of the apnea period of the current OSA event; determine the maximum value of blood oxygen saturation during the blood oxygen consumption period of the current OSA event and the minimum value of blood oxygen saturation during the apnea period of the current OSA event; calculate a first ratio between the maximum value and the minimum value, and a second ratio between the first fitting slope and the second fitting slope; and determine the dissolved oxygen demand coefficient of the current OSA event based on the first ratio, the second ratio, the duration, and the average amplitude of the respiratory effort wave during the apnea period of the current OSA event.

[0050] Specifically, the dissolved oxygen demand coefficient is calculated using the following formula in the embodiments of the present invention:

[0051]

[0052] In the above formula, This represents the dissolved oxygen demand coefficient. This represents the maximum blood oxygen saturation during the blood oxygen consumption period of the current OSA event. This represents the minimum blood oxygen saturation during the apnea period of the current OSA event. This indicates the length of the apnea period in the current OSA event. This represents the first fitted slope of blood oxygen saturation during the blood oxygen consumption period of the current OSA event. The second fitted slope represents the oxygen saturation during the apnea period of the current OSA event. This represents the average amplitude of the respiratory effort wave during the apnea period of the current OSA event. This represents the normalization function.

[0053] It should be noted that when calculating the dissolved oxygen demand coefficient, if the denominator of the calculation formula is 0, a zero-prevention parameter is added to the denominator before calculation. In this embodiment, the zero-prevention parameter is 0.001. In specific applications, implementers can set it according to specific circumstances.

[0054] Furthermore, when there is a difference in blood oxygen levels... The larger the value of l, the more oxygen is consumed in the blood due to ineffective ventilation, leading to a sustained and slow decline in blood oxygen saturation, reaching its lowest point during the period of minimal arousal. A larger value of l indicates a greater drop in blood oxygen levels. Because the body is in a state of extremely low energy consumption during sleep, oxygen is only absorbed through respiration when the body's blood oxygen saturation drops to a certain value; the slope of this decline is the rate of oxygen intake. During periods of sleep apnea, because the body's blood oxygen saturation reaches its lower limit, cells will excessively draw oxygen from the blood during this pause, leading to a further decrease in blood oxygen saturation. Greater than , The larger the value, the greater the oxygen demand coefficient; the average amplitude indicates the body's oxygen demand when it is hypoxic.

[0055] The first fitting slope can be obtained by fitting the blood oxygen saturation during the blood oxygen consumption period of the current OSA event using the least squares method, and the second fitting slope can be obtained by fitting the blood oxygen saturation during the apnea period of the current OSA event using the least squares method.

[0056] Furthermore, in OSA events, due to the lack of effective ventilation, blood oxygen is continuously consumed, leading to a sustained and slow decline in blood oxygen saturation. Hypoxia and hypercapnia stimulate the vagus nerve, causing a gradual decrease in heart rate. Meanwhile, the heart rate during micro-arousals due to the first-night effect may be faster throughout sleep due to anxiety. During micro-arousals, the heart rate suddenly increases without a prior gradual decrease, jumping directly from a relatively normal baseline level. Therefore, by analyzing the apnea periods preceding the micro-arousal period and the changes in heart rate during the micro-arousal period, combined with the dissolved oxygen demand coefficient of the current OSA event, the respiratory interruption performance of the current OSA event can be obtained. Therefore, as an optional embodiment of the present invention, the determining module 102 is further configured to: calculate a third difference between the standard deviation of the heart rate during the awakening period and the standard deviation of the heart rate during the apnea period of the current OSA event; determine a third fitting slope for the heart rate during the apnea period of the current OSA event; and determine the respiratory interruption performance of the current OSA event based on the third difference, the third fitting slope, and the dissolved oxygen demand coefficient.

[0057] Specifically, the embodiments of the present invention use the following formula to calculate the degree of respiratory interruption:

[0058]

[0059] In the above formula, This represents the degree of respiratory interruption in the current OSA event. This represents the standard deviation of the heart rate during the awakening period of the current OSA event. This represents the standard deviation of heart rate during the apnea period of the current OSA event. The slope fitted value of the heart rate during the apnea period of the current OSA event. This represents the normalization function. The third fitting slope can be obtained by fitting the heart rate during the apnea period of the current OSA event using the least squares method. In this embodiment, the maximum-minimum normalization method is used to normalize the data. This method is existing technology and will not be elaborated further here. In specific applications, other existing data normalization methods can also be used.

[0060] Furthermore, the interruption threshold in this embodiment can be determined according to the actual scenario; in this embodiment, it is set to 0.5. Based on the above embodiment, the respiratory interruption performance of each OSA event is obtained, and then an interruption threshold of 0.5 is preset. If the respiratory interruption performance of a historical OSA event is greater than this interruption threshold, then the OSA event is recorded as a credible OSA event.

[0061] Furthermore, as an optional embodiment of the present invention, the early warning module 103 is also used to: construct a coordinate system of the same dimension data of all PSG data in all trusted OSA events, with the horizontal axis of the coordinate system being time and the vertical axis being the value of the dimension data; and use the least squares method to fit all data points in the coordinate system to obtain an early warning model of each dimension data in the target user's PSG data.

[0062] Specifically, in this embodiment of the invention, all PSG data collected in each credible OSA event of the patient are statistically analyzed. This PSG data represents the patient's PSG data during the blood oxygen consumption period in each credible OSA event, i.e., data during bodily exertion. The PSG data contains multiple data points of different dimensions. A coordinate system is constructed from the data of the same dimension in all credible OSA events, and all data are mapped onto this coordinate system. Since PSG data consists of data sequences of points, the horizontal axis represents time, and the vertical axis represents the value of that dimension's data. Therefore, the least squares method is used to fit all the points, thereby obtaining an early warning model for each dimension of the patient's PSG data. This early warning model reflects the developmental pattern of the patient's PSG data indicative of sleep apnea.

[0063] Furthermore, this embodiment of the invention subsequently monitors the PSG data of each patient in real time at each moment. If the PSG data of a certain dimension in the monitored period is highly similar to the data of that dimension in the early warning model, for example, by using the Pearson correlation coefficient to calculate the correlation, when the correlation is greater than a certain threshold, it indicates that the patient may have a long-term sleep apnea in the future based on the PSG data of the blood oxygen consumption period, and therefore an early warning is issued to the patient. For example, this embodiment of the invention collects respiratory effort wave, airflow data, or other PSG data. If the Pearson correlation coefficient between the respiratory effort wave and the respiratory effort wave in the early warning model in the current period is greater than a certain threshold (such as 0.8), it indicates that the user's respiratory effort wave in the current period is highly similar to the user's respiratory effort wave in the blood oxygen consumption period in the early warning model, and sleep apnea may occur in the future, requiring a sleep apnea warning.

[0064] This invention, through joint analysis of respiratory effort wave spectrograms and airflow spectrograms, locates the awakening and apnea periods of OSA events, eliminating interference from non-OSA-related abnormal periods at the OSA event classification level. Furthermore, it constructs reliable OSA event judgment criteria using multi-dimensional physiological characteristics such as blood oxygen saturation during blood oxygen consumption periods, blood oxygen saturation during apnea periods, average amplitude of respiratory effort waves, and heart rate differences between awakening and apnea periods. Since micro-awakening caused by the first-night effect lacks a corresponding pathological chain of "apnea-blood oxygen change-heart rate adjustment," its related data is precisely filtered out, effectively avoiding interference from pseudo-data on the AHI index and other core monitoring indicators. This ensures that all identified OSA events are genuine pathological events, significantly improving the accuracy of OSA event identification. Furthermore, by employing a dual verification logic of accurately identifying genuine OSA events and screening credible OSA events, the reliability of the diagnostic basis is ensured from the data source: PSG data based on credible OSA events can accurately calculate core diagnostic indicators such as the AHI index, truly reflecting the severity of the patient's OSA condition; simultaneously, the joint judgment of multi-dimensional physiological characteristics can cover the pathological features of different types of OSA events, avoiding the limitations of single-indicator judgment and further improving the comprehensiveness and reliability of the diagnosis. Moreover, the personalized early warning model built based on credible OSA event data can accurately match the physiological characteristics of the target user with the OSA incidence pattern, improving the targeting of the early warning; by dynamically matching real-time collected PSG data with the early warning model, timely warnings can be issued when patients show signs of OSA events or when OSA events occur, facilitating rapid intervention by medical staff or patients' families and reducing the duration of OSA events and the resulting harm to the patient's body.

[0065] Example 2:

[0066] Corresponding to the sleep apnea real-time monitoring and early warning system provided in the above embodiments, based on the same technical concept, this invention also provides a sleep apnea real-time monitoring and early warning system. Figure 2 This is a schematic diagram of another real-time sleep apnea monitoring and early warning system provided in one embodiment of the present invention, as shown below. Figure 2As shown. Real-time sleep apnea monitoring and early warning systems can vary significantly due to differences in configuration or performance. They may include one or more processors 201 and memory 202. Memory 202 stores computer programs that can run on processor 201. Processor 201 executes the programs stored in memory 202 to perform the following steps: acquiring PSG data of the target user from the sleep time to the current time; determining all OSA events from the sleep time to the current time based on the spectrograms of respiratory effort waves and airflow corresponding to each time window in the PSG data, where each OSA event includes wakefulness periods and apnea periods; determining the blood oxygen consumption period of the current OSA event; determining reliable OSA events based on the blood oxygen saturation during the blood oxygen consumption period of the OSA event, the average amplitude of blood oxygen saturation and respiratory effort waves during the apnea period of the OSA event, the heart rate during the wakefulness period of the OSA event, and the heart rate during the apnea period of the OSA event; constructing an early warning model for the target user using the PSG data of the reliable OSA events; and providing sleep apnea early warning to the target user using the real-time PSG data of the target user and the early warning model.

[0067] The memory 202 can be either temporary or persistent storage. The application stored in the memory 202 can include one or more modules (not shown in the figure), each module including a series of computer-executable instructions for the real-time monitoring and early warning system for sleep apnea.

[0068] Furthermore, the processor 201 can be configured to communicate with the memory 202 and execute a series of computer-executable instructions stored in the memory 202 on the real-time sleep apnea monitoring and early warning system. The real-time sleep apnea monitoring and early warning system may also include one or more power supplies 203, one or more wired or wireless network interfaces 204, one or more input / output interfaces 205, and one or more keyboards 206.

[0069] Specifically, in this embodiment, the real-time sleep apnea monitoring and early warning system includes a processor, a communication interface, a memory, and a communication bus; wherein, the processor, communication interface, and memory communicate with each other via the bus; the memory stores computer programs; the processor executes the programs stored in the memory to achieve the following: "Acquire PSG data of the target user from the sleep time to the current time; determine all OSA events from the sleep time to the current time based on the spectrum of respiratory effort waves and the spectrum of airflow corresponding to each time window in the PSG data, wherein the OSA events include wakefulness periods and apnea periods; determine the current O..." The steps of identifying OSA events include: determining the oxygen consumption period during the OSA event; identifying credible OSA events based on the oxygen saturation during the OSA event's oxygen consumption period, the average amplitude of the oxygen saturation and respiratory effort wave during the OSA event's apnea period, the heart rate during the OSA event's awakening period, and the heart rate during the apnea period; constructing an early warning model for the target user using the PSG data of the credible OSA events; and providing sleep apnea early warning for the target user using the real-time PSG data of the target user collected in real time and the early warning model. This method, while possessing the beneficial effects of the above-described embodiments, will not be repeated here to avoid repetition.

[0070] It should be noted that the real-time sleep apnea monitoring and early warning system provided in this embodiment of the invention is based on the same application concept as the real-time sleep apnea monitoring and early warning system provided in this embodiment of the invention. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned real-time sleep apnea monitoring and early warning system, and has the same or similar beneficial effects. Repeated parts will not be described again.

[0071] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0072] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0073] This invention also proposes a computer-readable storage medium storing one or more programs, which, when executed by a real-time sleep apnea monitoring and early warning system comprising multiple applications, cause the real-time sleep apnea monitoring and early warning system to perform. Figure 1 The methods disclosed in the embodiments shown achieve the functions and beneficial effects of the methods in the preceding method embodiments, and will not be repeated here.

[0074] The computer-readable storage media include read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A real-time monitoring and early warning system for sleep apnea, characterized in that, include: The acquisition module is used to acquire PSG data of the target user from the sleep time to the current time; The determination module is configured to determine all OSA events from the sleep time to the current time based on the spectrograms of respiratory effort waves and airflow in the windows corresponding to each time moment in the PSG data, wherein the OSA events include wakefulness periods and apnea periods; determine the degree of respiratory disturbance at each time moment based on the spectrograms of respiratory effort waves and airflow in the windows corresponding to each time moment in the PSG data, and determine the degree of airflow disturbance at each time moment based on the spectrograms of airflow in the windows corresponding to each time moment in the PSG data; and determine the probability of micro-arousal at each time moment based on the degree of respiratory disturbance and the degree of airflow disturbance. The process involves: determining a target time when the micro-awakening probability is greater than or equal to an awakening threshold; defining the window preceding the target time as an awakening period; determining the difference coefficients for each moment in the window of consecutive moments prior to the target time based on the amplitude and airflow of the respiratory effort wave; calculating the average of the difference coefficients for all moments in the window of consecutive moments prior to the target time; defining the moment when the difference coefficient is greater than the average as a pause moment; defining the interval between the nearest pause moment before the current pause moment and the current pause moment as the apnea period of the target time; defining the awakening period and the apnea period as an OSA event; and determining all OSA events from the sleep time to the current time. The determining module is also used to determine the blood oxygen consumption period of the current OSA event; The determining module is further configured to determine a credible OSA event based on the blood oxygen saturation during the blood oxygen consumption period of the OSA event, the blood oxygen saturation and the average amplitude of the respiratory effort wave during the apnea period of the OSA event, the heart rate during the awakening period of the OSA event, and the heart rate during the apnea period. The early warning module is used to construct an early warning model for the target user using PSG data of the trusted OSA event, and to provide sleep apnea early warning for the target user using real-time collected PSG data of the target user and the early warning model. The determining module is further configured to: The period between the first moment of the apnea period of the current OSA event and the last moment of the apnea period of the previous OSA event is defined as the blood oxygen consumption period of the current OSA event. Determine the first fitted slope of blood oxygen saturation during the blood oxygen consumption period of the current OSA event, the second fitted slope of blood oxygen saturation during the apnea period of the current OSA event, and the duration of the apnea period of the current OSA event; determine the maximum value of blood oxygen saturation during the blood oxygen consumption period of the current OSA event and the minimum value of blood oxygen saturation during the apnea period of the current OSA event; calculate the first ratio between the maximum value and the minimum value, and the second ratio between the first fitted slope and the second fitted slope; based on the first ratio, the second ratio, the duration, and the average amplitude of the respiratory effort wave during the apnea period of the current OSA event, determine the dissolved oxygen demand coefficient of the current OSA event; Calculate the third difference between the standard deviation of heart rate during the awakening phase and the standard deviation of heart rate during the apnea phase of the current OSA event; determine the third fitting slope of heart rate during the apnea phase of the current OSA event; and determine the respiratory interruption performance of the current OSA event based on the third difference, the third fitting slope, and the dissolved oxygen demand coefficient. From all the OSA events, select credible OSA events whose respiratory interruption performance is greater than the interruption threshold.

2. The real-time monitoring and early warning system for sleep apnea according to claim 1, characterized in that, The determining module is further configured to: Calculate the absolute value of the first difference between the amplitude of each frequency in the spectrum of the respiratory effort wave of the window corresponding to the current time and the amplitude of each frequency in the spectrum of the respiratory effort wave of the window before the current time. The degree of respiratory disturbance at each moment is determined based on the absolute value of each of the first differences; Calculate the absolute value of the second difference between the amplitude of each frequency in the spectrum of the airflow of the window corresponding to the current moment and the amplitude of each frequency in the spectrum of the airflow of the window before the current moment; The degree of airflow turbulence at each moment is determined based on the absolute value of each of the second differences.

3. The real-time monitoring and early warning system for sleep apnea according to claim 1, characterized in that, The early warning module is also used for: Construct a coordinate system for all PSG data in all trusted OSA events with the same dimension, where the horizontal axis of the coordinate system is time and the vertical axis is the value of the data in the same dimension; The least squares method is used to fit all data points in the coordinate system to obtain the early warning model of each dimension of the PSG data of the target user.

4. The real-time monitoring and early warning system for sleep apnea according to claim 1, characterized in that, The PSG data includes: heart rate, respiratory effort wave, airflow, and blood oxygen saturation.

Images