Technologies for implementing respiratory therapy with digital twins

Abstract

The present technology pertains to simulating patient-specific medical conditions and treatment responses using computational modeling and / or machine learning techniques. Specifically, the present technology includes a therapy digital twin system (900) that simulates patient respiratory conditions and therapy responses. The digital twin system integrates a patient model (910) and a therapy engine (920) that iteratively adjust pressure settings based on real-world and simulated data related to therapy and respiratory events. The digital twin system also uses dynamic critical pressure signals to accurately reflect variations in airway patency influenced by sleep stage, position, and physiological factors. The digital twin system can provide personalized models for individual patients, which provides a time and resource efficient mechanism for diagnosis, treatment, and testing of therapy algorithms without requiring the presence of the patient.

Description

TECHNOLOGIES FOR IMPLEMENTING RESPIRATORY THERAPY WITH DIGITAL TWINS1 CROSS-REFERENCES TO RELATED APPLICATIONS

[0001] The present application claims the benefit of the filing date of U.S. Provisional Application No. 63 / 755,360, filed on February 7, 2025, the entire disclosure of which is hereby incorporated herein by reference.2 BACKGROUND OF THE TECHNOLOGY2.1 FIELD OF THE TECHNOLOGY

[0002] The present technology relates to one or more of the screening, diagnosis, monitoring, treatment, prevention and amelioration of respiratory-related disorders. The present technology also relates to medical devices or apparatus, and their use. The present technology also relates to the development and validation of personalized therapy models using digital twins. Specifically, the present technology pertains to systems and methods for simulating patient-specific respiratory conditions and therapy responses using computational modeling and / or machine learning techniques.2.2 DESCRIPTION OF THE RELATED ART2.2.1 Human Respiratory System and its Disorders

[0003] The respiratory system of the body facilitates gas exchange. The nose and mouth form the entrance to the airways of a patient.

[0004] The airways include a series of branching tubes, which become narrower, shorter and more numerous as they penetrate deeper into the lung. The prime function of the lung is gas exchange, allowing oxygen to move from the inhaled air into the venous blood and carbon dioxide to move in the opposite direction. The trachea divides into right and left main bronchi, which further divide eventually into terminal bronchioles. The bronchi make up the conducting airways, and do not take part in gas exchange. Further divisions of the airways lead to the respiratory bronchioles, and eventually to the alveoli. The alveolated region of the lung is where the gas exchange takes place, and is referred to as the respiratory zone. See “ Respiratory Physiology” , by John B. West, Lippincott Williams & Wilkins, 9th edition published 2012.

[0005] A range of respiratory disorders exist. Certain disorders may be characterized by particular events, e.g. apneas, hypopneas, and hyperpneas.

[0006] Examples of respiratory disorders include Obstructive Sleep Apnea (OSA), Cheyne-Stokes Respiration (CSR), respiratory insufficiency, Obesity Hypoventilation Syndrome (OHS), Chronic Obstructive Pulmonary Disease (COPD), Neuromuscular Disease (NMD) and Chest wall disorders.

[0007] OSA, a form of Sleep Disordered Breathing (SDB), is characterized by events including occlusion or obstruction of the upper air passage during sleep. It results from a combination of an abnormally small upper airway and the normal loss of muscle tone in the region of the tongue, soft palate and posterior oropharyngeal wall during sleep. The condition causes the affected patient to stop breathing for periods typically of 30 to 120 seconds in duration, sometimes 200 to 300 times per night. It often causes excessive daytime somnolence, and it may cause cardiovascular disease and brain damage. The syndrome is a common disorder, particularly in middle aged overweight males, although a person affected may have no awareness of the problem, e.g. see US Patent No. 4,944,310 (Sullivan).

[0008] Cheyne-Stokes Respiration (CSR) is another form of sleep disordered breathing. CSR is a disorder of a patient's respiratory controller in which there are rhythmic alternating periods of waxing and waning ventilation known as CSR cycles. CSR is characterized by repetitive deoxygenation and re-oxygenation of the arterial blood. It is possible that CSR is harmful because of the repetitive hypoxia. In some patients CSR is associated with repetitive arousal from sleep, which causes severe sleep disruption, increased sympathetic activity, and increased afterload, e.g. see US Patent No. 6,532,959 (Berthon-Jones).

[0009] Respiratory failure is an umbrella term for respiratory disorders in which the lungs are unable to inspire sufficient oxygen or exhale sufficient CO2 to meet the patient’s needs. Respiratory failure may encompass some or all of the following disorders.

[0010] A patient with respiratory insufficiency (a form of respiratory failure) may experience abnormal shortness of breath on exercise.

[0011] A range of therapies have been used to treat or ameliorate such conditions. Furthermore, otherwise healthy individuals may take advantage of such therapies to prevent respiratory disorders from arising. However, these have a number of shortcomings.2.2.2 Therapies

[0012] Various respiratory therapies, such as Continuous Positive Airway Pressure (CPAP) therapy, Non-invasive ventilation (NIV), Invasive ventilation (IV), and High Flow Therapy (HFT) have been used to treat one or more of the above respiratory disorders.2.2.2.1 Respiratory pressure therapies

[0013] Respiratory pressure therapy is the application of a supply of air to an entrance to the airways at a controlled target pressure that is nominally positive with respect to atmosphere throughout the patient’s breathing cycle (in contrast to negative pressure therapies such as the tank ventilator or cuirass).

[0014] Continuous Positive Airway Pressure (CPAP) therapy has been used to treat OSA. The mechanism of action is that continuous positive airway pressure acts as a pneumatic splint andmay prevent upper airway occlusion, such as by pushing the soft palate and tongue forward and away from the posterior oropharyngeal wall. Treatment of OSA by CPAP therapy may be voluntary, and hence patients may elect not to comply with therapy if they find devices used to provide such therapy one or more of: uncomfortable, difficult to use, expensive and aesthetically unappealing.

[0015] Non-invasive ventilation (NIV) provides ventilatory support to a patient through the upper airways to assist the patient breathing and / or maintain adequate oxygen levels in the body by doing some or all of the work of breathing. The ventilatory support is provided via a non-invasive patient interface. NIV has been used to treat CSR and respiratory failure, in forms such as OHS, COPD, NMD and Chest Wall disorders. In some forms, the comfort and effectiveness of these therapies may be improved.

[0016] Invasive ventilation (IV) provides ventilatory support to patients that are no longer able to effectively breathe themselves and may be provided using a tracheostomy tube or endotracheal tube. In some forms, the comfort and effectiveness of these therapies may be improved.2.2.3 Respiratory Therapy Systems

[0017] These respiratory therapies may be provided by a respiratory therapy system or device. Such systems and devices may also be used to screen, diagnose, or monitor a condition without treating it.

[0018] A respiratory therapy system may comprise a Respiratory Pressure Therapy Device (RPT device), an air circuit, a humidifier, a patient interface, an oxygen source, and data management.2.2.3.1 Patient Interface

[0019] A patient interface may be used to interface respiratory equipment to its wearer, for example by providing a flow of air to an entrance to the airways. The flow of air may be provided via a mask to the nose and / or mouth, a tube to the mouth or a tracheostomy tube to the trachea of a patient. Depending upon the therapy to be applied, the patient interface may form a seal, e.g., with a region of the patient's face, to facilitate the delivery of gas at a pressure at sufficient variance with ambient pressure to effect therapy, e.g., at a positive pressure of about 10 centimeters of water (cmH20) relative to ambient pressure. For other forms of therapy, such as the delivery of oxygen, the patient interface may not include a seal sufficient to facilitate delivery to the airways of a supply of gas at a positive pressure of about 10 cmH20. For flow therapies such as nasal HFT, the patient interface is configured to insufflate the nares but specifically to avoid a complete seal. One example of such a patient interface is a nasal cannula.2.2.3.2 Respiratory Pressure Therapy (RPT) Device

[0020] A respiratory pressure therapy (RPT) device may be used individually or as part of a system to deliver one or more of a number of therapies described above, such as by operating the device to generate a flow of air for delivery to an interface to the airways. The flow of air may be pressure-controlled (for respiratory pressure therapies) or flow-controlled (for flow therapies such as high flow therapy (HFT)). Thus, RPT devices may also act as flow therapy devices. Examples of RPT devices include a CPAP device and a ventilator.2.2.3.3 Air circuit

[0021] An air circuit is a conduit or a tube constructed and arranged to allow, in use, a flow of air to travel between two components of a respiratory therapy system such as the RPT device and the patient interface. In some cases, there may be separate limbs of the air circuit for inhalation and exhalation. In other cases, a single limb air circuit is used for both inhalation and exhalation.2.2.3.4 Humidifier

[0022] Delivery of a flow of air without humidification may cause drying of airways. The use of a humidifier with an RPT device and the patient interface produces humidified gas that minimizes drying of the nasal mucosa and increases patient airway comfort. In addition, in cooler climates, warm air applied generally to the face area in and about the patient interface is more comfortable than cold air.2.2.3.5 Vent technologies

[0023] Some forms of treatment systems may include a vent to allow the washout of exhaled carbon dioxide. The vent may allow a flow of gas from an interior space of a patient interface, e.g., the plenum chamber, to an exterior of the patient interface, e.g., to ambient.2.2.4 Critical Pressure

[0024] OSA is a form of sleep-disordered breathing characterized by repeated events of partial or complete collapse of a patient’s upper airway during sleep. This upper airway instability results in periodic hypoxic events which causes an imbalance in blood gases : reduced oxygen levels (oxygen desaturation) and increased carbon dioxide levels. In response, the patient is aroused briefly, and falls into sleep again. These repeated cycles of hypoxia and arousals not only disrupts the patient’s sleep quality and exacerbate symptoms like daytime sleepiness, but there is also a connection to a heightened risk of systemic consequences, such as cardiovascular and metabolic diseases through mechanisms of inflammation and increased sympathetic activation. The severity of OSA is typically assessed using the Apnea-Hypopnea Index (AHI), which quantifies the number of apnea (complete absence of airflow) and hypopnea (partial reduction in airflow) events occurring per hour of sleep.

[0025] One cause of OSA is the collapse of the pharyngeal airway due to anatomical and / or physiological factors. Pharyngeal collapsibility describes how likely it is for pharyngeal collapse to occur, which would then cause an apnea or hypopnea. Pharyngeal critical pressure (PCrit) is a metric used to quantify pharyngeal collapsibility. Pharyngeal Pcrit may describe the internal pressure at which the airway collapses. For a healthy individual without OSA, this would be a large negative pressure, like a vacuum which would cause the collapse. In contrast, a person with OSA would need a higher internal airway pressure to support the airway and prevent the collapse, and therefore, OSA patients typically have a higher PCrit value. In previous studies, high CPAP pressure was found to correlate with PCrit, and very high AHI indicated high PCrit values, whilst negative PCrit values were associated with low AHIs. In practice PCrit acts as a black box that encapsulates the numerous factors that can cause an apnea, including direct factors, such as pharyngeal muscle responsiveness, or indirect factors, such as body weight. Two likely factors that influence PCrit across a single night is sleep stage and body position. Both the supine body position and REM sleep stage have been shown to increase PCrit. As there is a link between PCrit and AHI, it is likely that many of the factors that increase a person’s AHI will also likely increase their PCrit.

[0026] OSA presents heterogeneously across patient populations, which may to group patients based on their specific traits (or "phenotypes"). Phenotyping enhances diagnosis by focusing on homogeneous subgroups to better understand their medical condition and prescribe the most effective treatment improving outcomes. The link between PCrit and AHI, and in turn the severity of OSA, may make PCrit a useful metric to describe a patient’s condition, such as characterizing OSA severity, and may guide phenotyping. Another advantage of using PCrit is that it can be easily measured and recorded in real time by an auto-adjusting positive airway pressure (APAP) device throughout the night.

[0027] Optimizing sleep therapy is a complex process traditionally reliant on time-consuming and resource-intensive clinical trials and physical bench testing. Thus, there is a need for an advanced computational system capable of dynamically optimizing therapy parameters based on patient-specific respiratory conditions. The present invention addresses this need by implementing a digital twin framework that models individual patient physiology and predicts therapy outcomes, streamlining treatment personalization.

[0028] 3 BRIEF SUMMARY OF THE TECHNOLOGY

[0029] The present technology is directed towards providing medical devices used in the screening, diagnosis, monitoring, amelioration, treatment, or prevention of respiratory disorders having one or more of improved comfort, cost, efficacy, ease of use and manufacturability. Thepresent technology addresses the issues discussed previously by implementing a pressure therapy digital twin framework that models individual patient conditions and predicts therapy outcomes.

[0030] A first aspect of the present technology relates to apparatus used in the screening, diagnosis, monitoring, amelioration, treatment, or prevention of a respiratory disorder.

[0031] Another aspect of the present technology relates to methods used in the screening, diagnosis, monitoring, amelioration, treatment, or prevention of a respiratory disorder.

[0032] An aspect of certain forms of the present technology is to provide a simulation framework to create digital twins for OSA patients during pressure therapy, such as PAP therapy and / or the like.

[0033] Some forms of the present technology may include a method of simulating a patient for treating respiratory-related disorders. The method may include obtaining, by one or more processors, a dataset including measured pressure data of the patient. The method may include generating, by the one or more processors, a critical pressure (Pcrit) signal based on the dataset, the Pcrit signal representing a variable Pcrit value over a period of time. The method may include generating, by the one or more processors, a flow signal based on the Pcrit signal and a first pressure value. The method may include generating, by the one or more processors, a second pressure value based on the flow signal.

[0034] In some forms, the method may include generating, by the one or more processors, a Pcrit model based on the dataset. In one form, the Pcrit signal is generated based on the Pcrit model.

[0035] In some forms, the method may include concatenating, by the one or more processors, the measured pressure data into a single pressure signal. The method may further include applying, by the one or more processors, a fast Fourier Transform to the single pressure signal to identify dominant frequency components in the single pressure signal. The method may further include performing, by the one or more processors, derivative processing on the single pressure signal to identify positive pressure increases. The method may further include generating, by the one or more processors, an impulse signal of positive pressure increases based on the identified positive pressure increases.

[0036] The method may further include segmenting, by the one or more processors, the single pressure signal into a set of segments based on the identified dominant frequency components. In one form, each segment of the set of segments has a same duration.

[0037] The method may further include fitting, by the one or more processors, an exponential distribution to interarrival times of each segment of the set of segments. In one form, fitting the exponential distribution to the interarrival times of each segment produces an array of rate parameters.

[0038] In some forms, the method may further include generating, by the one or more processors, an impulse vector based on the impulse signal of positive pressure increases and the array of rate parameters. In one form, the impulse signal of positive pressure increases is sampled by sampling with replacement to select increases that are below a threshold. In one form, the array of rate parameters is sampled by sampling with replacement.

[0039] In some forms, the method may further include generating, by the one or more processors, the Pcrit signal based on event timings of or from (e.g., associated with) the dataset by inserting impulses into the impulse vector based on event timings, wherein the event timings concern measured pressure attributable to respiratory events. In one form, the inserted impulses have a height equal to a maximum pressure at rates equal to a rate of a corresponding respiratory event (e.g., apneas).

[0040] In some forms, the flow signal is generated using a patient model. In one form, the patient model simulates human cardio-respiratory behaviors based on the first pressure value and the Pcrit signal. In one form, the patient model simulates human cardio-respiratory behaviors based on the second pressure value and the Pcrit signal. In one form, the patient model includes a plurality of compartments, each compartment of the plurality of compartments corresponds to a physiological system. In some forms, when a pressure introduced to the patient model drops below a Pcrit value of the Pcrit signal, the patient model generates the flow signal to indicate one or more respiratory events.

[0041] In some forms, the first pressure value is generated using a therapy engine model. In some forms, the second pressure value is generated using the therapy engine model. In one form, the second pressure value is larger than the first pressure value when the therapy engine model identifies at least one respiratory event in the flow signal. In some forms, the first pressure value and the second pressure value are stored as simulation data.

[0042] In some forms, the method may further include comparing, by the one or more processors, the simulation data with the measured pressure data. In some forms, the method may further include adjusting, by the one or more processors, at least one parameter for generating the Pcrit signal based on the comparison. In one form, the comparison is based on a Jensen Shannon divergence. In some implementations, the method may include iteratively applying one or more new pressure values for further generating the flow signal based on the Pcrit signal. The one or more new pressure values may be based on the flow signal from the further generating.

[0043] Some forms of the present technology may include a computer-readable medium encoded with computer-readable instructions, which when executed by one or more processors implement any of the methods described herein.

[0044] Some forms of the present technology may include a system for simulating a patient for treating respiratory-related disorders. The system may include a memory configured to store instructions related to a digital twin pipeline. The system may include at least one processor connected to the memory. The at least one processor may be configured to execute the instructions to obtain a dataset including measured pressure data of the patient; generate a critical pressure (Pcrit) signal based on the dataset, the Pcrit signal representing a variable Pcrit value over a period of time; generate a flow signal based on the Pcrit signal and a first pressure value; and generate a second pressure value based on the flow signal.

[0045] In some forms, the at least one processor is configured to execute the instructions to generate a Pcrit model based on the dataset. In one form, the Pcrit signal is generated based on the Pcrit model.

[0046] In some forms, the at least one processor is configured to execute the instructions to concatenate the measured pressure data into a single pressure signal. In one form, the at least one processor is configured to execute the instructions to apply a fast Fourier Transform to the single pressure signal to identify dominant frequency components in the single pressure signal. In one form, the at least one processor is configured to execute the instructions to perform derivative processing on the single pressure signal to identify positive pressure increases.

[0047] In some forms, the at least one processor is configured to execute the instructions to generate an impulse signal of positive pressure increases based on the identified positive pressure increases.

[0048] In some forms at least one processor is configured to execute the instructions to segment the single pressure signal into a set of segments based on the identified dominant frequency components. In one form, each segment of the set of segments has a same duration or length.

[0049] In some forms, the at least one processor is configured to execute the instructions to fit an exponential distribution to interarrival times of each segment of the set of segments. In one form, fitting the exponential distribution to the interarrival times of each segment produces an array of rate parameters.

[0050] In some forms, the at least one processor is configured to execute the instructions to generate an impulse vector based on the impulse signal of positive pressure increases and the array of rate parameters. In one form, the impulse signal of positive pressure increases is sampled by sampling with replacement to select samples of increases that are below a threshold. In one form, the array of rate parameters is sampled with replacement.

[0051] In some forms, the at least one processor is configured to execute the instructions to generate the Pcrit signal based on event timings of the dataset by inserting impulses into theimpulse vector based on the event timings, wherein the event timings concern measured pressure attributable to respiratory events. In one form, the inserted impulses have a height equal to a maximum pressure at rates equal to a rate of a corresponding respiratory event.

[0052] In some forms, the flow signal is generated using a patient model. In some forms, the patient model simulates human cardio-respiratory behaviors based on the first pressure value and the Pcrit signal. In some forms, the patient model simulates human cardio-respiratory behaviors based on the second pressure value and the Pcrit signal. In some forms, the patient model includes a plurality of compartments. In one form, each compartment of the plurality of compartments corresponds to a physiological system. In some forms, when a pressure introduced to the patient model drops below a Pcrit value of the Pcrit signal, the patient model generates the flow signal to indicate one or more respiratory events.

[0053] In some forms, the first pressure value is generated using a therapy engine model. In some forms, the second pressure value is generated using a therapy engine model. In some forms, the second pressure value is larger than the first pressure value when the therapy engine model identifies at least one respiratory event in the flow signal.

[0054] In some forms, the first pressure value and the second pressure value are stored as simulation data. In some forms, the at least one processor is configured to execute the instructions to compare the simulation data with the measured pressure data. In some forms, the at least one processor is configured to execute the instructions to adjust at least one parameter for generating the Pcrit signal based on the comparison. In one form, the comparison is based on a Jensen Shannon divergence. In some forms, the at least one processor of the system is configured to iteratively apply one or more new pressure values for further generating the flow signal based on the Pcrit signal, wherein the one or more new pressure values are based on the flow signal from the further generating

[0055] In some forms, the system includes a pressure device configured to deliver a flow of pressurised air to a patient interface that is, in use, connected to an airway of the patient. In one form, the at least one processor is a processor in a controller of a respiratory pressure therapy device.

[0056] In some forms, the system includes one or more sensors configured to monitor one or more characteristics of the pressurised air. In one form, the at least one processor is configured to execute the instructions to generate the dataset including measured pressure data of the patient based on the monitoring of the one or more characteristics of the pressurised air.

[0057] In some forms, the at least one processor is a processor of at least one server computer system. In other forms, the at least one processor is a processor of a client computer system.

[0058] Some forms of the present technology may include a system for simulating a patient for treating respiratory-related disorders, the system comprising: The system may include a memory configured to store instructions related to a digital twin pipeline. The system may include at least one processor, (e.g., one or more processors), connected to the memory. The at least one processor may be configured to execute the instructions to: obtain, from at least one medical device, time series pressure and respiratory event timing data related to a patient; operate a critical pressure (Pcrit) model to estimate a latent model related to the patient, the Pcrit model to adjust increases and decreases in auto-titrated pressure during therapy, and generate a Pcrit signal based on the latent model; operate a patient model to generate a flow signal based on a respiratory simulation of the patient, the respiratory simulation of the patient being based on the Pcrit signal and a pressure level; and operate a therapy engine model to generate a new pressure level based on the flow signal, and output new pressure level to the patient model for a next iteration of the respiratory simulation of the patient.

[0059] In some forms, the at least one processor is configured to execute the instructions to store each generated pressure level as simulation data.

[0060] In some forms, the at least one processor is configured to execute the instructions to compare the simulation data with the time series pressure and respiratory event timing data.

[0061] In some forms, the at least one processor is configured to execute the instructions to tune at least one parameter of at least one of the Pcrit model, the patient model, or the therapy engine model based on the comparison of the simulation data with the time series pressure and respiratory event timing data.

[0062] Some forms of the present technology may include a method of operating a digital twin pipeline to simulate a patient for treating respiratory-related disorders. The method may include obtaining, by at least one processor from at least one medical device, time series pressure and respiratory event timing data related to a patient. The method may include operating, by the at least one processor, a critical pressure (Pcrit) model to estimate a latent model related to the patient, the Pcrit model to adjust increases and decreases in auto-titrated pressure during therapy, and generate a Pcrit signal based on the latent model. The method may include operating, by the at least one processor, a patient model to generate a flow signal based on a respiratory simulation of the patient, the respiratory simulation of the patient being based on the Pcrit signal and a pressure level. The method may include operating, by the at least one processor, a therapy engine model to generate a new pressure level based on the flow signal, and output new pressure level to the patient model for a next iteration of the respiratory simulation of the patient.

[0063] In some forms, the method further includes storing, by the at least one processor, each generated pressure level as simulation data.

[0064] In some forms, the method further includes comparing, by the at least one processor, the simulation data with the time series pressure and respiratory event timing data.

[0065] In some forms, the method further includes tuning, by the at least one processor, at least one parameter of at least one of the Pcrit model, the patient model, or the therapy engine model based on the comparison of the simulation data with the time series pressure and respiratory event timing data.

[0066] Some forms of the present technology may include a computer-readable medium encoded with computer-readable instructions, which when executed by one or more processors implement any of the methods described herein.

[0067] The methods, systems, devices and apparatus described may be implemented so as to improve the functionality of a processor, such as a processor of a specific purpose computer, respiratory monitor and / or a respiratory therapy apparatus. Moreover, the described methods, systems, devices and apparatus can provide improvements in the technological field of automated management, monitoring, and / or treatment of respiratory conditions, including, for example, sleep disordered breathing.

[0068] Portions of the aspects may form sub-aspects of the present technology. Also, various ones of the sub-aspects and / or aspects may be combined in various manners and also constitute additional aspects or sub-aspects of the present technology. Other features of the technology will be apparent from consideration of the information contained in the following detailed description, abstract, drawings and claims.4 BRIEF DESCRIPTION OF THE DRAWINGS

[0069] The present technology is illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings, in which like reference numerals refer to similar elements including:4.1 RESPIRATORY THERAPY SYSTEMS

[0070] Fig. 1A shows a system including a patient 1000 wearing a patient interface 3000, in the form of nasal pillows, receiving a supply of air at positive pressure from an RPT device 4000. Air from the RPT device 4000 is humidified in a humidifier 5000, and passes along an air circuit 4170 to the patient 1000. A bed partner 1100 is also shown. The patient is sleeping in a supine sleeping position.

[0071] Fig. IB shows a system including a patient 1000 wearing a patient interface 3000, in the form of a nasal mask, receiving a supply of air at positive pressure from an RPT device 4000. Air from the RPT device is humidified in a humidifier 5000, and passes along an air circuit 4170 to the patient 1000.

[0072] Fig. 1C shows a system including a patient 1000 wearing a patient interface 3000, in the form of a full-face mask, receiving a supply of air at positive pressure from an RPT device 4000. Air from the RPT device is humidified in a humidifier 5000, and passes along an air circuit 4170 to the patient 1000. The patient is sleeping in a side sleeping position.4.2 RESPIRATORY SYSTEM AND FACIAL ANATOMY

[0073] Fig. 2A shows an overview of a human respiratory system including the nasal and oral cavities, the larynx, vocal folds, oesophagus, trachea, bronchus, lung, alveolar sacs, heart and diaphragm.

[0074] Fig. 2B shows a view of a human upper airway including the nasal cavity, nasal bone, lateral nasal cartilage, greater alar cartilage, nostril, lip superior, lip inferior, larynx, hard palate, soft palate, oropharynx, tongue, epiglottis, vocal folds, oesophagus and trachea.4.3 PATIENT INTERFACE

[0075] Fig. 3 shows an example patient interface in the form of a nasal mask in accordance with one form of the present technology.4.4 RPT DEVICE

[0076] Fig. 4A shows an RPT device in accordance with one form of the present technology.

[0077] Fig. 4B is a schematic diagram of the pneumatic path of an RPT device in accordance with one form of the present technology. The directions of upstream and downstream are indicated with reference to the blower and the patient interface. The blower is defined to be upstream of the patient interface and the patient interface is defined to be downstream of the blower, regardless of the actual flow direction at any particular moment. Items which are located within the pneumatic path between the blower and the patient interface are downstream of the blower and upstream of the patient interface.

[0078] Fig. 4C is a schematic diagram of the electrical components of an RPT device in accordance with one form of the present technology.

[0079] Fig. 4C-1 is a schematic diagram illustrating the interconnection of various electrical components of the RPT device.

[0080] Fig. 4D is a schematic diagram of the processing implemented in an RPT device in accordance with one form of the present technology.

[0081] Fig. 4E is a flow chart illustrating a method carried out by the therapy engine module of Fig. 4D in accordance with one form of the present technology.4.5 HUMIDIFIER

[0082] Fig. 5A shows an isometric view of a humidifier in accordance with one form of the present technology.

[0083] Fig. 5B shows an isometric view of a humidifier in accordance with one form of the present technology, showing a humidifier reservoir 5110 removed from the humidifier reservoir dock 5130.4.6 BREATHING WAVEFORMS

[0084] Fig. 6 shows a model typical breath waveform of a person while sleeping.4.7 SCREENING, DIAGNOSIS AND MONITORING SYSTEMS

[0085] Fig. 7 is a schematic diagram of the components of a screening / diagnosis / monitoring device that may be used to implement a respiratory polygraphy (RPG) headbox in an RPG screening / diagnosis / monitoring system concentrator in accordance with one form of the present technology.4.8 THERAPEUTIC PRESSURE WAVEFORMS

[0086] Fig. 8A shows a waveform of therapeutic pressure applied to a sleeping patient who is undergoing PAP therapy during a sleeping session.

[0087] Fig. 8B shows a waveform of therapeutic pressure applied to a sleeping patient who is having trouble tolerating their PAP therapy patient interface during a sleeping session.

[0088] Fig. 8C shows several waveforms of therapeutic pressure from subsequent sleep sessions of a patient whose PAP therapeutic pressure settings may be adjusted to make the therapy more effective.4.9 PRESSURE THERAPY DIGITAL TWINS

[0089] Fig. 9 depicts an example digital twin pipeline in accordance with aspects of the present technology.

[0090] Fig. 10 depicts an example patient model with various physiological compartments in accordance with aspects of the present technology.

[0091] Fig. 11 depicts a graph showing that, when PCrit is set to 8 cmH20, a CPAP of 8 cmH20 or more is needed to maintain the patient’s oxygen saturation, otherwise oxygen desaturation occurs.

[0092] Fig. 12 shows an example patient model simulation simulating when a pressure value is fixed below a Pcrit value for both flow limitation and apnea.

[0093] Fig. 13 depicts an example Pcrit signal generation pipeline in accordance with aspects of the present technology.

[0094] Fig. 14 depicts an example Pcrit signal in accordance with aspects of the present technology.

[0095] Fig. 15 depicts accuracy metrics AJS, calculated using Jensen Shannon (JS) divergence shows that data from a participant's digital twin was generally most similar to matched counterpart than other participants in the dataset.

[0096] Fig. 16 provides a comparison for three different patients and their digital twin counterparts, with each panel showing the collected and simulated pressure time series of different nights of data. This figure highlights the modelling pipeline’s ability to capture common temporal characteristics at difference scales of granularity which differ significantly across patients

[0097] Fig. 17 depicts an example digital twin process that may be used to practice aspects of the present technology.5 DETAILED DESCRIPTION OF EXAMPLES OF THE TECHNOLOGY

[0098] Before the present technology is described in further detail, it is to be understood that the technology is not limited to the particular examples described herein, which may vary. 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.

[0099] The following description is provided in relation to various examples which may share one or more common characteristics and / or features. It is to be understood that one or more features of any one example may be combinable with one or more features of another example or other examples. In addition, any single feature or combination of features in any of the examples may constitute a further example.5.1 THERAPY

[0100] In one form, the present technology comprises a method for treating a respiratory disorder comprising applying positive pressure to the entrance of the airways of a patient 1000.

[0101] In certain examples of the present technology, a supply of air at positive pressure is provided to the nasal passages of the patient via one or both nares.

[0102] In certain examples of the present technology, mouth breathing is limited, restricted or prevented.5.2 RESPIRATORY THERAPY SYSTEMS

[0103] In one form, the present technology comprises a respiratory therapy system for treating a respiratory disorder. The respiratory therapy system may comprise an RPT device 4000 for supplying a flow of air to the patient 1000 via an air circuit 4170 and a patient interface 3000 or 3800.5.3 PATIENT INTERFACE

[0104] A non-invasive patient interface 3000, such as that shown in Fig. 3, in accordance with one aspect of the present technology comprises the following functional aspects: a seal-forming structure 3100, a plenum chamber 3200, a positioning and stabilizing structure 3300, a vent 3400, one form of connection port 3600 for connection to air circuit 4170, and a forehead support 3700. In some forms a functional aspect 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 arranged to surround an entrance to the airways of the patient so as to maintain positive pressure at the entrance(s) to the airways of the patient 1000. The sealed patient interface 3000 is therefore suitable for delivery of positive pressure therapy.5.4 RPT DEVICE

[0105] An RPT device 4000 in accordance with one aspect of the present technology comprises mechanical, pneumatic, and / or electrical components and is configured to execute one or more algorithms 4300, such as any of the methods, in whole or in part, described herein. The RPT device 4000 may be configured to generate a flow of air for delivery to a patient’s airways, such as to treat one or more of the respiratory conditions described elsewhere in the present document.

[0106] In one form, the RPT device 4000 is constructed and arranged to be capable of delivering a flow of air in a range of -20 L / min to +150 L / min while maintaining a positive pressure of at least 4 cm FEO, or at least 10 cmEEO, or at least 20 cmEEO.

[0107] The RPT device may have an external housing 4010, formed in two parts, an upper portion 4012 and a lower portion 4014. Furthermore, the external housing 4010 may include one or more panel(s) 4015. The RPT device 4000 comprises a chassis 4016 that supports one or more internal components of the RPT device 4000. The RPT device 4000 may include a handle 4018.

[0108] The pneumatic path of the RPT device 4000 may comprise one or more air path items, e.g., an inlet air filter 4112, an inlet muffler 4122, a pressure generator 4140 capable of supplying air at positive pressure (e.g., a blower 4142), an outlet muffler 4124 and one or more transducers 4270, such as pressure sensors 4272 and flow rate sensors 4274.

[0109] One or more of the air path items may be located within a removable unitary structure which will be referred to as a pneumatic block 4020. The pneumatic block 4020 may be located within the external housing 4010. In one form a pneumatic block 4020 is supported by, or formed as part of the chassis 4016.

[0110] As shown in Fig. 4C, the RPT device 4000 may have an electrical 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, memory 4260, transducers 4270, data communication interface 4280 and one or more output devices 4290. Electrical components 4200 may be mounted on a single Printed Circuit Board Assembly (PCBA) 4202. In an alternative form, the RPT device 4000 may include more than one PCBA 4202.5.4.1 RPT device mechanical & pneumatic components[Oil 1] An RPT device may comprise one or more of the following components in an integral unit. In an alternative form, one or more of the following components may be located as respective separate units.5.4.1.1 Air filter (s)

[0112] An RPT device in accordance with one form of the present technology may include an air filter 4110, or a plurality of air filters 4110.

[0113] In one form illustrated in Fig. 4B, an inlet air filter 4112 is located at the beginning of the pneumatic path upstream of a pressure generator 4140.

[0114] In one form illustrated in Fig. 4B, an outlet air filter 4114, for example an antibacterial filter, is located between an outlet of the pneumatic block 4020 and a patient interface 3000 or 3800.5.4.1.2 Muffler(s)

[0115] An RPT device in accordance with one form of the present technology may include a muffler 4120, or a plurality of mufflers 4120.

[0116] In one form of the present technology (see e.g., Fig. 4B), an inlet muffler 4122 is located in the pneumatic path upstream of a pressure generator 4140.

[0117] In one form of the present technology, an outlet muffler 4124 is located in the pneumatic path between the pressure generator 4140 and a patient interface 3000 or 3800.5.4.1.3 Pressure generator

[0118] In one form of the present technology, a pressure generator 4140 for producing a flow, or a supply, of air at positive pressure is a controllable blower 4142. For example, the blower 4142 may include a brushless DC motor 4144 with one or more impellers. The impellers may be located in a volute. The blower may be capable of delivering a supply of air, for example at a rate of up to about 120 litres / minute, at a positive pressure in a range from about 4 cm FEO to about 20 cm FEO, or in other forms up to about 30 cmFEO when delivering respiratory pressure therapy. The blower may be as described in any one of the following patents or patent applications the contents of which are incorporated herein by reference in their entirety: U.S. PatentNo. 7,866,944; U.S. Patent No. 8,638,014; U.S. Patent No. 8,636,479; and PCT Patent Application Publication No. WO 2013 / 020167.

[0119] The pressure generator 4140 may be under the control of the therapy device controller 4240.

[0120] In other forms, a pressure generator 4140 may be a piston-driven pump, a pressure regulator connected to a high pressure source (e.g. compressed air reservoir), or a bellows.5.4.1.4 Transducer(s)

[0121] Transducers may be internal of the RPT device, or external of the RPT device. External transducers may be located for example on or form part of the air circuit, e.g., the patient interface. External transducers may be in the form of non-contact sensors such as a Doppler radar movement sensor that transmit or transfer data to the RPT device.

[0122] In one form of the present technology (see e.g., Fig. 4B), one or more transducers 4270 are located upstream and / or downstream of the pressure generator 4140. The one or more transducers 4270 may be constructed and arranged to generate signals representing properties of the flow of air such as a flow rate, a pressure or a temperature at that point in the pneumatic path.

[0123] In one form of the present technology, one or more transducers 4270 may be located proximate to the patient interface 3000 or 3800.

[0124] In one form, a signal from a transducer 4270 may be filtered, such as by low-pass, high-pass or band-pass filtering.5.4.1.4.1 Flow rate sensor

[0125] A flow rate sensor 4274 in accordance with the present technology may be based on a differential pressure transducer, for example, an SDP600 Series differential pressure transducer from SENSIRION.

[0126] In one form, a signal generated by the flow rate sensor 4274 and representing a flow rate is received by the central controller 4230.5.4.1.4.2 Pressure sensor

[0127] A pressure sensor 4272 in accordance with the present technology is located in fluid communication with the pneumatic path. An example of a suitable pressure sensor is a transducer from the HONEYWELL ASDX series. An alternative suitable pressure sensor is a transducer from the NPA Series from GENERAL ELECTRIC.

[0128] In one form, a signal generated by the pressure sensor 4272 and representing a pressure is received by the central controller 4230.5.4.1.4.3 Motor speed transducer

[0129] In one form of the present technology a motor speed transducer 4276 is used to determine a rotational velocity of the motor 4144 and / or the blower 4142. A motor speed signal from the motor speed transducer 4276 may be provided to the therapy device controller 4240. The motor speed transducer 4276 may, for example, be a speed sensor, such as a Hall effect sensor.5.4.1.5 Anti-spill back valve

[0130] As shown in Fig. 4B, one form of the present technology, an anti-spill back valve 4160 is located between the humidifier 5000 and the pneumatic block 4020. The anti-spill back valve is constructed and arranged to reduce the risk that water will flow upstream from the humidifier 5000, for example to the motor 4144.5.4.2 RPT device electrical components5.4.2.1 Power supply

[0131] A power supply 4210 may be located internal or external of the external housing 4010 of the RPT device 4000.

[0132] In one form of the present technology, power supply 4210 provides electrical power to the RPT device 4000 only. In another form of the present technology, power supply 4210 provides electrical power to both RPT device 4000 and humidifier 5000.

[0133] As illustrated in Fig. 4C-1, the power supply 4210 may provide electrical power to the input device 4220, the central controller 4230, the output device 4290, and the pressure generator 4140. The power supply 4210 may also provide electric energy to other components of the RPT device 4000 (or the humidifier 5000, as described above).5.4.2.1 Input devices

[0134] In one form of the present technology, an RPT device 4000 includes one or more input devices 4220 in the form of buttons, switches or dials to allow a person to interact with the device. The buttons, switches or dials may be physical devices, or software devices accessible via a touch screen. The buttons, switches or dials may, in one form, be physically connected to the external housing 4010, or may, in another form, be in wireless communication with a receiver that is in electrical connection to the central controller 4230.

[0135] In one form, the input device 4220 may be constructed and arranged to allow a person to select a value and / or a menu option.5.4.1.3 Central controller

[0136] In one form of the present technology, the central controller 4230 is one or a plurality of processors suitable to control an RPT device 4000. The central controller 4230 is show in Figs.4C and 4C-1.

[0137] Suitable processors may include an x86 instruction set architecture (ISA) processor such as any of those provided by Intel Corp, or Advanced Micro Devices, Inc. (AMD). Additional or alternative suitable processors may include a reduced instruction set computer (RISC) ISA-based processor such as an ARM® Cortex®-M processor from ARM Holdings. Examples of such an ARM® Cortex®-M processor may include a 32-bit RISC CPU such as an STM32 series microcontroller from STMicroelectronics or a 16-bit RISC CPU such as a processor from the MSP430 family of microcontrollers provided by Texas Instruments Inc.

[0138] In one form of the present technology, the central controller 4230 is a dedicated electronic circuit. For example, the central controller 4230 may be embodied as an applicationspecific integrated circuit (ASIC), field-programmable gate arrays (FPGA), digital signal processors (DSP), System on Chip (SoC), and / or other like device. In another form, the central controller 4230 comprises discrete electronic components. In some forms, the processors of the central controller 4230 may be located on the PCBA 4202. In another form, the central controller 4230 comprises discrete electronic components.

[0139] The central controller 4230 may be configured to receive input signal(s) from one or more transducers 4270, one or more input devices 4220, and / or the humidifier 5000, among others. The central controller 4230 may be configured to provide output signal(s) to one or more of an output device 4290, a pressure generator 4140, a therapy device controller 4240, a data communication interface 4280, and / or the humidifier 5000.

[0140] In some forms of the present technology, the central controller 4230 is configured to implement the one or more methodologies described herein, such as the one or more algorithms 4300 and / or the digital twin framework discussed infra with respect to Figs. 9-17, which may be implemented with processor-control instructions, expressed as computer programs stored in a non-transitory computer readable storage medium, such as memory 4260. In some forms of the present technology, the central controller 4230 may be integrated with an RPT device 4000. However, in some forms of the present technology, some methodologies may be performed by a remotely located device. For example, the remotely located device may determine control settings for a ventilator other respiratory therapy device such as an RPT device, or detect respiratory related events by analysis of stored data such as from any of the sensors described herein. In various forms of the present technology, the central controller 4230 may include or otherwise correspond to a control system configured for operations as discussed infra with respect to Figs. 8A, 8B and 8C. In some versions, one or more processors may be implemented with one or more servers or other computing devices, that may receive data from one or more RPT devices. The one or more processors, such as of one or more servers, may implement any one or more, or all, of the processes, models and / or machine learning of the digital twin framework(s) described herein.5.4.1.4 Clock

[0141] The RPT device 4000 may include a clock 4232 that is connected to the central controller 4230. The clock 4232 may be a physical device that is capable of providing a measurement of the passage of time, such as an atomic clock or clock generator. In one form, the clock 4232 may include a crystal oscillator or piezoelectric resonator such as quartz, polycrystalline ceramics, thin-film resonators, and / or the like.5.4.1.5 Therapy device controller

[0142] In one form of the present technology, therapy device controller 4240 is a therapy control module 4330 that forms part of the algorithms 4300 executed by the central controller 4230.

[0143] In one form of the present technology, therapy device controller 4240 is a dedicated motor control integrated circuit (IC), such as a motor bridge or three-phase inverter which is sometimes referred to as a triple half-bridge circuit, pulse-width modulation (PWM) controller, servo controller, stepper motor controller, variable frequency drive (VFD), and brushless DC(BLDC) motor controller, motor controller, among many others. In one example, therapy device controller 4240 is embodied as a MC33035 brushless DC motor controller manufactured by ONSEMI®. In another example, therapy device controller 4240 may be embodied as an L298N Dual Full Bridge Motor Driver manufactured by STMicroelectronics®. In some implementations, the therapy device controller 4240 may be implemented by a DRV8323x from TEXAS INSTRUMENTS. In some implementations, the therapy device controller may be implemented with integrated or external components that serve as an inverter bridge for the motor.5.4.1.6 Protection circuits

[0144] The one or more protection circuits 4250 in accordance with the present technology may comprise an electrical protection circuit, a temperature and / or pressure safety circuit.5.4.1.7 Memory

[0145] In accordance with one form of the present technology the RPT device 4000 includes memory 4260. The memory 4260 may be of any type capable of storing information accessible by a processor, such as the central controller 4230 or therapy device controller 4240, or other computing devices / components. Memory 4260 may be located on the PCBA 4202.

[0146] The memory 4260 may be a non -transitory medium, such as random-access memory (RAM), read only memory (ROM), a hard drive, memory card, optical disk, solid state drive, and / or other types of memory. In some forms, memory 4260 may include battery powered static RAM, dynamic RAM (DRAM), synchronous DRAM (SDRAM), and / or the like. In some forms, memory 4260 may include volatile RAM. In some forms, memory 4260 may be in the form of EEPROM and / or NAND flash. Additionally, or alternatively, RPT device 4000 includes a removable form of memory 4260, for example, a memory card made in accordance with the Secure Digital (SD) standard. Additionally, the memory 4260 may also include a memory controller, such as an integrated memory controller (IMC), northbridge-based memory controller, DDR memory controller, and the like, for interfacing with the processor(s) 202.

[0147] In one form of the present technology, the memory 4260 acts as a non -transitory computer readable storage medium on which is stored computer program instructions expressing the one or more methodologies described herein, such as the one or more algorithms 4300. The memory 4260 may include different combinations of the foregoing, whereby different portions of instructions and data are stored on different types of media. The instructions may be any set of instructions to be executed directly (such as machine code) or indirectly (such as scripts) by the processor(s). For example, the instructions may be stored as computing device code on the computing device-readable medium. In this regard, the terms “instructions”, “modules”, “programs”, and the like may be used interchangeably herein. The instructions may be stored in object code format for direct processing by the processor, or in any other computing devicelanguage including scripts or collections of independent source code modules that are interpreted on demand or compiled in advance.5.4.1.8 Data communication systems

[0148] In one form of the present technology, a data communication interface 4280 is provided, and is connected to the central controller 4230 (see e.g., Fig. 4C). Data communication interface 4280 may be connectable to a remote external communication network 4282 and / or a local external communication network 4284. The remote external communication network 4282 may be connectable to a remote external device 4286. The local external communication network 4284 may be connectable to a local external device 4288.

[0149] In one form, data communication interface 4280 is part of the central controller 4230. In another form, data communication interface 4280 is separate from the central controller 4230, and may comprise an integrated circuit or a processor.

[0150] In one form, remote external communication network 4282 is the Internet. The data communication interface 4280 may use wired communication (e.g. via Ethernet, or optical fibre) or a wireless protocol (e.g. CDMA, GSM, LTE) to connect to the Internet.

[0151] In one form, local external communication network 4284 utilises one or more communication standards, such as Bluetooth, or a consumer infrared protocol.

[0152] In one form, remote external device 4286 is one or more computers, such as one or more servers, for example a cluster of networked computers. In one form, remote external device 4286 may be virtual computers, rather than physical computers. In either case, such a remote external device 4286 may be accessible to an appropriately authorised person such as a clinician.

[0153] The local external device 4288 may be a personal computer, mobile phone, tablet or remote control.5.4.1.9 Output devices including optional display, alarms

[0154] An output device 4290 in accordance with the present technology may take the form of one or more of a visual, audio, and / or haptic unit.5.4.2.9.1 Display driver

[0155] A display driver 4292 receives as an input the characters, symbols, or images intended for display on the display 4294, and converts them to commands that cause the display 4294 to display those characters, symbols, or images.5.4.2.9.2 Display

[0156] A visual display device, such as 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 a segment display, such as an eight-segment display in which case the display driver 4292 converts each character or symbol, such as the figure “0”, toeight logical signals indicating whether the eight respective segments are to be activated to display a particular character or symbol.

[0157] In other examples, the display 4294 may be a liquid crystal display (LCD), light emitting diode (LED) display, quantum dot display, electroluminescent display (ELD), electronic ink (e-ink), and / or any other type of display devices. In these examples, the display 4294 may also be a touchscreen input device. Other visual output devices may also be included, such as individual LED indicators and the like.5.4.2.10 Interconnect

[0158] Although not shown, the components of the RPT device 4000 may be connected to one another using one or more interconnect technologies (IX). The IX may be a communication pathway that facilitates data transfer and connection between different components, circuits, devices, and / or systems in electronic and computing architectures, such as RPT device 4000. The IX may be capable of supporting various communication protocols and design configurations.

[0159] The IX may comprise one or more electrical interconnects (e.g., conductors such as metal, metal alloys, graphene, etc.), optical interconnects (e.g., fiber optics), on-chip or network-on-chip (NoC) interconnects, chip-to-chip interconnects, backplane interconnects, bus interconnects or bus systems, point-to-point interconnects, parallel interconnects, serial interconnects, and / or some other interconnection means that permit communication among the various components of the RPT device 4000. The IX may include any number of interconnect and / or interface technologies such as, for example, Inter-Integrated Circuit (I2C), Inter-Integrated Circuit Sound (I2S), serial peripheral interface (SPI), Serial Advanced Technology Attachment (SATA), peripheral component interconnect (PCI) such as PCI express (PCIe) or PCI extended (PCIx), Universal Asynchronous Receiver / Transmitter (UART), Advanced Microcontroller Bus Architecture (AMBA), Mobile Industry Processor Interface (MIPI) 13 C, General -Purpose Input / Output (GPIO), Low-Voltage Differential Signaling (LVDS), Pulse-Density Modulation (PDM) bus, PWM bus, Universal Serial Bus (USB), HyperTransport, controller area network (CAN) bus, PROFIBUS, Infinity Fabric (IF), power management bus (PMBus), and / or any number of other bus or interconnect technologies including proprietary buses or interconnects.5.4.3 RPT device algorithms

[0160] As mentioned previously, in some forms of the present technology, the central controller 4230 may be configured to implement one or more processes according to algorithms 4300 that are expressed as computer programs stored in a non-transitory computer readable storage medium, such as memory 4260. The algorithms 4300 are generally grouped into groups referred to as modules, models, engines, components, scripts, methods, functions, and / or the like. In some forms of the present technology, at least some of the algorithms 4300 may be implemented asDynamic Link Library (DLL) integrated or called by Python scripts within a Python programming environment.

[0161] In other forms of the present technology, some portion or all of the processes of algorithms 4300 may be implemented by a controller of an external device such as the local external device 4288 or the remote external device 4286. In such forms, data representing the input signals and / or intermediate algorithm outputs necessary for the portion of the processes of algorithms 4300 to be executed at the external device may be communicated to the external device via the local external communication network 4284 or the remote external communication network 4282. In such forms, the portion of the processes of algorithms 4300 to be executed at the external device may be expressed as computer programs, such as with processor control instructions to be executed by one or more processor(s), stored in a non-transitory computer readable storage medium accessible to the controller of the external device. Such programs configure the controller of the external device to execute the portion of the processes of algorithms 4300.

[0162] In such forms, the therapy parameters generated by the external device via the therapy engine module 4320 (if such forms part of the portion of the algorithms 4300 executed by the external device) may be communicated to the central controller 4230 to be passed to the therapy control module 4330.5.4.3.1 Pre-processing module

[0163] A pre-processing module 4310 in accordance with one form of the present technology receives as an input a signal from a transducer 4270, for example a flow rate sensor 4274 or pressure sensor 4272, and performs one or more process steps to calculate one or more output values that will be used as an input to another module, for example a therapy engine module 4320.

[0164] In one form of the present technology, the output values include the interface pressure Pm, the vent flow rate Qv, the respiratory flow rate Qr, and the leak flow rate QI.

[0165] In various forms of the present technology, the pre-processing module 4310 comprises one or more of the following algorithms: interface pressure estimation 4312, vent flow rate estimation 4314, leak flow rate estimation 4316, and respiratory flow rate estimation 4318.5.4.3.1.1 Interface pressure estimation

[0166] In one form of the present technology, an interface pressure estimation algorithm 4312 receives as inputs a signal from the pressure sensor 4272 indicative of the pressure in the pneumatic path proximal to an outlet of the pneumatic block (the device pressure Pd) and a signal from the flow rate sensor 4274 representative of the flow rate of the airflow leaving the RPT device 4000 (the device flow rate Qd). The device flow rate Qd, absent any supplementary gas 4180, may be used as the total flow rate Qt. The interface pressure algorithm 4312 estimates the pressure drop AP through the air circuit 4170. The dependence of the pressure drop AP on the total flow rate Qtmay be modelled for the particular air circuit 4170 by a pressure drop characteristic AP(Q). The interface pressure estimation algorithm, 4312 then provides as an output an estimated pressure, Pm, in the patient interface 3000 or 3800. The pressure, Pm, in the patient interface 3000 or 3800 may be estimated as the device pressure Pd minus the air circuit pressure drop AP.5.4.3.1.2 Vent flow rate estimation

[0167] In one form of the present technology, a vent flow rate estimation algorithm 4314 receives as an input an estimated pressure, Pm, in the patient interface 3000 or 3800 from the interface pressure estimation algorithm 4312 and estimates a vent flow rate of air, Qv, from a vent 3400 in a patient interface 3000 or 3800. The dependence of the vent flow rate Qv on the interface pressure Pm for the particular vent 3400 in use may be modelled by a vent characteristic Qv(Pm).5.4.3.1.3 Leak flow rate estimation

[0168] In one form of the present technology, a leak flow rate estimation algorithm 4316 receives as an input a total flow rate, Qt, and a vent flow rate Qv, and provides as an output an estimate of the leak flow rate QI. In one form, the leak flow rate estimation algorithm estimates the leak flow rate QI by calculating an average of the difference between total flow rate Qt and vent flow rate Qv over a period sufficiently long to include several breathing cycles, e.g. about 10 seconds.

[0169] In one form, the leak flow rate estimation algorithm 4316 receives as an input a total flow rate Qt, a vent flow rate Qv, and an estimated pressure, Pm, in the patient interface 3000 or 3800, and provides as an output a leak flow rate QI, by calculating a leak conductance, and determining a leak flow rate QI to be a function of leak conductance and pressure, Pm. Leak conductance is calculated as the quotient of low pass filtered non-vent flow rate equal to the difference between total flow rate Qt and vent flow rate Qv, and low pass filtered square root of pressure Pm, where the low pass filter time constant has a value sufficiently long to include several breathing cycles, e.g. about 10 seconds. The leak flow rate QI may be estimated as the product of leak conductance and a function of pressure, Pm.5.4.3.1.4 Respiratory flow rate estimation

[0170] In one form of the present technology, a respiratory flow rate estimation algorithm 4318 receives as an input a total flow rate, Qt, a vent flow rate, Qv, and a leak flow rate, QI, and estimates a respiratory flow rate of air, Qr, to the patient, by subtracting the vent flow rate Qv and the leak flow rate QI from the total flow rate Qt.5.4.3.1 Therapy Engine Module

[0171] In one form of the present technology, a therapy engine module 4320 receives as inputs, such as any one or more of a pressure, Pm, in a patient interface 3000 or 3800, and a respiratory flow rate of air to a patient, Qr, and provides as an output one or more therapyparameters. In some, the therapy parameters include one or more of a treatment pressure Pt, an amplitude of a pressure variation, a base pressure, a target ventilation, a critical pressure (Pcrit), and / or leak parameters, or other therapy parameter concerning a provision of a therapy, such as a flow rate parameter of a flow therapy such as a high flow therapy.

[0172] In various forms, the therapy engine module 4320 comprises one or more of the following algorithms: phase determination 4321, waveform determination 4322, ventilation determination 4323, inspiratory flow limitation determination 4324, apnea / hypopnea determination 4325, snore determination 4326, airway patency determination 4327, target ventilation determination 4328, and therapy parameter determination 4329. In some forms, the therapy engine module 4320 may also include, or correspond with, any of the aspects of the processes of the digital twin framework described herein, such as those discussed with respect to Figs. 9-17.5.4.3.2.1 Phase determination

[0173] In one form of the present technology, the RPT device 4000 does not determine phase.

[0174] In one form of the present technology, a phase determination algorithm 4321 receives as an input a signal indicative of respiratory flow rate, Qr, and provides as an output a phase cp of a current breathing cycle of a patient 1000.

[0175] In some forms, known as discrete phase determination, the phase output cp is a discrete variable. One implementation of discrete phase determination provides a bi-valued phase output cp with values of either inhalation or exhalation, for example represented as values of 0 and 0.5 revolutions respectively, upon detecting the start of spontaneous inhalation and exhalation respectively. RPT devices 4000 that “trigger” and “cycle” effectively perform discrete phase determination, since the trigger and cycle points are the instants at which the phase changes from exhalation to inhalation and from inhalation to exhalation, respectively. In one implementation of bi-valued phase determination, the phase output cp is determined to have a discrete value of 0 (thereby “triggering” the RPT device 4000) when the respiratory flow rate Qr has a value that exceeds a positive threshold, and a discrete value of 0.5 revolutions (thereby “cycling” the RPT device 4000) when a respiratory flow rate Qr has a value that is more negative than a negative threshold. The inhalation time Ti and the exhalation time Te may be estimated as typical values over many respiratory cycles of the time spent with phase cp equal to 0 (indicating inspiration) and 0.5 (indicating expiration) respectively.

[0176] Another implementation of discrete phase determination provides a tri -valued phase output (p with a value of one of inhalation, mid-inspiratory pause, and exhalation.

[0177] In other forms, known as continuous phase determination, the phase output cp is a continuous variable, for example varying from 0 to 1 revolutions, or 0 to 277 radians. RPT devices4000 that perform continuous phase determination may trigger and cycle when the continuous phase reaches 0 and 0.5 revolutions, respectively. In one implementation of continuous phase determination, a continuous value of phase > Dis determined using a fuzzy logic analysis of the respiratory flow rate Qr. A continuous value of phase determined in this implementation is often referred to as “fuzzy phase”. In one implementation of a fuzzy phase determination algorithm 4321, the following rules are applied to the respiratory flow rate Qr:1. If Qr is zero and increasing fast then cp is 0 revolutions.2. If Qr is large positive and steady then cp is 0.25 revolutions.3. If Qr is zero and falling fast, then cp is 0.5 revolutions.4. If Qr is large negative and steady then cp is 0.75 revolutions.5. If Qr is zero and steady and the 5-second low-pass filtered absolute value of Qr is large then cp is 0.9 revolutions.6. If Qr is positive and the phase is expiratory, then cp is 0 revolutions.7. If Qr is negative and the phase is inspiratory, then cp is 0.5 revolutions.8. If the 5-second low-pass filtered absolute value of Qr is large, cp is increasing at a steady rate equal to the patient’s breathing rate, low-pass filtered with a time constant of 20 seconds.

[0178] The output of each rule may be represented as a vector whose phase is the result of the rule and whose magnitude is the fuzzy extent to which the rule is true. The fuzzy extent to which the respiratory flow rate is “large”, “steady”, etc. is determined with suitable membership functions. The results of the rules, represented as vectors, are then combined by some function such as taking the centroid. In such a combination, the rules may be equally weighted, or differently weighted.

[0179] In another implementation of continuous phase determination, the phase cp is first discretely estimated from the respiratory flow rate Qr as described above, as are the inhalation time Ti and the exhalation time Te. The continuous phase (p at any instant may be determined as the half the proportion of the inhalation time Ti that has elapsed since the previous trigger instant, or 0.5 revolutions plus half the proportion of the exhalation time Te that has elapsed since the previous cycle instant (whichever instant was more recent).5.4.3.2.2 Waveform determination

[0180] In one form of the present technology, the therapy parameter determination algorithm 4329 provides an approximately constant treatment pressure throughout a respiratory cycle of a patient.

[0181] In other forms of the present technology, the therapy control module 4330 controls the pressure generator 4140 to provide a treatment pressure Pt that varies as a function of phase cp of a respiratory cycle of a patient according to a waveform template II((p).

[0182] In one form of the present technology, a waveform determination algorithm 4322 provides a waveform template II((p) with values in the range [0, 1] on the domain of phase values (p provided by the phase determination algorithm 4321 to be used by the therapy parameter determination algorithm 4329.

[0183] In one form, suitable for either discrete or continuously -valued phase, the waveform template II((p) is a square-wave template, having a value of 1 for values of phase up to and including 0.5 revolutions, and a value of 0 for values of phase above 0.5 revolutions. In one form, suitable for continuously-valued phase, the waveform template II((p) comprises two smoothly curved portions, namely a smoothly curved (e.g. raised cosine) rise from 0 to 1 for values of phase up to 0.5 revolutions, and a smoothly curved (e.g. exponential) decay from 1 to 0 for values of phase above 0.5 revolutions. In one form, suitable for continuously-valued phase, the waveform template II((p) is based on a square wave, but with a smooth rise from 0 to 1 for values of phase up to a “rise time” that is less than 0.5 revolutions, and a smooth fall from 1 to 0 for values of phase within a “fall time” after 0.5 revolutions, with a “fall time” that is less than 0.5 revolutions.

[0184] In some forms of the present technology, the waveform determination algorithm 4322 selects a waveform template n(q>) from a library of waveform templates, dependent on a setting of the RPT device. Each waveform template n(q>) in the library may be provided as a lookup table of values Pt against phase values cp. In other forms, the waveform determination algorithm 4322 computes a waveform template If(q>) “on the fly” using a predetermined functional form, possibly parametrised by one or more parameters (e.g. time constant of an exponentially curved portion). The parameters of the functional form may be predetermined or dependent on a current state of the patient 1000.

[0185] In some forms of the present technology, suitable for discrete bi-valued phase of either inhalation (cp = 0 revolutions) or exhalation (cp = 0.5 revolutions), the waveform determination algorithm 4322 computes a waveform template II “on the fly” as a function of both discrete phase (p and time t measured since the most recent trigger instant. In one such form, the waveform determination algorithm 4322 computes the waveform template II(cp, t) in two portions (inspiratory and expiratory) as follows:<J> = 00 = 0.5

[0186] where Hi (t) and Ile(t) are inspiratory and expiratory portions of the waveform template II(cp, t). In one such form, the inspiratory portion Ili(t) of the waveform template is a smooth risefrom 0 to 1 parametrised by a rise time, and the expiratory portion He(t) of the waveform template is a smooth fall from 1 to 0 parametrised by a fall time.5.4.3.2.3 Ventilation determination

[0187] In one form of the present technology, a ventilation determination algorithm 4323 receives an input a respiratory flow rate Qr, and determines a measure indicative of current patient ventilation, Vent.

[0188] In some implementations, the ventilation determination algorithm 4323 determines a measure of ventilation Vent that is an estimate of actual patient ventilation. One such implementation is to take half the absolute value of respiratory flow rate, Qr, optionally filtered by low-pass filter such as a second order Bessel low-pass filter with a corner frequency of 0.11 Hz.

[0189] In other implementations, the ventilation determination algorithm 4323 determines a measure of ventilation Vent that is broadly proportional to actual patient ventilation. One such implementation estimates peak respiratory flow rate Qpeak over the inspiratory portion of the cycle. This and many other procedures involving sampling the respiratory flow rate Qr produce measures which are broadly proportional to ventilation, provided the flow rate waveform shape does not vary very much (here, the shape of two breaths is taken to be similar when the flow rate waveforms of the breaths normalised in time and amplitude are similar). Some simple examples include the median positive respiratory flow rate, the median of the absolute value of respiratory flow rate, and the standard deviation of flow rate. Arbitrary linear combinations of arbitrary order statistics of the absolute value of respiratory flow rate using positive coefficients, and even some using both positive and negative coefficients, are approximately proportional to ventilation. Another example is the mean of the respiratory flow rate in the middle K proportion (by time) of the inspiratory portion, where 0 < K < 1. There is an arbitrarily large number of measures that are exactly proportional to ventilation if the flow rate shape is constant.5.4.3.2.4 Determination of Inspiratory Flow Limitation

[0190] In one form of the present technology, the central controller 4230 executes an inspiratory flow limitation determination algorithm 4324 for the determination of the extent of inspiratory flow limitation.

[0191] In one form, the inspiratory flow limitation determination algorithm 4324 receives as an input a respiratory flow rate signal Qr and provides as an output a metric of the extent to which the inspiratory portion of the breath exhibits inspiratory flow limitation.

[0192] In one form of the present technology, the inspiratory portion of each breath is identified by a zero-crossing detector. A number of evenly spaced points (for example, sixty-five), representing points in time, are interpolated by an interpolator along the inspiratory flow rate-time curve for each breath. The curve described by the points is then scaled by a scalar to have unitylength (duration / period) and unity area to remove the effects of changing breathing rate and depth. The scaled breaths are then compared in a comparator with a pre-stored template representing a normal unobstructed breath, similar to the inspiratory portion of the breath shown in Fig. 6. Breaths deviating by more than a specified threshold (typically 1 scaled unit) at any time during the inspiration from this template, such as those due to coughs, sighs, swallows and hiccups, as determined by a test element, are rejected. For non-rejected data, a moving average of the first such scaled point is calculated by the central controller 4230 for the preceding several inspiratory events. This is repeated over the same inspiratory events for the second such point, and so on. Thus, for example, sixty-five scaled data points are generated by the central controller 4230, and represent a moving average of the preceding several inspiratory events, e.g., three events. The moving average of continuously updated values of the (e.g., sixty -five) points are hereinafter called the "scaled flow rate ", designated as Qs(t). Alternatively, a single inspiratory event can be utilised rather than a moving average.

[0193] From the scaled flow rate, two shape factors relating to the determination of partial obstruction may be calculated.

[0194] Shape factor 1 is the ratio of the mean of the middle (e.g. thirty -two) scaled flow rate points to the mean overall (e.g. sixty -five) scaled flow rate points. Where this ratio is in excess of unity, the breath will be taken to be normal. Where the ratio is unity or less, the breath will be taken to be obstructed. A ratio of about 1.17 is taken as a threshold between partially obstructed and unobstructed breathing, and equates to a degree of obstruction that would permit maintenance of adequate oxygenation in a typical patient.

[0195] Shape factor 2 is calculated as the RMS deviation from unit scaled flow rate, taken over the middle (e.g. thirty -two) points. An RMS deviation of about 0.2 units is taken to be normal. An RMS deviation of zero is taken to be a totally flow-limited breath. The closer the RMS deviation to zero, the breath will be taken to be more flow limited.

[0196] Shape factors 1 and 2 may be used as alternatives, or in combination. In other forms of the present technology, the number of sampled points, breaths and middle points may differ from those described above. Furthermore, the threshold values can be other than those described.5.4.3.2.5 Determination of apneas and hypopneas

[0197] In one form of the present technology, the central controller 4230 executes an apnea / hypopnea determination algorithm 4325 for the determination of the presence of apneas and / or hypopneas.

[0198] In one form, the apnea / hypopnea determination algorithm 4325 receives as an input a respiratory flow rate signal Qr and provides as an output a flag that indicates that an apnea or a hypopnea has been detected.

[0199] In one form, an apnea will be said to have been detected when a function of respiratory flow rate Qr falls below a flow rate threshold for a predetermined period of time. The function may determine a peak flow rate, a relatively short-term mean flow rate, or a flow rate intermediate of relatively short-term mean and peak flow rate, for example an RMS flow rate. The flow rate threshold may be a relatively long-term measure of flow rate.

[0200] In one form, a hypopnea will be said to have been detected when a function of respiratory flow rate Qr falls below a second flow rate threshold for a predetermined period of time. The function may determine a peak flow, a relatively short-term mean flow rate, or a flow rate intermediate of relatively short-term mean and peak flow rate, for example an RMS flow rate. The second flow rate threshold may be a relatively long-term measure of flow rate. The second flow rate threshold is greater than the flow rate threshold used to detect apneas.5.4.3.2.6 Determination of snore

[0201] In one form of the present technology, the central controller 4230 executes one or more snore determination algorithms 4326 for the determination of the extent of snore.

[0202] In one form, the snore determination algorithm 4326 receives as an input a respiratory flow rate signal Qr and provides as an output a metric of the extent to which snoring is present.

[0203] The snore determination algorithm 4326 may comprise the step of determining the intensity of the flow rate signal in the range of 30-300 Hz. Further, the snore determination algorithm 4326 may comprise a step of filtering the respiratory flow rate signal Qr to reduce background noise, e.g., the sound of airflow in the system from the blower.5.4.3.2.7 Determination of airway patency

[0204] In one form of the present technology, the central controller 4230 executes one or more airway patency determination algorithms 4327 for the determination of the extent of airway patency.

[0205] In one form, the airway patency determination algorithm 4327 receives as an input a respiratory flow rate signal Qr, and determines the power of the signal in the frequency range of about 0.75 Hz and about 3 Hz. The presence of a peak in this frequency range is taken to indicate an open airway. The absence of a peak is taken to be an indication of a closed airway.

[0206] In one form, the frequency range within which the peak is sought is the frequency of a small forced oscillation in the treatment pressure Pt. In one implementation, the forced oscillation is of frequency 2 Hz with amplitude about 1 cmH20.

[0207] In one form, airway patency determination algorithm 4327 receives as an input a respiratory flow rate signal Qr, and determines the presence or absence of a cardiogenic signal. The absence of a cardiogenic signal is taken to be an indication of a closed airway.5.4.3.2.8 Determination of target ventilation

[0208] In one form of the present technology, the central controller 4230 takes as input the measure of current ventilation, Vent, and executes one or more target ventilation determination algorithms 4328 for the determination of a target value Vtgt for the measure of ventilation.

[0209] In some forms of the present technology, there is no target ventilation determination algorithm 4328, and the target value Vtgt is predetermined, for example by hard-coding during configuration of the RPT device 4000 or by manual entry through the input device 4220.

[0210] In other forms of the present technology, such as adaptive servo-ventilation (ASV), the target ventilation determination algorithm 4328 computes a target value Vtgt from a value Vtyp indicative of the typical recent ventilation of the patient.

[0211] In some forms of adaptive servo-ventilation, the target ventilation Vtgt is computed as a high proportion of, but less than, the typical recent ventilation Vtyp. The high proportion in such forms may be in the range (80%, 100%), or (85%, 95%), or (87%, 92%).

[0212] In other forms of adaptive servo-ventilation, the target ventilation Vtgt is computed as a slightly greater than unity multiple of the typical recent ventilation Vtyp.

[0213] The typical recent ventilation Vtyp is the value around which the distribution of the measure of current ventilation Vent over multiple time instants over some predetermined timescale tends to cluster, that is, a measure of the central tendency of the measure of current ventilation over recent history. In one implementation of the target ventilation determination algorithm 4328, the recent history is of the order of several minutes, but in any case should be longer than the timescale of Cheyne-Stokes waxing and waning cycles. The target ventilation determination algorithm 4328 may use any of the variety of well-known measures of central tendency to determine the typical recent ventilation Vtyp from the measure of current ventilation, Vent. One such measure is the output of a low-pass filter on the measure of current ventilation Vent, with time constant equal to one hundred seconds.5.4.3.2.9 Determination of therapy parameters

[0214] In some forms of the present technology, the central controller 4230, or other processor described herein, executes one or more therapy parameter determination algorithms 4329 for the determination of one or more therapy parameters using the values returned by one or more of the other algorithms in the therapy engine module 4320.

[0215] In one form of the present technology, the therapy parameter is an instantaneous treatment pressure Pt. In one implementation of this form, the therapy parameter determination algorithm 4329 determines the treatment pressure Pt using the equation (1).Pt = AIl(d>, t) + PQ(1)

[0216] In equation (1), A is the amplitude, II((p,t) is the waveform template value (in the range 0 to 1) at the current value cp of phase and t of time, and Po is a base pressure.

[0217] If the waveform determination algorithm 4322 provides the waveform template n(cp,t) as a lookup table of values II indexed by phase cp, the therapy parameter determination algorithm 4329 applies equation (Error! Reference source not found.) by locating the nearest lookup table entry to the current value cp of phase returned by the phase determination algorithm 4321, or by interpolation between the two entries straddling the current value cp of phase.

[0218] The values of the amplitude A and the base pressure PO may be set by the therapy parameter determination algorithm 4329 depending on the chosen respiratory pressure therapy mode in the manner described below.

[0219] The therapy parameter determination algorithms may include any or all of the algorithms for adjusting therapy that are described herein, such as those described in relation to Figs. 8Ato 16.5.4.3.2.10 Mask-off determination

[0220] It can be helpful to assess whether a patient is compliant with respiratory pressure therapy, i.e. whether they wear or remove the patient interface while sleeping. Accordingly, the therapy control module 4330 may periodically query the leak flow rate estimation algorithm 4316, which uses an estimate of pressure at the patient interface to determine how much air is leaking from the interface. In case the patient is not wearing / using the patient interface while the RPT device is operating, the leak flow rate will exceed a threshold value, and the therapy control module 4330 may assess a “mask-off’ event. It should be understood that “mask-off’ is a term of convenience and that the determination equally is applicable to other patient interfaces, e.g., nasal pillows. A mask-on event may be determined similarly such as by detecting a low leak flow rate or patient respiratory flow via the mask.5.4.3.3 Therapy Control module

[0221] The therapy control module 4330 in accordance with one aspect of the present technology receives as inputs the therapy parameters from the therapy parameter determination algorithm 4329 of the therapy engine module 4320, and controls the pressure generator 4140 to deliver a flow of air in accordance with the therapy parameters.

[0222] In one form of the present technology, the therapy parameter is a treatment pressure Pt, and the therapy control module 4330 controls the pressure generator 4140 to deliver a flow of air whose interface pressure Pm at the patient interface 3000 or 3800 is equal to the treatment pressure Pt.5.4.3.4 Detection of fault conditions

[0223] In one form of the present technology, the central controller 4230 executes one or more methods 4340 for the detection of fault conditions. The fault conditions detected by the one or more methods 4340 may include at least one of the following: power failure (no power, or insufficient power); transducer fault detection; failure to detect the presence of a component; operating parameters outside recommended ranges (e.g. pressure, flow rate, temperature, PaO2); and / or failure of a test alarm to generate a detectable alarm signal.

[0224] Upon detection of the fault condition, the corresponding algorithm 4340 signals the presence of the fault by one or more of the following: initiation of an audible, visual & / or kinetic (e.g. vibrating) alarm; sending a message to an external device; and / or logging of the incident 5.5 AIR CIRCUIT

[0225] An air circuit 4170 in accordance with an aspect of the present technology is a conduit or a tube constructed and arranged to allow, in use, a flow of air to travel between two components such as RPT device 4000 and the patient interface 3000 or 3800.

[0226] In particular, the air circuit 4170 may be in fluid connection with the outlet of the pneumatic block 4020 and the patient interface. The air circuit may be referred to as an air delivery tube. In some cases there may be separate limbs of the circuit for inhalation and exhalation. In other cases a single limb is used.

[0227] In some forms, the air circuit 4170 may comprise one or more heating elements configured to heat air in the air circuit, for example to maintain or raise the temperature of the air. The heating element may be in a form of a heated wire circuit, and may comprise one or more transducers, such as temperature sensors. In one form, the heated wire circuit may be helically wound around the axis of the air circuit 4170. The heating element may be in communication with a controller such as a central controller 4230. One example of an air circuit 4170 comprising a heated wire circuit is described in U.S. Pat. No. 8,733,349, which is hereby incorporated by reference in its entirety and for all purposes.5.5.1 Supplementary gas delivery

[0228] In one form of the present technology, supplementary gas, e.g. oxygen, 4180 is delivered to one or more points in the pneumatic path, such as upstream of the pneumatic block 4020, to the air circuit 4170, and / or to the patient interface 3000 or 3800.5.6 HUMIDIFIER5.6.1 Humidifier overview

[0229] In one form of the present technology there is provided a humidifier 5000 (e.g. as shown in Fig. 5A) to change the absolute humidity of air or gas for delivery to a patient relative toambient air. Typically, the humidifier 5000 is used to increase the absolute humidity and increase the temperature of the flow of air (relative to ambient air) before delivery to the patient’s airways.

[0230] The humidifier 5000 may comprise a humidifier reservoir 5110, a humidifier inlet 5002 to receive a flow of air, and a humidifier outlet 5004 to deliver a humidified flow of air. In some forms, as shown in Fig. 5A and Fig. 5B, an inlet and an outlet of the humidifier reservoir 5110 may be the humidifier inlet 5002 and the humidifier outlet 5004 respectively. The humidifier 5000 may further comprise a humidifier base 5006, which may be adapted to receive the humidifier reservoir 5110 and comprise a heating element 5240.5.7 BREATHING WAVEFORMSFig. 6 shows a model typical breath waveform of a person while sleeping. The horizontal axis is time, and the vertical axis is respiratory flow rate. While the parameter values may vary, a typical breath may have the following approximate values: tidal volume Vt 0.5L, inhalation time Ti 1.6s, peak inspiratory flow rate Qpeak 0.4 L / s, exhalation time Te 2.4s, peak expiratory flow rate Qpeak -0.5 L / s. The total duration of the breath, Ttot, is about 4s. The person typically breathes at a rate of about 15 breaths per minute (BPM), with Ventilation Vent about 7.5 L / min. A typical duty cycle, the ratio of Ti to Tiol, is about 40%.5.8 SCREENING, DIAGNOSIS, MONITORING SYSTEMS5.8.1 Respiratory poly raphy

[0231] Fig. 7 is a block diagram illustrating a screening / diagnosis / monitoring device 7200 that may be used to implement an RPG headbox in an RPG screening / diagnosis / monitoring system. The screening / diagnosis / monitoring device 7200 receives the three RPG channels mentioned above (a signal indicative of thoracic movement, a signal indicative of nasal flow rate, and a signal indicative of oxygen saturation) at a data input interface 7260. The screening / diagnosis / monitoring device 7200 also contains a processor 7210 configured to carry out encoded instructions. The screening / diagnosis / monitoring device 7200 also contains a non-transitory computer readable memory / storage medium 7230.

[0232] Memory 7230 may be the screening / diagnosis / monitoring device 7200's internal memory, such as RAM, flash memory or ROM. In some implementations, memory 7230 may also be a removable or external memory linked to screening / diagnosis / monitoring device 7200, such as an SD card, server, USB flash drive or optical disc, for example. In other implementations, memory 7230 can be a combination of external and internal memory. Memory 7230 includes stored data 7240 and processor control instructions (code) 7250 adapted to configure the processor 7210 to perform certain tasks. Stored data 7240 can include RPG channel data received by data input interface 7260, and other data that is provided as a component part of an application. Processor control instructions 7250 can also be provided as a component part of an applicationprogram. The processor 7210 is configured to read the code 7250 from the memory 7230 and execute the encoded instructions. In particular, the code 7250 may contain instructions adapted to configure the processor 7210 to carry out methods of processing the RPG channel data provided by the interface 7260. One such method may be to store the RPG channel data as data 7240 in the memory 7230. Another such method may be to analyse the stored RPG data to extract features. The processor 7210 may store the results of such analysis as data 7240 in the memory 7230.

[0233] The screening / diagnosis / monitoring device 7200 may also contain a communication interface 7220. The code 7250 may contain instructions configured to allow the processor 7210 to communicate with an external computing device (not shown) via the communication interface 7220. The mode of communication may be wired or wireless. In one such implementation, the processor 7210 may transmit the stored RPG channel data from the data 7240 to the remote computing device. In such an implementation, the remote computing device may be configured to analyse the received RPG data to extract features. In another such implementation, the processor 7210 may transmit the analysis results from the data 7240 to the remote computing device.

[0234] Alternatively, if the memory 7230 is removable from the screening / diagnosis / monitoring device 7200, the remote computing device may be configured to be connected to the removable memory 7230. In such an implementation, the remote computing device may be configured to analyse the RPG data retrieved from the removable memory 7230 to extract the features.

[0235] In some forms, the monitoring device 7200 is implemented as part of a pressure therapy device, such as RPT device 4000 and / or an HFT device, which may include, for example, CPAP devices, BiPAP devices, APAP devices, PEP devices, oxygen concentrators, portable ventilators, and / or the like. Examples of such forms are disclosed in U.S. Pat. No. 11,844,907, U.S. Pat. No. 11,759,595, U.S. Pat. No. 11,892,000, and U.S. Pat. No. 11,712,529, the entire disclosures of each of which are hereby incorporated herein by reference for all purposes. In other forms, the monitoring device 7200 is implemented as part of another therapy device such as, for example, nerve or muscle stimulator such as a nebulizer, dialysis device, compression therapy devices, ultrasound therapy devices, traction devices, cryotherapy devices, and / or the like.

[0236] In some forms, the monitoring device 7200 is implemented as a standalone medical device that is separate from a pressure therapy device, such as RPT device 4000. In these forms, the monitoring device 7200 may be or include, for example, a smart watch, ECGZEKG device, finger sensor device (such as a photoplethysmogram (PPG), pulse oximeter, and / or peripheral arterial tone sensor), biopotential measurement device, health tracker, fitness tracker, a blood monitor (e.g., a glucose meter, lactic acid meter / analyzer, and / or the like), flow sensors and / or flow rate sensors, pressure sensors, motion sensors, image capture devices (e.g., cameras), sonarsensors, and / or microphones, among other possibilities. In some examples, the medical devices may be or include any of the devices discussed in EP Pat. No. 3,593,707, U.S. Pat. Pub.2020 / 0015737, U.S. Provisional App. No. 63 / 662,455, U.S. Provisional App. No. 63 / 747,604, U.S. Pat. No. 7,785,265, U.S. Pat. No. 9,629,572, U.S. Pat. No. 9,687,177, U.S. Pat. No. 10,376,670, U.S. Pat. No. 11,364,362, U.S. Pat. No. 11,779,268, U.S. Pat. Pub. 2021 / 0275056, U.S. Pat. Pub.2022 / 0007965, U.S. Pat. Pub. 2014 / 0024917, Int’l App. No. PCT / EP2017 / 070773, U.S. Pat. Pub.2018 / 0239014, and U.S. Pat. No. 11,033,196, the entire disclosures of each of which are hereby incorporated herein by reference for all purposes.5.9 RESPIRATORY THERAPY MODES

[0237] Various respiratory therapy modes may be implemented by the disclosed respiratory therapy system.5.9.1 CPAP therapy

[0238] In some implementations of respiratory pressure therapy, the central controller 4230 sets the treatment pressure Pt according to the treatment pressure equation (Error! Reference source not found.) as part of the therapy parameter determination algorithm 4329. In one such implementation, the amplitude A is identically zero, so the treatment pressure Pt (which represents a target value to be achieved by the interface pressure Pm at the current instant of time) is identically equal to the base pressure P0 throughout the respiratory cycle. Such implementations are generally grouped under the heading of CPAP therapy. In such implementations, there is no need for the therapy engine module 4320 to determine phase cp or the waveform template H(cp,t).

[0239] In CPAP therapy, the base pressure P0 may be a constant value that is hard-coded or manually entered to the RPT device 4000. Alternatively, the central controller 4230 may repeatedly compute the base pressure P0 as a function of indices or measures of sleep disordered breathing returned by the respective algorithms in the therapy engine module 4320, such as one or more of flow limitation, apnea, hypopnea, patency, and snore. This alternative is sometimes referred to as APAP therapy.

[0240] Fig. 4E is a flow chart illustrating a method 4500 carried out by the central controller 4230 to continuously compute the base pressure P0 as part of an APAP therapy implementation of the therapy parameter determination algorithm 4329, when the pressure support A is identically zero.

[0241] The method 4500 starts at step 4520, at which the central controller 4230 compares the measure of the presence of apnea / hypopnea with a first threshold, and determines whether the measure of the presence of apnea / hypopnea has exceeded the first threshold for a predetermined period of time, indicating an apnea / hypopnea is occurring. If so, the method 4500 proceeds to step 4540; otherwise, the method 4500 proceeds to step 4530. At step 4540, the centralcontroller 4230 compares the measure of airway patency with a second threshold. If the measure of airway patency exceeds the second threshold, indicating the airway is patent, the detected apnea / hypopnea is deemed central, and the method 4500 proceeds to step 4560; otherwise, the apnea / hypopnea is deemed obstructive, and the method 4500 proceeds to step 4550.

[0242] At step 4530, the central controller 4230 compares the measure of flow limitation with a third threshold. If the measure of flow limitation exceeds the third threshold, indicating inspiratory flow is limited, the method 4500 proceeds to step 4550; otherwise, the method 4500 proceeds to step 4560.

[0243] At step 4550, the central controller 4230 increases the base pressure P0 by a predetermined pressure increment AP, provided the resulting treatment pressure Pt would not exceed a maximum treatment pressure Pmax. In one implementation, the predetermined pressure increment AP and maximum treatment pressure Pmax are 1 cmfhO and 25 cmfhO respectively. In other implementations, the pressure increment AP can be as low as 0.1 cmfhO and as high as 3 cmfhO, or as low as 0.5 cmfhO and as high as 2 crnThO. In other implementations, the maximum treatment pressure Pmax can be as low as 15 cmfhO and as high as 35 cmfhO, or as low as 20 cmfhO and as high as 30 crnThO. The method 4500 then returns to step 4520.

[0244] At step 4560, the central controller 4230 decreases the base pressure P0 by a decrement, provided the decreased base pressure P0 would not fall below a minimum treatment pressure Pmin. The method 4500 then returns to step 4520. In one implementation, the decrement is proportional to the value of PO-Pmin, so that the decrease in P0 to the minimum treatment pressure Pmin in the absence of any detected events is exponential. In one implementation, the constant of proportionality is set such that the time constant T of the exponential decrease of P0 is 60 minutes, and the minimum treatment pressure Pmin is 4 cmThO. In other implementations, the time constant T could be as low as 1 minute and as high as 300 minutes, or as low as 5 minutes and as high as 180 minutes. In other implementations, the minimum treatment pressure Pmin can be as low as 0 cmfhO and as high as 8 cmfhO, or as low as 2 cmfhO and as high as 6 crnThO. Alternatively, the decrement in P0 could be predetermined, so the decrease in P0 to the minimum treatment pressure Pmin in the absence of any detected events is linear.5.9.2 Bi-level therapy

[0245] In other implementations of this form of the present technology, the value of amplitude A in equation (Error! Reference source not found.) may be positive. Such implementations are known as bi-level therapy, because in determining the treatment pressure Pt using equation (Error! Reference source not found.) with positive amplitude A, the therapy parameter determination algorithm 4329 oscillates the treatment pressure Pt between two values or levels in synchrony with the spontaneous respiratory effort of the patient 1000. That is, based on the typicalwaveform templates n(cp,t) described above, the therapy parameter determination algorithm 4329 increases the treatment pressure Pt to Po + A (known as the IPAP) at the start of, or during, or inspiration and decreases the treatment pressure Pt to the base pressure PO (known as the EPAP) at the start of, or during, expiration.

[0246] In some forms of bi-level therapy, the IPAP is a treatment pressure that has the same purpose as the treatment pressure in CPAP therapy modes, and the EPAP is the IPAP minus the amplitude A, which has a “small” value (a few cmEEO) sometimes referred to as the Expiratory Pressure Relief (EPR). Such forms are sometimes referred to as CPAP therapy with EPR, which is generally thought to be more comfortable than straight CPAP therapy. In CPAP therapy with EPR, either or both of the IPAP and the EPAP may be constant values that are hard-coded or manually entered to the RPT device 4000. Alternatively, the therapy parameter determination algorithm 4329 may repeatedly compute the IPAP and / or the EPAP during CPAP with EPR. In this alternative, the therapy parameter determination algorithm 4329 repeatedly computes the EPAP and / or the IPAP as a function of indices or measures of sleep disordered breathing returned by the respective algorithms in the therapy engine module 4320 in analogous fashion to the computation of the base pressure P0 in APAP therapy described above.

[0247] In other forms of bi-level therapy, the amplitude A is large enough that the RPT device 4000 does some or all of the work of breathing of the patient 1000. In such forms, known as pressure support ventilation therapy, the amplitude A is referred to as the pressure support, or swing. In pressure support ventilation therapy, the IPAP is the base pressure P0 plus the pressure support A, and the EPAP is the base pressure P0.

[0248] In some forms of pressure support ventilation therapy, known as fixed pressure support ventilation therapy, the pressure support A is fixed at a predetermined value, e.g. 10 cmEEO. The predetermined pressure support value is a setting of the RPT device 4000, and may be set for example by hard-coding during configuration of the RPT device 4000 or by manual entry through the input device 4220.

[0249] In other forms of pressure support ventilation therapy, broadly known as servoventilation, the therapy parameter determination algorithm 4329 takes as input some currently measured or estimated parameter of the respiratory cycle (e.g. the current measure Vent of ventilation) and a target value of that respiratory parameter (e.g. a target value Vtgt of ventilation) and repeatedly adjusts the parameters of equation (Error! Reference source not found.) to bring the current measure of the respiratory parameter towards the target value. In a form of servoventilation known as adaptive servo-ventilation (ASV), which has been used to treat CSR, the respiratory parameter is ventilation, and the target ventilation value Vtgt is computed by the targetventilation determination algorithm 4328 from the typical recent ventilation Vtyp, as described above.

[0250] In some forms of servo-ventilation, the therapy parameter determination algorithm 4329 applies a control methodology to repeatedly compute the pressure support A so as to bring the current measure of the respiratory parameter towards the target value. One such control methodology is Proportional -Integral (PI) control. In one implementation of PI control, suitable for ASV modes in which a target ventilation Vtgt is set to slightly less than the typical recent ventilation Vtyp, the pressure support A is repeatedly computed as shown by equation (2).A = G J (Vent — Vtgt)dt (2)

[0251] In equation (2), G is the gain of the PI control. Larger values of gain G can result in positive feedback in the therapy engine module 4320. Smaller values of gain G may permit some residual untreated CSR or central sleep apnea. In some implementations, the gain G is fixed at a predetermined value, such as -0.4 cmH2O / (L / min) / sec. Alternatively, the gain G may be varied between therapy sessions, starting small and increasing from session to session until a value that substantially eliminates CSR is reached. Conventional means for retrospectively analysing the parameters of a therapy session to assess the severity of CSR during the therapy session may be employed in such implementations. In yet other implementations, the gain G may vary depending on the difference between the current measure Vent of ventilation and the target ventilation Vtgt.

[0252] Other servo-ventilation control methodologies that may be applied by the therapy parameter determination algorithm 4329 include proportional (P), proportional-differential (PD), and proportional-integral-differential (PID).

[0253] The value of the pressure support A computed via equation (2) may be clipped to a range defined as [Amin, Amax], In this implementation, the pressure support A sits by default at the minimum pressure support Amin until the measure of current ventilation Vent falls below the target ventilation Vtgt, at which point A starts increasing, only falling back to Amin when Vent exceeds Vtgt once again.

[0254] The pressure support limits Amin and Amax are settings of the RPT device 4000, set for example by hard-coding during configuration of the RPT device 4000 or by manual entry through the input device 4220.

[0255] In pressure support ventilation therapy modes, the EPAP is the base pressure P0. As with the base pressure P0 in CPAP therapy, the EPAP may be a constant value that is prescribed or determined during titration. Such a constant EPAP may be set for example by hard-coding during configuration of the RPT device 4000 or by manual entry through the input device 4220. This alternative is sometimes referred to as fixed-EPAP pressure support ventilation therapy. Titration of the EPAP for a given patient may be performed by a clinician during a titration sessionwith the aid of PSG, with the aim of preventing obstructive apneas, thereby maintaining an open airway for the pressure support ventilation therapy, in similar fashion to titration of the base pressure P0 in constant CPAP therapy.

[0256] Alternatively, the therapy parameter determination algorithm 4329 may repeatedly compute the base pressure P0 during pressure support ventilation therapy. In such implementations, the therapy parameter determination algorithm 4329 repeatedly computes the EPAP as a function of indices or measures of sleep disordered breathing returned by the respective algorithms in the therapy engine module 4320, such as one or more of respiratory events of flow limitation, apnea, hypopnea, patency, and snore. Because the continuous computation of the EPAP resembles the manual adjustment of the EPAP by a clinician during titration of the EPAP, this process is also sometimes referred to as auto-titration of the EPAP, and the therapy mode is known as auto-titrating EPAP pressure support ventilation therapy, or auto-EPAP pressure support ventilation therapy.

[0257] In some forms of the present technology, a respiratory therapy system, which may comprise an RPT device 4000 for supplying a flow of air to the patient 1000 via an air circuit 4170 and a patient interface 3000 or 3800, is configured or otherwise embodied to incorporate further potential changes to therapy settings, such as those discussed in Int’l App. No. PCT / AU2024 / 050572 (WO 2024 / 243640 Al) filed on 31 May 2024 (hereinafter referred to as‘572), the entire contents of which are hereby incorporated by reference in its entirety and for all purposes.

[0258] In some forms of the present technology, a respiratory therapy system, which may comprise an RPT device 4000 for supplying a flow of air to the patient 1000 via an air circuit 4170 and a patient interface 3000 or 3800, is configured or otherwise embodied as an auto-adjusting pressure device (or self-adjusting device) comprising advanced event detection and auto-adjusting mechanisms, such as those discussed in ‘572.5.10 FURTHER POTENTIAL CHANGES TO THERAPY SETTINGS

[0259] Certain situations may arise in positive airway pressure therapy, such as CPAP or bilevel therapy, that would make it desirable to change therapy parameters. For example, there may be an excessive number of adverse events (e.g., apnea, hypopnea, snoring, mask-off) that suggests that at least some of the therapy parameters are not appropriate (i.e., set properly) for the patient. Accordingly, it may be helpful to implement (e.g., in the therapy control module 4330) one or more processes or algorithms in one or more processors (e.g., in any processor described herein such as a controller of the RPT, in a server and / or other external or remote computer) for recommending, such as to a clinician, or even implementing control to automate making changes to one or more therapy parameters, based on detections of any of various treatment issues. Suchparameters may include, but are not limited to minimum therapeutic pressure (minCPAP), maximum therapeutic pressure (maxCPAP), ramp time and / or rate of pressure increase from StartCPAP pressure at initiation of therapy to minCPAP minimum therapeutic pressure, bridge time from minimum therapeutic pressure to a critical range of therapeutic pressures Pcrit, etc. For example, in some implementations of the technology, the process(es) described herein for detection of treatment issues, may use treatment information from one or more previous sessions to recommend and / or adjust parameters of an RPT controller (e.g., the therapy control module 4330) so that they may be better set to respond to respiratory events such as detections of snore, flow limitation and / or apnea, such as how fast or slow pressure changes (e.g., increase or decrease) should be made in light of detections of such respiratory events. Alternatively, the adjustment of one or more parameters may provide a more comfortable experience for the user, thus potentially improving therapy or even improving the user’s compliance with the treatment.

[0260] For example, such process(es) or algorithm(s) may be configured to detect if there are treatment issues (e.g., observed repetitively over a couple of sessions of therapy (e.g., a couple of nights of treatment for that patient with the RPT) which are likely to be a result of incorrect control parameter settings, such as Start CPAP, MinCPAP, MaxCPAP, ramp time / ramp rate. Such detections may serve as a basis for the process(s) to generate one more messages, such as to provide the user or a clinician, with an explanation of the treatment issues that have been detected and how the therapy parameters or settings may be adjusted in order to resolve or reduce the occurrences, or negative effects thereof, of those treatment issues. Such recommendations may be generated as, for example, as textual or graphic indication. In some implementations of the technology, the process(es) may be configured to automatically or semi-automatically implement the adjustment of such settings, such as by remotely communicating commands for the new param eter / setting adjustments to the RPT and / or setting them in the controller of the RPT, such as if the recommended adjustments are also approved by manual input or confirmation of the user / clinician.

[0261] In this regard, such detections of treatment issues described herein may be made by evaluation of sensor data (e.g., pressure and / or flow rate data from transducer(s) described herein), and / or event data (e.g., respiratory events) determined from the sensor data, that may be generated by a controller of the RPT from therapy sessions with the RPT. Additionally, such detection of issues may also be made with an evaluation of additional information such as, for example, sleep information such as sleep time, scored sleep stage information, body position information, head position information, which may be input or otherwise detected from sensor data, such as by analysis of motion data from an accelerometer or other non-contact motion sensor such as radio frequency motion sensors. Any of such data may be received remotely from the RPT at one or more servers that store, in one or more database(s), the sensor data and / or event data so that theevaluations may be made, at least in part, by a server or computer (e.g., local or remote) accessing such database data. Thus, the evaluations to detect the treatment issues described herein may be made by any one or both of the controller of the RPT, such as without remote processing, and / or by a remote computer and / or server accessing the sensor and / or event data of the one or more databases. For example, the controller of the RPT may record sensor data, detect respiratory events from multiple sessions and evaluate the treatment issues as discussed herein. Similarly, a server or computer (local or remote) may access the sensor data such as from the database(s), detect respiratory events therefrom, and further evaluate the data for detecting the treatment issues described herein. Still further a server or computer (local or remote) may access the sensor data and / or detected respiratory events, such as from the database(s) and / or RPT, and evaluate the data for detecting the treatment issues described herein. The time during which the data is collected and processed, may include one or more treatment sessions.

[0262] For example, in some implementations of the technology, the process(es) may be configured to detect a ramp related treatment issue, such as if it occurs repeatedly in a plurality of sessions. For example, the respiratory event data from one prior session or each of a plurality of previous sessions may be evaluated to detect whether a count of apnea, (and / or hypopnea), events that occur particularly during a ramp period of each session may exceed a threshold value. Such a ramp period typically occurs during initiation of a therapy session (e.g., a “mask on” event) when lower pressures are provided to help a patient fall asleep using lower pressures of the ramp period, but may occur more than once in a session, such as every time the patient interrupts and then renews the treatment session. The pressure during a predetermined ramp period will gradually change to a predetermined minimum therapeutic pressure at the conclusion of the ramp period. Such a detection may be considered in relation to the example graph of FIG. 8 A. FIG. 8 A depicts a therapeutic pressure waveform 8100 from a sleep session. The waveform 8100 shows a ramp 8102 having a ramp time. If a number of apnea (and / or hypopnea) events occur in one or more ramp sessions, a ramp related treatment issue is detected. Based upon such a detection, the process(es) may generate a recommendation message of a change to a setting time for the ramp (e.g., an increase in rate of pressure increase during the ramp time, a reduction to a ramp time or removing the ramp time altogether). Similarly, based upon such a detection, the process(es) may generate a recommendation for a change (e.g., an increase) to the initial pressure used for the ramp (i.e., the startCPAP pressure). Moreover, the process(es) may be further configured to automatically or semi -automatically (such as with a manual input confirmation of a recommended change) implement any of such adjustments to such settings (e.g., the initial pressure and / or ramp period or rate of increase), such as by remotely communicating command(s) for the new param eter / setting adjustments to the RPT for the next session.

[0263] Another example may be related to a “critical pressure range” (shown in Fig. 8C at 8301), which is a treatment pressure range in which generally no adverse respiratory events or no more than a predetermined maximum number of adverse respiratory events (e.g., apnea or hypopnea events, snoring, or flow limitation events) occur for a particular patient. This range is deemed a “critical” range because it is generally efficacious for the particular patient over multiple sessions. The PCrit range may be defined by a minimum PCrit ( crit-min) 8305 (which can be above a minimum CPAP setting such as at or above a pressure range 8307 between the MinCPAP setting and the minimum Pcrit), and a maximum PCnt ( crit-max) 8303, which can be below a maximum CPAP setting. This Pcrit range of therapeutic pressures can typically maintain upper airway patency for the particular patient and the peak of the range may be designated as Pcrit or Pcriticai and may optionally be determined by evaluation of density of breathing events per pressure range such as from one or a plurality of treatment sessions. Thus, data from several previous sessions of treatment, such as with an auto-titrating device (also referred to as an auto-adjusting device), may be utilized to establish such a critical range of therapeutic pressures Pcrit. The process(es) may be configured to detect related treatment issue (e.g., in the form of one or more respiratory events), such as if it occurs repeatedly in a plurality of sessions during a period 8104 wherein pressure increases from a minimum therapeutic pressure into a critical range of therapeutic pressures (see FIG. 8A). In such a situation, it may be desirable to increase the minimum therapeutic pressure setting (MinCPAP), and / or to increase the rate of climb from the MinCPAP towards the Pcrit. range of pressures. For example, any processor described herein, such as a processor of controller (e.g., the therapy control module 4330), may have access to or be configured to identify a critical therapeutic pressure range. Data from several previous sessions of treatment, such as with an autotitrating device, may be utilized to establish such a critical range of therapeutic pressures Pcrit. See e.g., ‘572.

[0264] In relation to such a range, a treatment issue may be detected, for example, if an evaluation of the respiratory event data from each of a plurality of sessions detects the occurrence of a predetermined number of respiratory events, or that a count of apnea (and / or hypopnea) events exceeds a threshold, during a time period of the session between the end of a ramp period and a first time when the delivered pressure reaches the Pcrit range, as pressure changes due to autotitration (e.g., pressure responses to respiratory events). Based upon such a detection, the process(es) may generate a recommendation message of a change to a pressure setting parameter concerning the ramp approaching the Pcrit, (see treatment pressure range 8104 in Fig. 8 A). For example, based upon such a detection, the process(es) may generate a recommendation for a change (e.g., an increase) to the end pressure (i.e., the MinCPAP pressure.) that is achieved by the aforementioned ramp. Moreover, the process(es) may be further configured to automatically orsemi-automatically (such as with a manual input confirmation of a recommended change) implement such adjustment to such setting (e.g., the MinCPAP pressure), such as by remotely communicating command(s) for the new param eter / setting adjustments to the RPT for the next session.

[0265] In another example, the process(es) may be configured to detect a maximum pressure related treatment issue, such as if it occurs repeatedly in a plurality of sessions. For example, the data from each of a plurality of sessions may be evaluated to detect whether the therapy pressure, such as during an auto-titrating therapy, repeatedly (e.g., a number of occurrences exceeding a threshold) reaches the limit of a maximum therapy pressure threshold setting (i.e., MaxCPAP) and / or maintains pressure at such a limit for an amount time that exceeds a time threshold that may be deemed significant. Moreover, such detection may be further conditioned on a count of apnea (and / or hypopnea) events that exceeds a threshold, that may be deemed significant. Based upon such a detection, the process(es) may generate a recommendation message of a change to a pressure setting parameter concerning the therapy. For example, based upon such a detection, the process(es) may generate a recommendation for a change (e.g., an increase) to the maximum pressure limit (i.e., the MaxCPAP pressure). Moreover, the process(es) may be further configured to automatically or semi-automatically (such as with a manual input confirmation of a recommended change) implement such an adjustment to such settings (e.g., the MaxCPAP pressure), such as by remotely communicating command(s) for the new param eter / setting adjustments to the RPT for the next session.

[0266] In another example, the process(es) may be configured to detect a mask on / off related treatment issue, such as if it occurs repeatedly in a plurality of sessions. For example, as illustrated in FIG. 8B, the graph depicts a therapeutic pressure waveform 8200 during a sleep session in which the patient removes their mask repeatedly (e.g., at 8202, 8204, 8206), which may be detected by detecting a drop in mask pressure. During such a sleep session, a count of such mask-off events may exceed a threshold value. As an additional parameter, one can observe if such mask-off events may occur in a similar pressure range. In such a situation, it may be desirable to reduce the maximum therapeutic pressure setting since the pressure may be waking the patient. For example, the data from each of a plurality of sessions may be evaluated to detect whether the number of detected mask-off events exceeds a threshold, such as that occur when the delivered therapy pressure is the same or similar pressure (e.g., within plus or minus a margin (e.g., 2 cm H2O) of each other). Based upon such a detection, the process(es) may generate a recommendation message of a change to a pressure setting parameter concerning the therapy. For example, based upon such a detection, the process(es) may generate a recommendation for a change (e.g., a decrease) to the maximum pressure limit (i.e., the maximum of the Pcrit. Range (Pcrit-max 8303) and / or theMaxCPAP pressure). Moreover, the process(es) may be further configured to automatically or semi-automatically (such as with a manual input confirmation of a recommended change) implement such an adjustment to such settings (e.g., the MaxCPAP pressure), such as by remotely communicating command(s) for the new param eter / setting adjustments to the RPT for the next session.

[0267] In another similar example, the process(es) may be configured to detect a mask on / off related treatment issue, such as if it occurs repeatedly in a plurality of sessions. If a count of detected mask-off events in a session exceeds a corresponding threshold value and the RPT device is set to use a ramp period after each such event, it may be desirable to disable the ramp or increase the ramp rate of climb. For example, the process(es) may be configured to detect whether a count of mask-off events exceeds a threshold and that ramp function is enabled. Based upon such a detection, the process(es) may generate a recommendation message of a change to the ramp (e.g., disabling the ramp, such as after detected mask off events, decreasing a ramp time or increasing the ramp end pressure (MinCPAP) or increasing a rate of a ramp pressure increase). Moreover, the process(es) may be further configured to automatically or semi-automatically (such as with a manual input confirmation of a recommended change) implement any of such adjustments to such settings (e.g., disabling the ramp, such as after detected mask off events, decreasing a ramp time or increasing the ramp end pressure (MinCPAP) or increasing rate of ramp pressure increase), such as by remotely communicating command(s) for the new param eter / setting adjustments to the RPT for the next session. In some cases, the above recommendations / changes may be made only if there is an indication that the mask-off events and / or the subsequent ramp time increase the number of adverse respiratory events.

[0268] Additional processes for avoiding treatment issues may be considered in relation to FIG. 8C. Such processes may be configured to evaluate data (e.g., sensor data, such as pressure and / or flow data, and / or respiratory event data) from previous sessions (e.g., previous nights of treatment) for modifying automated therapy pressure changes in response to detections of respiratory events, such as snore or flow limitation or apnea, and / or how fast or slow pressure increase or decreases. Such processes may be based on the aforementioned determination of a critical pressure (Pcrit) range.

[0269] For example, FIG. 8C depicts a graph 8300 of therapeutic pressure waveforms from a plurality of sessions and a determined Pcrit range. The previously determined Pcrit range may be implemented as a control input(s), such as the therapy parameter determination module 4329. The Pcrit and / or the Pcrit range or a value therein, determined from one or more prior sessions, may serve as an additional control parameter or pressure control parameter for controlling therapy of a subsequent session. For example, a value of the Pcrit or Pcrit range may serve as a soft limitto pressure increases or decreases, or modify the rate of change of pressure within certain pressure ranges. The system may be programmed to focus on operating within this range and to resist moving the treatment pressure out of the range, such as by implementing of one or more biasing functions as described in more detail herein.

[0270] For example, in some implementations, the controller of the RPT may increase pressure generally in response to a detection of a respiratory event, such as any one or more flow limitation, snore, and apnea as previously described. However, if such a change in pressure AP would increase the treatment pressure above the maximum Pcrit (PcritMax 8303), the Pcrit range may serve as a further constraint to resist such a change. For example, such a change in pressure AP may be reduced, such as by some proportion, if the resulting pressure would be outside the Pcrit range. Similarly, application of such a change in pressure AP may be delayed until of another similar respiratory event is detected. In this way, a value of the Pcrit or Pcrit range may serve as a soft limit to pressure increases. Such increases may occur but the controller operates to maintain the pressure within, or closer to, the Pcrit range. By modifying the controller’ s response to detected respiratory events so as to resist pressure changes above the Pcrit, the controller may, in part, permit such increases but may minimize or avoid unnecessary overshoot events 8302 that could or might lead to any of patient discomfort, leak and / or patient arousal.

[0271] Similarly, the controller of the RPT may decrease pressure over time generally in the absence of a detection of a respiratory event, such an absence of any one or more flow limitation, snore, and apnea as previously described. However, if such a change in pressure AP would decrease the treatment pressure below the Pcrit range, the Pcrit range may serve as a further constraint to resist such a change. For example, such a change in pressure AP may be reduced, such as by some proportion, if the resulting pressure would fall below the Pcrit range. Similarly, application of such a change in pressure AP due to an absence of events may be delayed for a longer period of time without detection of respiratory events. In this way, a value of the minimum Pcrit (Pcrit-Min 8305) or Pcrit range may serve as a soft limit to pressure decreases. Such decreases may occur but the controller operates to maintain the pressure closer to or within the Pcrit range.

[0272] Similarly, the Pcrit range may serve as a “magnet” for changes to the treatment pressure, with the controller in some cases being programmed to more rapidly approach the Pcrit range when the treatment pressure is not within the range. For example, for a detection of a respiratory event requiring an increase in pressure if detected when the treatment pressure is below the range, the change in pressure AP for such event may be larger, such as by a factor, when compared to an increase in pressure for an event detected when the treatment pressure is within the Pcrit range. Similarly, in an absence of detections of respiratory events that requires a decrease in treatment pressure, if occurring when the treatment pressure is above the range, the change inpressure AP for such an absence of events may be larger, such as by a factor, when compared to a decrease in pressure for an absence of events when the treatment pressure is within the Pcrit range.

[0273] In another example, such as during a ramp period, if respiratory events are detected, the controller may modify one or more control parameters of the ramp so that the pressure of the ramp approaches a value in the Pcrit range, such as a minimum pressure value of the Pcrit range. FIG. 8C depicts an example of such a ramp 8304 in which a processor of the controller of the RPT detects a number of apnea (and / or hypopnea) events, which number during the ramp exceeds or meets a threshold value. In response, the controller adjusts one or more of the control parameter(s) of the ramp (e.g., the rate of increase and / or end pressure for the ramp) so that the ramp approaches a value of the Pcrit range, which is larger than the end pressure target of the ramp that had been set at the beginning of the ramp (i.e., MinCPAP), or that would have been set otherwise.5.11 PRESSURE THERAPY DIGITAL TWINS

[0274] In some forms of the present technology, a respiratory therapy system, which may comprise an RPT device 4000 for supplying a flow of air to the patient 1000 via an air circuit 4170 and a patient interface 3000 or 3800, is configured or otherwise embodied as an auto-adjusting pressure device (or self-adjusting device) comprising advanced event detection and auto-adjusting mechanisms. The event detection mechanism may detect one or more respiratory events from a flow signal measured and recorded during a respiratory therapy period (RTP), such as a night of sleep, or portions thereof. The auto-adjusting mechanism may automatically adjust respiratory therapy parameters, such as the pressure settings used by the respiratory therapy system to supply pressure to the patient during the RTP. The event detection and auto-adjusting mechanisms may be based on pressure therapy digital twin technologies discussed in relation to Figs. 9-17.

[0275] Various forms of the present technology include development and potential applications of personalized digital twins, which may be used in the context of providing pressure therapy and / or treating sleep apnea, and may be implemented for optimizing therapy for patients who are modeled by the personalized digital twins. A digital twin may be developed to simulate a patient's physiological data, allowing for testing and validation of therapy treatment protocols or therapy algorithms, such as with simulated models of the protocol algorithm, without needing the subject to be physically present.

[0276] The pressure therapy digital twin, in one form of the present technology, may be a virtual representation of patients, such as OSA patients, on pressure therapy, such as PAP, CPAP, and / or APAP therapy, connected by device sensor data. The pressure therapy digital twin allows for (i) the development and testing of therapy prototypes in an efficient and cost-effective manner, (ii) making improved therapy-related decisions for patients, (iii) predicting how patients willrespond to new pressure therapies, and / or (iv) determining appropriate therapy configurations or parameters for patients.5.11.1 Digital Twin Pipeline

[0277] Fig. 9 shows an example digital twin pipeline 900. The digital twin pipeline 900 is a pipeline for creating patient digital twins based on a virtual patient model 910, therapy engine model 920, and dataset(s) 950. Through simulating the patient model 910 and therapy engine model 920 with the dataset 950, digital twins may be created as virtual representations of individual patients. The pipeline 950 allows for the creation of multiple Pcrit vectors, such as Pcrit signal 935, for individual patients in less time than would be needed to measure Pcrit in a laboratory setting or during home sleep studies. In one form, one pipeline 900 may be implemented or instantiated for a single patient. In another form, one pipeline 900 may be implemented or instantiated for multiple patients that have similar symptoms or patient characteristics.

[0278] In one form of the present technology, the digital twin pipeline 900 may be instantiated in therapy engine code of the therapy engine 4320. For example, some of the depicted components of the digital twin pipeline 900 may be instantiated as individual components or algorithms 4300 of the therapy engine 4320, and other components of the digital twin pipeline 900 may be instantiated as part of one or more components or algorithms 4300 of the pre-processing module 4310 shown by Fig. 4D. Additionally or alternatively, some of the depicted components of the digital twin pipeline 900 may be instantiated as individual components or modules of the central controller 4230, therapy device controller 4240, and / or the like.

[0279] In the example of Fig. 9, the patient model 910 is connected to the therapy engine model 920 and a Pcrit model 930 and may be implemented with suitable software processes. For example, in some implementations, the models 910, 920, and 930 may be implemented as individual DLLs, where the patient DLL is coupled with the therapy engine DLL and / or Pcrit generator DLL in a python script. In another implementation, models 910, 920, and 930 may be encapsulated as components, microservices, or containerized applications that communicate through an API, inter-process communication such as remote procedure call (RPC), message queues, middleware, shared memory, or some other communication mechanism within a Pythonbased framework using shared object libraries. In other forms, the digital twin pipeline 900 may be implemented in other ways using various frameworks.

[0280] Additionally or alternatively, in some forms, one or more of the models 910, 920, 930 may be implemented using a machine learning model, such as one or more neural networks or the like. In one form, the machine learning model may be a recurrent neural network (RNN), such as a Long Short-Term Memory (LSTM) model, gated recurrent unit (GRU) model, a convolutional neural network (CNN), a transformer model, and / or the like.5.11.1.1 Clinical Trial Data

[0281] In some forms of the present technology, the dataset 950 in accordance with one form of the present technology may include previously recorded data related to a set of patients. In one form, the dataset 950 includes data collected from medical devices, such as diagnostic devices and / or therapy devices. Such medical devices may include any number or combination of the examples of monitoring device 7200 discussed previously with respect to Fig. 7.

[0282] In another form, the dataset 950 includes data from one or more previous clinical trials can be used. In one example, data from a previous clinical trial studying the use of PAP therapy at home was used for this analysis. In this example, sensor data stored on the SD cards of CPAP devices was obtained from 18 (XX female) participants who all had an auto-titrating algorithm (AutoSet) enabled on their CPAP devices. Data was used for an average of 7 nights per participant. The 18 participants were manually selected from a wider pool of participants based on the characteristics of their intra-night pressure requirement. For the purpose of a simple night by night comparison of the simulated and real pressure data, subjects were selected if the variation in their average titrated pressure level across nights of therapy was low. The data of each selected participant was then used to create their own digital twin counterpart.5.11.1.2 Pcrit Model

[0283] In some forms of the present technology, the Pcrit model 930, which may also be referred to as Pcrit signal generator 930 or digital twin generator 930, generates virtual representations of a patient, or specific aspects of a patient, based on data related to that patient in the dataset 950. In one form, the Pcrit model 930 models or estimates a patient's latent airway patency dynamics based on their pressure data in the dataset 950 and regulates fluctuations or changes in auto-titrated pressure during therapy. Additionally, the Pcrit model 930 may represent the time-varying Pcrit of the patient’ s upper airway. For example, the Pcrit model 930 may perform digital signal processing and feature extraction on pressure data from dataset 950 to derive or otherwise create a latent Pcrit model of pressure data / signal(s). This model may form the basis of the patient’s digital twin. The Pcrit model 930 may synthesize a Pcrit vector, such as Pcrit signal 935, from the latent Pcrit model. The size of the Pcrit signal 935 may be based on how long the patient received therapy. This process and its associated hyperparameters are tuned through an iterative process described infra.5.11.1.3 Pcrit Signal

[0284] Fig. 13 illustrates an example Pcrit vector generation pipeline 1300 for generating a Pcrit vector 935 from dataset 950 in accordance with aspects of the present technology. In some forms, pipeline 1300 is an example of how the Pcrit model 930 generates the Pcrit signal 935. ThePcrit vector generation pipeline 1300 involves multiple data processing and transformation steps or stages to model a patient's upper airway patency variations.

[0285] The pipeline 1300 may begin with pressure signals (or traces) 1301 being extracted from the dataset 950. These pressure signals 1301 may be recorded over several sleep periods (nights) using a medical device. The pressure signals 1301 may be concatenated 1303 into a single multi-night pressure signal 1305 for further processing. The multi-night pressure signal 1305 may be a single continuous time-series representation of the signals 1301. In some forms, the multinight pressure signal 1305 may include or indicate event timings, such as apneas, hypopneas, airway closures, flow limitations, etc., that were extracted from the dataset 950.

[0286] Dominant frequency components 1310 are identified in the multi -night pressure signal 1305. A Fast Fourier Transform (FFT) 1307 may be applied to the multi-night pressure signal 1305 to identify the dominant frequency components 1310. Derivative processing 1309 may also be performed on the multi-night pressure signal 1305 to identify positive pressure increases, which may be used to model airway patency changes. The derivative processing 1309 operation may produce an impulse signal of positive increases 1315. The multi-night pressure signal 1305 also is segmented 1311 into segments of length (or duration) T. In some forms, the segments 1311 may correspond to the dominant frequency components 1310.

[0287] The dominant frequency signal 1310 may then be segmented 1311 into a segmented signal 1320 with length of T (time, such as a value between 10 and 30 minutes), or each segment may have a length of T. Additionally, derivative processing 1309 may be performed on the multinight pressure signal 1305 to produce an impulse signal of positive increases 1315. The impulse signal 1315 may be combined with the segmented signal 1320.

[0288] An exponential distribution is fit 1323 to the interarrival times of each segment of the segmented signal 1320. Here, each segment may be analyzed to determine the interarrival times of pressure fluctuations. Fitting 1323 the exponential distribution to each segment’s interarrival times may produce an array of rate parameters 1325.

[0289] The impulse signal 1315 may be sampled 1317 (e.g., using a statistical sampling technique such as sampling with replacement) and the sampling may select samples that are positive increases below a threshold (e.g., 0.2). Each such sample may be added to the previous value (i.e., to create a cumulative sum). The array of rate parameters 1325 may be sampled 1327 (e.g., by sampling with replacement). A new array of impulses, such as impulse vector 1330, may be created based on the sampled 1317 impulse signal 1315 and the sampled 1327 array of rate parameters 1325. The impulse vector 1330 may be a vector of increasing pressures and / or represent dynamic pressure changes.

[0290] Event timing information 1350 are inserted 1355 into the impulse vector 1330. Specifically, impulses with height equal to the maximum pressure at rates equal to rate of particular respiratory events (e.g., apneas, hypopneas, flow limitations, etc.) are inserted into the impulse vector 1330 to produce the final Pcrit vector 935.

[0291] In some forms, stages 1305 to 1330 may be implemented using a machine learning model. In one form, the machine learning model may be an RNN, such as an LSTM model, GRU model, CNN, a transformer model, and / or the like. In such forms, the dataset 950 may be labeled with events (apneas, hypopneas, etc.) and their timings within the dataset 950. Once trained, the machine learning model may predict the Pcrit vector 935 for new patients.

[0292] Fig. 14 shows different characteristics of an example Pcrit vector 935. The Pcrit vector 935 may be divided into a set of epochs 1405 of Pcrit activity, such as in relation to changes in Pcrit values. Each epoch 1405 includes a cumulative increase in Pcrit activity followed by a drop or decrease in Pcrit activity, which delineates a full epoch 1405. Fig. 14 shows multiple epochs 1405, although not all epochs 1405 are labeled in Fig. 14. Within each epoch 1405, there may be one or more impulses. Thus, the Pcrit signal 935 may not be a continuous signal, and instead may be a collection of smaller impulses. Each of these impulses may include an amount of activity and a drop in activity back down to some baseline.

[0293] Additionally, each epoch 1405 also includes different spacings between the impulses. As shown by Fig. 14, some epochs 1405 have impulses that are more sparsely spread out from one another, and other epochs 1405 with more densely packed impulses. That density or sparsity is defined by a rate parameter 1410, which is part of the pipeline 900, 1300, such as the array of rate parameters 1325. This may be learned from the patient's data. Each epoch 1405 has its own rate parameter 1410, which determines the density or sparsity of impulses. Some epochs, such as epoch 1420, have a rate parameter of 0, which is defined as a period where there is no Pcrit activity. Furthermore, the Pcrit signal 935 may also include points in the signal where relatively large impulses of activity are inserted, such as large impulses 1415. The impulses 1415 may be regions where events, such as apneas, are stimulated in the signal.5.11.1.4 Patient Model

[0294] In some forms of the present technology, the patient model 910 is a physiological model that can simulate human physiological systems and functions, such as human cardiorespiratory behaviors in a physiologically-credible manner for a target OSA patient population. For example, the patient model 910 may model an OSA patient’s respiratory system, cardiac system, neural functions, and / or other physical systems, which are all brought together to simulate the patient’s physiological functions, such as sleeping behaviors and the like. In one form, thepatient model 910 may be created using Visual Solutions for embedded systems (VisSim / Embed) provided by Altair Engineering Inc. One example patient model 910 is shown by Fig. 10.

[0295] Fig. 10 shows an example patient model 1000 with various physiological compartments. The patient model 1000 may correspond to the patient model 910 of Fig. 9. The patient model 1000 includes several compartments, such as central control compartment 1010, respiratory mechanics compartment 1020, cardiovascular system compartment 1030, gas exchange compartment 1040, upper airway compartment, and the like. The compartments are interconnected forming an integrated patient model 1000 to simulate human physiological behaviors.

[0296] In the example of Fig. 10, the compartments of the patient model 1000 may be configured such that the responses of the patient model 1000 are representative of sleep apnea, such as OSA. For example, the compartments in Fig. 10 may be adapted from those discussed in Armitstead et al., “A study of the bifurcation behavior of a model of flow through a collapsible tube,” Bulletin of Mathematical Biology, vol. 58, no. 4, pp. 611-641 (1996), Blaxland, “The effect of CPAP on the pulsatile dynamics of the heart,” M.S. Thesis, School of Biomedical Engineering, Univ. New South Wales, Sydney, NSW, Australia (2005), Duffin et al., “A model of the chemoreflex control of breathing in humans: Model parameters measurement. Respiration Physiology,” Respiration Physiology, vol. 120, no. 1, pp. 13-26 (2000), Khoo et al., “Sleep-induced periodic breathing and apnea: a theoretical study,” Journal of applied physiology, vol. 70, no. 5, pp. 2014-2024 (1991), Lu et al., “Whole-body gas exchange in human predicted by a cardiopulmonary model. Cardiovascular Engineering,” Cardiovascular Engineering, vol. 3, no. 1, pp. 1-19 (2003), and Ursino et al., “An integrated model of the human ventilatory control system: the response to hypercapnia,” Clinical Physiology, vol. 21, no. 4, pp. 447-64 (2001), the contents of each of which are hereby incorporated by reference in their entireties and for all purposes.

[0297] In other forms, the compartments of the patient model 1000 may be configured such that the responses of the patient model 1000 are representative of other diseases, such as respiratory disease populations including chronic obstructive pulmonary disease (COPD), neuromuscular disease (NMD), and / or Obesity Hypoventilation Syndrome (OHS). In such forms, the compartments in Fig. 10 may be adapted from those discussed in Dickens et al., “Development of unique online e-leaming to teach non-invasive ventilation, featuring high-fidelity physiology simulation,” American Journal of Respiratory and Critical Care Medicine, vol. 189, pp. A4554-A4554 (2014) and Dickens et al., “Virtual patients can assist in the development of closed-loop ventilation algorithms,” 2nd Annual BMES / FDA Frontiers in Medical Devices Conference, Washington DC (2016), the contents of each of which are hereby incorporated by reference in their entireties and for all purposes.

[0298] Given that PCrit is a valuable metric for characterizing the condition of OSA patients, the patient model 1000 may incorporate an input to control PCrit within the upper airway compartment, such as the collapsible upper airway compartment in the respiratory mechanics compartment 1020. To ensure the patency of the upper airway, the applied pressure (which may be understood as a treatment pressure, such as the CPAP pressure, an IPAP pressure, an EPAP pressure, or an EEP pressure, etc.) should be equal to or exceed the specified PCrit value. For instance, as illustrated in Fig. 11, when the model's PCrit was set to 8 cmEEO, a pressure of at least 8 cmEEO was required to prevent airway collapse and maintain the patient's oxygen saturation.

[0299] Thus, when the pressure 925 dips below the Pcrit value, the patient simulated by the patient model 910 will typically start to experience one or more events, such as apneas, hypopneas, narrowing of the airway, and / or the like. An example of such events are shown by Fig. 12.

[0300] Fig. 12 shows an example patient model simulation 1200 when a pressure value is fixed below a Pcrit value for both flow limitation simulation 1202 and an apnea simulation 1204. In the example of Fig. 12, when the set pressure dips below the Pcrit, the patient model simulation 1200 simulates a flow limitation 1202 and / or an apnea 1204. The flow limitation simulation 1202 shows that, when the Pcrit is below the pressure level, the patient model 910 will simulate a narrowed airway function followed by some sort of termination event, which may be an arousal that results in the airway opening again. The apnea simulation 1204 simulates a full closure of the airway for a certain number of seconds (e.g., 10 seconds or more) and then an opening of the airway. These simulations may continue in a loop or for a set number of iterations.

[0301] Referring back to Fig. 9, in some forms, the patient model 910 may accept one or more pressure values 925 as an input. The pressure values 925 may also be referred to as pressure levels, pressure settings, set pressure, or the like. The pressure value 925 is generated by the therapy engine model 920.

[0302] The patient model 910 also accepts a latent Pcrit signal 935 as an input. As alluded to previously, the Pcrit may be a specific pressure value or amount of pressure necessary to keep the patient’s airway open. Conventionally, the Pcrit value was set at a fixed value, such as 8 cmEEO. However, Pcrit typically varies throughout the night for most patients due to factors such as body position, sleep stage, airway dynamics, alcohol consumption, hormonal aspects, etc. Thus, the Pcrit signal 935 represents the variable Pcrit values that may be experienced by a patient over a period of time, such as a sleep period of 6 or 8 hours. Pcrit can be understood to represent airway patency, and thus the Pcrit signal 935 may be understood to represent the variable airway patency that may be experienced by the patient over a period of time. Therefore, in various forms of the present technology, the Pcrit signal 935 represents a dynamic Pcrit value that changes over time in a similar manner as an OSA patient. The Pcrit signal 935, which may also be referred to as aPcrit vector, may be a series of evenly spaced impulses of varying height. For example, the Pcrit signal 935 may be randomly sampled Pcrit values from multiple overlapping and / or nonoverlapping distributions.

[0303] Optionally, the patient model 910 may accept one or more event types 905 as an input. The event types 905 may indicate the types of events that the patient model 910 is to simulate. For example, the event types 905 may include apneas, hypopneas, narrowing of the airway, flow limitation, and / or any other type of medical events, such as sleep related events and the like.

[0304] The patient model 910 generates and outputs a flow signal 915 based at least one the Pcrit signal 935 and the pressure value 925, and optionally based on the event type(s) 905. The flow signal 915 is then provided to the therapy engine model 9205.11.1.5 Therapy Model

[0305] In some forms of the present technology, the therapy engine model 920 is a model of a therapy device, such as RPT device 4000, that simulates various operations of the therapy device. In one form of the present technology, the therapy engine model 920 is a model of one or more components of an RPT device 4000, such as central controller 4230, therapy device controller 4240, therapy engine module 4320, and / or the like.

[0306] The flow signal 915 generated by the patient model 910 and represents patient respiratory airflow. The flow signal 915 is provided to the therapy engine model 920, which produces a pressure value 925 based on the flow signal 915. For instance, the therapy engine model 920 may be configured to analyze the flow signal 915 to detect respiratory events from the flow signal 915. The therapy engine model 920 may also be configured to respond to the detected respiratory events by increasing or decreasing the pressure 925 accordingly. The pressure value 925 is then fed back into the patient model 910, and is used during the next iteration of the patient model 910. Thus, in various forms of the present technology, the patient model 910 and the therapy engine model 920 may be run in a loop, or run for a predefined or preconfigured number of iterations or epochs. For example, the patient model 910 may output a flow signal 915 that simulates a narrowing airway as the pressure drops, and the therapy engine model 920 may increase the pressure 925 when respiratory events are detected or predicted based on the simulated narrowing airway.

[0307] In one example, the therapy engine model 920 is configured to simulate a virtual (AutoSet) APAP therapy device. CPAP therapy typically involves a "titration study," during which the optimal pressure required to alleviate the patient's OSA symptoms is determined. In contrast, APAP therapy eliminates the need for this study by automatically adjusting the pressure to provide a comparable therapeutic effect to a correctly titrated CPAP therapy. The titrated CPAP pressure can be used to approximate PCrit. However, PCrit can vary throughout the night due to factorssuch as body position, sleep stage, airway dynamics, alcohol consumption, and / or different hormonal features, the real-time adjustments made by APAP therapy offer an approximation of the patient's fluctuating Pcrit. Thus, in this example, the virtual (AutoSet) APAP therapy device may determine the appropriate pressure 925 by monitoring or detecting sleep disordered breathing events, such as indicators of: apneas, flow limitations, and snoring in the flow signal. When any of these indicators are detected, the AutoSet APAP therapy algorithm increases the CPAP pressure accordingly and may decrease pressure periodically in the absence of detected respiratory events.5.11.1.6 Simulation Data

[0308] In some forms of the present technology, a collection of pressure values 925 that have been generated over a period of time, such as a six hour time period, may be stored and recorded as a set of simulation data 940. Various analyses can be performed on the set of simulation data 940 after the simulation is run for a period of time. In one example, a waveform may be produced using the simulation data 940.

[0309] In some forms, the simulations may be run much faster than typical bench models. In one example implementations, a simulation of six hours of sleep can be run in 24 minutes. Thus, such simulations may provide time savings and reduce resource consumption in comparison to running conventional sleep study experiments or even therapy sessions with device used by the patient.5.11.2 Digital Twin Modelling Pipeline Operation

[0310] Fig. 9 also shows how the pipeline 900 can be used for developing, modelling, and validating PAP therapy digital twins. In one form of the present technology, the modelling pipeline 900 may include several steps or stages and is implemented separately for each participant and their associated digital twin.

[0311] At stage 1, patient data 950, such as previous clinical study, is injected into the modelling pipeline 900 to create digital twins for individual patients. Specifically, the dataset 950 includes time series pressure, and may optionally include other event data, such as respiratory event timing data obtained from medical devices and forms the input to the digital twin pipeline 900 at the Pcrit model 930. In one form, a pressure signal may be generated based on the dataset 950. The pressure signal may be a time-varying titrated pressure level determined by an autotitrating algorithm designed to set pressure or other therapy parameter to maintain the upper airway unobstructed based on detected respiratory events from device detection algorithms.

[0312] At stage 2, the Pcrit model 930 estimates the patient’s underlying latent model responsible for the increases and decreases in auto-titrated pressure during therapy. This latent model corresponds to the time varying patency of a patient’s upper airway, sometimes referred to as their critical pressure (Pcrit). The input pressure data is passed through a series of digital signalprocessing and feature extraction techniques to estimate a latent critical model. The estimated latent critical model forms the basis of the patient’s digital twin. This process and its associated hyperparameters are tuned through an iterative process described with respect to stage 5.

[0313] At stage 3, once the latent Pcrit model is estimated, a Pcrit signal 935 is generated by the Pcrit model 930. In one form, the Pcrit signal 935 is time-varying signal that reflects the patency of the patient’s upper airway during sleep. The Pcrit signal 935 dictates the pressure necessary to keep the upper airway unobstructed at the patient model 910. The underlying model, or digital twin, can generate signals of any specified length. In one example, during each iteration of the pipeline tuning, seven Pcrit signals 935, each having a 6-hour duration, were generated per digital twin to reflect each patient’s seven nights of sleep using PAP therapy.

[0314] At stage 4, the Pcrit signal 935 is input to the patient model 910, which runs in a continuous loop with the therapy engine model 920. In one form, the therapy engine model 920 is a digital implementation of an APAP algorithm, and therefore, the therapy engine model 920 may be referred to as an APAP model or the like. Within this closed loop simulation, the patient model 910 outputs the flow signal 915 to the therapy engine model 920 and the therapy engine model 920 outputs a titrated pressure level 925 to the patient model 910. Should the pressure level fall below the current value of the Pcrit signal 935, the patient model 910 will begin to simulate respiratory events including apneas and flow limitation. This in turn is detected by the APAP model, which may prescribe increases in the pressure 925 until the virtual patient’s breathing has stabilized (which may be indicated by a flow signal 915 of a subsequent iteration). The titrated pressure level may also be recorded over time and used for the tuning process described with respect to stage 5.

[0315] At stage 5, the digital twins are tested to iterate and improve the pipeline. Specifically, the therapy pressure 925 output from the APAP model is recorded and compared 945 with the real-world data in the dataset 950 for the purposes of pipeline evaluation and further tuning. In one form, the comparison and tuning is conducted at the level of the full dataset 950 with each participant being compared to their digital twin counterpart. For example, summary metrics of mean pressure, standard deviation of pressure, 95th percentile pressure, and device-scored AHI may be obtained from both physical and digital devices and compared using root mean square error. Time-series pressure signals may also be inspected visually for comparison.

[0316] The pipeline 900 may be tuned over one or multiple iterations. At each iteration, the various hyperparameters associated with the modelling and digital signal processing tasks for producing the latent critical pressure vectors, such as Pcrit signal 935, may be adjusted based on the comparison 945. At each iteration, the similarity of the digital twins and their real -worldcounterparts may be assessed 945, and learnings from the assessment may inform the next iteration of pipeline tuning.

[0317] The comparison 945 may involve extracting one or more metrics or features from the simulation data / signal 940, extracting the same metrics or features from the dataset 950, and determining the similarity between each of those metrics / features. Such metrics / features may include, for example, pressure level, standard deviation of the pressure, mean pressure, the residual AHI, and / or the like. For example, the pressure level extracted from the simulation data / signal 940 may be compared with the pressure level extracted from the dataset 950, and so forth. The tuning may involve applying any technique(s) that may reduce the divergence between each the metrics / features. For example, one or more parameters of models 910, 920, 930 may be adjusted to get the mean pressure from the simulated data 940 as close as possible to the mean pressure from the dataset 950.

[0318] In one form, evaluation of the pipeline 900 may involve comparing or otherwise evaluating the similarity of a participant and their digital twin against the similarity of that participant with all other twins. This similarity between participants and twin pressure distributions may be measured using Jensen Shannon (JS) Divergence, a symmetric variation of Kullback-Leibler Divergence, which may be expressed using equation (3).< <<

[0319] In equation (3), DJSP || Q) is the JS divergence for probability distributions P and Q, P(x) is a first probability distribution over a set of events x and represents the pressure, Q(x) is a second probability distribution over the same set of events x and represents the flow, M(x) is an average mixture distribution, the set of events x represent possible outcomes or elements of theprobability distributions, and represent the relative entropy (or Kullback-Leibler divergence) between each probability distribution and the mixture distribution. The size of the JS divergence may indicate the similarity of the two probability distributions. For example, when the JS divergence is zero (DjSP || Q) = 0), then the two probability distributions P(x) and Q(x) may be identical. The larger the JS divergence, the more dissimilar the two probability distributions P(x) and Q(x) may be from one another.

[0320] The pipeline accuracy in modelling a participant’s digital twin A, may be scored by subtracting the JS divergence between the participant and their digital twin counterpart from an average JS divergence between the participant and all other digital twins, which may be expressed using equation (4).

[0321] In equation (4), Aj is the pipeline accuracy score for participant i, N is the total number of participants, Ptis the real probability distribution of participant i’s data, Ttis the digital twin probability distribution for participant i, Tj is the digital twin probability distribution for another participant j, DJSPL|| 7}) is the JS divergence between participant i’s real data distribution and their digital twin and represents how well the model replicates participant i, and DjS(Pt || 7}) is the JS divergence between participant i’s real data distribution and the other participant j’s digital twin and represents different participant i is from participant j.

[0322] As alluded to previously, in some forms, a machine learning model may be used to learn the relationship between real -world data 950 and the latent critical pressure signal 935 used for digital twin modeling 930. For example, a digital twin machine learning model may be trained on historical clinical study data, where patient pressure signals from the dataset 950 are used as an input to create the latent Pcrit model of the pressure signals and predict (synthesize) the Pcrit vector 935. In another form, a machine learning model may be trained on historical data 950 and the simulated data 940 to perform the comparison 945. For example, such a machine learning model may assess the error between the real data 950 and the simulated data 940, which may then be fed back into the digital twin machine learning model to change the weights in order to try and minimize that error. In some forms, the training and / or the inference generation may be performed using a gradient descent optimization algorithm or the like.5.11.3 Example Results

[0323] Table 1 shows the root mean square error (RMSE) between metrics derived from physical and virtual PAP devices. Table 1 shows a comparable difference between real patient and digital twin produced therapy data. Such metrics are typically used by patients, providers, and clinicians for understanding a patient’s overnight of PAP therapy.Table 1: RMSE for device reported metrics between physical and digital data

[0324] RMSE shows relatively good agreement between a participant’s real data and data reconstructed from their digital twin. However, additional or alternative metrics may be used better understand the significance of this comparison.

[0325] Figure 15 shows a graph 1500 of an average pipeline accuracy score (AJS) 1510 and standard deviation 1520 across participants. The accuracy reflects the difference between the similarity of a matched twin pair and the similarity of non-matched twin pairs, with more positivevalues indicating better distinction. A one-sample t-test shows that AJS was significantly greater than 0 (T = 13.1 p < 5 x l 0l0, df = 18). This indicates that data from a patient was more similar to data from their own digital twin rather than data from the digital twins of other participants.

[0326] Figure 16 shows a comparison for three different patients and their digital twin counterparts. Each panel shows the collected and simulated pressure time series of different nights of data. Figure 16 highlights the ability of the modelling pipeline 900 to capture common temporal characteristics at difference scales of granularity which differ significantly across patients.

[0327] The results shown by Figs. 15-16 demonstrate that the digital twins of the present technology achieve relatively high accuracy compared to corresponding real-life counterparts. High reconstruction accuracy of device reported metrics was achieved allowing patients and clinicians to understand patient treatment, such as AHI and average pressure. Furthermore, the digital twin pipeline 900 is capable of reconstructing data that is very similar to the input patient, which allows multiple simulations to be run that can be used to diagnose and treat medical conditions, such as OSA or other sleep disorders. Moreover, a visual inspection of the pressure traces shows that the digital twins create a pressure signal that looks realistic to what is typically observed from device outputs. Mask leak is a common issue faced by PAP users, and including models for mask fit and / or mask leak may further enhance the accuracy of the digital twin pipeline.5.11.4 Example Methods

[0328] Fig. 17 shows an example process 1700 of simulating a patient for treating respiratory-related disorders. Process 1700 may be operated by any of the computing devices mentioned herein. Process 1700 begins at operation 1710 where a dataset including measured pressure data of the patient is obtained. At operation 1720, a Pcrit signal is generated based on the dataset. The Pcrit signal may represent a variable Pcrit value over a period of time. At operation 1730, a flow signal is generated based on the Pcrit signal and a first pressure value. At operation 1740, a second pressure value is generated based on the flow signal. The second pressure value may be used in a next iteration for generating a next flow signal based on the Pcrit signal (or a portion thereof) and the second pressure value.5.12 GLOSSARY

[0329] For the purposes of the present technology disclosure, in certain forms of the present technology, one or more of the following definitions may apply. In other forms of the present technology, alternative definitions may apply.5.12.1 General

[0330] Air '. In certain forms of the present technology, air may be taken to mean atmospheric air, and in other forms of the present technology air may be taken to mean some other combination of breathable gases, e.g. oxygen enriched air.

[0331] Ambient. In certain forms of the present technology, the term ambient will be taken to mean (i) external of the treatment system or patient, and (ii) immediately surrounding the treatment system or patient. For example, ambient humidity with respect to a humidifier may be the humidity of air immediately surrounding the humidifier, e.g. the humidity in the room where a patient is sleeping. Such ambient humidity may be different to the humidity outside the room where a patient is sleeping. In another example, ambient pressure may be the pressure immediately surrounding or external to the body. In certain forms, ambient (e.g., acoustic) noise may be considered to be the background noise level in the room where a patient is located, other than for example, noise generated by an RPT device or emanating from a mask or patient interface. Ambient noise may be generated by sources outside the room.

[0332] Automatic Positive Airway Pressure (APAP) therapy. CPAP therapy in which the treatment pressure is automatically adjustable, e.g. from breath to breath, between minimum and maximum limits, depending on the presence or absence of indications of SDB events.

[0333] Continuous Positive Airway Pressure (CPAP) therapy. Respiratory pressure therapy in which the treatment pressure is approximately constant through a respiratory cycle of a patient. In some forms, the pressure at the entrance to the airways will be slightly higher during exhalation, and slightly lower during inhalation. In some forms, the pressure will vary between different respiratory cycles of the patient, for example, being increased in response to detection of indications of partial upper airway obstruction, and decreased in the absence of indications of partial upper airway obstruction. For purposes of the present disclosure, the term “Continuous Positive Airway Pressure” and “CPAP” may also refer to the type of therapy provided by an APAP device, albeit where that therapy can adjust, and therefore, the terms “CPAP” and “APAP” may be understood herein to be used interchangeably herein unless explicitly stated otherwise.

[0334] Flow rate. The volume (or mass) of air delivered per unit time. Flow rate may refer to an instantaneous quantity. In some cases, a reference to flow rate will be a reference to a scalar quantity, namely a quantity having magnitude only. In other cases, a reference to flow rate will be a reference to a vector quantity, namely a quantity having both magnitude and direction. Flow rate may be given the symbol Q. ‘Flow rate’ is sometimes shortened to simply ‘flow’ or ‘airflow’. In the example of patient respiration, a flow rate may be nominally positive for the inspiratory portion of a breathing cycle of a patient, and hence negative for the expiratory portion of the breathing cycle of a patient. Device flow rate, Qd, is the flow rate of air leaving the RPT device. Total flow rate, Qt, is the flow rate of air and any supplementary gas reaching the patient interface via the air circuit. Vent flow rate, Qv, is the flow rate of air leaving a vent to allow washout of exhaled gases. Leak flow rate, QI, is the flow rate of leak from a patient interface system or elsewhere. Respiratory flow rate, Qr, is the flow rate of air that is received into the patient's respiratory system.

[0335] Flow therapy. Respiratory therapy comprising the delivery of a flow of air to an entrance to the airways at a controlled flow rate referred to as the treatment flow rate that is typically positive throughout the patient’s breathing cycle.

[0336] Humidifier'. The word humidifier will be taken to mean a humidifying apparatus constructed and arranged, or configured with a physical structure to be capable of providing a therapeutically beneficial amount of water (H2O) vapour to a flow of air to ameliorate a medical respiratory condition of a patient.

[0337] Leak'. The word leak will be taken to be an unintended flow of air. In one example, leak may occur as the result of an incomplete seal between a mask and a patient's face. In another example leak may occur in a swivel elbow to the ambient.

[0338] Noise, conducted (acoustic)'. Conducted noise in the present document refers to noise which is carried to the patient by the pneumatic path, such as the air circuit and the patient interface as well as the air therein. In one form, conducted noise may be quantified by measuring sound pressure levels at the end of an air circuit.

[0339] Noise, radiated (acoustic)'. Radiated noise in the present document refers to noise which is carried to the patient by the ambient air. In one form, radiated noise may be quantified by measuring sound power / pressure levels of the object in question according to ISO 3744.

[0340] Noise, vent (acoustic)'. Vent noise in the present document refers to noise which is generated by the flow of air through any vents such as vent holes of the patient interface.

[0341] Oxygen enriched air'. Air with a concentration of oxygen greater than that of atmospheric air (21%), for example at least about 50% oxygen, at least about 60% oxygen, at least about 70% oxygen, at least about 80% oxygen, at least about 90% oxygen, at least about 95% oxygen, at least about 98% oxygen, or at least about 99% oxygen. “Oxygen enriched air” is sometimes shortened to “oxygen”.

[0342] Medical Oxygen'. Medical oxygen is defined as oxygen enriched air with an oxygen concentration of 80% or greater.

[0343] Patient'. A person, whether or not they are suffering from a respiratory condition.

[0344] Pressure: Force per unit area. Pressure may be expressed in a range of units, including centimeters of water (cmFFO), grams-force per square centimeter (g-f / cm2), hectopascal (hPa), Newtons per square meter (N / m2), millibars (mb), atmosphere (atm), pounds per square inch (psi), and the like. 1 cmFFO is equal to 1 g-f / cm2and is approximately 0.98 hPa (1 hPa = 100 Pascals (Pa) = 100 N / m2= 1 mb ~ 0.001 atm). In this specification, unless otherwise stated, pressure is given in units of cmFhO. The pressure in the patient interface is given the symbol Pm, while the treatment pressure, which represents a target value to be achieved by the interface pressure Pm at the current instant of time, is given the symbol Pt.

[0345] Respiratory Pressure Therapy. The application of a supply of air to an entrance to the airways at a treatment pressure that is typically positive with respect to atmosphere.

[0346] Ventilator'. A mechanical device that provides pressure support to a patient to perform some or all of the work of breathing.

[0347] Seal'. May be a noun form ("a seal") which refers to a structure, or a verb form (“to seal”) which refers to the effect. Two elements may be constructed and / or arranged to ‘seal’ or to effect ‘sealing’ therebetween without requiring a separate ‘seal’ element per se.

[0348] Shell'. A shell will be taken to mean a curved, relatively thin structure having bending, tensile and compressive stiffness. For example, a curved structural wall of a mask may be a shell. In some forms, a shell may be faceted. In some forms a shell may be airtight. In some forms a shell may not be airtight.

[0349] Sleep Stages'. Sleep occurs in five stages, including wake / alert, Nl, N2, N3, and rapid eye movement (REM). Stages Nl to N3 are considered non-rapid eye movement (NREM) sleep, with each stage leading to progressively deeper sleep. Sleep staging events may refer to events related to one or more sleep stages and / or features within biosignals and / or biopotential signals that are indicative of one or more sleep stages.

[0350] Stiffener. A stiffener will be taken to mean a structural component designed to increase the bending resistance of another component in at least one direction.

[0351] Strut'. A strut will be taken to be a structural component designed to increase the compression resistance of another component in at least one direction.

[0352] Swivel (noun): A subassembly of components configured to rotate about a common axis, preferably independently, preferably under low torque. In one form, the swivel may be constructed to rotate through an angle of at least 360 degrees. In another form, the swivel may be constructed to rotate through an angle less than 360 degrees. When used in the context of an air delivery conduit, the sub-assembly of components preferably comprises a matched pair of cylindrical conduits. There may be little or no leak flow of air from the swivel in use.5.12.2 Respiratory cycle

[0353] Apnea. According to some definitions, an apnea is said to have occurred when flow falls below a predetermined threshold for a duration, e.g. 10 seconds. An obstructive apnea will be said to have occurred when, despite patient effort, some obstruction of the airway does not allow air to flow. A central apnea will be said to have occurred when an apnea is detected that is due to a reduction in breathing effort, or the absence of breathing effort, despite the airway being patent. A mixed apnea occurs when a reduction or absence of breathing effort coincides with an obstructed airway.

[0354] Breathing rate '. The rate of spontaneous respiration of a patient, usually measured in breaths per minute.

[0355] Duty cycle '. The ratio of inhalation time, Ti to total breath time, Ttot.

[0356] Effort (breathing): The work done by a spontaneously breathing person attempting to breathe.

[0357] Expiratory portion of a breathing cycle'. The period from the start of expiratory flow to the start of inspiratory flow.

[0358] Flow limitation'. Flow limitation will be taken to be the state of affairs in a patient's respiration where an increase in effort by the patient does not give rise to a corresponding increase in flow. Where flow limitation occurs during an inspiratory portion of the breathing cycle it may be described as inspiratory flow limitation. |Where flow limitation occurs during an expiratory portion of the breathing cycle it may be described as expiratory flow limitation.

[0359] Types of flow limited inspiratory waveforms:(i) Flattened: Having a rise followed by a relatively flat portion, followed by a fall. (ii) M-shaped: Having two local peaks, one at the leading edge, and one at the trailing edge, and a relatively flat portion between the two peaks.(iii) Chair-shaped: Having a single local peak, the peak being at the leading edge, followed by a relatively flat portion.(iv) Reverse-chair shaped: Having a relatively flat portion followed by single local peak, the peak being at the trailing edge.

[0360] Efypopnea'. According to some definitions, a hypopnea is taken to be a reduction in flow, but not a cessation of flow. In one form, a hypopnea may be said to have occurred when there is a reduction in flow below a threshold rate for a duration. A central hypopnea will be said to have occurred when a hypopnea is detected that is due to a reduction in breathing effort. In one form in adults, either of the following may be regarded as being hypopneas:(i) a 30% reduction in patient breathing for at least 10 seconds plus an associated 4% desaturation; or(ii) a reduction in patient breathing (but less than 50%) for at least 10 seconds, with an associated desaturation of at least 3% or an arousal.

[0361] Hyperpnea'. An increase in flow to a level higher than normal.

[0362] Inspiratory portion of a breathing cycle: The period from the start of inspiratory flow to the start of expiratory flow will be taken to be the inspiratory portion of a breathing cycle.

[0363] Patency (airway): The degree of the airway being open, or the extent to which the airway is open. A patent airway is open. Airway patency may be quantified, for example with a value of one (1) being patent, and a value of zero (0), being closed (obstructed).

[0364] Positive End-Expiratory Pressure (PEEP)'. The pressure above atmosphere in the lungs that exists at the end of expiration.

[0365] Peak flow rate (Qpeak): The maximum value of flow rate during the inspiratory portion of the respiratory flow waveform.

[0366] Respiratory flow rate, patient airflow rate, respiratory airflow rate (Qr): These terms may be understood to refer to the RPT device’s estimate of respiratory flow rate, as opposed to “true respiratory flow rate” or “true respiratory flow rate”, which is the actual respiratory flow rate experienced by the patient, usually expressed in litres per minute.

[0367] Tidal volume (Vt): The volume of air inhaled or exhaled during normal breathing, when extra effort is not applied. In principle the inspiratory volume Vi (the volume of air inhaled) is equal to the expiratory volume Ve (the volume of air exhaled), and therefore a single tidal volume Vt may be defined as equal to either quantity. In practice the tidal volume Vt is estimated as some combination, e.g. the mean, of the inspiratory volume Vi and the expiratory volume Ve.

[0368] Inhalation Time (Ti): The duration of the inspiratory portion of the respiratory flow rate waveform.

[0369] Exhalation Time (Te): The duration of the expiratory portion of the respiratory flow rate waveform.

[0370] Total Time (Ttot): The total duration between the start of one inspiratory portion of a respiratory flow rate waveform and the start of the following inspiratory portion of the respiratory flow rate waveform.

[0371] Typical recent ventilation'. The value of ventilation around which recent values of ventilation Vent over some predetermined timescale tend to cluster, that is, a measure of the central tendency of the recent values of ventilation.

[0372] Upper airway obstruction (UAO): includes both partial and total upper airway obstruction. This may be associated with a state of flow limitation, in which the flow rate increases only slightly or may even decrease as the pressure difference across the upper airway increases (Starling resistor behaviour).

[0373] Ventilation (Vent): A measure of a rate of gas being exchanged by the patient’s respiratory system. Measures of ventilation may include one or both of inspiratory and expiratory flow, per unit time. When expressed as a volume per minute, this quantity is often referred to as “minute ventilation”. Minute ventilation is sometimes given simply as a volume, understood to be the volume per minute.5.12.3 Ventilation

[0374] Adaptive Servo-Ventilator (AS V) : A servo-ventilator that has a changeable, rather than fixed target ventilation. The changeable target ventilation may be learned from some characteristic of the patient, for example, a respiratory characteristic of the patient.

[0375] Backup rate. A parameter of a ventilator that establishes the minimum breathing rate (typically in number of breaths per minute) that the ventilator will deliver to the patient, if not triggered by spontaneous respiratory effort.

[0376] Cycled. The termination of a ventilator's inspiratory phase. When a ventilator delivers a breath 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 the breath.

[0377] Expiratory positive airway pressure (EPAP): a base pressure, to which a pressure varying within the breath is added to produce the desired interface pressure which the ventilator will attempt to achieve at a given time.

[0378] End expiratory pressure (EEP): Desired interface pressure which the ventilator will attempt to achieve at the end of the expiratory portion of the breath. If the pressure waveform template n(cp) is zero-valued at the end of expiration, i.e. n(cp) = 0 when cp = 1, the EEP is equal to the EPAP.

[0379] Inspiratory positive airway pressure (IPAP): Maximum desired interface pressure which the ventilator will attempt to achieve during the inspiratory portion of the breath.

[0380] Pressure support'. A number that is indicative of the increase in pressure during ventilator inspiration over that during ventilator expiration, and generally means the difference in pressure between the maximum value during inspiration and the base pressure (e.g., PS = IPAP -EPAP). In some contexts, pressure support means the difference which the ventilator aims to achieve, rather than what it actually achieves.

[0381] Servo-ventilator'. A ventilator that measures patient ventilation, has a target ventilation, and which adjusts the level of pressure support to bring the patient ventilation towards the target ventilation.

[0382] Spontaneous / Timed (S / T): A mode of a ventilator or other device that attempts to detect the initiation of a breath of a spontaneously breathing patient. If however, the device is unable to detect a breath within a predetermined period of time, the device will automatically initiate delivery of the breath.

[0383] Swing'. Equivalent term to pressure support.

[0384] Triggered. When a ventilator, or other respiratory therapy device such as an RPT device or portable oxygen concentrator, delivers a volume of breathable gas to a spontaneouslybreathing patient, it is said to be triggered to do so. Triggering usually takes place at or near the initiation of the respiratory portion of the breathing cycle by the patient's efforts.5.12.4 Data Processing, Data Analytics, and Machine Learning

[0385] Classifier', any system or methodology that performs classification, wherein classification involves categorizing or assigning objects or data points to specific classes or categories, which may sometimes be called “targets” or “labels.” Classifiers may include ML classifiers or non-ML classifiers. Non-ML classifiers typically use predefined rules, heuristics, or algorithms to perform classification. Examples of non-ML classifiers include rules-based classifiers, heuristics classifiers, pattern matching classifiers, knowledge-based classifiers, and hand-crafted or manually designed classifiers. ML classifiers typically learn patterns from data. More specifically, an ML classifier is an ML algorithm or ML model that takes input data, which can be in the form of features or attributes, and assigns each data point to one of several classes or categories. The classes or categories may be predefined or learned. The process of classification involves training the classifier on labeled training data, where each data point is associated with a known class label. During training, the classifier learns to identify patterns, trends, and decision boundaries in the input data that differentiate between different classes. Once trained, the classifier can then be used to predict the class labels of new, unseen data points (e.g., referred to as an “inference dataset” or the like) based on the learned patterns and relationships. Examples of ML classifiers include linear classifiers such as logistic regression and perceptrons, k-nearest neighbor (kNN), decision trees, random forests, support vector machines (SVMs), Bayesian classifiers, CNNs, RNNs, among many others (note that some of these algorithms can be used for other ML tasks as well).

[0386] Digital twin. A digital representation or virtual replica of a physical object, person, system, or process contextualized in a digital version of its environment. The digital representation or virtual replica may be generated using real-time data and simulations. Digital twins are typically used to simulate real-world situations and their outcomes to optimize performance, predict outcomes, and support decision-making.

[0387] Epoch'. One cycle through a full training dataset during an ML training process. An epoch may also be a full training pass over an entire training dataset such that each training example has been seen once. In some examples, an epoch represents N / batch size training iterations, where N is the total number of examples.

[0388] Feature. A measurable and / or quantifiable property, and / or a characteristic of a phenomenon being observed. Additionally or alternatively, a feature may refer to an input variable used in making predictions / inferences. Features may be represented using numbers / numerals (e.g., integers), strings, variables, ordinals, real-values, categories, and / or the like.

[0389] Filterbank-, a layer in a neural network that applies a set of filters or kernels to input data, usually in parallel. Each filter captures different aspects or features of the input, resulting in a multi-channel output. The filters in a filterbank layer can have different sizes, shapes, and properties, allowing the network to capture diverse features at different spatial or frequency scales.

[0390] Hyperparameter-. Characteristics, properties, and / or parameters for an ML process that are not learned during a training process. Hyperparameters are usually set before training takes place, and may be used in processes to help estimate model parameters. Examples of hyperparameters include model size (e.g., in terms of memory space, bytes, number of layers, and the like); training data shuffling (e.g., whether to do so and by how much); number of evaluation instances, iterations, epochs (e.g., a number of iterations or passes over the training data), or episodes; number of passes over training data; regularization; learning rate (e.g., the speed at which the algorithm reaches (converges to) optimal weights); learning rate decay (or weight decay); momentum; number of hidden layers; size of individual hidden layers; weight initialization scheme; dropout and gradient clipping thresholds; the C value and sigma value for SVMs; the k in k-nearest neighbors; number of branches in a decision tree; number of clusters in a clustering algorithm; vector size; word vector size for NLP and NLU; and / or the like.

[0391] Model parameter-. Values, characteristics, and / or properties that are learnt during ML training. A model parameter may be a configuration variable that is internal to the model and whose value can be estimated from the given data. Model parameters are usually required by a model when making predictions or inferences, and their values define the skill of the model on a particular problem. Examples of such model parameters include weights, biases, constraints, classes or categories, support vectors in an SVM, coefficients in a linear regression and / or logistic regression, and / or the like.

[0392] Inference '. The process of utilizing a trained ML model to extract meaningful insights, make predictions, and / or take actions based on new data. For example, inference generation in supervised learning involves making predictions based on labeled data, inference generation in unsupervised learning involves uncovering patterns in unlabeled data, and inference generation in reinforcement learning involves decision-making based on learned policies and environmental feedback. For purposes of the present disclosure, the term “inference”, “inference generation”, or the like, refers to the process of using trained ML model(s) to generate statistical inferences, predictions, decisions, probabilities, probability distributions, actions, configurations, policies, data analytics, outcomes, optimizations, and / or the like based on new, unseen data (e.g., “input inference data”).

[0393] Iteration'. The repetition of a process in order to generate a sequence of outcomes, wherein each repetition of the process is a single iteration, and the outcome of each iteration is thestarting point of the next iteration. Additionally or alternatively, an iteration may refer to a single update of a model’s weights during training.

[0394] Optimal', the best or most desirable, favorable, functional, satisfactory, advantageous, efficient, or effective design(s), configuration, arrangement, operation(s), decision(s), condition(s), criteria, parameter(s), allocation(s), variable(s), solution(s), result(s), and / or output(s) considering a set of constraints, criteria, and / or parameters. An “optimum” may refer to an amount or degree to which something is optimal. The term “optimization” may refer to an act, process, algorithm, or methodology of making something as optimal as possible. Optimization typically includes mathematical procedures, such as finding the maximum or minimum of a function, or solving one or more loss functions and / or objective functions.

[0395] Pipeline', a series of interconnected steps or stages through which data flows to be transformed, processed, and delivered in a desired format and / or for a specific purpose. In a pipeline, the output of at least one step or stage becomes the input for at least one other step or stage.

[0396] Softmax'. a function that converts a vector of numerical values into a probability distribution. The softmax function is often used in ML models, such as those involving classification tasks where the model needs to output probabilities for multiple classes. The softmax function may be used as the last activation function of a neural network, to normalize the output of the ML model to a probability distribution over predicted output classes.5.13 OTHER REMARKS

[0397] A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in Patent Office patent files or records, but otherwise reserves all copyright rights whatsoever.

[0398] Unless the context clearly dictates otherwise and where a range of values is provided, it is understood that each intervening value, to the tenth of the unit of the lower limit, between the upper and lower limit of that range, and any other stated or intervening value in that stated range is encompassed within the technology. The upper and lower limits of these intervening ranges, which may be independently included in the intervening ranges, are also encompassed within the technology, subject to any specifically excluded limit in the stated range. Where the stated range includes one or both of the limits, ranges excluding either or both of those included limits are also included in the technology.

[0399] Furthermore, where a value or values are stated herein as being implemented as part of the technology, it is understood that such values may be approximated, unless otherwise stated,and such values may be utilized to any suitable significant digit to the extent that a practical technical implementation may permit or require it.

[0400] Furthermore, “approximately”, “substantially”, “about”, or any similar term used herein means + / - 5-10% of the recited value.

[0401] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this technology belongs. Although any methods and materials similar or equivalent to those described herein can also be used in the practice or testing of the present technology, a limited number of the exemplary methods and materials are described herein.

[0402] When a particular material is identified as being used to construct a component, obvious alternative materials with similar properties may be used as a substitute. Furthermore, unless specified to the contrary, any and all components herein described are understood to be capable of being manufactured and, as such, may be manufactured together or separately.

[0403] It must be noted 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.

[0404] All publications mentioned herein are incorporated herein by reference in their entirety to disclose and describe the methods and / or materials which are the subject of those publications. The publications discussed herein are provided solely for their disclosure prior to the filing date of the present application. Nothing herein is to be construed as an admission that the present technology is not entitled to antedate such publication by virtue of prior invention. Further, the dates of publication provided may be different from the actual publication dates, which may need to be independently confirmed.

[0405] The terms "comprises" and "comprising" should be interpreted as referring to elements, components, or steps in a non-ex elusive manner, indicating that the referenced elements, components, or steps may be present, or utilized, or combined with other elements, components, or steps that are not expressly referenced.

[0406] The subject headings used in the detailed description are included only for the ease of reference of the reader and should not be used to limit the subject matter found throughout the disclosure or the claims. The subject headings should not be used in construing the scope of the claims or the claim limitations.

[0407] Although the technology herein has been described with reference to particular examples, it is to be understood that these examples are merely illustrative of the principles and applications of the technology. In some instances, the terminology and symbols may imply specific details that are not required to practice the technology. For example, although the terms "first" and "second" may be used, unless otherwise specified, they are not intended to indicate any order butmay be utilised to distinguish between distinct elements. Furthermore, although process steps in the methodologies may be described or illustrated in an order, such an ordering is not required. Those skilled in the art will recognize that such ordering may be modified and / or aspects thereof may be conducted concurrently or even synchronously.

[0408] It is therefore to be understood that numerous modifications may be made to the illustrative examples and that other arrangements may be devised without departing from the spirit and scope of the technology.

Claims

6 CLAIMS1. A method of simulating a patient for treating respiratory -related disorders, the method comprising:obtaining, by one or more processors, a dataset including measured pressure data of the patient;generating, by the one or more processors, a critical pressure (Pcrit) signal based on the dataset, the Pcrit signal representing a variable Pcrit value over a period of time;generating, by the one or more processors, a flow signal based on the Pcrit signal and a first pressure value; andgenerating, by the one or more processors, a second pressure value based on the flow signal.

2. The method of claim 0, further comprising:generating, by the one or more processors, a Pcrit model based on the dataset.

3. The method of claim 2, wherein the Pcrit signal is generated based on the Pcrit model.

4. The method of any one of claims 0 to 3, further comprising:concatenating, by the one or more processors, the measured pressure data into a single pressure signal.

5. The method of claim 4, further comprising:applying, by the one or more processors, a fast Fourier Transform to the single pressure signal to identify dominant frequency components in the single pressure signal.

6. The method of claim 5, further comprising:performing, by the one or more processors, derivative processing on the single pressure signal to identify positive pressure increases.

7. The method of claim 6, further comprising:generating, by the one or more processors, an impulse signal of positive pressure increases based on the identified positive pressure increases.

8. The method of any one of claims 5 to 7, further comprising:segmenting, by the one or more processors, the single pressure signal into a set of segments based on the identified dominant frequency components.

9. The method of claim 8, wherein each segment of the set of segments has a same duration.

10. The method of any one of claims 8 to 9, further comprising:fitting, by the one or more processors, an exponential distribution to interarrival times of each segment of the set of segments.

11. The method of claim 10, wherein fitting the exponential distribution to the interarrival times of each segment produces an array of rate parameters.

12. The method of claim 11, further comprising:generating, by the one or more processors, an impulse vector based on the impulse signal of positive pressure increases and the array of rate parameters.

13. The method of claim 12, wherein the impulse signal of positive pressure increases is sampled by sampling with replacement to select samples of increases that are below a threshold.

14. The method of any one of claims 12 to 13, wherein the array of rate parameters is sampled with replacement.

15. The method of any one of claims 12 to 14, further comprising:generating, by the one or more processors, the Pcrit signal based on event timings of the dataset by inserting impulses into the impulse vector based on the event timings, wherein the event timings concern measured pressure attributable to respiratory events.

16. The method of claim 15, wherein the inserted impulses have a height equal to a maximum pressure at rates equal to a rate of a corresponding respiratory event.

17. The method of any one of claims 0 to 16, wherein the flow signal is generated using a patient model.

18. The method of claim 17, wherein the patient model simulates human cardiorespiratory behaviors based on the first pressure value and the Pcrit signal.

19. The method of claim 17 or claim 18, wherein the patient model simulates human cardio-respiratory behaviors based on the second pressure value and the Pcrit signal.

20. The method of claim 18, wherein the patient model includes a plurality of compartments, each compartment of the plurality of compartments corresponding to a physiological system.

21. The method of any one of claims 17 to 20, wherein when a pressure introduced to the patient model drops below a Pcrit value of the Pcrit signal, the patient model generates the flow signal to indicate one or more respiratory events.

22. The method of any one of claims 0 to 21, wherein the first pressure value is generated using a therapy engine model.

23. The method of claim 22, wherein the second pressure value is generated using the therapy engine model.

24. The method of any one of claims 0 to 21, wherein the second pressure value is generated using a therapy engine model.

25. The method of claim 24, wherein the second pressure value is larger than the first pressure value when the therapy engine model identifies at least one respiratory event in the flow signal.

26. The method of any one of claims 0 to 25, wherein the first pressure value and the second pressure value are stored as simulation data.

27. The method of claim 26, further comprising:comparing, by the one or more processors, the simulation data with the measured pressure data.

28. The method of claim 27, further comprising:adjusting, by the one or more processors, at least one parameter for generating the Pcrit signal based on the comparison.

29. The method of claim 28, wherein the comparison is based on a Jensen Shannon divergence.

30. The method of any one of claims 1 to 29, wherein the method comprises iteratively applying one or more new pressure values for further generating the flow signal based on the Pcrit signal, wherein the one or more new pressure values are based on the flow signal from the further generating.

31. A computer-readable medium encoded with computer-readable instructions, which when executed by one or more processors implement the method of any one of claims 0 to 2930.

32. A system for simulating a patient for treating respiratory-related disorders, the system comprising:a memory configured to store instructions related to a digital twin pipeline; and at least one processor connected to the memory, the at least one processor is configured to execute the instructions to:obtain a dataset including measured pressure data of the patient; generate a critical pressure (Pcrit) signal based on the dataset, the Pcrit signal representing a variable Pcrit value over a period of time;generate a flow signal based on the Pcrit signal and a first pressure value; andgenerate a second pressure value based on the flow signal.

33. The system of claim 32, wherein the at least one processor is configured to execute the instructions to: generate a Pcrit model based on the dataset.

34. The system of claim 33, wherein the Pcrit signal is generated based on the Pcrit model.

35. The system of any one of claims 32 to 34, wherein the at least one processor is configured to execute the instructions to: concatenate the measured pressure data into a single pressure signal.

36. The system of claim 35, wherein the at least one processor is configured to execute the instructions to: apply a fast Fourier Transform to the single pressure signal to identify dominant frequency components in the single pressure signal.

37. The system of claim 34 or claim 36, wherein the at least one processor is configured to execute the instructions to: perform derivative processing on the single pressure signal to identify positive pressure increases.

38. The system of claim 37, wherein the at least one processor is configured to execute the instructions to: generate an impulse signal of positive pressure increases based on the identified positive pressure increases.

39. The system of any one of claims 36 to 38, wherein the at least one processor is configured to execute the instructions to: segment the single pressure signal into a set of segments based on the identified dominant frequency components.

40. The system of claim 39, wherein each segment of the set of segments has a same duration.

41. The system of any one of claims 39 to 40, wherein the at least one processor is configured to execute the instructions to: fit an exponential distribution to interarrival times of each segment of the set of segments.

42. The system of claim 41, wherein fitting the exponential distribution to the interarrival times of each segment produces an array of rate parameters.

43. The system of claim 42, wherein the at least one processor is configured to execute the instructions to: generate an impulse vector based on the impulse signal of positive pressure increases and the array of rate parameters.

44. The system of claim 43, wherein the impulse signal of positive pressure increases is sampled by sampling with replacement to select samples of increases that are below a threshold.

45. The system of any one of claims 43to 44, wherein the array of rate parameters is sampled with replacement.

46. The system of any one of claims 43 to 45, wherein the at least one processor is configured to execute the instructions to: generate the Pcrit signal based on event timings of the dataset by inserting impulses into the impulse vector based on the eventtimings, wherein the event timings concern measured pressure attributable to respiratory events.

47. The system of claim 46, wherein the inserted impulses have a height equal to a maximum pressure at rates equal to a rate of a corresponding respiratory event.

48. The system of any one of claims 32 to 47, wherein the flow signal is generated using a patient model.

49. The system of claim 48, wherein the patient model simulates human cardiorespiratory behaviors based on the first pressure value and the Pcrit signal.

50. The system of claim 48 or claim 49, wherein the patient model simulates human cardio-respiratory behaviors based on the second pressure value and the Pcrit signal.

51. The system of any one of claims 48 to 50, wherein the patient model includes a plurality of compartments, each compartment of the plurality of compartments corresponding to a physiological system.

52. The system of any one of claims 48 to 51, wherein when a pressure introduced to the patient model drops below a Pcrit value of the Pcrit signal, the patient model generates the flow signal to indicate one or more respiratory events.

53. The system of any one of claims 32 to 52, wherein the first pressure value is generated using a therapy engine model.

54. The system of claim 53, wherein the second pressure value is generated using the therapy engine model.

55. The system of any one of claims 32 to 52, wherein the second pressure value is generated using a therapy engine model.

56. The system of claim 55, wherein the second pressure value is larger than the first pressure value when the therapy engine model identifies at least one respiratory event in the flow signal.

57. The system of any one of claims 32 to 56, wherein the first pressure value and the second pressure value are stored as simulation data.

58. The system of claim 57, wherein the at least one processor is configured to execute the instructions to: compare the simulation data with the measured pressure data.

59. The system of claim 58, wherein the at least one processor is configured to execute the instructions to: adjust at least one parameter for generating the Pcrit signal based on the comparison.

60. The system of claim 59, wherein the comparison is based on a Jensen Shannon divergence.

61. The system of any one of claims 32 to 60, wherein the at least on processor is configured to iteratively apply one or more new pressure values for further generating the flow signal based on the Pcrit signal, wherein the one or more new pressure values are based on the flow signal from the further generating.

62. The system of any one of claims 32 to 61, further comprising: a pressure device configured to deliver a flow of pressurised air to a patient interface that is, in use, connected to an airway of the patient.

63. The system of claim 62, wherein the at least one processor is a processor in a controller of a respiratory pressure therapy device.

64. The system of any one of claims 62 to 63, further comprising: one or more sensors configured to monitor one or more characteristics of the pressurised air.

65. The system of claim 64, wherein the at least one processor is configured to execute the instructions to generate the dataset including measured pressure data of the patient based on the monitoring of the one or more characteristics of the pressurised air.

66. The system of any one of claims 32 to 61, wherein the at least one processor is a processor of at least one server computer system.

67. The system of any one of claims 32 to 61, wherein the at least one processor is a processor of a client computer system.

68. A system for simulating a patient for treating respiratory-related disorders, the system comprising:a memory configured to store instructions related to a digital twin pipeline; andat least one processor connected to the memory, the at least one processor is configured to execute the instructions to:obtain, from at least one medical device, time series pressure and respiratory event timing data related to a patient;operate a critical pressure (Pcrit) model to estimate a latent model related to the patient, the Pcrit model to adjust increases and decreases in auto-titrated pressure during therapy, and generate a Pcrit signal based on the latent model;operate a patient model to generate a flow signal based on a respiratory simulation of the patient, the respiratory simulation of the patient being based on the Pcrit signal and a pressure level; andoperate a therapy engine model to generate a new pressure level based on the flow signal, and output new pressure level to the patient model for a next iteration of the respiratory simulation of the patient.

69. The system of claim 68, wherein the at least one processor is configured to execute the instructions to: store each generated pressure level as simulation data.

70. The system of claim 69, wherein the at least one processor is configured to execute the instructions to: compare the simulation data with the time series pressure and respiratory event timing data.

71. The system of claim 70, wherein the at least one processor is configured to execute the instructions to: tune at least one parameter of at least one of the Pcrit model, the patient model, or the therapy engine model based on the comparison of the simulation data with the time series pressure and respiratory event timing data.

72. A method of operating a digital twin pipeline to simulate a patient for treating respiratory-related disorders, the method comprising:obtaining, by at least one processor from at least one medical device, time series pressure and respiratory event timing data related to a patient;operating, by the at least one processor, a critical pressure (Pcrit) model to estimate a latent model related to the patient, the Pcrit model to adjust increases and decreases in auto-titrated pressure during therapy, and generate a Pcrit signal based on the latent model;operating, by the at least one processor, a patient model to generate a flow signal based on a respiratory simulation of the patient, the respiratory simulation of the patient being based on the Pcrit signal and a pressure level; andoperating, by the at least one processor, a therapy engine model to generate a new pressure level based on the flow signal, and output new pressure level to the patient model for a next iteration of the respiratory simulation of the patient.

73. The method of claim 72, further comprising: storing, by the at least one processor, each generated pressure level as simulation data.

74. The method of claim 73, further comprising: comparing, by the at least one processor, the simulation data with the time series pressure and respiratory event timing data.

75. The method of claim 74, further comprising: tuning, by the at least one processor, at least one parameter of at least one of the Pcrit model, the patient model, or the therapy engine model based on the comparison of the simulation data with the time series pressure and respiratory event timing data.

76. A computer-readable medium encoded with computer-readable instructions, which when executed by one or more processors implement the method of any one of claims 72 to 75.

Images