Method and apparatus for detecting sleep disorder events - Patents.com

Abstract

An apparatus and method for detecting a sleep disorder event. The apparatus may be configured to access one or more physiological signals generated by one or more sensors. The apparatus may be configured to detect a seed event indicative of a sleep disorder event from the one or more physiological signals. The apparatus may be configured to calculate features indicative of a pattern within a portion of the one or more physiological signals associated with the detected seed event. The apparatus may be configured to apply the calculated features indicative of the pattern of the seed event to a classifier. The classifier may be trained to calculate a goodness of fit of the calculated features to a learned recurring pattern of sleep disorder events. The apparatus may be configured to output an identification of a sleep disorder event(s) corresponding to the seed event based on the calculated goodness of fit determined by the classifier.

Description

[Technical field]

[0001] (CROSS REFERENCE TO RELATED APPLICATIONS) This application claims the benefit of European Patent Application No. 21173562.6, filed May 12, 2021, the disclosure of which is incorporated herein by reference.

[0002] The present technology generally relates to one or more of screening, detection, diagnosis, monitoring, treatment, prevention, and improvement of disorders such as respiratory related disorders.The present technology also relates to medical devices or apparatuses and their uses.In some implementations, the present technology also relates to screening, detection diagnosis, monitoring, treatment, prevention, and / or improvement of sleep disorder events such as sleep disordered breathing, including central or obstructive apnea. [Background technology]

[0003] (Therapy) A variety of therapies have been used to treat or ameliorate such diseases, and in some cases, otherwise healthy individuals may successfully utilize such treatments to prevent the onset of respiratory disorders.

[0004] Nasal continuous positive airway pressure (CPAP) therapy is used to treat obstructive sleep apnea (OSA). Its mechanism of action is that the CPAP acts as a pneumatic splint and can prevent the upper airway from closing by pushing the soft palate and tongue forward or backward against the posterior oropharyngeal wall.

[0005] High-flow therapy (HFT) does not rely on positive airway pressure in a sealed patient interface, but on delivering air at a therapeutic flow rate near the patient's airway inlet(s) through an unsealed patient interface that may be obviously open to the atmosphere. High-flow therapy is used to treat OSA, CSR, and COPD. Air delivered to the airway at a high flow rate relative to typical respiratory flows flushes the patient's anatomical dead space, reduces the amount of CO2 rebreathed, and improves the efficiency of gas exchange. High-flow therapy can be used in combination with respiratory pressure therapy.

[0006] Non-invasive ventilation (NIV) provides ventilatory support to a patient through the upper airway to assist the patient in breathing deeply and / or maintain adequate oxygen levels in the body by performing some or all of the respiratory functions. NIV is delivered through a non-invasive patient interface. NIV is used to treat CSR, OHS, COPD, NMD, and chest wall disorders.

[0007] (Diagnostic and treatment systems) These treatments may be provided by a treatment system or device. Such systems and devices may also be used to diagnose a condition without treating it.

[0008] The treatment system may include a respiratory therapy device (RT device), such as a respiratory flow therapy device, such as a high flow therapy device (HFT device), or a respiratory pressure therapy device (RPT device), an air circuit, a humidifier, and a patient interface.

[0009] (Patient Interface) A patient interface may be used to interface a respiratory device to its user, for example by providing a flow of breathable gas. The flow of breathable gas may be provided to the user's nose and / or mouth via a mask, to the mouth via a tube, or to the user's trachea via a tracheotomy tube. Depending on the treatment being applied, the patient interface may facilitate gas delivery at a pressure sufficiently different from ambient pressure, for example, about 10 cmH2O positive pressure relative to ambient pressure, by forming a seal with, for example, a portion of the patient's face, to effectively perform the treatment. For other forms of treatment, such as oxygen delivery, the patient interface may not include a sufficient seal to facilitate delivery of a gas supply to the airways at a positive pressure of about 10 cmH2O.

[0010] (Respiratory Pressure Therapy (RPT) Devices) Air pressure generators are known in a wide range of applications (e.g., industrial-scale ventilation systems). However, air pressure generators for medical applications have specific requirements that are not met by more generalized air pressure generators, such as the reliability, size, and weight requirements of medical devices. In addition, even devices designed for medical use may have shortcomings, including one or more of comfort, noise, ease of use, effectiveness, size, weight, manufacturability, cost, and reliability.

[0011] One known RPT device used to treat sleep disordered breathing is the S9 Sleep Therapy System manufactured by ResMed, which has proven CPAP therapy. Another example of an RPT device is a ventilator. Ventilators, such as the ResMed Stellar® series of adult and pediatric ventilators, can provide invasive and non-invasive non-addictive ventilation therapy to a variety of patients to treat many conditions, including but not limited to NMD, OHS, and COPD.

[0012] The RPT device typically includes a pressure generator, such as a motor-driven blower or compressed gas reservoir, and is configured to supply pressurized air to the patient's airway. The outlet of the RPT device is connected via an air circuit to a patient interface, such as those described above.

[0013] There may still be a need for systems that not only monitor a user's health, but also help them understand symptoms, triggers, reactions to medications, and how to improve their own quality of life.

[0014] Additionally, it is desirable for the system to be able to access and process as much data as possible associated with the user's situation, and in this regard it is also desirable for the system to be able to monitor the patient as much as possible at all times of the day and night, and not just when the patient is in bed.

[0015] (Screening, diagnostic and surveillance systems) Screening and diagnosis generally refer to the identification of a disease from its signs and symptoms. Screening usually produces true or false results, such as whether a patient's impairment is severe enough to warrant further investigation, while diagnosis may provide clinically actionable information. Screening and diagnosis are often one-time processes, whereas monitoring the progression of a disease can continue indefinitely. Some screening / diagnostic systems apply only to screening / diagnosis, while others also apply to surveillance.

[0016] Polysomnography (PSG) is a conventional system for diagnosing / monitoring cardiopulmonary disorders, and application of the system usually involves specialized clinical staff. PSG typically involves placing 15-20 contact sensors on a person to record various biosignals, such as electroencephalography (EEG), electrocardiography (ECG), electrooculography (EOG), and electromyography (EMG). PSG for sleep-disordered breathing involves observing the patient for two nights in a clinic, with one night of pure diagnosis and the second night of titration of treatment parameters by a clinician. A clinical expert may be able to adequately diagnose or monitor the patient based on visual observation of the PSG signal. However, in some cases, a clinical expert may not be available or may be unable to afford it. Therefore, PSG is expensive and has low convenience. It is especially unsuitable for home diagnosis / monitoring.

[0017] In the field of sleep analysis, scoring or detection of sleep disorder events such as obstructive apneas, central apneas, obstructive hypopneas, central hypopneas, and respiratory effort-related arousals (RERAs) is usually achieved by such polysomnography, where various bodily functions are monitored, such as brain monitoring by electroencephalography (EEG), eye movement monitoring by electrooculography (EOG), muscle or skeletal muscle activity monitoring by electromyography (EMG), heart rate monitoring by electrocardiography (ECG), and respiratory airflow monitoring. Despite being considered the gold standard, the technology has several drawbacks. First, the test itself may require setup by medical professionals in inpatient care or in a home environment. Second, interpretation of the test is often not fully automated and requires the sleep technician to manually analyze the recorded signals, resulting in inter-scorer variability. Third, the test may disturb a person's sleep due to complex wiring and overall overhead, which may affect clinical parameters such as supine sleep time, sleep onset, and awakening during sleep. Fourth, sleep disorders such as sleep apnea are known to have large nocturnal variations, and current diagnostic systems are not suitable for multi-night studies due to clinical shortcomings, lack of convenience, and high costs per test.

[0018] Other convenient home screening / diagnostic / monitoring systems include a nasal cannula, a pressure sensor, a processing device, and a recording device. The nasal cannula is a device that includes two hollow, open prongs configured for small, non-invasive insertion into the patient's nares to impede the patient's breathing as little as possible. The hollow prongs are in fluid communication with the pressure sensor via a Y-tube. The pressure sensor provides a data signal representative of the pressure at the patient's nostril entrance (nasal pressure). The nasal pressure signal has been shown to be a sufficient substitute for the nasal flow signal generated by a sealed nasal mask and an in-line flow transducer in that the shape of the nasal pressure signal is comparable to the nasal flow signal. The processing device may be configured to analyze the nasal pressure signal from the pressure sensor in real-time or near real-time to detect and classify SDB events to monitor the patient's condition. Similar analysis may be required for screening or diagnosis, but not necessarily in real-time or near real-time. Thus, the recording means is configured to record the nasal pressure signal from the pressure sensor for later offline or "batch" analysis by the processing device for filter / diagnostic purposes.

[0019] Recently, scoring of sleep events has also been performed using another technique based on measurements by photoplethysmography (PPG). PPG measurements at one or more wavelengths make it possible to monitor changes in peripheral arterial tone, indicative of peripheral vasodilation or vasoconstriction, to derive oxygen saturation (SpO2) and to derive the pulse rate. For example, an advantageous embodiment of a system that makes it possible to derive the occurrence of sleep events from changes in peripheral arterial tone (PAT) is disclosed in EP 3 593 707 A1. This system and the measurement technique used in it allow the detection of sleep disorder events with minimal impact or disturbance on the sleeping subject, which is a major advantage compared to polysomnography.

[0020] In both polysomnography and PAT-based approaches, the scoring of sleep disorder events is generally based on the simultaneous occurrence of many physiological phenomena. In polysomnography, when one or more phenomena appear simultaneously, for example, reduced or absent airflow, awakening detectable by EEG, oxygen desaturation detected by a drop in SpO2 graph, and then returning to baseline, the event can be qualified as a breathing-related sleep disorder event, also called a respiratory event. For each of these phenomena, a threshold value is determined, for example, by the American Academy of Sleep Medicine (AASM), and measurements above or below this threshold value can be qualified as an indication of a respiratory event. In PAT-based methods, the simultaneous occurrence of oxygen desaturation, heart rate increase, and / or PAT increase is considered an indicator of a respiratory event.

[0021] A problem associated with current approaches is that some sleep disorder events may be missed in the scoring process, for example if one of the measurements does not reach a predefined threshold for that feature. Similarly, the scoring process may result in false positive detection of sleep disorder events. Furthermore, home sleep testing devices, such as PAT-based devices and cardiopulmonary polygraphy devices, are at an implicit disadvantage in detecting sleep disorder events because they do not specifically utilize gold standard signals from polygraphy (e.g., airflow or EEG). This is because they must infer respiratory events from surrogate signal patterns.

[0022] Improvements are needed to provide cost-effective or more readily available screening or monitoring tools that can help raise awareness of the condition of SDB and / or promote treatment for SDB. Realizing such reliable and cost-effective screening devices, such as home-use screening devices, remains a significant technical challenge. Thus, the objective of some versions of the present technology is to solve or at least mitigate one or more of the problems mentioned above. Thus, the present invention aims to provide improved methods and devices for detecting sleep disorder events with relatively high accuracy. Summary of the Invention [Problem to be solved by the invention]

[0023] The present technology aims to provide medical devices having one or more of improved comfort, cost, effectiveness, ease of use, and manufacturability for use in screening, monitoring, diagnosing, ameliorating, treating, or preventing respiratory diseases. [Means for solving the problem]

[0024] Some implementations of the present technology may include a processor-implemented method for detecting a sleep disorder event. The method may include accessing a plurality of physiological signals generated by one or more sensors. The method may include detecting a seed event from the plurality of physiological signals that is indicative of a sleep disorder event. The method may include calculating a feature indicative of a pattern within a portion of the plurality of physiological signals associated with the detected seed event. The method may include applying the calculated feature indicative of the pattern of the seed event to a classifier. The classifier may be trained, for example, with a machine learning classification algorithm, to calculate how well the calculated feature fits a learned pattern of sleep disorder events. The method may include outputting an identification of one or more sleep disorder events corresponding to the seed event based on the calculated goodness of fit determined by the classifier.

[0025] In some implementations, the classifier may be one or more of a machine learning classifier, a decision tree model, a machine learning classifier model, a logistic regression classifier model, a neural network, a naive Bayes classifier model, and a support vector machine. The physiological signals may include a peripheral arterial tone (PAT) signal and one or more of an oxygen saturation signal, a pulse rate signal, a respiratory effort signal, a motion signal, and an airflow signal (e.g., a flow signal). In some implementations, each seed event of the seed events may include one or more of a decrease in amplitude from baseline in the peripheral arterial tone (PAT) signal, a desaturation of oxygen in the oxygen saturation (SpO2) signal, an increase in amplitude from baseline in the pulse rate (PR) signal, a change in amplitude from baseline (e.g., an increase or decrease) in the respiratory effort signal, and a change in amplitude from baseline (e.g., an increase or decrease) in the airflow signal. The pattern may include a morphological pattern. The pattern may include a temporal pattern.

[0026] In some implementations, the features of the calculated features may include one or more of the following: duration of the seed event, duration of the seed event, intensity of the seed event, derived gradient or slope of the seed event, morphological asymmetry derived from the varying slope around the seed event, depth of the seed event, variance of the signal amplitude of the seed event, mean value of the signal amplitude of the seed event, degree of skew of the seed event, characterization of the morphological shape of the seed event. The one or more features of the calculated features may include one or more of a determined start point of the seed event, a determined end point of the seed event, a determined point of maximum or minimum intensity, and a determined feature point of the seed event. The detected seed event may include a first seed event of a first signal of the plurality of physiological signals and a second seed event of a second signal of the plurality of physiological signals. The second signal may be a physiological signal different from the first signal, and the one or more features of the calculated features may characterize the first seed event and the second seed event. For example, one or more of the calculated features may characterize the first seed event in relation to the second seed event. Such characterization may be used to correlate one seed event from one physiological signal with another seed event from a different physiological signal, such as, for example, comparing one seed event from one physiological signal with another seed event from a different physiological signal. The one or more features characterizing the calculated features of the first seed event in relation to the second seed event include (a) the amount of time that the desaturation nadir follows or precedes the peak pulse rate increase and / or PAT signal amplitude decrease, and / or (b) the time difference between the detected peak of the pulse rate (PR) surge and the decreasing trough of the PAT signal.

[0027] The detected seed event may include a third seed event of the first signal, and one or more of the calculated features may characterize the first seed event and the third seed event. For example, one or more of the calculated features may characterize the first seed event in relation to the third seed event. Such characterization may be used, for example, to associate the seed event with the physiological signal. Optionally, the first seed event and the third seed event may be a pair of adjacent seed events. In some implementations, the one or more features characterizing the calculated features of the first seed event and the third seed event may include (a) a duration between the first seed event and the third seed event, (b) a stability of a calculation of a period between the first seed event and the third seed event, and (c) a stability of a calculation of the first seed event and the third seed event. The calculation of the duration may include detecting a feature point in each of the first seed event and the third seed event and determining the duration based on an interval associated with the detected feature point. The detected feature point may include one or more of a local amplitude minimum and a local amplitude maximum. The calculated stability may be derived from multiple seed events and may include one or more of depth, mean, and variance, and / or may include stability, which may be, for example, event depth.

[0028] In some implementations, the detected seed events may include a fourth seed event of the second signal, and one or more of the calculated features may characterize (a) the first and third seed events of the first signal, and (b) the second and fourth seed events of the second signal. For example, one or more of the calculated features may characterize (a) the first and third seed events of the first signal in relation to (b) the second and fourth seed events of the second signal. Such characterization may be useful for correlating seed events of one physiological signal with seed events of another, different physiological signal, e.g., detecting overlapping sequences of seed events from the different physiological signals. For example, the one or more features characterizing the calculated features of (a) the first seed event and the third seed event of the first signal, and (b) the second seed event and the fourth seed event of the second signal may include a temporal correspondence of (a) a pulse rate (PR) peak of the detected seed event of the PR signal, and (b) a decrease in peripheral arterial tone (PAT) to a minimum point of the valley of the seed event of the PAT signal.

[0029] In some implementations, the method may further include generating an output of the identification information as feedback in response to a user input on a user interface for a selection of at least one seed event detected by the detection implemented by the one or more processors. The method may further include generating a signal for controlling operation of a respiratory therapy device based on the output or application. The generating may include transmitting the identification information of the one or more sleep disorder events to a remote computing system or server. The generating may include transmitting the signal to the respiratory therapy device via a network communication link.

[0030] Some implementations of the present technology may include a controller that may include at least one processor and at least one memory that includes processor control instructions. The at least one memory and the processor control instructions may be configured to cause the controller to execute a method, including any one or more aspects of the methods described herein, utilizing the at least one processor.

[0031] Some implementations of the present technology may include an apparatus for detecting a sleep disorder event. The apparatus may include one or more sensors. The apparatus may include a controller. The controller may include one or more processors and at least one memory including processor control instructions. The controller may be configured to access a plurality of physiological signals generated by the one or more sensors. The controller may be configured to detect a seed event indicative of a sleep disorder event from the plurality of physiological signals. The controller may be configured to calculate features indicative of a pattern within a portion of the plurality of physiological signals associated with the detected seed event. The controller may be configured to apply the calculated features indicative of the pattern of the seed event to a classifier. The classifier may be trained to calculate a fit of the calculated features to a learned pattern of sleep disorder events. The controller may be configured to output an identification of one or more sleep disorder events corresponding to the seed event based on the calculated fit determined by the classifier.

[0032] Some implementations of the technology may include a processor-readable storage medium that includes processor-executable instructions for performing methods, including any one or more aspects of the methods described herein, when executed by one or more processors.

[0033] Some implementations of the present technology may include a processor-readable medium having stored thereon processor-executable instructions that, when executed by one or more processors, cause the one or more processors to detect a sleep disorder event. The processor-executable instructions may include instructions for accessing a plurality of physiological signals generated by one or more sensors. The processor-executable instructions may include instructions for detecting a seed event indicative of a sleep disorder event from the plurality of physiological signals. The processor-executable instructions may include instructions for computing features indicative of a pattern within a portion of the plurality of physiological signals associated with the detected seed event. The processor-executable instructions may include instructions for applying the computed features indicative of the pattern of the seed event to a classifier, the classifier being trained to compute a fit of the computed features to a learned pattern of sleep disorder events. The processor-executable instructions may include instructions for outputting an identification of one or more sleep disorder events corresponding to the seed event based on the computed fit determined by the classifier.

[0034] In some implementations, the processor-executable instructions may further include instructions for generating a signal to control operation of a respiratory therapy device based on the output or application. The control operation may include controlling pressure or flow therapy of a blower of the respiratory therapy device.

[0035] Some implementations of the technology may include a server that can access any of the processor-readable media described herein. The server may be configured to receive requests to download the processor-executable instructions of the processor-readable media over a network to a processing device.

[0036] Some implementations of the present technology may include a processing device. The processing device may include one or more processors and (a) a processor-readable medium as described herein, or (b) may be configured to access the processor-executable instructions using a server as described herein. The processing device may be a respiratory therapy device. The processing device may be configured to generate a pressure therapy or a flow therapy.

[0037] Some implementations of the present technology may include a method of a server accessing any of the processor-readable media described herein. The server method may include receiving, via a network, a request at the server to download processor-executable instructions of the processor-readable media to an electronic processing device. The server method may include transmitting the processor-executable instructions to the electronic processing device in response to the request.

[0038] Some implementations of the present technology may include a method in one or more processors for detecting sleep disorder breathing events. The method may include accessing, with one or more processors, any processor-readable medium described herein. The method may include executing, in one or more processors, processor-executable instructions of the processor-readable medium.

[0039] Of course, some of the above aspects may form sub-aspects of the technology, and various of the sub-aspects and / or aspects may be combined in various ways to form further aspects or sub-aspects of the technology.

[0040] Other features of the technology will become apparent from consideration of the information contained in the following detailed description, abstract, drawings, and claims. [Brief description of the drawings]

[0041] The present technology is illustrated by way of example, and not by way of limitation, in the accompanying drawings in which like reference numerals represent similar elements and in which: [Figure 1] 1 shows a first schematic graph illustrating multiple physiological signals over time. [Diagram 2] 1 shows a second schematic graph illustrating multiple physiological signals over time. [Diagram 3] 2 shows a second schematic graph illustrating one implementation of the method according to the present technology. [Figure 4] 13 shows a third schematic graph illustrating multiple physiological signals over time. [Diagram 5] 13 shows a fourth schematic graph illustrating multiple physiological signals over time. [Figure 6A] 1 shows an example of a finger sensor device on a patient suitable for implementation of the present technology. [Figure 6B] 6A illustrates an example computing system or processing device suitable for performing various steps according to an example implementation of the present technology, such as the sensor apparatus of FIG. 6A. [Figure 7] For example, an example of processing of one or more processors in a system of the present technology having a method for detecting a sleep disorder event of the present technology. [Figure 8] 13 is another exemplary process of one or more processors of the method for detecting a sleep disorder event of the present technology, for example, in the system of the present technology. [Figure 8A] FIG. 1 shows an exemplary system according to the present technology. A patient 1000 wearing a patient interface 3000 is supplied with compressed air from an RPT device 4000. Air from the RPT device 4000 is humidified in a humidifier 5000 and sent to the patient 1000 through an air circuit 4170. A bed companion 1100 is also shown. [Figure 8B] The RPT device 4000 is shown in use on a patient 1000 wearing a nasal mask 3000. [Figure 8C] The RPT device 4000 is shown in use on a patient 1000 wearing a full face mask 3000. [Figure 9]An example of a non-invasive patient interface 3000 in the form of a nasal mask is shown. [Figure 10A] 4 shows an RPT device 4000 in accordance with one form of the present technology. [Figure 10B] FIG. 40 shows a schematic diagram of a pneumatic circuit for an RPT device 4000 in accordance with one form of the present technology, showing upstream and downstream directions. [Figure 10C] A schematic diagram of electrical components of an RPT device 4000 according to one aspect of the present technology is shown. [Figure 10D] 10D shows a schematic diagram of an algorithm 4300 implemented in an RPT device 4000 according to one aspect of the present technology. In FIG. 10D, the solid arrows indicate the actual flow of information, for example via electronic signals. [Figure 10E] 10D, in accordance with one aspect of the present technology. [Figure 11] A humidifier 5000 is shown. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0042] Before describing the present technology in further detail, it is to be understood that the present technology is not limited to the particular examples described herein, as such may vary, and it is also to be understood that the terminology used in this disclosure is for the purpose of describing only the particular examples discussed herein, and is not intended to be limiting.

[0043] The following description is provided in conjunction with various examples that may share one or more common characteristics and / or features. It should be understood that one or more features of any one example may be combined with one or more features of another example or other examples. In addition, any single feature or combination of features in any example may constitute an additional example.

[0044] In an exemplary implementation, a processor- or computer-implemented process detects a sleep disorder event. The method may include obtaining at least one physiological signal that may include a pointer to a sleep disorder event as known to those skilled in the art, identifying seed events from the at least one physiological signal that are indicative of the sleep disorder event, determining a pattern among at least some of the seed events, determining a fit of the seed event to the pattern, and detecting the sleep disorder event based on the determined fit. Various aspects of such a method may be considered in conjunction with the graphs of FIGS. 1-5.

[0045] For example, FIG. 1 shows a first schematic graph depicting an example of multiple physiological signals over time, any one or more of which may be utilized in the present technology. In this graph, the physiological signals are an oxygen saturation (SpO2) signal 1, a PAT signal 2, a pulse rate signal 3, and a limb activity signal 4. The PAT signal or PAT measurement is a measurement of changes in pulsatile arterial blood volume reflecting changes in arterial tension, and may be measured, for example, by attaching an air pressure or optical sensor to the patient's fingertip. The pulse rate signal may be obtained in many different ways known to those skilled in the art. The oxygen saturation (SpO2) may be measured, for example, by a pulse oximeter attached to the patient's finger, nostril, wrist, or earlobe. The oxygen saturation, PAT, and pulse rate signals 1, 2, 3 may be derived, for example, by measurements with a single device, such as, for example, by measuring arterial pulsatile volume changes at the fingertip, such as the devices disclosed in EP3,593,707, or U.S. Patent Application Publication No. 2020 / 0015737, the entire disclosure of which is incorporated herein by reference. Such a device may be configured to perform photoplethysmography (PPG), preferably dual wavelength photoplethysmography, from which PAT, SpO2, and pulse signals can be derived. FIG. 6A illustrates an example of such a finger sensor apparatus. However, measurements may also be made by separate devices, such as any of the devices described herein.

[0046] The activity signal 4 may be derived, for example, from a motion sensor such as an accelerometer or a non-contact sensor (e.g., an RF motion sensor or a sonar-based sensor) connected to the patient's arm or leg. For example, a radio frequency (RF) transceiver of an RF sensor may be used as the non-contact portion of the sensing device. Such an RF sensing device may be integrated or coupled with the processing device described herein and may be implemented using any of the techniques and sensor components described in U.S. Patent Application Publication No. 2014 / 0024917, filed July 18, 2013, entitled "Range-Gated Radio Frequency Physiological Sensor," International Patent Application No. PCT / EP2017 / 070773, filed August 16, 2017, entitled "Digital Radio Frequency Motion Detection Sensor," and U.S. Patent Application Publication No. 2018 / 0239014, filed August 16, 2016, entitled "Digital Range-Gated Radio Frequency Sensor," each of which is incorporated herein by reference. The sonar-based sensor may be, for example, a processing device implemented by a microphone and a speaker, such as any of the processing devices described in U.S. Patent Application Publication No. 2021 / 0275056, filed September 19, 2017, or U.S. Patent Application Publication No. 2022 / 0007965, filed November 19, 2019, the entire disclosures of which are incorporated herein by reference. Limb movements may be obtained, for example, from an accelerometer connected to the patient's arm or leg. For sleep analysis, a device configured to connect to the fingertip or wrist may be better than a device connected to the head or chest, as the latter device may disrupt natural sleep and affect the results. Other sensors and physiological signals may also be implemented.

[0047] The selection of a given combination of physiological signals may be determined, for example, by the type of sleep disorder event that needs to be identified or by the potential impact that the acquisition of the signals has on the human sleep. As the skilled person will appreciate, many more combinations of physiological signals are possible. For example, both respiratory effort and airflow signals can be realized. Respiratory effort can be derived, inter alia, from an accelerometer, a bioimpedance measurement, or an inductance plethysmography belt. Airflow can be derived from measurements by pressure, flow, or air temperature sensors attached to the patient's nares, also known to the skilled person. In general, such signals are obtained within a common time range. Such a common time range does not mean that the measurements of the multiple physiological signals must all have the same duration, or that they all start and stop at the same time. By "common time range" is meant that the physiological signals must include the same or identical time range for the multiple signals, and must at least partially overlap in time, such that the common time range includes the seed event.

[0048] In conventional methods, the magnitude of the SpO2 drop observed in the saturation of the oxygen signal 1 is primarily compared to thresholds set, for example, by the AASM, to determine whether the SpO2 drop can be evaluated as a sleep disorder event. A similar procedure is used for the PAT amplitude drop in the PAT signal 2, or the pulse rate and activity spikes in the respective signals 3 and 4. If all or more than one of these four indicators point to a potential sleep disorder event, the conventional approach concludes that a single apnea or sleep disorder event has been detected. In these prior art methods for detecting sleep disorder events, the potential indicators of a sleep disorder event in the physiological signal are primarily processed individually, i.e., compared individually to thresholds, without considering contextual information from the rest of the physiological signal. In this regard, the prior art algorithms may threshold the indicators in combination to be more or less consistent, such that the simultaneous occurrence 5 of SpO2 below a predefined threshold, PAT amplitude below a predefined threshold, pulse rate spikes above a predefined threshold, and limb activity spikes above a predefined threshold are observed, as shown in FIG. 1.

[0049] Unlike conventional methods, implementations of the present technology consider the larger context of such signals, such as pattern identification to identify events within a given signal and / or across multiple signals. For example, a second schematic graph is shown representing multiple physiological signals over a common time range. The physiological signals include an oxygen saturation (SpO2) signal 1, a PAT signal 2, a pulse rate signal 3, and a limb activity signal 4, as in the first graph. In one method, e.g., a method implemented by one or more processors, the present technology can evaluate these or any of the aforementioned physiological signals, e.g., SpO2 signal 1 alone, or any combination of at least two or more of these signals to detect a sleep disorder event. According to this method, a seed event 6 is selected that is indicative of a sleep disorder event. The seed event 6 can be, for example, detection of one or more of an oxygen desaturation 7, a decrease from baseline in PAT signal amplitude 8, an increase from baseline in pulse signal amplitude 9, a decrease from baseline in respiratory effort signal amplitude 10, an increase in limb activity 10, or any other known physiological signal that may be an indicator of a sleep disorder event. Seed event 6 suggests only sleep disorder events, but does not necessarily point to sleep disorder events as they may have another underlying cause.

[0050] Thus, the seed events can be, for example, at least one of a decrease in PAT signal amplitude from baseline, an oxygen saturation decrease, an increase in pulse rate signal amplitude from baseline, a decrease in respiratory effort signal amplitude from baseline, a decrease in airflow signal amplitude from baseline, and an increase in limb activity from baseline. These seed events are indicative of a sleep disorder event, meaning that they may be signs or symptoms of a sleep disorder event, but are not necessarily associated with a sleep disorder event. In other words, a seed event is an event that may be, but is not necessarily, assessed as being associated with a sleep disorder event. When other physiological signals are acquired, the seed events may be defined differently depending on the signals.

[0051] Optionally, a seed event from one physiological signal may be typically associated with similarly timed seed events from the other physiological signal(s), and a set of associated seed events (e.g., one from each signal) within a common time frame of multiple physiological signals may be considered indicative of a candidate respiratory event. An example of a set of associated seed events of a common time scale may be a desaturating seed event of a PPG signal that coincides in time with an amplitude decrease of a PAT signal, an increase of a pulse rate signal, and an increase in activity counts of a motion signal.

[0052] It has been observed that sleep disorder events, which may be breathing-related sleep disorder events such as apnea events, snoring, RERA sickness, or non-breathing-related sleep disorder events such as spontaneous autonomic arousals, generally occur in relatively regular intervals grouped together or consecutively. This context information can be utilized by the present technology. Since the scoring is less dependent on the simple pre-defined thresholds of the prior art scoring, the accuracy of sleep disorder event detection can be improved by limiting events to sleep disorder events based at least in part on their fit with patterns identified in the data. In the prior art, sleep disorder events are typically analyzed based on single events or temporal separation, determining whether two or more events of a single measurement or one or more physiological signals that occur simultaneously are higher or lower than a pre-defined threshold to identify the event as a sleep disorder event without looking at the context of the event. As a result of such single event methods, a sleep disorder event may be missed, for example, if at least one seed event associated with the sleep disorder event is just below or above such threshold. Unlike the prior art, the present method does not simply compare each seed event with a predefined threshold value when determining whether a seed event is associated with a sleep disorder. The present method can also consider the entire data set to determine patterns in the measurement data, and then incorporate the patterns into identifying whether a seed event is associated with a sleep disorder event, which allows the present technology to achieve higher accuracy in detecting sleep disorder events than the prior art methods.

[0053] Thus, in the next step of the process, unlike the prior art methods, a pattern, e.g. a repeating pattern or a substantially regular pattern, is determined within at least a portion(s) of each signal that includes the seed event(s) 6 selected in the physiological signal. That is, a particular portion (e.g., a sample) representative of the signal in the vicinity of each seed event (e.g., a sample of the signal representing a desaturation period from its start to its end) is further evaluated, e.g., by determining or calculating one or more features having the particular portion to evaluate whether one or more patterns are present and / or are shared by multiple seed events. These patterns can be related to morphological patterns, i.e., shapes of the representation of the physiological signal, and can be associated with the amplitude of the signal corresponding to the intensity of the seed event, or the slope, skew, or morphological asymmetry of the gradient change (e.g., increase or decrease) in the vicinity of the seed event in the signal.

[0054] For example, this seed-based determination of morphological features of the SpO2 signal can include, for example, the length of desaturation, the depth of desaturation, and the slope asymmetry between the desaturation and resaturation phases. Seed-based features of signal morphology can also be determined or calculated from seed event-related portions of other physiological signal(s). Thus, determining the morphological pattern can include determining features that provide characteristics of the seed event morphology using signal processing and characteristic signal data samples and / or statistical functions that provide a numerical value(s) that serve as a characteristic of the morphology (e.g., signal temporal changes) in the vicinity of each seed event. Additionally and / or alternatively, the process of determining or calculating the pattern can optionally include calculating or determining points of maximum or minimum intensity and / or determining any other characteristic points of the seed events that can be used to characterize the feature(s) of the seed events.

[0055] Thereby, the pattern may include a morphological pattern. The morphological pattern may be associated with, for example, typical shape features of the representation of the seed events in the acquired signal. For example, SpO2 drops may show a relatively slow drop and a relatively rapid regular rise. Determining such a pattern may include, for example, at least one of determining the duration of the seed events, determining the intensity of the seed events, determining the morphological asymmetry of the seed events, and / or characterizing the morphological shape of the seed events. The template morphology may be derived, for example, from a plurality of seed events that are adjacent in time or spaced apart in time. The shape features of the representation of the seed events may then be compared to the template. Determining the pattern may also include determining features of the pattern that may characterize the pattern, such as, for example, shape features that may characterize the shape of the pattern, such as the average morphological asymmetry or the average size of the seed events in a substantially regular pattern.

[0056] The patterns may also include temporal patterns, e.g., as described with reference to Figure 3 or Figure 4. Such time-dependent patterns are particularly well identified in, e.g., the SpO2 signal 1 and the PAT signal 2, but are also present in other physiological signals. Thus, the process may further include a seed-based determination of temporal features representative of particular portions (e.g., samples) of the signal in the vicinity of each seed event, e.g., using signal processing and characteristic signal data samples and / or statistical functions that provide a numerical value(s) characterizing the temporal pattern of features in the vicinity of each seed event.

[0057] Thus, the pattern may include a temporal pattern, such as repeating a seed event at a predefined interval. Thus, determining such a pattern may include at least one of determining a start point of the seed event, determining an end point of the seed event, determining a most intense or least intense point, and / or determining any other characteristic point of the seed event. In this way, a temporal pattern may be identified in the repetition of the seed event, for example in the time difference between the start points of the seed events or between the end points of the seed events. The temporal distance or time difference between local maxima or minima in the acquired signal may also indicate a substantially regular temporal pattern. Derivation of the pattern may also include determining features including, for example, the interquartile range of the (normalized) difference between adjacent seed events. Determining the pattern may also include determining pattern features, for example temporal features, that may characterize the pattern, for example determining a number of events in the repeating pattern. Determining the pattern may also include determining the quality of the substantially regular pattern, for example by determining the periodic stability or variability between the characteristic points of the events represented by the substantially regular pattern, using the standard deviation or any other known parameter. Such temporal pattern features may be calculated or determined to provide temporal ordering information of or between seed events, such as adjacent events of the same signal. Further, such temporal pattern(s) may be calculated or determined to provide temporal alignment information across or regarding seed events of different physiological signals.

[0058] For example, for a set of seed events (e.g., one per signal) in a common time frame, the process can determine a relative juxtaposed temporal alignment to characterize the signal features of the seed events from multiple different physiological signals. For example, the process can determine how long (e.g., in seconds) the desaturation nadir of the SpO2 signal is after or before the peak of the increase in the pulse rate (PR) signal and the decrease in amplitude of the PAT signal. The degree of temporal alignment between the pulse rate (PR) surge peak and the PAT decrease trough is also an important temporal feature, as perfect alignment is an important indicator of sympathetic activation, an autonomic arousal that typically occurs near the end of a respiratory event.

[0059] As another example, two seed events from different signals, preceded by a gradual drop in peripheral oxygen saturation followed by a minimum in oxygen saturation and a sudden rise in oxygen saturation, can be considered to be a time-aligned pattern when, for example, the temporal characteristics of the signal features including the highest point of the pulse rate spike of the seed event match the temporal characteristics of the minimum PAT signal amplitude of the other seed event. The time difference between the onset of the seed events in the different signals turns out to be relatively constant. The order between the seed events may vary from person to person but may remain relatively constant for a given patient. Other features may intervene in the grouping of seed events from multiple physiological signals into a set, for example, substantial temporal alignment of local minima or maxima of the physiological signals, in particular the local minima of the oxygen desaturation and / or the local minima or maxima of the pulse rate preceding the PAT signal. Another example of a pattern in the co-occurrence of seed events may be the detection of a local minimum in the PAT signal and / or the pulse rate towards the end of a flow reduction event.

[0060] As a further example, a temporal (and / or morphological) pattern may be evaluated with respect to seed events of the same physiological signal. Thus, the process may determine ordering-related features, for example, with a pair or pairs of adjacent seed events. A second schematic graph of FIG. 2 is shown illustrating a further aspect of the process of the present technology. In this regard, a process of determining a repetitive or temporal pattern (e.g., a substantially regular temporal pattern such as a time ordering for a seed event of one signal to another seed event (e.g., an adjacent pair) of the same signal) may include determining features including a duration or time range 6d between subsequent seed events 6, for example, by determining feature points 6s such as local minima or maxima of the amplitude of the seed event 6, and determining subsequent feature points 6e of subsequent seed events. This process may be performed in one or more physiological signals, respectively. Such processing may be useful when respiratory events and their associated seed event pairs tend to occur in order with respect to the repetitive seed events or repetition of their signal characteristics.

[0061] In addition, to further characterize each such sequence (e.g., a portion of a physiological signal beginning and ending in a neighborhood of a first seed event and a neighborhood of a second seed event, such as where the first and second seed events are adjacent seed events), the process can determine additional sequencing-related features for each sequence associated with multiple seed events (e.g., adjacent pairs of signals) that can be considered with respect to the aforementioned pattern evaluation. For example, the process can determine features based on the stability of the period of each adjacent seed event pair to characterize the period or period of each adjacent seed event pair. Such stability of the physiological signal between adjacent seed event pair(s) can be characterized by one or more statistical functions, such as, for example, variance. For example, calculated by determining the variance or other stability for the seed event features, such as the depth of each seed event. Similarly, the process can further characterize the seed events in the sequence by calculating other statistical functions, such as, for example, the average event morphology. In addition, the process can evaluate as a feature the average number of signal samples between the neighborhood of the seed event. Similarly, the process can characterize the stability of the seed events in the sequence, for example, by calculating the stability of the event morphology of the signal samples in the neighborhood of the seed event. A sequence of highly repetitive seed events with very similar morphology (i.e., with a highly stable period) is considered to be more likely to be associated with a sequence of respiratory events. Furthermore, a calculated feature of the pattern for the sequence may be the interquartile range of the (normalized) difference between adjacent seed events.

[0062] The evaluation of each sequence of such signals can be extended by determining or calculating or determining additional features (e.g., temporal features associated with a temporal pattern) that characterize the sequence in relation to other sequences (e.g., sequences associated with multiple seed events from other physiological signals, e.g., adjacent pairs or multiple adjacent pairs of seed events). This process provides information that characterizes the relationship of different physiological signal sequences (e.g., sequences that overlap in time). In this respect, synchronization between such sequences of different seed event signal types contains a wealth of information. For example, a PR sequence in which all PR peaks are perfectly aligned with the valleys of PAT decrease events in a PAT sequence indicates that all seed events in these sequences are likely to be representations of a (series of) respiratory events. These features can include the timing of the feature points of the seed events in the different signals (e.g., PR, PAT, and activity sequences) to confirm the temporal alignment. Similarly, the feature(s) can characterize the alignment delay of a sequence of seed events of a PPG signal relative to a sequence of other signals, since the desaturation nadir usually tracks the change timing of the other signals in a very systematic way.

[0063] For example, as indicated by the combining rectangle 11, the multiple seed events in the physiological signals 1, 2, 3, 4 shown in FIG. 3 are grouped or combined in the respective physiological signals 1, 2, 3, 4 based on similar event duration 6d. These seed events or combined sequences of seed events may indicate recurring sleep disorder events. Additionally and / or alternatively, for combining seed events 6 within a single physiological signal, the process further determines features that associate sequences of seed events from multiple different physiological signals into sets 12 of aligned / correlated sequences, where the seed events grouped in the sets 12 indicate the same sleep disorder event. In other words, these sets are sets of seed events of different physiological signals, where these sets include seed events spanning the same time range and may indicate the same sleep disorder event. To explain more visually, the sets 12 are "vertical" groups or associated sequences of seed events from different signals, whereas the combining of the aforementioned seed events is done "horizontally", which are associated seed events of a single signal. Grouping the seed events into sets may include determining a substantially regular pattern in the co-occurrence of the seed events of the set. For example, the median duration or onset time of the seed events may be similar in different physiological signals. In FIG. 3, such a set 12 of seed events is shown.

[0064] Thus, utilizing any combination of the aforementioned determined / calculated features associated with at least one seed event (e.g., seed events exhibiting the aforementioned pattern, aligned / correlated seed events from different signals, sequences of seed events from each signal, and / or temporal and / or morphological features from aligned / correlated sequences of seed events from different signals) may then be applied in combination (e.g., in feature vectors or other data structures of such features) to a processing algorithm in one or more digital processor(s), such as a classifier (e.g., a training model generated by a machine learning classifier, such as a rule-based model or decision tree) or other machine learning algorithm. A machine learning classifier, such as a classifier or a training model from such a classifier, may be configured to determine a goodness of fit, e.g., probability, between a particular seed event and pattern-related features (e.g., a data structure or a set of the aforementioned features, such as the aforementioned temporal and / or morphological features) associated with features of a known sleep disorder event-related pattern to determine whether the particular seed event is indicative of a sleep disorder event or not. In other words, based on the pattern, the classifier can determine that one or more sets of seed events may be associated with breathing-related sleep disorder events and another set may be associated with non-breathing-related sleep disorder events. The classifier(s), e.g., models thereof, may be any one or more of a decision tree classifier or model, a machine learning classifier or model, a logistic regression classifier or model, a neural network, a naive Bayes classifier or model, and a support vector machine, etc. Thus, the determined goodness of fit(s) may include, for example, a goodness of fit activation value(s) or an objective fit metric(s), and may include output calculation(s) of one or more activation functions when applied to the values ​​of the input features, and / or calculation of a probability value(s) determining the similarity of the feature(s) to the pattern, which may be compared to one or more thresholds to ascertain the discriminatory information(s) (e.g., respiratory events or non-respiratory events).

[0065] Optionally, multiple classifiers can be implemented. For example, a first classifier may be trained to perform an ordered correlation pattern evaluation of pre-described seed events from multiple physiological signals associated with the seed events, and a second classifier may be trained to detect seed events associated with breathing-related sleep disorder events or seed events associated with non-breathing-related sleep disorder events from a sequence of seed events. The distinction between breathing-related and non-breathing-related sleep disorder events can be based, for example, on certain features of the patterns and sequences themselves. For example, a periodic limb movement sequence of a seed event (non-breathing-related sleep disorder event) does not have oxygen desaturation and has a relatively short time between subsequent events. Alternatively, a single classifier can be trained to perform the grouping step and then the detection step. The classifier may also be trained to perform the detection step without grouping the seed events into sets. The classifier may be trained, for example, with a marker data set from a clinical trial. Alternatively, the classifier may be trained or developed as a rule-based system based on thresholds of any of the above features. The classifier can be any known classifier, such as a neural network, a decision tree, or a support vector machine.

[0066] In this way, a particular seed event is selected by the processing algorithm based on a determined fit with the aforementioned pattern (e.g., the aforementioned features), for example based on the substantially regular duration of the seed event in the physiological signal, and based on the significance of the fit as a detection of a sleep disorder event, and thus can detect a sleep disorder event with a relatively high degree of confidence to improve on prior detection techniques, such as detection techniques for home study devices that are sensitive based on PPG. The classifier can be trained or developed by designing and inputting a set of rules for detection. These rules can include, for example, rules based on one or more measurement thresholds. These rules can further include rules based on one or more decision values ​​associated with time and / or morphological patterns. For example, the classifier can use as input the features of the seed event, how well the seed event fits into a pattern (e.g., a substantially regular pattern), and information associated with the features of the pattern. In some implementations, the classifier can also be trained using tag data based on previous measurements. The training dataset can be built, for example, from empirical data of a clinical trial, where the locations of sleep disorder events can be scored or annotated by manual or computer-assisted scoring, and the seed events can be associated therewith. Alternatively or additionally, the training dataset may be constructed from empirical data of a clinical trial in which seed events have been marked as belonging or not belonging to sleep disorder events, for example by manual or computer-assisted scoring.

[0067] Such an improved process can be considered with reference to Figures 4 and 5. Figure 4 shows a third schematic graph representing multiple physiological signals over time. As in the previous figures, the flags represent multiple physiological signals over a common time range, including an oxygen saturation (SpO2) signal 1, a PAT signal 2, a pulse rate signal 3, and a limb activity signal 4. In the SpO2 signal 1, five desaturations of oxygen 7, 7a can be identified, which can be selected as five sub-events. In the prior art method, the amplitudes of these desaturations are determined and compared to a threshold to determine whether these events are related to a sleep disorder event. In the current graph, the first desaturation of oxygen 7a may be disqualified because it differs by only 2% from the previous baseline. However, in an exemplary process according to the present technology, features regarding, for example, a substantially regular pattern in the onset of desaturation and / or duration of seed events lead to a reliable detection of a sleep disorder event, such as an apnea, despite the relatively small oxygen desaturation values. Furthermore, the process of evaluating features for grouping seed events among different physiological signals into groups (e.g., sequences) of said seed events enhances the reliability of detection since there is a substantially regular pattern in the co-occurrence of the seed events in the group, in particular a decrease in PAT signal amplitude from baseline 8a, an increase in pulse rate signal amplitude from baseline 9a, and an increase in limb activity 10a.

[0068] FIG. 5 shows a fourth schematic graph representing multiple physiological signals 1, 2, 3, 4 over time. Similarly, in the SpO2 signal 1, three desaturations of oxygen 7b, 7c, 7d can be distinguished, which can be selected as three seed events potentially indicative of a sleep disorder event. In conventional methods, the magnitude of these desaturations is determined and compared to a threshold. Since all oxygen desaturation events 7c, 7d, 7b exceed the threshold, the method determines that a sleep disorder event has been detected. However, in the process of the present technology, features relating to a substantially regular pattern are to be evaluated in at least some of the seed events. This is difficult because pattern recognition is difficult even in the oxygen desaturation signal 1, and the overall shapes of the desaturation signals 7b, 7c, 7d are different (thus resulting in a low or only slight fit). The situation is still compelling, since in the other physiological signals 2, 3, and 4, there appears to be no substantially regular pattern, despite some drops in PAT amplitude, spikes in pulse rate, or spikes in limb activity. Thus, an exemplary process of the present technology indicates that a sleep disorder event may not be detected and that the desaturation may be artifactual and / or due to another reason other than a sleep disorder event.

[0069] Thus, in some implementations of such a process, the device may be configured to generate feedback regarding detection or non-detection of a sleep disorder event in response to an input to the system of selecting at least one seed event, e.g., when a selection is input to a user interface of such a system. In response, the process may output an indication that the seed event is not recognized as a cause of a sleep disorder event (e.g., a sleep disorder breathing event), e.g., a "pattern lacking basic regularity." The feedback may be presented, preferably, in natural language. For example, positive feedback may include "detection based on low desaturation but high match with regular pattern," which may be applied to the physiological signals described in FIG. 4 when applied to such a process.

[0070] In this regard, the method may include feedback on the detection or non-detection of a sleep disorder event by the selection of at least one seed event. The feedback may be presented in natural language. The feedback may include, for example, information on why a seed event is evaluated as being related to a sleep disorder event or not being related to a sleep disorder event. This information may be based on, for example, a decision tree classifier. Thus, the following additional examples may also be considered as inference examples of a decision tree classifier, where the feedback may be displayed on a screen, for example, as a coded sentence.

[0071] Example: A sleep disorder event associated with respiration may be detected, for example, and feedback may include information such as "Even though the oxygen desaturation associated with this event is less than 3%, it is part of a sequence of desaturation events with relatively large temporal and morphological similarity." Two adjacent oxygen desaturations are part of this sequence, averaging more than 3%. Six adjacent oxygen desaturations are part of this sequence, averaging more than 3.5%. The desaturations overlap significantly with a pulse rate spike and PAT decrease sequence with very similar periodicity.

[0072] Example: For example, a sleep disorder event associated with respiration may be detected and the feedback may include information such as "The desaturation associated with this event is part of a sequence of desaturation events with relatively large temporal and morphological similarity. Two adjacent desaturations are part of this sequence, averaging greater than 3%. Six adjacent desaturations are part of this sequence, averaging greater than 3.5%. Limb movement around this event is very low."

[0073] For example, if a seed event is not selected or retained as associated with a sleep disturbance event, the feedback may include information such as "even though 2 adjacent desaturations are greater than 3% on average and 6 adjacent desaturations are greater than 3.5% on average, there is too much limb movement around the location and the surrounding desaturations have low temporal and morphological similarity."

[0074] Of course, the language or expression of such feedback information need not be the same as in the present application. However, it will be clear to one skilled in the art that "temporal and morphological similarity" refers to the degree of fit of the pattern in the seed event. Other ways of presenting this information are possible.

[0075] As shown in the above examples, the present method for detecting sleep disorder events can improve the accuracy of detecting sleep disorder events based on single or multiple physiological signals because the process relies on contextual information about a single seed event in the form of determining a pattern, such as a substantially regular pattern, and determining a goodness of fit to the pattern, rather than relying solely on a decision tree based on a threshold of an isolated seed event, as is done in prior art methods.

[0076] In this regard, such methods of the technology described above may be further considered in conjunction with the system process flow diagrams of FIG. 7 and FIG. 8. In FIG. 7, process 700 may begin at 710. At 710, one or more processors may access or receive one or more physiological signals generated by one or more sensors, e.g., any one or more signals and sensors described herein, e.g., sensors including PAT signals. At 720, one or more processors may detect a seed event indicative of a sleep disorder event from the one or more physiological signals. At 730, one or more processors may calculate features indicative of a pattern or repeating pattern within a portion of the one or more physiological signals associated with the detected seed event. At 740, one or more processors may apply the calculated features indicative of the seed event pattern to a processing algorithm (e.g., a machine learning classifier or other classifier, or a hard-coded process or a deterministic process such as a decision tree). The processing algorithm may be trained to calculate how well the calculated features fit a learned pattern of sleep disorder events. At 750, the one or more processors may output an identification (e.g., a score) of one or more sleep disorder events corresponding to the seed event based on the calculated goodness of fit determined by the processing algorithm. Alternatively, at 760, one or more processors of the system may generate an output based on the output identification and / or application. Such output may include, for example, a signal, such as a control signal for a setting or setting (e.g., pressure or flow setting) having a respiratory therapy action, such as an action or treatment action described in further detail herein. For example, the signal may optionally be transmitted from the one or more processors to a respiratory therapy device, for example, via a communication network (e.g., the Internet) and / or other intermediate device(s) (e.g., one or more servers).

[0077] An example of such a process may be further considered in connection with the flowchart of FIG. 8. Similar to FIG. 7, at 820 of process 800, one or more processors may detect seed events indicative of sleep disorder events from one or more physiological signals. At 822, one or more processors may select a set of aligned or associated signal seed events, e.g., one from each signal, and may compute or determine one or more features of each event and / or associated these seed events. At 824, one or more processors may select one or more sequence sets of seed events and may compute or determine one or more sequence-related pattern features. At 826, one or more processors may select one or more aligned or associated sets of sequence sets of seed events and may compute or determine overlapping sequence-related pattern features. At 830, one or more processors may then merge the aforementioned pattern-related features from the seed events and sequences. At 840, one or more processors may then utilize the merged features to score sleep disorder events by applying the features to a classifier algorithm, such as any one or more of the classifier algorithms described above. Further outputs based on the scores may be generated as described above.

[0078] 6B illustrates a suitable computing system 600 including circuitry that enables performing any one or more of these processes or any one or more step(s) according to the described examples of the present technology utilizing a sensor device described herein, such as the finger sensor device of FIG. 6A. The computing system 600 may generally be formed as a suitable general-purpose computer and may include a bus 610, a processor 602, a local memory 604, one or more optional input interfaces 614, one or more optional output interfaces 616, a communication interface 612, a storage element interface 606, and one or more storage elements 608. The bus 610 may include one or more conductors that enable communication between the components of the computing system 600. The processor 602 may include any type of conventional processor or microprocessor that interprets and executes programming instructions. The local memory 604 may include a random access memory (RAM) or another type of dynamic storage device that stores information and instructions executed by the processor 602, and / or a read only memory (ROM) or another type of static storage device that stores static information and instructions used by the processor 602. The input interface 614 may include one or more conventional mechanisms, such as a keyboard 620, a mouse 630, a pen, a voice recognition and / or biometric mechanism, a camera, etc., that allow an operator or user to input information into the computing device 600. The output interface 816 may include one or more mechanisms, such as a display 640, for outputting information to an operator or user. The communication interface 612 may include a transceiver-like mechanism, such as one or more Ethernet interfaces that allow the computing system 600 to communicate with other devices and / or systems, such as other computing devices 681, 682, 683.The communication interface 612 of the computing system 600 may be connected to such other computing systems via a local area network (LAN) or a wide area network (WAN), for example, the Internet. The storage element interface 606 may include a storage interface, such as a Serial Advanced Technology Attachment (SATA) interface or a Small Computer System Interface (SCSI), to connect the bus 610 to one or more storage elements 608, for example, one or more local disks, such as SATA disk drives, and to control the reading and writing of data to and / or from these storage elements 608. Although the storage element(s) above are described as local disks, in general any other suitable computer readable medium may be used, such as a removable magnetic disk, an optical storage medium, such as a CD or DVD, a -ROM disk, a solid state drive, a flash memory card, etc.

[0079] As used herein, the term "circuitry" may refer to one or more or all of the following: (a) A purely hardware circuit implementation, such as an implementation using only analog and / or digital circuitry; (b) Combinations of hardware circuitry and software, such as, as applicable, (i) combinations of analog and / or digital hardware circuitry(s) and software / firmware;

[0080] (ii) a hardware processor(s) with software (including digital signal processor(s), software, and memory that work together to cause a device such as a mobile phone or server to perform various functions); and

[0081] (c) Hardware circuitry and / or processors, such as a microprocessor or parts of a microprocessor, that require software (e.g., firmware) in order to operate but may not be present if not necessary for the software to operate.

[0082] This definition of circuitry applies to all uses of the term in this application, including any claims. As a further example, as used in this application, the term circuitry may also cover a hardware circuitry alone, or a processor (or processors) alone, or a portion of a hardware circuitry or processor and (their) associated software and / or firmware implementation. The term circuitry may also include, for example, a baseband or processor integrated circuit of a mobile device, or a similar integrated circuit in a server, cellular network device, or other computing or network device, if applicable to a particular claim element.

[0083] Thus, a processing device capable of implementing any of the above processes may include an integrated chip, memory and / or other control instructions, data or information storage medium. For example, programming instructions including any of the evaluation / signal processing methods described herein may be encoded on an integrated chip in the memory of a device or apparatus to form an application specific integrated chip (ASIC). Such instructions having such processes may also be loaded using a suitable processor readable medium(s), data storage medium, or memory, or alternatively as software or firmware. Alternatively, such processing instructions may be downloaded, for example, from a server via a network (e.g., the Internet) to the processing device, which functions as a screening, monitoring device and / or therapy device when the instructions are executed.

[0084] Thus, such a processing device, e.g., a server, may include multiple components such as interfaces linked to the aforementioned sensors and / or receiving data representative of these signals from these sensors. In addition to other components, the processing device may include a processor(s), any display interface, any user control / input interface, and a memory / data memory 312 such as processing instructions having the processing methods / modules described herein. One or more sensors may be integrally or operatively coupled to the processing device. For example, the sensor(s) may be integrally or coupled to the processing device via a wired or wireless link (e.g., Bluetooth, Wi-Fi, etc.). Thus, the processing device may include a data communication interface. In some implementations, the system of the present technology may include a server. The server may be configured to access any of the processor-readable medium(s) or data storage medium(s) described herein that includes processor control instructions for any of the aforementioned processes. The server may be configured to receive requests to download the processor-executable instructions of the processor-readable medium to these processing devices via a network. In some implementations, the processing device may be a smartphone, tablet, or other smart device, or other computing device. The processing device may be a respiratory processing device, such as those described herein, that uses a blower controller to provide processing.

[0085] 4.1 Optional sample processing system As previously mentioned, in one form, the present technique may include an apparatus or device for treating and / or monitoring respiratory disorders, which may be associated with and / or configured for use as or in communication with any of the processes previously described as the computing system of FIG. 6. The apparatus or device may be a respiratory therapy device (RT), such as an RPT device 4000 for providing a pressurized air flow to the patient 1000 via an air circuit 4170 leading to a patient interface 3000. The air flow may be pressure controlled (for respiratory pressure therapy) or flow controlled (for flow therapy, such as high flow therapy HFT). Thus, the RPT device may be configured to function as a flow therapy device, for example, when using a patient interface that does not use a sealed seal with the patient's respiratory system. In the following description, the RT or RPT device may be considered with reference to FIGS. 8-11.

[0086] 4.2 Patient Interface A non-invasive patient interface 3000 according to one aspect of the present technology may optionally include any of the functional aspects of a seal-forming structure 3100, a plenum chamber 3200, a positioning and stabilizing structure 3300, a vent 3400, a connection port 3600 for connecting to an air circuit 4170, and a forehead support 3700. In some forms, the functional aspects may be provided by one or more physical components. In some forms, one physical component may provide one or more functional aspects. In use, the seal-forming structure 3100 is positioned to surround an entrance to the patient's airway to facilitate the delivery of pressurized air to the airway.

[0087] (4.3 RPT Device) An RPT device 4000 according to one aspect of the present technology comprises mechanical and pneumatic components 4100, electronic components 4200, and is programmed to execute one or more algorithms 4300. The RPT device 4000 may have an outer housing 4010 formed as two parts, an upper portion 4012 and a lower portion 4014. Further, the outer housing 4010 may include one or more panel(s) 4015. The RPT device 4000 may comprise a chassis 4016 that supports one or more internal components of the RPT device 4000. The RPT device 4000 may include a handle 4018.

[0088] The air pressure path of the RPT device 4000 may include, for example, an air inlet filter 4112, an inlet muffler 4122, a pressure generator 4140 (e.g., blower 4142) that can supply pressurized air, an outlet muffler 4124, and one or more sensors 4270, such as a pressure sensor 4272 and a flow sensor 4274.

[0089] One or more pneumatic path items may be disposed within a movable, unitary structure referred to as a pneumatic block 4020. The pneumatic block 4020 may be disposed within the external housing 4010. In one form, the pneumatic block 4020 is supported by or formed as part of the chassis 4016.

[0090] The RPT device 4000 can have a power supply 4210, one or more input devices 4220, a central controller 4230, a therapy device controller 4240, a pressure generator 4140, one or more protection circuits 4250, a memory 4260, a transducer 4270, a data communication interface 4280, and one or more output devices 4290. The electronic components 4200 can be mounted on a single printed circuit board assembly (PCBA) 4202. In the alternative, the RPT device 4000 can include two or more PCBAs 4202.

[0091] (4.3.1 Mechanical and Pneumatic Components of the RPT Device) The RPT device 4000 may include, in an overall unit, one or more of the following components: In another form, one or more of the following components may be arranged as separate units.

[0092] (4.3.1.1 Air filter(s)) An RPT device 4000 in accordance with one form of the present technology may include an air filter 4110 or multiple air filters 4110.

[0093] In one form, the air inlet filter 4112 is located at the beginning of the air pressure path upstream of the pressure generator 4140 .

[0094] In one form, the air outlet filter 4114, for example an antibacterial filter, is located between the outlet of the pneumatic block 4020 and the patient interface 3000.

[0095] (4.3.1.2 Muffler(s)) An RPT device 4000 in accordance with one form of the present technology may include a muffler 4120 or multiple mufflers 4120.

[0096] In one form of the present technology, an inlet muffler4122 is positioned in the pneumatic path upstream of a pressure generator4140.

[0097] In one form of the present technology, the outlet muffler 4124 is placed in the pneumatic path between the pressure generator 4140 and the patient interface 3000.

[0098] (4.3.1.3 Pressure Generator) In one form of the present technology, the pressure generator 4140 for supplying compressed air is a controllable blower 4142. For example, the blower 4142 may include a brushless DC motor 4144 having one or more impellers housed within a volute housing. The pressure generator 4140 may generate an air supply or air flow of, for example, about 120 liters / minute, at a positive pressure within a range of about 4 cmh2O to about 20 cmh2O, or in other forms up to about 30 cmH2O.

[0099] The pressure generator 4140 is under the control of the treatment device controller 4240 .

[0100] In other forms, the pressure generator 4140 can be a piston-driven pump, a pressure regulator connected to a high pressure source (eg, a compressed air container), or a bellows.

[0101] (4.3.1.4 Transducer(s)) The transducer may be internal to the RPT device or external to the RPT device, for example a finger PPG sensor or PAT sensing apparatus as described above, including a finger sensor device as shown in Figure 6A. The external sensor may be located on the air circuit, for example, or may form part of the air circuit, such as a patient interface. The external transducer may be in the form of a non-contact sensor, such as a Doppler radar motion sensor, that transmits or transfers data to the RPT device.

[0102] In one form of the present technology, the one or more sensors 4270 are positioned upstream and / or downstream of the pressure generator 4140. The one or more transducers 4270 are constructed and positioned to generate data representative of a corresponding attribute of the airflow, such as flow rate, pressure or temperature, at that point in the airflow.

[0103] In one form of the present technology, one or more sensors 4270 are positioned proximate the patient interface 3000.

[0104] In one form, the signal from the transducer 4270 can be filtered, for example, by low pass, high pass, or band pass.

[0105] (4.3.1.5 Spillback prevention valve) In one form of the present technology, the anti-spillback valve 4160 is located between the humidifier 5000 and the pneumatic block 4020. The anti-spillback valve is constructed and arranged to reduce the risk of water flowing upstream from the humidifier 5000, for example to the motor 4144.

[0106] (4.3.1.6 Air Circuit) The air circuit 4170, according to one aspect of the present technology, is a conduit or tube constructed and arranged to allow air flow to travel between two components, such as the pneumatic block 4020 and the patient interface 3000, in use.

[0107] 4.3.1.7 Oxygen Delivery In one form of the present technology, supplemental oxygen 4180 is delivered to the air circuit 4170 and / or patient interface 3000 at one or more points in the pneumatic path, such as upstream of the pneumatic block 4020.

[0108] (4.3.2 Electronic Components of the RPT Device) (4.3.2.1 Power supply) In one form of the present technology, the power source 4210 is located inside the external housing 4010 of the RPT device 4000. In another form of the present technology, the power source 4210 is located outside the external housing 4010 of the RPT device 4000.

[0109] In one form of the present technology, the power supply 4210 powers only the RPT device 4000. In another form of the present technology, the power supply 4210 powers both the RPT device 4000 and the humidifier 5000.

[0110] (4.3.2.2 Input Devices) In one form of the present technology, the RPT device 4000 includes one or more input devices 4220 in the form of buttons, switches, or turntables that allow a person to interact with the device. The buttons, switches, or dials may be physical or software devices accessible via a touch screen. The buttons, switches, or turntables may be physically connected to the external housing 4010 in one form, or in another form may communicate wirelessly with a receiver electrically connected to the central controller 4230.

[0111] In one form, the input device 4220 may be constructed and arranged to allow a human to select values ​​and / or menu options.

[0112] (4.3.2.3 Central Controller) In one form of the present technology, the central controller 4230 is a processor adapted to control the RPT device 4000, for example an x86 INTEL processor.

[0113] A central controller 4230 suitable for controlling an RPT device 4000 in accordance with another aspect of the present technology includes a processor based on an ARM Cortex-M processor from ARM Holdings. For example, an STM32 series microcontroller from STMicroelectronics can be used.

[0114] According to another alternative form of the present technology, another central controller 4230 adapted to control the RPT device 4000 includes a selected member of the ARM9-based 32-bit RISC CPU series. For example, the STR9 series microcontroller from STMicroelectronics can be used.

[0115] In some alternative forms of the present technology, a 16-bit RISC CPU can function as the central controller 4230 of the RPT device 4000. For example, a processor from the Texas Instruments MSP430 series microcontrollers can be used.

[0116] In another form of the present technology, the central controller 4230 is a dedicated electronic circuit. In another form, the central controller 4230 is an application specific integrated circuit (ASIC). In another form, the central controller 4230 includes discrete electronic components.

[0117] The central controller 4230 is configured to receive input signal(s) from the one or more sensors 4270, the one or more input devices 4220, and the humidifier 5000.

[0118] The central controller 4230 is configured to provide output signal(s) to one or more of the output device 4290, the therapy device controller 4240, the data communication interface 4280, and the humidifier 5000.

[0119] In some forms of the present technology, the central controller 4230 is configured to implement one or more methods described herein, e.g., one or more algorithms 4300 represented as a computer program stored in a non-transitory computer-readable storage medium, such as the memory 4260 or other memory described herein. In some forms of the present technology, the central controller 4230 may be integrated with the RPT device 4000, as previously described. However, in some forms of the present technology, some methods may be performed by a remote device or server, such as the server described above. For example, the remote device or server may determine, for example, by detecting a breathing-related event and analyzing stored data (e.g., from any of the sensors described herein) to identify control settings by type for transmission to the ventilator or other RT device.

[0120] The central controller 4230 may include a single controller that interacts with the various sensors 4270, data communication interface 4280, memory 4260, and other devices, although the functions of the controller 4230 may be distributed among multiple controllers. Thus, the term "central" as used herein is not meant to limit the architecture to a single controller or processor controlling other devices. For example, alternative architectures may include a distributed controller architecture including one or more controllers or processors, which may optionally communicate electronically (wired or wirelessly) directly or indirectly with the finger sensors described above, such as to perform any of the methods described herein, or may include a server in communication with the finger sensors. This may include, for example, a separate local (i.e., within the RPT device 4000) that executes several algorithms 4300, or a remotely located controller that executes one or more local or remote memories that store several algorithms. Furthermore, when expressed as a computer program, the algorithms may include high-level human readable code (e.g., C++, Visual Basic, other object-oriented languages, etc.) or low-level / machine level instructions (assembler, Verilog, etc.). Depending on the function of the algorithm(s), such code or instructions may be written into a controller, for example an ASIC or DSP, or may be a run-time executable ported to a DSP or general-purpose processor, specially programmed to perform the tasks required by the algorithm(s).

[0121] (4.3.2.4 Clock) The RPT device 4000 may include a clock 4232 connected to the central controller 4230 .

[0122] (4.3.2.5 Therapy Device Controller) In one form of the present technology, the therapy device controller 4240 is a therapy control module 4330 that forms part of the algorithm 4300 executed by the central controller 4230.

[0123] In one form of the present technology, the therapy device controller 4240 is a dedicated motor control integrated circuit. For example, in one form, the MC33035 brushless DC motor controller manufactured by ONSEMI is used.

[0124] (4.3.2.6 Protection circuit) An RPT device 4000 in accordance with the present technology may include one or more protection circuits 4250.

[0125] One form of protection circuit 4250 according to the present technology is an electrical protection circuit.

[0126] One form of protection circuit 4250 in accordance with the present technology is a temperature or pressure safety circuit.

[0127] (4.3.2.7 Memory) In accordance with one form of the present technology, the RPT device 4000 includes a memory 4260, such as a non-volatile memory. In some forms, the memory 4260 may include battery-powered static RAM. In some forms, the memory 4260 may include volatile RAM.

[0128] The memory 4260 may reside on the PCBA 4202. The memory 4260 may be in the form of an EEPROM or NAND flash memory.

[0129] Additionally or alternatively, the RPT device 4000 includes a removable form of memory 4260, such as a memory card made in accordance with the Secure Digital (SD) standard.

[0130] In one form of the present technology, memory 4260, e.g., one of the memories described above, operates as a non-transitory computer-readable storage medium that stores computer program instructions, e.g., one or more algorithms 4300, that represent one or more methods described herein.

[0131] (4.3.2.8 Transducers) The transducer may be internal to the RPT device 4000 or external to the RPT device 4000. The external sensor may be located on the air circuit 4170 or may form part of the air circuit 4170, for example in the patient interface 3000. The external transducer may be in the form of a non-contact sensor, such as a Doppler radar motion sensor, that transmits or transfers data to the RPT device 4000.

[0132] (4.3.2.8.1 Flow rate) The flow sensor 4274 according to the present technology can be based on a differential pressure sensor, such as SENSIRION's SDP600 series. The differential pressure sensor is in fluid communication with the air circuit, with one of the pressure sensors being connected to a corresponding first and second point of the flow restricting element.

[0133] In one example, the central controller 4230 receives a signal representing the total flow Qt from the flow sensor 4274.

[0134] (4.3.2.8.2 Pressure) A pressure sensor 4272 according to the present technology is positioned in fluid communication with the pneumatic path. An example of a suitable pressure sensor 4272 is a HONEYWELL ASDX series sensor. Another suitable pressure sensor is GE's NPA series sensor.

[0135] In use, the signal from the pressure sensor 4272 is received by the central controller 4230. In one form, the signal from the pressure sensor 4272 is filtered before being received by the central controller 4230.

[0136] (4.3.2.8.3 Motor Speed) In one form of the present technology, the motor speed sensor 4276 is used to determine the rotational speed of the motor 4144 and / or the blower 4142. A motor speed signal from the motor speed sensor 4276 can be provided to the therapy device controller 4240. The motor speed sensor 4276 may be a speed sensor such as, for example, a Hall effect sensor.

[0137] (4.3.2.9 Data communication systems) In one form of the present technology, a data communications interface 4280 is provided and connected to a central controller 4230. The data communications interface 4280 can be connected to a remote external communications network 4282 and / or a local external communications network 4284. The remote external communications network 4282 can be connected to a remote external device 4286. The local external communications network 4284 can be connected to a local external device 4288.

[0138] In one form, the data communication interface 4280 is part of the central controller 4230. In another form, the data communication interface 4280 can be separate from the central controller 4230 and include an integrated circuit or processor.

[0139] In one form, the remote external communications network 4282 is the Internet. The data communications interface 4280 can connect to the Internet using wired communications (e.g., via Ethernet or optical fiber) or wireless protocols (e.g., CDMA, GSM, LTE).

[0140] In one form, the local external communications network 4284 utilizes one or more communications standards, such as Bluetooth or consumer infrared protocols, and may optionally communicate with any of the sensors described herein.

[0141] In one form, the remote external device 4286 is one or more computers, such as a cluster of computers and / or servers connected to a network, as described herein. In one form, the remote external device 4286 may be a virtual computer rather than a physical computer. In either case, such a remote external device 4286 may be accessed by appropriately authorized persons (e.g., a clinician).

[0142] The local external device 4288 may be a personal computer, a mobile phone, a tablet, or a remote control.

[0143] (4.3.2.10 Output Devices (including optional displays and alerts)) The output device 4290 according to the present technology can take the form of one or more of a visual, auditory and tactile unit. The visual display can be a liquid crystal display (LCD) or a light emitting diode (LED) display.

[0144] (4.3.2.10.1 Display Drivers) The display driver 4292 receives as input characters, symbols or images intended to be displayed on the display 4294 and converts them into commands that cause the display 4294 to display them.

[0145] (4.3.2.10.2 Display) The display 4294 is configured to visually display characters, symbols, or images in response to commands received from the display driver 4292. For example, the display 4294 may be an eight-segment display, in which case the display driver 4292 converts each character or symbol (e.g., the number "0") into eight logical signals indicating whether eight corresponding segments are activated to display the particular character or symbol.

[0146] (4.3.3 RPT Device Algorithm) (4.3.3.1 Pre-processing module) The pre-processing module 4310 of the present technology receives raw data as input from a sensor 4270 (e.g., a flow sensor 4274 or a pressure sensor 4272) and performs one or more processing steps to calculate one or more output values ​​that are used as inputs to other modules (e.g., a therapy engine module 4320).

[0147] In one form of the present technology, the output values ​​include interface pressure Pm, respiratory flow Qr, and leak flow Ql.

[0148] In various forms of the present technology, the pre-processing module 4310 comprises one or more of the following algorithms: pressure compensation 4312, ventilation flow estimation 4314, leak flow estimation 4316, respiratory flow estimation 4317, ventilation volume determination 4311, target ventilation volume determination 4313, respiratory rate estimation 4318, and backup rate determination 4319.

[0149] (4.3.3.1.1 Pressure compensation) In one form of the present technology, a pressure compensation algorithm 4312 receives as an input a signal indicative of the pressure in the pneumatic path near the outlet of the pneumatic block 4020. The pressure compensation algorithm 4312 then provides the estimated pressure Pm as an output to the patient interface 3000.

[0150] (4.3.3.1.2 Ventilation flow estimation algorithm) In one form of the present technology, a ventilation flow estimation algorithm 4314 takes as input the estimated pressure Pm at the patient interface 3000 from the pressure compensation algorithm 4312 and estimates the ventilation flow Qv of air from the vent 3400 in the patient interface 3000.

[0151] (4.3.3.1.3 Leak flow rate estimation algorithm) In one form of the present technology, a leak flow estimation algorithm 4316 receives as inputs the total flow Qt and the ventilation flow Qv and provides as an output an estimate of the leak flow Ql. In one form, the leak flow estimation algorithm estimates the leak flow Ql by calculating the average value of the difference between the total flow Qt and the ventilation flow Qv over a period of time long enough to include several respiratory cycles (e.g., 10 s).

[0152] In one form, the leak flow estimation algorithm 4316 receives as input the total flow Qt, ventilation flow Qv, and estimated pressure Pm in the patient interface 3000, estimates the leak flow Ql by calculating the leak conductance, and determines that the leak flow Ql is a function of the leak conductance and pressure Pm. The leak conductance is calculated as the quotient of the low-pass filtered unventilated flow, which is equal to the difference between the total flow Qt and the ventilation flow Qv, and the low-pass filtered square root of the pressure Pm, with the low-pass filter time constant having a value long enough to include several respiratory cycles, for example, about 10 seconds. The leak flow Ql can be estimated as a function of the product of the leak conductance and the pressure, Pm.

[0153] (4.3.3.1.4 Respiratory Flow Estimation Algorithm) In one form of the present technology, the respiratory flow estimation algorithm 4317 takes as inputs the total flow Qt, the ventilation flow Qv, and the leak flow Ql, and estimates the air respiratory flow Qr for the patient by subtracting the ventilation flow Qv and the leak flow Ql from the total flow Qt.

[0154] In another form of the present technology, the respiratory flow estimation algorithm 4317 provides a surrogate value for respiratory flow Qr. Possible surrogates for respiratory flow include: - 1000 respiratory movements of the patient's chest, - the current drawn by the pressure generator 4140; - the motor speed of the pressure generator 4140, -Patient transthoracic impedance 1000 is included.

[0155] The respiratory flow proxy value may be provided, for example, by a sensor 4270 within the RPT device 4000, such as a motor speed sensor 4276, or by a sensor external to the RPT device 4000, such as a respiratory motion sensor or a transthoracic impedance sensor.

[0156] (4.3.3.1.5 Ventilation volume determination algorithm) In one form of the present technology, a ventilation determination algorithm 4311 takes as input the respiratory flow, Qr, and determines a measurement, Vent, indicative of the current patient ventilation.

[0157] In some implementations, the ventilation determination algorithm 4311 determines a ventilation measurement that is an estimate of the actual patient ventilation, Vent.

[0158] In one such implementation, the measurement of ventilation volume Vent is half the absolute respiratory flow Qr, optionally filtered by a low-pass filter such as a second-order Bezier low-pass filter with a corner frequency of 0.11 Hz.

[0159] In one such implementation, the measurement of ventilation Vent is an estimate of total alveolar ventilation (i.e., non-anatomical dead space ventilation). This requires an estimation of anatomical dead space. Patient height (or arm spread in cases of severe skeletal deformity) can be used as a good predictor of anatomical dead space. Total alveolar ventilation is equal to the measurement of actual patient ventilation, e.g., as determined above, minus the product of the estimated anatomical dead space and the estimated spontaneous breathing rate Rs.

[0160] In other embodiments, the ventilation determination algorithm 4311 determines a measurement of ventilation Vent that is substantially proportional to the actual patient ventilation. In such an embodiment, the peak respiratory flow rate Qpeak of the estimated cycle inspiration period is realized. Many other programs, including this program and the sampling of the respiratory flow rate Qr, generate measurements that are approximately proportional to the ventilation volume on the assumption that the shape of the flow rate waveform does not change much (here, when the flow rate waveforms of breaths that are normalized temporally and amplitude-wise are similar, the shapes of both breaths are considered to be similar). Simple examples include the median of the positive respiratory flow rate, the median of the absolute value of the respiratory flow rate, and the standard deviation of the flow rate. Any linear combination of any order statistics of the statistics of the absolute value of the respiratory flow rate using positive coefficients, or even those using both positive and negative coefficients, is approximately proportional to the ventilation volume. Another example is the average value of the respiratory flow velocity at the intermediate KK ratio (time) of the inspiration part, where 0 < K < 1. If the shape of the flow velocity waveform does not change, there can be any number of measurements that are exactly proportional to the ventilation volume.

[0161] In other forms, the ventilation determination algorithm 4311 does not base on the ventilation flow rate Qr, but determines a measurement of ventilation Vent, which is an agent of the current patient's ventilation, obtained from a suitable sensor connected to the patient 1000, such as the oxygen saturation (SaO2) or the partial pressure of carbon dioxide (PCO2).

[0162] (4.3.3.1.6 Target Ventilation Determination Algorithm) In one form of the present technology, the central controller 4230 takes the current measurement of ventilation Vent as an input and executes one or more target ventilation determination algorithms 4328 to determine the target value Vtgt of the ventilation measurement.

[0163] In some forms of the present technology, the target ventilation determination algorithm 4328 does not exist, and the target value Vtgt is predetermined, for example, by hard-coding it in the configuration of the RPT device 4000 or by manually inputting it via the input device 4220.

[0164] In other forms of the present technology, for example adaptive servo ventilation (ASV) therapy (described below), the target ventilation determination algorithm 4313 calculates the target ventilation Vtgt from a value Vtyp indicative of the patient's 1000 typical recent ventilation.

[0165] In some forms of adaptive servo ventilation, the target ventilation Vtgt is calculated as a high percentage of the typical recent ventilation Vtyp, but less than the typical recent ventilation Vtyp. Such a high percentage may be in the range of (80%, 100%), (85%, 95%), (87%, 92%).

[0166] In other forms of adaptive servo ventilation, the target ventilation, Vtgt, is calculated as a unit multiple slightly larger than the typical current ventilation, Vtyp.

[0167] Typical recent ventilation Vtyp is the value to which the distribution of current ventilation Vent at multiple time instants in a given time scale tends to converge, i.e., a measure of the convergence tendency of current ventilation in recent history. In one implementation of the target ventilation determination algorithm 4313, the recent history is on the order of minutes, but in any case must be longer than the time scale of a Cheyne-Stokes rise-fall cycle. The target ventilation determination algorithm 4313 can use any of a variety of known convergence tendency measures to determine the typical recent ventilation Vtyp from a measurement of current ventilation Vent. One such measure is the output of a low pass filter on a measurement of current ventilation Vent, with a time constant equal to 100 seconds.

[0168] (4.3.3.1.7 Respiratory flow rate estimation) In one form of the present technology, a respiratory flow estimation algorithm 4318 takes as input the patient's 1000 respiratory flow Qr and generates an estimate of the patient's spontaneous respiratory flow Rs.

[0169] The respiratory rate estimation algorithm 4318 can estimate the spontaneous breathing rate Rs during spontaneous breathing of the patient 1000, i.e., when the RPT device 4000 is not delivering a "pre-breath" (described below). In some forms of the present technology, the respiratory flow estimation algorithm 4318 estimates respiratory flow during periods of low servo assist (defined as pressure support minus minimum pressure support) at a value less than 4 cmH2O in one implementation, as this is more likely to reflect spontaneous respiratory effort.

[0170] In some forms of the present technology, the respiratory flow estimation algorithm 4318 estimates the respiratory flow during sleep breathing, since the respiratory flow during these periods can be substantially different from the respiratory flow during wakefulness. Anxiety typically causes a higher respiratory rate than during sleep. While the patient is concentrating on their breathing process, the respiratory rate typically is lower than the normal awake or sleep respiratory rate. Patent application number PCT / AU2010 / 000894, disclosed as WO 2011 / 006199, is incorporated herein by reference and may be used to identify periods of wakefulness breathing from the respiratory flow Qr.

[0171] In some forms of the present technology, the respiratory flow estimation algorithm 4318 estimates the spontaneous respiratory flow Rs as the reciprocal of one of various well-known statistical measures of the concentration tendency of the breath duration Ttot during the period of interest. It is desirable for such measures to reject outliers or at least be robust to outliers. A metric that trims the average value where a percentage of the lower and higher K of the ranked breath durations are discarded and calculates the average value from the remaining breath durations is robust to outliers. For example, if K is 0.25, this corresponds to discarding the upper and lower quartiles of the breath duration Ttot. The median is another reliable measure of concentration tendency, but may not produce satisfactory results when the distribution is strongly bimodal. Although the simple average is sensitive to outliers, it can also be used as a measure of concentration tendency. An initial interval filtering phase may be used in which consecutive time intervals corresponding to impossible breath rates (e.g., greater than 45 breaths / min or less than 6 breaths / min) are excluded from the average calculation as outliers. Other filtering mechanisms that can be used alone or in combination with interval filtering are to eliminate any breaths that do not belong to a sequence of N consecutive spontaneous breaths, where N is some small integer (e.g., 3), and to eliminate early and late breaths of a sequence of consecutive spontaneous breaths, e.g., the first and last breaths of a sequence of four breaths. The basic principle of the latter mechanism is that the first and last breaths of a sequence of spontaneous breaths, as well as early and late breaths in general, may be atypical. For example, the first spontaneous breath may be the result of arousal, while the last spontaneous breath may be longer because of reduced respiratory drive and a reserve breath terminating the spontaneous breath sequence.

[0172] In some forms of the present technology, the respiratory flow estimation algorithm 4318 uses an initial estimation period to make an initial estimate of the spontaneous breathing flow Rs to allow for the initiation of subsequent processing in the therapy engine module 4320, and then continuously updates the estimate of the spontaneous breathing flow Rs using a longer estimation period to improve statistical robustness. For example, the estimated initial period may be 20 minutes of adequate spontaneous breathing, but the estimated period may then be gradually increased to some maximum duration, e.g., 8 hours. This estimation can use a low pass filter on the breathing duration, instead of using a low pass filter on the breathing duration, which has a gradually longer response time (or more precisely, a gradually lower angular frequency) as the session progresses.

[0173] In some forms, a suitably processed short-term (e.g., 10 minute) measurement of concentration trend, such as a trimmed average, can be input into a suitable low pass filter to give an estimated Rs that varies over the hourly or longer time scale. The advantage is that it does not require the storage and processing of large amounts of breath duration data that may result if a trimmed average needs to be calculated over a moving window of breath duration data spanning hours or days.

[0174] In some forms of the present technology, respiratory rate measured for a short period of time, particularly within a single breath, may be used in place of respiratory duration in the concentration tendency measurements described above, giving substantially similar but different results.

[0175] (4.3.3.1.8 Backup rate determination) In one form of the present technology, a backup rate determination algorithm 4319 takes as input the spontaneous breathing rate estimate Rs provided by the respiratory rate estimation algorithm 4318 and returns a "backup rate" Rb. The backup rate Rb is the rate at which the RPT device 4000 continues to deliver backup breaths, i.e., provide ventilatory support, to the patient 1000 without significant spontaneous breathing effort.

[0176] In one form of the pre-processing module 4310, the backup rate determination algorithm 4319 is not present and the backup rate Rb is manually provided to the RPT device 4000, for example via the input device 4220, or is hard-coded when the RPT device 4000 is configured.

[0177] In what is called an adaptive backup rate, the backup rate determination algorithm 4319 determines the backup rate Rs as a function of the spontaneous breathing rate Rb. In one implementation, the function determines the backup rate Rb as the spontaneous breathing flow Rs minus a constant, for example, 2 breaths per minute. In another implementation, the function determines the backup rate Rb as the spontaneous breathing flow Rs multiplied by a constant that is slightly less than one.

[0178] In a form called variable backup rate, the backup rate determination algorithm 4319 determines the backup rate Rb as a function of time. The backup rate Rb is initialized to a value called spontaneous backup rate (SBR) that is a fraction of the final target backup rate called sustained planned backup rate (STBR). The fraction can be two-thirds, three-quarters, or any other positive value less than one. The SBR is the breath for which the most recent inspiration was spontaneous (e.g., patent trigger). The STBR can be predetermined (e.g., by manual entry or hard coding as described above) or can be set to a typical respiratory rate, such as 15 bpm. The backup rate Rb increases from the SBR toward the STBR over time since the last spontaneous breath. The increase can depend on a predefined profile, such as a series of steps, or a continuous linear profile. The profile is selected such that the backup rate Rb reaches the STBR after a predefined interval. This interval can be measured in time units, such as 30 seconds, or it can be measured in terms of patient breaths, such as 5 breaths.

[0179] In some forms of variable backup rate, the predetermined interval during which the backup rate Rb increases from SBR towards STBR may be a function of the sufficiency of the current ventilation. In implementations suitable for servo ventilation where there is a target value for the ventilation measurement Vtgt, the standby rate approaches STBR more quickly as the current ventilation measurement Vent is less than the target ventilation Vtgt.

[0180] In one form of variable backup rate, called adaptive variable backup rate, the backup rate determination algorithm 4319 determines the backup rate Rb as a function of the currently estimated spontaneous breathing rate Rs provided by the respiratory rate estimation algorithm 4318 and as a function of time. As in the case of variable backup rate determination, the adaptive variable backup rate determination increases the backup rate Rb from SBR to STBR at a predetermined interval, which may be a function of the current ventilation adequacy. The STBR may be initialized to a standard respiratory flow, e.g., 15 bpm. Once a reliable estimate of the spontaneous breathing flow Rs is obtained from the respiratory flow estimation algorithm 4318, the STBR may be set to multiply the currently estimated spontaneous breathing flow Rs by a constant. The SBR may be set to a fraction of the STBR, such as the variable backup rate. In one form, during the initial period of the estimated spontaneous breathing rate Rs, a fraction, e.g., two-thirds, may be set to a low value, such as 0.55, to accommodate the occasional long breath duration of a patient with a relatively low respiratory rate, e.g., 12 breaths per minute.

[0181] In some forms, the constant by which the current estimated spontaneous breathing rate Rs is multiplied to obtain the STBR may be slightly greater than 1, e.g., 1.1, to provide more aggressive ventilation during apneas, which may be desirable during short apneas. The constant may also be slightly less than 1, e.g., 0.8, if difficulty in resynchronizing with the patient, especially upon return of patient effort, proves to be a problem in certain patients. The lower the backup rate, the easier it may be to resynchronize by extending the expiratory pause, during which time resynchronization occurs more frequently.

[0182] (4.3.3.2 Therapy Engine Module) In one form of the present technology, the therapy engine module 4320 receives as inputs one or more of the pressure Pm in the patient interface 3000, the patient's air breathing flow Qr, and an estimate of the intrinsic breathing rate Rs, and provides as output one or more therapy parameters. In various forms, the therapy engine module 4320 includes one or more algorithms for phase determination 4321, waveform determination 4322, inspiratory flow limitation determination 4324, apnea / hypopnea determination 4325, snoring detection 4326, airway patency determination 4327, and therapy parameter determination 4329, including, for example, the central and obstructive type determinations discussed above.

[0183] (4.3.3.2.1 Phase Determination Algorithm) In one form of the present technology, a phase determination algorithm 4321 receives as an input a signal indicative of respiratory flow, Qr, and provides as an output the phase Φ of the current respiratory cycle of the patient 1000.

[0184] In some forms known as discrete phase determination, the phase output Φ is a discrete variable. One implementation of discrete phase determination provides a binary phase output Φ with a value of inspiration or expiration, e.g., a value represented by 0 and 0.5 revolutions, respectively, upon detecting the start of spontaneous inspiration and expiration, respectively. The "trigger" and "loop" RPT device 4000 can efficiently perform discrete phase determination because the trigger and loop points are the moments when the phase changes from expiration to inspiration and from inspiration to expiration, respectively. In one embodiment of binary phase determination, the phase output Φ is determined to a discrete value of 0 (thus "triggers" the RPT device 4000) when the respiratory flow Qr has a value above a positive threshold, and is determined to a discrete value of 0.5 revolutions (thus "cycles" the RPT device 4000) when the respiratory flow Qr has a value more negative than a negative threshold.

[0185] Another implementation of the discrete phase determination provides a ternary phase output Φ having one of the following values: inspiration, inspiration pause, expiration.

[0186] In another form, called continuous phase determination, the phase output Φ is a continuous variable, such as 0 to 1 revolution or 0 to 2π radian change. An RPT device 4000 with continuous phase determination can trigger and loop when the continuous phase is 0 revolutions and 0.5 revolutions, respectively. In one implementation of continuous phase determination, a fuzzy logic analysis of the respiratory flow rate Qr is used to determine continuous values ​​of the phase Φi. The continuous phase values ​​determined in this implementation are commonly referred to as "fuzzy phases." In one implementation of the fuzzy phase determination algorithm 4321, the following rules are applied to the respiratory flow rate Qr: 1.1. If respiratory flow is zero and then increases suddenly, the phase will be 0 rotations. 2.2. If respiratory flow is stable and positive, the phase should be 0.25 rotations. 3.3. If respiratory flow is zero and falls rapidly, the phase is 0.5 rotation. 4.4. If respiratory flow is large and stable, the phase should be 0.75 rpm. 5.5. When respiratory flow is stable at zero and the 5 second low pass filtered absolute value of respiratory flow is large, the phase is 0.9 rpm. 6.6. If respiratory flow is positive and the phase is expiration, the phase is 0 rotations. 7.7. If respiratory flow is negative and the phase is inspiration, the phase is 0.5 rotation. 8. 8. 5 second low pass filtered If the absolute value of respiratory flow is large, the phase increases at a steady rate equal to the patient's respiratory flow, low pass filtered with a time constant of 20 seconds.

[0187] The output of each rule can be represented as a vector with a phase, which is the result of that rule, and a magnitude, which is the degree of ambiguity that the rule is true. Respiratory flow is a blurriness of "large", "stable", etc., as determined by appropriate membership functions. The results of the rules are represented as vectors and combined by some function, such as centroid. In this combination, the rules may be weighted equally or differently.

[0188] In another implementation of sequential phase determination, the inspiration time Ti and expiration time Te are first estimated from the respiratory flow Qr, and the phase Φ is determined to be half the percentage of the inspiration time Ti that has elapsed since the previous trigger time, or half the percentage of the expiration time Te that has elapsed since the previous cycle time per 0.5 revolutions, whichever is closer.

[0189] In some forms of the present technology suitable for pressure support ventilation therapy (described below), the phase determination algorithm 4321 is configured to be triggered even when respiratory flow Qr is not significant, such as during apnea. As a result, the RPT device 4000 provides a "backup breath" when the patient 1000 is not making a spontaneous respiratory effort. In this form, referred to as the spontaneous / timed (S / T) pattern, the phase determination algorithm 4321 can utilize the backup rate Rb provided by the backup rate determination algorithm 4319.

[0190] The phase decision algorithm 4321 using "fuzzy phase" can achieve the S / T pattern by including a "momentum" rule in the fuzzy phase rule, using a backup rate Rb. The momentum rule acts to advance the successive phase from expiration to inspiration at the backup rate Rb when the characteristics of the respiratory flow Qr do not advance the successive phase by other rules. In one implementation, the further the measured value of ventilation Vent (described below) is below the target value Vtgt of ventilation (described below), the higher the weight of the momentum rule in the combination. However, a sudden increase in pressure support in response to mild to moderate hypoventilation (relative to the target ventilation) may bring ventilation very close to the target ventilation. When ventilation is close to the target, the desired momentum rule is given a low weight, and at other times (when the patient is not centrally apneic) it allows the patient to breathe at a rate significantly lower than the respiratory flow without being unnecessarily prompted by the ventilator to breathe at a higher rate. However, if a low weighting is given to the momentum rule when ventilation is below the target ventilation but above a value close to the target ventilation, adequate ventilation may be easily achieved with relatively high pressure support at a rate well below the backup rate. Because the spare breath allows the target breath to be delivered with lower pressure support, it is desirable to deliver it at a higher rate. This is ideal for many reasons, one important of which is the reduction of mask leak.

[0191] In summary, in the fuzzy phase decision algorithm 4321 implementing the S / T pattern, when selecting the weight of the momentum rule including the backup rate Rb, there is a dilemma that if the weight is too high, the patient may feel "pushed" by the backup rate. If it is too low, the pressure support may be too large. Therefore, it is desirable to provide a method for realizing the S / T pattern that does not depend on the momentum rule described above.

[0192] The phase determination algorithm 4321 (discrete or continuous without momentum regulation) can implement the S / T pattern using the backup rate Rb in a manner called scheduled backup. Scheduled backup can be achieved by the phase determination algorithm 4321 attempting to detect the start of inspiration due to spontaneous breathing effort, for example by monitoring respiratory flow Qr, as described above. The phase determination algorithm 4321 sets the phase output Φ to an inspiration value (thus triggering the RPT device 4000) if the start of inspiration due to spontaneous breathing effort is not detected within a period of time since the last trigger time whose duration is equal to the inverse of the backup rate Rb (an interval called the backup timing threshold). When the RPT device 4000 is triggered to begin delivering a spare breath, the phase determination algorithm 4321 attempts to detect the start of spontaneous expiration, such as by monitoring respiratory flow Qr, and based on this sets the phase output Φ to an expiration value (thus cycling the RPT device 4000).

[0193] As in the variable backup rate scheme described above, as the backup rate Rb increases over time from SBR to STBR, the backup timing threshold begins to lengthen and gradually shortens. That is, the RPT device 4000 is initially less alert and becomes increasingly more alert to the lack of spontaneous breathing effort as more reserve breaths are delivered. Such an RPT device 4000 is less likely to make the patient feel "pushed" while providing reserve breaths as needed if the patient prefers to breathe at a lower rate than the standard.

[0194] Like the adaptive variable backup rate system described above, if the STBR of the variable backup rate system matches the patient's estimated spontaneous breathing rate Rs, then reserve breaths are delivered at a rate that matches the patient's own recent spontaneous breathing effort.

[0195] (4.3.3.2.2 Waveform Determination Algorithm) In one form of the present technology, the therapy control module 4330 controls the pressure generator 4140 to provide a therapy pressure Pt that varies as a function of the phase Φ of the patient's respiratory cycle according to a waveform template Π(Φ).

[0196] In one form of the present technology, the waveform determination algorithm 4322 provides a waveform template Π(Φ) having values ​​in the range [0, 1] over the domain of phase values ​​Φ for use by the treatment parameter determination algorithm 4329.

[0197] In one form, applied to discrete or continuous-valued phases, the waveform template Π(Φ) is a square wave template with a value of 1 for phase values ​​less than 0.5 revolutions and a value of 0 for phase values ​​equal to or greater than 0.5 revolutions. In a form suitable for continuous-valued phases, the waveform template Π(Φ) has two smoothly curved sections: a smoothly curved section (e.g., a rising cosine) that rises from 0 to 1 for phase values ​​up to 0.5 revolutions, and a smooth bend (e.g., exponentially for phase values ​​above 0.5 revolutions) that decays from 1 to 0. One example of this "smooth and comfortable" waveform template is the "shark fin" waveform template, where the rise is a rising cosine and the smooth decay is subexponential (so that the limit of Π is just zero as Φ approaches one revolution).

[0198] In some forms of the present technology, the waveform determination algorithm 4322 selects a waveform template Π(Φ) from a waveform template library according to the settings of the RPT device. Each waveform template Π(Φ) in the library may be provided as a look-up table of values ​​Φ versus phase values ​​Π. In other forms, the waveform determination algorithm 4322 uses a predefined functional form (which may be parameterized by one or more parameters (e.g., a time constant of an exponential curve portion)) to "dynamically" calculate the waveform template Π(Φ). The parameters of the functional form may be predefined or may be dependent on the current state of the patient 1000.

[0199] In some forms of the present technology, applied to a discrete binary phase of inspiration (Φ=0 revolutions) or expiration (Φ=0.5 revolutions), the waveform determination algorithm 4322 calculates a waveform template Π "dynamic" as a function of the discrete phase Φ and time t measured since the most recent trigger instant. In one such form, the waveform determination algorithm 4322 calculates the waveform template Π(Φ,t) as two parts (inspiration and expiration) as follows:

[0200]

number

[0201] Here, Π i (t) and Π e (t) is the inhalation-exhalation portion of the waveform template Π(Φ,t) and Ti is the inhalation time. i (t) is the smooth rise from 0 to 1 parameterized by the rise time, which is the expiratory portion of the waveform template, Π e (t) is the smooth fall from 1 to 0 parameterized by the fall time.

[0202] (4.3.3.2.3 Inspiratory Flow Limit Determination) In one form of the present technology, the processor executes one or more algorithms 4324 for detecting inspiratory flow limitation (partial obstruction).

[0203] In one form, the inspiratory flow limitation determination algorithm 4324 receives as an input the respiratory flow signal Qr and provides as an output a measure of the extent to which the inspiratory portion of the breath exhibits inspiratory flow limitation.

[0204] In one form of the technology, the inspiration portion of each breath is identified based on the phase Φ estimated at each instant. For example, the inspiration portion of a breath is a respiratory flow value with a phase Φ of 0.5 or less. A number of equally spaced points (e.g., 65) representing time points are interpolated by an interpolator along the inspiration flow time curve of each breath. The curve described by the points is then scaled by a scaler to have unit length (period / period) and unit area to remove the effects of varying respiratory rate and depth. The scaled breaths are then compared in a comparator to a pre-stored template representing normal undisturbed breathing. Breaths that deviate from this template by more than a pre-defined threshold (usually one proportional unit) at any time during inspiration, e.g., breaths determined by the test element to be due to a cough, sigh, swallow, burp, are rejected. For data that is not rejected, a running average of the first such scaled points for the first few inspiration events is calculated by the central controller 4230. For the second such point, this is repeated for the same inspiration event, and so on. Thus, for example, 65 scaled data points are generated by the central controller 4230 to represent a moving average of several preceding inhalation events, for example, three events. The moving average of the continuously updated (for example, 65) point values ​​is hereinafter referred to as the "scaled flow rate" and identified as Qs(t). Alternatively, a single inhalation event may be used in place of the moving average.

[0205] From the scaled flow rates, two shape factors associated with determining partial occlusion can be calculated.

[0206] The shape factor 1 is the ratio of the average of the middle (e.g. 32) scaled flow points to the average of the total (e.g. 65) scaled flow points. If this ratio exceeds 1, the breath is considered normal. If the ratio is less than or equal to 1, the breath is considered obstructed. A ratio of approximately 1.17 is used as the threshold between partial obstruction and unobstructed breaths, and is equivalent to a degree of obstruction that allows adequate oxygenation to be maintained in a typical patient.

[0207] Shape factor 2 is calculated as the root mean square deviation from unity scaling flow and is taken to be above a midpoint (e.g. 32). A root mean square deviation of about 0.2 units is considered normal. A root mean square deviation of zero is considered to be a completely flow limited breath. The closer the root mean square deviation is to zero, the more flow limited the breath is.

[0208] Shape factors 1 and 2 may be used alternatively or in combination. In other forms of the present technology, the number of sampling points, breathing points, and midpoints may be different than those described above. Also, the thresholds may be different from those described.

[0209] 4.3.3.2.4 Determining Apnea and Hypopnea In one form of the present technology, the central controller 4230 executes one or more algorithms 4325 for detecting apnea and / or hypopnea.

[0210] In one form, the apnea / hypopnea decision algorithm 4325 receives as an input the respiratory flow signal Qr and provides as an output a flag indicating that an apnea or hypopnea has been detected.

[0211] In one form, apnea is detected when a function of respiratory flow Qr falls below a flow threshold for a predetermined period of time. The function may determine a peak flow, a relative short-term average flow, or a flow intermediate the relative short-term average and peak flow, e.g., RMS flow. The flow threshold may be a relatively long-term measure of flow.

[0212] In one embodiment, hypopnea is detected when a function of respiratory flow Qr falls below a second flow threshold for a predetermined period of time. The function may determine a peak flow, a relative short-term average flow, or a flow intermediate the relative short-term average and peak flow, e.g., RMS flow. The second flow threshold may be a relatively long-term measure of flow. The second flow threshold is greater than the flow threshold for detecting apnea.

[0213] In one form, such respiratory events may be characterized as central or obstructive based at least in part on the finger sensor PPG-based type detection described above.

[0214] (4.3.3.2.5 Snoring Detection) In one form of the present technology, the central controller 4230 executes one or more snore determination algorithms 4326 to determine the degree of snoring.

[0215] In one form, the snore determination algorithm 4326 receives as an input the respiratory flow signal Qr and provides as an output a measure of the degree of the presence of snoring.

[0216] The snore determination algorithm 4326 may include determining the strength of the flow signal in the range of 30-300 Hz. Additionally, the snore determination algorithm 4326 may include filtering the respiratory flow signal Qr to reduce background noise, for example the sound of airflow in the system from a blower.

[0217] (4.3.3.2.6 Airway Patency Determination) In one form of the present technology, the central controller 4230 executes one or more airway patency determination algorithms 4327 to determine the degree of airway patency.

[0218] In one form, the airway patency determination algorithm 4327 receives as input the respiratory flow signal Qr and determines the power of the signal in the frequency range of about 0.75 Hz and about 3 Hz. The appearance of peaks in this frequency range indicates an open airway. The absence of peaks is considered to indicate airway closure.

[0219] In one form, the frequency range for which the peak is sought is the frequency of a small forced oscillation in the treatment pressure Pt. In one implementation, the forced oscillation has a frequency of 2 Hz and an amplitude of approximately 1 cmH2O.

[0220] In one form, the airway patency determination algorithm 4327 receives as input the respiratory flow signal Qr and determines the presence or absence of a cardiogenic signal, the absence of which is considered an indication of airway obstruction.

[0221] 4.3.3.2.7 Determination of Treatment Parameters In some forms of the present technology, the central controller 4230 executes one or more treatment parameter determination algorithms 4329 to determine one or more treatment parameters using values ​​returned by one or more other algorithms in the therapy engine module 4320.

[0222] In one form of the present technology, the treatment parameter is the instantaneous treatment pressure Pt. In one implementation of this form, the treatment parameter determination algorithm 4329 determines the treatment pressure Pt using the formula:

[0223]

number

[0224] Where: -A is the amplitude, -Φ is the current value of the phase, -Π(Φ) is the waveform template value (range 0 to 1) at the current value of phase Φ; -P0 is the base pressure.

[0225] If the waveform determination algorithm 4322 provides the waveform template Π(Φ) as a lookup table of values ​​indexed by the phase Φ, the treatment parameter determination algorithm 4329 applies equation (1) by locating the lookup table entry closest to the current value Φ of the phase returned by the phase determination algorithm 4321, or by interpolating between two entries that intersect the current value Φ of the phase.

[0226] Depending on the respiratory pressure treatment pattern selected as described below, the treatment parameter determination algorithm 4329 may set the values ​​of amplitude A and base pressure P0 as follows:

[0227] (4.3.3.3 Therapy Control Module) A therapy control module 4330 according to one aspect of the present technology receives therapy parameters as input from a therapy parameter determination algorithm 4329 of the therapy engine module 4320 and controls the pressure generator 4140 to deliver a gas flow in accordance with the therapy parameters.

[0228] In one form of the present technology, the therapy parameter is a therapy pressure Pt, and the therapy control module 4330 controls the pressure generator 4140 to deliver an airflow such that the interface pressure Pm at the patient interface 3000 is equal to the therapy pressure Pt.

[0229] (4.3.3.4 Detecting Fault Conditions) In one form of the present technology, the processor executes one or more methods 4340 for detecting a fault condition. The fault condition detected by the one or more methods may include at least one of the following: - Power failure (no power or insufficient power) - Transducer failure detection - Inability to detect the presence of a component - Operating parameters are outside the recommended range (e.g. pressure, flow, temperature, PaO2) - The test alarm is not capable of generating a detectable alarm signal.

[0230] When a fault condition is detected, the corresponding algorithm signals the presence of a fault by one or more of the following: Initiating audio, visual, and / or dynamic (e.g. vibration) alarms - Send messages to external devices - Event logging

[0231] (4.4 Humidifier) In one form of the present technology, a humidifier 5000 (e.g., as shown in FIG. 10) is provided for modifying the absolute humidity of air or gas delivered to a patient relative to the ambient air. Typically, the humidifier 5000 is used to increase the absolute humidity and raise the temperature of the airflow (relative to the ambient air) before it is delivered to the patient's airways.

[0232] (4.5 Terminology) For purposes of this disclosure, in some aspects of the technology, one or more of the following definitions may apply. In other aspects of the technology, alternative definitions may apply.

[0233] (4.5.1 General) Air: In certain forms of the present technology, air may refer to atmospheric air, while in other forms of the present technology, air may refer to combinations of other breathable gases (e.g., oxygen-rich atmospheric air).

[0234] Respiratory Pressure Therapy (RPT): Air is delivered to the entrance of the airways at a therapeutic pressure, usually positive relative to the atmosphere.

[0235] Continuous Positive Airway Pressure (CPAP) Therapy: Respiratory pressure therapy in which the therapeutic pressure is approximately constant throughout the patient's respiratory cycle. In some forms, the pressure at the entrance to the airway increases slightly during expiration and decreases slightly during inspiration. In some forms, the pressure varies during different respiratory cycles of the patient (e.g., increased in response to detection of an indication of partial upper airway obstruction and decreased in the absence of notification of partial upper airway obstruction).

[0236] Patient: A person, whether or not suffering from a respiratory condition.

[0237] Automatic Positive Airway Pressure (APAP) Therapy: A CPAP therapy that is capable of automatically adjusting therapeutic pressure between minimum and maximum limits, for example, between breaths, depending on the presence or absence of signs of an SDB episode.

[0238] (4.5.2 Aspects of the Respiratory Cycle) Apnea: According to some definitions, apnea is said to occur when flow falls below a certain threshold for a period of time, e.g., 10 seconds. Obstructive apnea is said to occur when some airway obstruction does not allow airflow despite patient effort. Central apnea is said to refer to a condition in which apnea is detected due to reduced or absent respiratory effort despite a patent airway.

[0239] Respiratory rate: the patient's spontaneous breathing rate, usually measured in breaths per minute.

[0240] Duty cycle: The ratio of inspiration time Ti to total breathing time Ttot.

[0241] Effort (breathing): The effort made by a spontaneous breather to breathe.

[0242] Expiratory portion of the respiratory cycle: the period from the start of the expiratory flow to the start of the inspiratory flow.

[0243] Flow limitation: Flow limitation is considered to be an event in a patient's breathing where an increase in the patient's effort does not result in a corresponding increase in flow. If the flow limitation occurs during the inspiratory portion of the respiratory cycle, it can be referred to as inspiratory flow limitation. If the flow limitation occurs during the expiratory portion of the respiratory cycle, it can be referred to as expiratory flow limitation.

[0244] Hypopnea: Flow is reduced but does not cease. In one form, hypopnea is said to occur when flow is reduced below a threshold rate for a sustained period of time. In one form, hypopnea may be considered hypopnea if any of the following occur in adults: (i) A 30% decrease in patient respiration for at least 10 seconds plus an associated 4% desaturation, or

[0245] (ii) A reduction in patient respiration (but less than 50%) for at least 10 seconds and associated desaturation or agitation of at least 3%.

[0246] Inspiration portion of the respiratory cycle: The period from the start of the inspiratory flow to the start of the expiratory flow is taken as the inspiration portion of the respiratory cycle.

[0247] Patency (Airway): The degree to which the airway is open, or the extent to which the airway is open. The patient's airway is open. Airway patency may be quantified, for example, as a value of one (1) for an open state and a value of zero (0) for a closed (occluded) state.

[0248] Positive end-expiratory pressure (PEEP): The pressure in the lungs above atmosphere that exists at the end of expiration.

[0249] Peak Flow (Qpeak): The maximum value of flow during the inspiratory portion of the respiratory flow waveform.

[0250] Respiratory flow, patient airflow rate, respiratory airflow rate (Qr): These terms may be understood to refer to the estimate of respiratory flow by an RPT device, as distinct from "true respiratory flow" or "true respiratory flow", which is the actual respiratory flow experienced by the patient, usually expressed in liters per minute.

[0251] Tidal Volume (Vt): The volume of air inspired or expired during normal breathing when no extra effort is exerted.

[0252] (Inspiration) Time (Ti): The duration of the inspiratory portion of the respiratory flow waveform.

[0253] (Expiratory) Time (Te): The duration of the expiratory portion of the respiratory flow waveform.

[0254] (Total) Time (Ttot): The total duration between the start of one inspiratory portion of the respiratory flow waveform and the start of the next inspiratory portion of the respiratory flow waveform.

[0255] Upper Airway Obstruction (UAO): Includes both partial and total upper airway obstruction. This can be associated with a state of flow limitation where flow increases slightly or may even decrease as the pressure difference across the upper airway increases (Starling resistor behavior).

[0256] Vent: A measurement of the rate of gas exchange performed by a patient's respiratory system. Measurements of ventilation may include either or both inspiratory and expiratory flows per unit of time. When expressed as volume per minute, this amount is often called "minute ventilation." Minute ventilation is sometimes expressed simply as volume, understood as volume per minute.

[0257] (4.5.3 RPT Device Parameters) Flow rate: The instantaneous amount (or mass) of air delivered per unit time. Flow rate and ventilation have the same volumetric or mass dimensions per unit time, but flow rate is measured over a shorter period of time. In the example of patient breathing, flow rate may be nominally positive for the inhalation portion of the patient's breathing cycle and therefore negative for the exhalation portion of the patient's breathing cycle. In some cases, when we refer to flow rate, we refer to a scalar quantity (i.e., a quantity that has only a magnitude). In other cases, when we refer to flow rate, we refer to a vector quantity (i.e., a quantity that has both a magnitude and a direction). The flow rate is given with the symbol Q. "Flow rate" is sometimes simply referred to as "flow." Total flow rate Qt is the flow rate of air leaving the RPT device. Ventilation flow rate Qv is the flow rate of air leaving the exhaust port to expel exhaled gases. Leak flow rate Ql, is the flow rate of leak from the patient interface system or elsewhere. Respiratory flow rate Qr is the flow rate of air entering the patient's respiratory system.

[0258] Leak: The term "leak" is taken as an unintended air flow. In one example, a leak may occur as a result of an imperfect seal between the mask and the patient's face. In another example, a leak may occur at the elbow to the perimeter.

[0259] Pressure: force per unit area. Pressure is measured in cmH2O, gf / cm 2 , and a series of units including hectopascals. 1 cmH2O is 1 g-f / cm 2 , which is approximately 0.98 hectopascals. In this specification, pressures are given in units of cmH2O unless otherwise stated. The pressure at the patient interface is labeled Pm and the treatment pressure, which indicates the target value that the mask pressure Pm should currently achieve, is labeled Pt.

[0260] (4.5.4 Ventilator Terminology) Adaptive servo ventilator (ASV): A servo ventilator that has a variable, rather than fixed, target ventilation that can be learned from some characteristics of the patient, for example the patient's breathing characteristics.

[0261] Backup Rate: A ventilator parameter that sets the minimum number of breaths (typically breaths per minute) that the ventilator will deliver to the patient unless triggered by spontaneous breathing effort.

[0262] Cycled: The end of the inspiratory phase of the ventilator. When a ventilator delivers breaths to a spontaneously breathing patient, at the end of the inspiratory portion of the breathing cycle, the ventilator is said to be cycled to stop delivering breaths.

[0263] Expiratory Positive Airway Pressure (EPAP): The base pressure to which varying pressures are applied within a breath to produce the desired interface pressure the ventilator attempts to achieve at a given time.

[0264] End-Expiratory Pressure (EEP): The desired interface pressure that the ventilator attempts to achieve at the end of the expiratory portion of an exhalation. If Φ=1, then EEP is equal to EPAP when the pressure waveform template Π(Φ) is taken to be zero at the end of exhalation, i.e., Π(Φ)=0.

[0265] Inspiratory Positive Airway Pressure (IPAP): The maximum desired interface pressure that the ventilator attempts to achieve during the inspiratory portion of the breath.

[0266] Pressure Support: A numerical value indicating the rise in pressure during ventilator inspiration above the pressure during ventilator expiration, generally referring to the difference in pressure between peak and base pressure during inspiration (e.g., PS=IPAP-EPAP). In some situations, pressure support refers to the difference that the ventilator attempts to achieve, rather than the difference that it actually achieves.

[0267] Servo Ventilator: A ventilator that measures patient ventilation, has a target ventilation, and adjusts the level of pressure support to move patient ventilation toward the target ventilation.

[0268] Servo Assist: Pressure Support minus Minimum Pressure Support.

[0269] Spontaneous / Timed (S / T): A pattern of a ventilator or other device that attempts to detect the onset of a breath in a spontaneously breathing patient. If the device does not detect a breath within a predetermined period of time, the device automatically begins delivering a breath.

[0270] Swing: A term equivalent to pressure support.

[0271] Triggered: When a ventilator delivers breathing air to a spontaneously breathing patient, it is said to do so by being triggered at the start of the respiratory portion of the breathing cycle by patient effort.

[0272] Typical Recent Ventilation: Typical recent ventilation Vtyp is the value around which recent ventilation measurements over a given timescale tend to cluster, i.e. a measure of the clustering tendency of ventilation measurements over recent history.

[0273] Ventilator: A mechanical device that provides pressure support to a patient during some or all of the work of breathing.

[0274] (4.5.5 Structure of the Respiratory System) Diaphragm: sheet of muscle that spans the base of the thoracic cavity. It separates the thoracic cavity, which contains the heart, lungs, and ribs, from the abdominal cavity. When the diaphragm contracts, it increases the volume of the thoracic cavity and draws air into the lungs.

[0275] Larynx: The larynx, or voice box, houses the vocal cords and connects the lower part of the pharynx (hypopharynx) with the trachea.

[0276] Lungs: the human respiratory system. The conducting region of the lungs includes the trachea, bronchi, bronchioles and terminal bronchioles. The respiratory region includes the respiratory bronchioles, alveolar ducts and alveoli.

[0277] Nasal cavity: The nasal cavity (or nasal fossa) is a large air-filled space in the center of the face above and behind the nose. The nasal cavity is divided into two by a vertical fin called the nasal septum. On either side of the nasal cavity are three horizontal extensions called the nasal turbinates (conchae) or turbinates. At the front of the nasal cavity is the nose, which merges dorsally into the nasopharynx via the posterior nares.

[0278] Pharynx: part of the throat just below the nasal cavity and above the esophagus and larynx. The pharynx is usually divided into three parts: the epipharynx (nasal part of the pharynx), the mesopharynx (oral part of the pharynx), and the hypopharynx.

[0279] (4.6 Other Notes) A portion of the disclosure of this patent document contains material that is copyrighted. The copyright owner has no objection to anyone copying the patent document or the patent disclosure for purposes of disclosure in the Patent and Trademark Office patent file or records, but reserves all copyright rights thereto for all other purposes.

[0280] Unless otherwise clearly indicated from the context and unless a range of values ​​is provided, it is understood that each intervening value, between the upper and lower limits of the range and any other stated or intervening value in the stated range, is encompassed within the technology. The upper and lower limits of these intervening ranges, which may be independently included within the intervening ranges, are also encompassed within the technology, subject to any specifically excluded limit in the stated range. Where a stated range includes one or both of the limits, ranges excluding one or both of those included limits are also encompassed within the technology.

[0281] Furthermore, when a value or values ​​are described herein as being implemented as part of the present technology, unless otherwise noted, such values ​​may be approximate, and it is understood that such values ​​may be utilized to any appropriate significant figure to the extent that may be permitted or required in practical technical implementations.

[0282] Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this technology belongs.Any method and material similar or equivalent to the method and material described herein can be used in the practice or testing of this technology, but a limited number of exemplary methods and materials are described herein.

[0283] Although particular materials are described as being preferred for use in the construction of a component, obvious alternative materials having similar properties may be substituted. Further, unless stated to the contrary, any and all components described herein are understood to be manufacturable and therefore may be manufactured collectively or separately.

[0284] Please note that as used herein and in the appended claims, the singular forms "a," "an," and "the" include their plural equivalents unless the context clearly dictates otherwise.

[0285] All publications mentioned herein are incorporated by reference to disclose and describe the methods and / or materials that are the subject of these publications. Publications mentioned herein are provided solely for their disclosure prior to the filing date of this application. Nothing herein should be construed as an admission that the technology is not entitled to antedate such publications by prior invention. Also, the publication dates provided may be different from the actual publication dates, which may need to be independently confirmed.

[0286] Moreover, in interpreting this disclosure, all terms should be interpreted in the broadest reasonable manner according to the context. The terms "comprises" and "comprising" should be interpreted as referring to elements, components, or steps in a non-exclusive sense, indicating that a described element, component, or step may be present in, utilized with, or combined with other elements, components, or steps not specifically recited.

[0287] The headings used in the detailed description are for the convenience of the reader and should not be used to limit the content found in the disclosure or claims as a whole. These headings should not be used in interpreting the scope of the claims or the limitations of the claims.

[0288] Although the technology herein has been described with reference to specific embodiments, it should be understood that these embodiments are merely illustrative of the principles and applications of the technology. In some cases, terms and symbols may indicate specific details that are not necessary for the implementation of the technology. For example, the terms "first" and "second" may be used, but unless otherwise specified, these terms are not intended to indicate any order and may be utilized to distinguish separate elements. Furthermore, although process steps in a methodology may be described or illustrated in a sequence, such ordering is not required. Those skilled in the art will recognize that such order can be changed and / or aspects can be performed simultaneously or even synchronously.

[0289] It is therefore to be understood that numerous modifications may be made in the illustrative embodiments and other arrangements may be devised without departing from the spirit and scope of the present technology.

[0290] Although the present invention has been described with reference to specific implementations, it will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary implementations described above, and that the present invention can be embodied with various changes and modifications without departing from its scope. The present implementations are therefore considered to be illustrative and not restrictive in all respects, and the scope of the present invention is indicated by the appended claims, rather than by the foregoing description, and therefore all changes that come within the meaning and range of equivalence of the claims are intended to be embraced by the present invention. In other words, any changes, modifications, or equivalents that are within the scope of the basic underlying principles and whose essential attributes are claimed in this patent application are intended to be covered. Furthermore, readers of this patent application will understand that the terms "comprising" or "comprise" do not exclude other elements or steps, and that the terms "a" or "an" do not exclude a plurality, and that a single element, such as a computer system, a processor, or another integrated unit, may realize the functions of several means recited in the claims. Any reference signs in the claims should not be construed as limiting the respective associated claims. Terms such as "first," "second," "third," "a," "b," "c," and the like, when used in the specification or claims, are introduced to distinguish between similar elements or steps and do not necessarily describe an order or chronology. Similarly, terms such as "top," "bottom," "upper," "lower," and the like, are introduced for purposes of explanation and do not necessarily indicate a relative location. It should be understood that the terms so used are interchangeable under appropriate circumstances and that implementations of the invention are capable of operating in accordance with the invention in other sequences or with orientations other than those described or illustrated above.

[0291] (Another example of this technology) The following paragraphs provide further examples of the techniques described herein. Example 1. A computer-implemented method for detecting a sleep disorder event, comprising: - acquiring at least one physiological signal; - identifying an event indicative of a sleep disorder event from at least one physiological signal seed; - determining a substantially regular pattern within at least a portion of the seed events; - determining the extent to which the seed events conform to a substantially regular pattern; - detecting sleep disorder events by selecting seed events based on the determined fitness.

[0292] Example 2 The method of Example 1, wherein the at least one physiological signal is at least one of a PAT signal, an oxygen saturation signal, a pulse rate signal, a respiratory effort signal, and an airflow signal.

[0293] (Example 3) A method as described in any of the above examples, wherein the seed event is at least one of a decrease in PAT signal amplitude from baseline, a decrease in oxygen saturation, an increase in pulse rate signal amplitude from baseline, a decrease in respiratory effort signal amplitude from baseline, and a decrease in airflow signal amplitude from baseline.

[0294] Example 4: The method according to any of the above examples, wherein the substantially regular pattern comprises a substantially regular morphological pattern.

[0295] Example 5. The method according to any of the above examples, wherein the substantially regular pattern comprises a substantially regular temporal pattern.

[0296] Example 6: A method as described in any of the above examples, wherein the steps of determining the substantially regular pattern and determining the goodness of fit include at least one of determining a duration of a seed event, determining an intensity of a seed event, deriving a morphological asymmetry, and / or characterizing a morphological shape of the seed event.

[0297] (Example 7) A method as described in any of the above examples, wherein the steps of determining the substantially regular turns and determining the fitness include at least one of determining a start point of a seed event, determining an end point of a seed event, determining a most intense point or a least intense point, and / or determining any other characteristic point of the seed event.

[0298] Example 8. The method according to any of the above examples, wherein detecting a sleep disorder event includes using a trained or developed classifier for the detection.

[0299] Example 9: A method according to any of the above examples, wherein a plurality of physiological signals are acquired over a common time range that includes a seed event.

[0300] (Example 10) A method as described in any of the above examples, further comprising a step of grouping seed events from a plurality of physiological signals into sets, the seed events grouped into the sets being indicative of the same sleep disorder event, and the step of grouping the seed events into sets comprising a step of determining a substantially regular pattern in the co-occurrence of the seed events in the sets.

[0301] Example 11 The method of Examples 8 and 10, wherein the classifier is configured to detect a breathing related sleep disorder event or a non-breathing related sleep disorder event within the set of seed events.

[0302] (Example 12) A method according to any of the above examples, further comprising feeding back the detection of a sleep disorder event by selecting at least one seed event, the feedback being preferably presented in natural language.

[0303] (Example 13) A controller including at least one processor and at least one memory containing computer program code, and configured to use the at least one processor to cause the controller to execute a method described in any of Examples 1 to 12 above.

[0304] Example 14 A computer program product comprising computer-executable instructions for carrying out the method according to any of the preceding Examples 1-12 when the program is run on a computer.

[0305] Example 15: A computer-readable storage medium, comprising computer-executable instructions for performing the method according to any one of Examples 1-12 above, when the program is executed on a computer.

Claims

1. accessing a plurality of physiological signals generated by one or more sensors; detecting a seed event indicative of the sleep disorder event from the plurality of physiological signals; calculating a feature indicative of a pattern within a portion of the plurality of physiological signals associated with the detected seed event; applying the calculated features indicative of a pattern of seed events to a classifier, the classifier being trained to calculate a goodness of fit of the calculated features to learned patterns of sleep disorder events; outputting an identification of one or more sleep disorder events corresponding to the seed event based on the calculated goodness of fit determined by the classifier; A method implemented by one or more processors for detecting a sleep disorder event, comprising:

2. 2. The method of claim 1, wherein the classifier comprises one or more of a machine learning classifier, a decision tree model, a machine learning classifier model, a logistic regression classifier model, a neural network, a naive Bayes classifier model, and a support vector machine.

3. 3. The method of claim 1 or 2, wherein the plurality of physiological signals comprises a peripheral arterial tone (PAT) signal, and one or more of an oxygen saturation signal, a pulse rate signal, a respiratory effort signal, a motion signal, and an airflow signal.

4. Each of the seed events is determined by a decrease in amplitude from baseline in a peripheral arterial tone (PAT) signal, an increase in oxygen saturation (SpO ２ 3. The method of claim 1 or 2, wherein the changes in amplitude from baseline in the respiratory effort signal include one or more of a decrease in oxygen saturation in the pulmonary artery (PBA) signal, an increase in amplitude from baseline in the pulse rate (PR) signal, a change in amplitude from baseline in the respiratory effort signal, and a change in amplitude from baseline in the airflow signal.

5. The method of claim 1 or 2, wherein the pattern comprises a morphological pattern.

6. The method of claim 1 or 2, wherein the pattern comprises a temporal pattern.

7. 3. The method of claim 1 or 2, wherein one of the calculated features comprises one or more of the following: duration of a seed event, intensity of a seed event, a derived gradient or slope of the seed event, a morphological asymmetry derived from varying slope around a seed event, a depth of the seed event, a variance of signal amplitude of the seed event, a mean value of signal amplitude of the seed event, a degree of skewness of the seed event, a characterization of the morphological shape of the seed event.

8. 3. The method of claim 1 or 2, wherein the one or more of the calculated features comprises one or more of a determined start point of the seed event, a determined end point of the seed event, a determined point of maximum or minimum intensity, and a determined feature point of the seed event.

9. 3. The method of claim 1, wherein the detected seed events include a first seed event in a first signal of the plurality of physiological signals and a second seed event in a second signal of the plurality of physiological signals, the second signal being a different physiological signal from the first signal, and one or more of the calculated features characterize the first seed event and the second seed event.

10. 10. The method of claim 9, wherein one or more of the calculated features characterizing the first seed event and the second seed event include (a) the amount of time that a desaturation nadir follows or precedes a peak pulse rate increase and / or a PAT signal amplitude decrease, and / or (b) the time difference between a detected pulse rate (PR) surge peak and a PAT signal decrease valley.

11. 10. The method of claim 9, wherein the detected seed event includes a third seed event of the first signal, and one or more of the calculated features characterize the first seed event and the third seed event.

12. The method of claim 11 , wherein the first seed event and the third seed event comprise a pair of adjacent seed events.

13. 12. The method of claim 11, wherein one or more of the calculated features characterizing the first seed event and the third seed event include: (a) a duration between the first seed event and the third seed event; (b) a calculated stability of a period between the first seed event and the third seed event; and (c) a calculated stability of the first seed event and the third seed event.

14. 14. The method of claim 13, wherein calculating the duration comprises detecting feature points in each of the first and third seed events and determining the duration based on intervals associated with the detected feature points.

15. The method of claim 14 , wherein the detected feature points include one or more of local amplitude minima and local amplitude maxima.

16. The method of claim 13 , wherein the calculated stability is derived from a plurality of seed events and includes one or more of a depth, a mean, and a variance.

17. 12. The method of claim 11 , wherein the detected seed events include a fourth seed event of the second signal, and one or more of the calculated features characterize (a) the first and third seed events of the first signal, and (b) the second and fourth seed events of the second signal.

18. 18. The method of claim 17, wherein one or more of the calculated features characterizing (a) the first and third seed events of the first signal, and (b) the second and fourth seed events of the second signal, include a temporal correspondence of (a) a detected pulse rate (PR) peak of a seed event of a PR signal, and (b) a peripheral arterial tone (PAT) decrease to a minimum point of a valley of a seed event of a PAT signal.

19. The method of claim 1 or 2, further comprising the step of generating an output of identification information as feedback in response to a user input on a user interface for a selection of at least one seed event detected by the detection performed by the one or more processors.

20. 3. The method of claim 1 or 2, further comprising generating a signal for controlling operation of a respiratory therapy device based on the output or application.

21. 21. The method of claim 20, wherein the generating includes transmitting an identification of the one or more sleep disorder events to a remote computing system or server.

22. 21. The method of claim 20, wherein the generating includes transmitting the signal to the respiratory therapy device over a network communications link.

23. A controller comprising at least one processor and at least one memory containing processor control instructions, the at least one memory and the processor control instructions configured to cause the controller, together with the at least one processor, to perform a method according to claim 1 or 2.

24. One or more sensors; a controller comprising one or more processors and at least one memory containing processor control instructions; The controller: Accessing a plurality of physiological signals generated by one or more sensors; detecting a seed event indicative of the sleep disorder event from the plurality of physiological signals; calculating features indicative of patterns within portions of the plurality of physiological signals associated with the detected seed events; applying the calculated features indicative of the pattern of seed events to a classifier, the classifier being trained to calculate a fit of the calculated features to the learned pattern of sleep disturbance events; and outputting an identification of one or more sleep disorder events corresponding to the seed event based on the calculated goodness of fit determined by the classifier. An apparatus for detecting sleep disorder events.

25. A processor-readable storage medium comprising processor-executable instructions for performing the method of claim 1 or 2 when executed by the one or more processors.

26. 1. A processor-readable medium having stored thereon processor-executable instructions that, when executed by one or more processors, cause the one or more processors to detect a sleep disorder event, the processor-executable instructions comprising: instructions for accessing a plurality of physiological signals generated by one or more sensors; instructions for detecting a seed event indicative of the sleep disorder event from the plurality of physiological signals; instructions for calculating features indicative of a pattern within portions of the plurality of physiological signals associated with the detected seed event; instructions for applying the calculated features indicative of a pattern of seed events to a classifier, the classifier being trained to calculate a goodness of fit of the calculated features to learned patterns of sleep disorder events; outputting an identification of one or more sleep disorder events corresponding to a seed event based on the calculated goodness of fit determined by the classifier; A processor-readable medium comprising:

27. 27. The processor readable medium of claim 26, wherein the processor executable instructions further include instructions for generating a signal to control operation of a respiratory therapy device based on the output or application.

28. 30. The processor-readable medium of claim 27, wherein the control operation includes controlling a pressure or flow therapy of a blower of the respiratory therapy device.

29. A server accessible to a processor-readable medium of any one of claims 25 to 28, the server being configured to receive, via a network, a request for downloading the processor-executable instructions of the processor-readable medium to a processing device.

30. A processing device comprising one or more processors and (a) the processor-readable medium of any one of claims 25 to 28; or (b) configured to access processor-executable instructions of the processor-readable medium of any one of claims 25 to 28.

31. 31. The processing device of claim 30, which is a respiratory therapy device.

32. 32. The processing device of claim 31 configured to generate a pressure therapy or a flow therapy.

33. A method for a server to access a processor-readable medium according to any one of claims 25 to 28, comprising the steps of receiving at the server a request to download processor-executable instructions of the processor-readable medium to an electronic processing device over a network, and transmitting the processor-executable instructions to the electronic processing device in response to the request.

34. 1. A method, by one or more processors, for detecting sleep disorder breathing events, comprising: using said one or more processors to access a processor-readable medium according to any one of claims 25 to 28; and executing processor-executable instructions of the processor-readable medium on the one or more processors.

Images