System and method for controlling pulse oximetry sensor

Abstract

A system and method for controlling a pulse oximetry sensor. The pulse oximetry sensor is activated and deactivated in accordance with one or more respiratory characteristics of the subject's respiration such that the pulse oximetry sensor acquires pulse oximetry data only when the one or more respiratory characteristics indicate that a reduction in oxygen saturation is likely to occur.

Description

Technical Field

[0001] The present invention relates to the field of pulse oximetry and in particular to controlling a pulse oximetry sensor. Background Art

[0002] Obstructive sleep-disordered breathing (SDB) is an increasingly common and concerning syndrome of upper airway dysfunction during sleep characterized by snoring and / or respiratory effort secondary to high upper airway resistance and pharyngeal collapse.

[0003] A subset of SDB patients suffer from obstructive sleep apnea (OSA) that requires treatment. Traditionally, OSA diagnosis is achieved by performing a PSG in a sleep laboratory or a home sleep apnea test (HSAT), which involves measuring multiple parameters, including respiratory airflow, chest movement, and heart rate, which involves a dedicated measuring device. This test is usually prescribed by a sleep physician based on the symptoms of the patient requesting the visit.

[0004] However, this approach is cumbersome, time-consuming, and resource-intensive. Moreover, many people do not seek medical care for their illnesses because they are unaware and fail to take action.

[0005] Higher coverage could be achieved by applying focused screening techniques that use measurements from medical or consumer wearable devices to indicate to people that they may be suffering from OSA.

[0006] To this end, wearable devices have been developed to measure sleep-disordered breathing. A common method of screening for OSA and similar conditions is to monitor oxygen desaturation events during sleep and, when necessary, provide feedback to the user to guide them to the correct diagnosis.

[0007] One conventional method for measuring oxygen desaturation events or levels is to utilize pulse oximetry. Traditionally, pulse oximetry sensors are able to estimate the oxygen content in the blood by optically probing the peripheral capillary bed, such as using transmission or reflection photoplethysmography (PPG). The principles of measuring pulse oximetry using this technology are well known in the art.

[0008] There is a continuing desire to improve the efficiency of pulse oximetry sensors and the clinical relevance of pulse oximetry data.

[0009] US2020 / 330030A1 discloses a physiological state monitoring device, including a motion sensor, an event detector and an estimator.

[0010] US2017 / 087050A1 discloses a device for measuring signs of snoring and sleep apnea and generating audible sounds to wake up the subject.

[0011] Ng et al. published “Could formant frequencies of snore signals be an alternative means for the diagnosis of obstructive sleep apnea?” in 2008, Sleep Medicine 9(8):894–898, which disclosed the difference in resonant peak frequency between apneic snoring and benign snoring. Summary of the invention

[0012] The invention is defined by the claims.

[0013] According to an example of one aspect of the invention, there is provided a processing system for controlling a pulse oximetry sensor, the processing system being configured to: acquire respiration data responsive to respiration of a subject from a detector distinct from the pulse oximetry sensor; process the respiration data to determine one or more respiration characteristics; and control the pulse oximetry sensor responsive to the one or more respiration characteristics.

[0014] The inventors have recognized that a subject's breathing pattern during sleep, particularly whether and how the subject snores, can be indicative of sleep disordered breathing. By controlling the pulse oximetry sensor based on the subject's breathing pattern, the pulse oximetry sensor can be turned on (e.g., collect pulse oximetry data) only when the subject's breathing pattern indicates obstructive sleep disordered breathing.

[0015] This pulse oximetry sensor control method reduces the on / active time of the pulse oximetry sensor during sleep, thereby reducing power consumption while still reliably detecting oxygen desaturation events. The reduced power consumption allows the pulse oximetry sensor to collect pulse oximetry data for a longer period of time (e.g., throughout sleep) using a smaller battery.

[0016] The proposed method also implies and / or ensures that the recorded pulse oximetry data is relevant for assessing OSA, as such data will be directly correlated with snoring events and other breathing patterns indicative of sleep-disordered breathing. For example, this may enable more efficient post-processing or analysis of pulse oximetry data collected or sampled by a pulse oximetry sensor.

[0017] For example, the subject's breathing is the subject's breathing during a sleep session.

[0018] In some examples, the detector includes at least one of: a microphone, a resonator, an accelerometer, a gyroscope, an inertial measurement unit, a pressure sensor, a flow sensor, a seismocardiograph sensor, and / or a ballistocardiograph sensor.

[0019] For example, respiration data may be obtained from one or more contact microphones (i.e., in contact with the subject, such as located on the subject's neck or chest) or contactless microphones (positioned near the subject, such as a microphone in the subject's smartphone, placed on a nightstand next to the subject). Resonators (such as those used in stethoscopes) may also be used to obtain respiration data (e.g., a resonator comprising a diaphragm that contacts a transducer such as a piezoelectric crystal when the diaphragm vibrates, or a resonator comprising an electromagnetic diaphragm and a conductive plate that together form a capacitive sensor). Other sensors that detect motion / vibration associated with respiration may also be used to detect respiration data (e.g., one or more accelerometers, gyroscopes, inertial measurement units (IMUs), pressure sensors, flow sensors, seismocardiography sensors, and / or ballistocardiography sensors).

[0020] In some examples, the processing system is configured to control the pulse oximetry sensor by: processing one or more respiratory characteristics to determine whether the subject is snoring; and in response to determining that the subject is snoring, controlling the pulse oximetry sensor to begin acquiring pulse oximetry data.

[0021] This provides a simple mechanism for controlling a pulse oximetry sensor based on the subject's snoring. The inventors have recognized that snoring typically occurs during periods of hypopnea, immediately prior to an obstructive apnea event, and / or shortly after the obstruction resolves. By controlling the pulse oximetry sensor to begin collecting pulse oximetry data (i.e., turning on / activating the pulse oximetry sensor) as soon as possible after determining that the subject has begun snoring, pulse oximetry data can be collected when oxygen desaturation is likely to occur. Not collecting pulse oximetry data before the subject begins snoring reduces the turn-on / activation time of the pulse oximetry sensor, thereby reducing power consumption of the pulse oximetry sensor, and, since oxygen desaturation is unlikely to occur before the subject begins snoring, a pulse oximetry sensor using this control mechanism is less likely to miss detection of an oxygen desaturation event. In this way, power consumption of the pulse oximetry sensor is reduced throughout the sleep process while maintaining the accuracy of the pulse oximetry data.

[0022] In some examples, the processing system is configured to control the pulse oximetry sensor by: processing one or more respiratory characteristics to determine whether the subject is snoring; and, in response to determining that the subject is snoring: processing the one or more respiratory characteristics to identify a snoring type; and controlling the pulse oximetry sensor based on the identified snoring type.

[0023] The inventors have recognized that a subject's snoring has different characteristics during regular breathing, during hypopnea events, and before and after obstructive apnea events. By identifying which of these types the subject's current snoring indicates, the processing system can determine the most appropriate time to acquire even more accurate pulse oximetry data (i.e., when oxygen desaturation may occur), thereby further reducing the power consumption of the pulse oximetry sensor without reducing the accuracy of the acquired data.

[0024] In some examples, the processing system is configured to, in response to the identified snoring type being a first snoring type occurring during hypopnea: control the pulse oximetry sensor to begin acquiring pulse oximetry data.

[0025] Snoring generally persists during a hypopnea event; such snoring has different characteristics than snoring that occurs before and after an apnea event. By identifying the occurrence of a hypopnea event based on the subject's breathing characteristics and controlling the pulse oximetry sensor to collect pulse oximetry data during a hypopnea event, pulse oximetry data can be acquired at a time when an oxygen desaturation event may occur.

[0026] In some examples, the processing system is configured to: in response to the identified snoring type being a second snoring type that occurs before an obstructive apnea event, process one or more respiratory features to determine a time since a most recent snoring; and in response to determining that the time since the most recent snoring exceeds a first predetermined time period, control the pulse oximetry sensor to begin acquiring pulse oximetry data.

[0027] In other words, when the subject's snoring indicates an impending obstructive apnea event, the pulse oximetry sensor can be controlled to begin collecting pulse oximetry data as soon as the subject stops snoring. Obstructive apnea is a cessation of breathing, so the subject does not snore during an obstructive apnea. Therefore, the cessation of snoring can indicate the beginning of an obstructive apnea event, and is therefore an appropriate time to begin collecting pulse oximetry data.

[0028] The first predetermined time period for determining that snoring has ceased due to an obstructive apnea event may be, for example, shorter than the duration of a respiratory cycle, such as half the duration of a respiratory cycle. For example, assuming that the subject's respiratory rate is 12 breaths per minute, the first predetermined time period may be 2.5 seconds. In some examples, the processing system may process the subject's respiratory data to determine the subject's respiratory rate, and determine the subject-specific first predetermined time period based on the determined respiratory rate. In other examples, the first predetermined time period may be determined based on a typical respiratory rate (e.g., a population average).

[0029] In some examples, the processing system is configured to, in response to the identified snoring type being a third snoring type that occurs after resolution of the obstruction, control the pulse oximetry sensor to begin acquiring pulse oximetry data.

[0030] After the obstruction resolves, it will take some time for the subject's oxygen saturation level to return to normal levels. By beginning to collect pulse oximetry data as soon as it is determined that the subject's snoring indicates that the obstruction has resolved, pulse oximetry data can be obtained before the oxygen saturation returns to normal levels.

[0031] In some examples, the processing system is configured to identify the snoring type based on subject-specific data acquired during at least one previous sleep session for the subject.

[0032] The breathing characteristics indicative of each snoring type may differ between subjects. By knowing the breathing characteristics of each snoring type for a subject, the accuracy of snoring type identification can be improved.

[0033] The subject-specific data may include characteristics of one or more breathing features of the subject for each of a plurality of snoring types, wherein the plurality of snoring types include one or more of: a first snoring type exhibited by the subject during a hypopnea event; a second snoring type exhibited by the subject prior to an obstructive apnea event; a third snoring type exhibited by the subject after the obstruction resolves; and / or a fourth snoring type exhibited by the subject during regular breathing. The first snoring type, the second snoring type, and the third snoring type are associated with sleep disordered breathing.

[0034] For example, during a diagnostic sleep session for a subject, respiratory data can be acquired throughout the session, and obstructive apnea and hypopnea events that occur during the session can be detected (e.g., using PSG in a sleep lab or using a home sleep apnea test). The respiratory data and the detected obstructive apnea and hypopnea events can then be processed to determine one or more respiratory characteristics (e.g., spectral properties) of the subject's breathing during the hypopnea events and before and after the apnea events.

[0035] In some examples, the processing system is configured to utilize a machine learning algorithm to identify the snoring type, wherein the machine learning algorithm has been trained using subject-specific data acquired during at least one previous sleep session for the subject. The machine learning algorithm may be an active reinforcement algorithm.

[0036] For example, a pulse oximetry sensor may continuously collect pulse oximetry data at the beginning of the machine learning algorithm training, and active reinforcement learning may be used to maximize the detection of oxygen desaturation events while minimizing the time that the pulse oximetry sensor is activated. In other words, the machine learning algorithm may be trained to recognize snoring types associated with regular breathing, hypopnea events, and before and after obstructive apnea events by rewarding the machine learning algorithm if the pulse oximetry sensor is activated only when oxygen desaturation occurs and penalizing the machine learning algorithm if the pulse oximetry sensor collects pulse oximetry data during a period when the subject's oxygen saturation level remains high.

[0037] In some embodiments, additionally or alternatively, the processing system is configured to control the pulse oximetry sensor by: processing one or more respiratory characteristics to determine whether the subject is snoring; and, in response to determining that the subject is snoring: processing the one or more respiratory characteristics to determine an operating variance of one or more snoring characteristics; in response to determining that each operating variance exceeds a predetermined operating variance threshold, controlling the pulse oximetry sensor to collect pulse oximetry data; and in response to determining that the operating variance of the one or more snoring characteristics does not exceed the predetermined threshold, controlling the pulse oximetry sensor to not collect pulse oximetry data.

[0038] In other words, the pulse oximetry sensor may be turned on and off (or activated / deactivated) depending on whether the operating variance exceeds a threshold.

[0039] Irregular snoring (and thus an increased running variance of the time intervals between consecutive snores) has been found to be characteristic of apneic snoring. Therefore, using a running variance threshold to control the pulse oximetry sensor provides a simple mechanism to collect pulse oximetry data during apneic snoring, but not during regular snoring.

[0040] This method of controlling a pulse oximetry sensor may be particularly effective for heavy snorers who have long duration apnea phases.

[0041] In some examples, the processing system is configured to control the pulse oximetry sensor by: processing one or more respiratory characteristics to determine a time since a most recent snoring; and in response to determining that the time since the most recent snoring exceeds a second predetermined time period, controlling the pulse oximetry sensor to stop acquiring pulse oximetry data.

[0042] In other words, once an interval has passed since the end of the snoring cycle, the pulse oximetry sensor can stop collecting pulse oximetry data (i.e., turn off / deactivate). Once the interval has passed, an oxygen desaturation event is unlikely to occur, so the pulse oximetry data can be turned off / deactivated, thereby saving power without missing the detection of an oxygen desaturation event.

[0043] The second predetermined time period should be long enough not only to determine that the subject has stopped snoring, but also to allow pulse oximetry data to be collected immediately after the subject has stopped snoring, as this period may coincide with an obstructive apnea event. For example, the second predetermined time period may be 15 seconds to 30 seconds (e.g., 20 seconds to 25 seconds) in length.

[0044] In some examples, the processing system is configured to control the pulse oximetry sensor by: acquiring pulse oximetry data from the pulse oximetry sensor; processing the pulse oximetry data to determine an oxygen saturation measurement; and in response to determining that the oxygen saturation measurement exceeds an oxygen saturation threshold throughout a third predetermined time period, controlling the pulse oximetry sensor to stop acquiring the pulse oximetry data.

[0045] In other words, if a decrease in oxygen saturation is not detected during the period in which the decrease in oxygen saturation is expected to occur during an obstructive apnea or hypopnea event, the pulse oximetry sensor can be turned off / deactivated.

[0046] If the oxygen saturation measurement drops below an oxygen saturation threshold, a drop in oxygen saturation is detected. The oxygen saturation threshold may be a predetermined threshold (eg, 95%), or may be a subject-specific threshold determined based on oxygen saturation measurements for the subject measured during regular breathing.

[0047] The length of the third predetermined time period may depend on the mechanism used to turn on / activate the pulse oximetry sensor. For example, if the pulse oximetry sensor collects pulse oximetry data whenever the subject snores, the third predetermined time period may be longer because snoring may precede an obstructive apnea event, in which case the oxygen saturation level may not drop below the oxygen saturation threshold until a few seconds after the snoring stops. Where the pulse oximetry sensor is controlled to begin collecting pulse oximetry data based on the identified type of snoring, the pulse oximetry sensor is more likely to collect pulse oximetry data immediately after an obstructive apnea or hypopnea event begins, and thus the third predetermined time period may be shorter (e.g., less than 15 seconds, such as 10 seconds).

[0048] A system for measuring oxygen saturation is also proposed, comprising: a pulse oximetry sensor; a detector capable of acquiring respiration data responsive to the subject's respiration during the subject's sleep process; and the above-mentioned processing system.

[0049] According to another aspect of the invention, a computer-implemented method for controlling a pulse oximetry sensor is provided, the computer-implemented method comprising: acquiring respiratory data responsive to the subject's respiration during a sleep session of the subject from a detector distinct from the pulse oximetry sensor; processing the respiratory data to determine one or more respiratory characteristics; and controlling the pulse oximetry sensor in response to the one or more respiratory characteristics.

[0050] A computer program product comprising instructions is also proposed, which, when executed on a computer device having a processing system, causes the processing system to execute all the steps of the above method.

[0051] A processing system for controlling a pulse oximetry sensor is also presented, the processing system being configured to: acquire respiratory data responsive to the respiration of a subject from a detector different from the pulse oximetry sensor; process the respiratory data to determine one or more respiratory characteristics; and control the pulse oximetry sensor responsive to the one or more respiratory characteristics by processing the one or more respiratory characteristics to determine whether the subject is snoring; and, responsive to determining that the subject is snoring: process the one or more respiratory characteristics to determine operating variances of one or more snoring characteristics; responsive to determining that each operating variance exceeds a predetermined operating variance threshold, control the pulse oximetry sensor to acquire pulse oximetry data; and responsive to determining that the operating variance of the one or more snoring characteristics does not exceed the predetermined threshold, control the pulse oximetry sensor to not acquire pulse oximetry data.

[0052] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] For a better understanding of the present invention, and in order to more clearly show how the same may be carried into effect, reference will now be made, by way of example only, to the accompanying drawings, in which:

[0054] Figure 1 A system for measuring oxygen saturation according to an embodiment of the present invention is shown; and

[0055] Figure 2 A computer-implemented method for controlling a pulse oximetry sensor according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0056] The present invention will be described below with reference to the accompanying drawings.

[0057] It should be understood that the detailed description and specific examples, while indicating exemplary embodiments of the apparatus, system, and method, are intended for illustrative purposes only and are not intended to limit the scope of the invention. These and other features, aspects, and advantages of the apparatus, system, and method of the invention may be better understood in conjunction with the following description, the appended claims, and the accompanying drawings. It should be understood that the drawings are schematic only and are not drawn to scale. It should also be understood that the same reference numerals represent the same or similar components throughout the drawings.

[0058] A system and method for controlling a pulse oximetry sensor is provided. The pulse oximetry sensor is activated and deactivated based on one or more respiratory characteristics of a subject's breathing during sleep, such that the pulse oximetry sensor acquires pulse oximetry data only when the one or more respiratory characteristics indicate that oxygen desaturation may have occurred.

[0059] Various embodiments are based at least in part on the recognition that breathing characteristics, particularly those associated with snoring, are indicative of sleep-disordered breathing, and that controlling a pulse oximetry sensor to collect pulse oximetry data only during periods of sleep-disordered breathing can reduce power consumption of the pulse oximetry sensor during the entire sleep process without reducing the accuracy or reliability of the acquired data by missing oxygen desaturation events.

[0060] The illustrative embodiments may be used, for example, in systems that use a pulse oximetry sensor to measure oxygen saturation, particularly systems that use a wearable wireless pulse oximetry sensor that is powered by a battery.

[0061] Figure 1 A system 100 for measuring oxygen saturation according to an embodiment of the present invention is shown. The system 100 includes: a pulse oximetry sensor 110; a (different) detector 120 capable of acquiring respiration data 125; and a processing system 130 for controlling the pulse oximetry sensor. The processing system may be located in the same device as the pulse oximetry sensor and / or the detector, or may be located in a separate device. The processing system itself is an embodiment of the present invention.

[0062] exist Figure 1In the present invention, pulse oximetry sensor 110 is a wireless pulse oximetry sensor worn on a subject's fingertip; however, the pulse oximetry sensor may be any sensor capable of collecting pulse oximetry data of a subject during the subject's sleep. For example, the pulse oximetry sensor may be incorporated into or mounted on a finger ring worn on a subject's finger or thumb, a bracelet or band worn on a subject's wrist, or an ear clip worn on a subject's ear. The present invention is particularly suitable for battery-powered pulse oximetry sensors, as such sensors benefit from the reduced power consumption provided by the present invention.

[0063] Detector 120 acquires breathing data 125 of subject 140. The breathing data is in response to the subject's breathing during the sleep process. Figure 1 In the embodiment, the detector is a microphone placed on the subject's neck that detects the sound of the subject's breathing; however, the detector may be any type of detector capable of acquiring data responsive to the subject's breathing. For example, the detector may include at least one of: a (contact and / or contactless) microphone, a resonator, an accelerometer, a gyroscope, an inertial measurement unit (IMU) (e.g., a MEMS IMU), a pressure sensor, a flow sensor, a cardiotocography sensor, and / or an impact tracing sensor. Other examples of suitable detectors for acquiring respiratory data will be apparent to those skilled in the art.

[0064] A contact microphone placed on / against the subject's body (e.g., the subject's neck or chest), and / or a contactless microphone near the subject (e.g., a microphone in a smartphone held close to the subject's bed) can be used to detect breathing sounds. In some examples, the microphone can be included in the same device as the pulse oximetry sensor.

[0065] Similarly, one or more resonators placed on / against the subject's body (e.g., resonators of the type suitable for use in a stethoscope) can be used to detect breathing sounds. Suitable resonators include resonators having a diaphragm that, when vibrated, is connected to a transducer (e.g., a piezoelectric crystal); and / or resonators having a diaphragm made of an electromagnetic material that, together with a conductive plate, forms a capacitive sensor.

[0066] Other detectors (such as accelerometers, gyroscopes, IMUs, pressure sensors, flow sensors, and / or seismic cardiotography sensors placed on or against the subject's body, and / or bedside ballistocardiotography sensors) can be used to detect vibrations and / or other movements associated with breathing. The use of such detectors to obtain corresponding data for the subject's breathing is well known. Other suitable detectors for obtaining data responsive to the subject's breathing will be apparent to those skilled in the art.

[0067] The processing system 130 acquires the respiratory data 125 from the detector 120 and processes the respiratory data to determine one or more respiratory features. The one or more respiratory features may include features that can be used to determine whether the subject is snoring, such as the power in a predefined frequency band (e.g., a high frequency band) corresponding to typical snoring, the total power, and / or the amplitude of the signal envelope within a predetermined time period (e.g., a few seconds). The one or more respiratory features may also include features that can be used to identify the type of snoring, such as the fundamental frequency of the signal carrying the data, the amplitude of the signal envelope, the power of the signal, and / or the changes in frequency, amplitude, and / or power within the duration of an identified snoring and / or between consecutive snorings.

[0068] Then, in response to the one or more respiratory characteristics, processing system 130 controls pulse oximetry sensor 110. Processing system controls pulse oximetry sensor 110 to acquire pulse oximetry data 115 only when the one or more respiratory characteristics indicate that an obstructive apnea or hypopnea event may occur.

[0069] In some examples, processing system 130 may control pulse oximetry sensor 110 to determine whether subject 140 is breathing by processing one or more breathing characteristics. For example, processing system may control pulse oximetry sensor 110 to collect pulse oximetry data 115 in response to determining that the subject has stopped breathing for more than a predetermined respiratory cessation period (e.g., pulse oximetry data may be controlled to begin collecting pulse oximetry data in response to determining that no breathing is detected in the respiratory data for a period of at least 10 seconds). The period during which no breathing occurs (e.g., at least 10 seconds) is characteristic of an obstructive apnea event, during which there is no airflow.

[0070] In some examples, processing system 130 may control pulse oximetry sensor 110 by processing one or more breathing characteristics to determine whether subject 140 is snoring.

[0071] Methods of detecting snoring based on data responsive to a subject's breathing are well known and therefore methods for performing the steps of determining one or more appropriate breathing characteristics and determining based on the one or more breathing characteristics will be apparent to those skilled in the art.See, for example: Jiali Xie et al., (2021), “Audio-based snore detection using deep neural networks”, Computer Methods and Programs in Biomedicine, 200:105917; Eliran Dafna et al., (2013), “Automatic detection of whole night snoring events using non-contact microphone”, PloS One, 8(12):e84139; Erna Arnardottir et al., (2016), “How to measure snoring? A comparison of the microphone, cannula and piezoelectric sensor”, Journal of Sleep Research, 25(2):158–168; D Sánchez Morillo et al., (2007), “Monitoring and analysis of cardio respiratory and snoring signals by using an accelerometer”, Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2007:3942–5; Hyo-Ki Lee et al., (2013), “Automatic snoring detection from nasal pressure data”, Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2013:6870–2; and Su Hwan Hwang et al., (2015), “Polyvinylidene fluoride sensor-based method for unconstrained snoring detection”, Physiological Measurement, 36(7):1399–1414。

[0072] In response to determining that the subject is snoring, processing system 130 may control pulse oximetry sensor 110 to begin acquiring pulse oximetry data 115 (i.e., without further processing one or more respiratory features), or to process one or more respiratory features to identify a snoring type and / or determine a running variance of time intervals between consecutive snorings. The snoring type and / or running variance may be used to control the pulse oximetry sensor.

[0073] In other words, in some examples, processing system 130 may turn on or activate pulse oximetry sensor 110 immediately upon determining that subject 140 is snoring. If it is determined that the subject is not snoring, processing system 130 may continue to process one or more breathing characteristics to determine whether the subject is snoring.

[0074] If subject 140 has not started snoring (i.e., if processing system has not turned on / activated pulse oximetry sensor 110, or has turned off / deactivated pulse oximetry sensor 110 since the most recent snoring), processing system 130 may control pulse oximetry sensor 110 to not collect pulse oximetry data.

[0075] Throughout the specification, references to a subject starting to snore and stopping to snore refer to the start and end of a snoring cycle, respectively, rather than the start and end of a single snoring episode. A snoring cycle is characterized by the (relatively short) intervals between consecutive snoring episodes: a snoring episode having an interval greater than a predetermined snoring interval threshold compared to the previous snoring episode is considered to belong to a snoring cycle different from the previous snoring episode. Suitable thresholds for determining the end of a snoring cycle are described below.

[0076] In some examples, processing system 130 controls pulse oximetry sensor 110 based on a snoring type identified by processing one or more respiratory characteristics. The time (for a snoring period) when an oxygen desaturation event is more likely to occur will depend on whether the snoring occurs before or after an obstructive apnea event, or during a hypopnea event. Subject 140 may also snore during regular (i.e., non-obstructed) breathing, when the likelihood of an oxygen desaturation event is very low. By identifying which of these snoring types the subject's current snoring corresponds to, processing system 130 can control pulse oximetry sensor 110 to collect pulse oximetry data 115 only during periods when an oxygen desaturation event is likely to occur.

[0077] In some examples, a snoring type may be identified based on reference data including typical characteristics of one or more breathing features for each of a plurality of snoring types. The plurality of snoring types may include: a first snoring type occurring during a hypopnea event; a second snoring type occurring before an obstructive apnea event (i.e., immediately before an obstructive apnea event occurs); a third snoring type occurring after obstruction resolves (i.e., immediately after an obstructive apnea event ends); and / or a fourth snoring type occurring during regular (i.e., non-obstructed breathing).

[0078] In some examples, the type of snoring may be identified based on known breathing / snoring characteristics. For example, during an obstructive apnea event, subject 140 may stop breathing, and no breathing sounds or vibrations / other movements associated with breathing may be detected. Thus, periods (e.g., at least 10 seconds) during which no breathing is detected in the respiratory data indicate an obstructive apnea event. In particular, a burst of consecutive apnea events may be characterized by a snoring pattern interspersed with short-term (e.g., at least 10 seconds) snoring / breathing stops.

[0079] Resolution of an obstructive apnea or hypopnea event is often accompanied by a short period of snoring, gasping, and / or other vocalizations that are different from regular (usually fixed frequency) snoring. Therefore, features of these vocalizations (e.g., frequency changes) can be used to identify the type of snoring that occurs after the obstruction resolves.

[0080] During a hypopnea event, airflow is reduced and snoring characteristics differ from those of regular snoring and often change throughout a hypopnea event. For example, an increase in the energy / envelope of a signal carrying respiratory data corresponding to a progressive increase in respiratory effort may indicate that snoring occurs during a hypopnea event.

[0081] In some examples, a machine learning algorithm may be used to identify the snoring type. A machine learning algorithm is any self-training algorithm that is capable of processing input data to produce or predict output data. Here, the input data includes one or more breathing features (or simply, breathing data for periods during which snoring is detected), and the output data includes an identification of the snoring type.

[0082] Suitable machine learning algorithms for the present invention will be apparent to those skilled in the art. Examples of suitable machine learning algorithms include decision tree algorithms and artificial neural networks. Other machine learning algorithms such as logistic regression, support vector machines, or naive Bayes models are suitable alternative algorithms.

[0083] The structure of an artificial neural network (or simply, a neural network) is inspired by the human brain. A neural network consists of multiple layers, each layer consists of multiple neurons. Each neuron contains a mathematical operation. In particular, each neuron can contain different weighted combinations of a single transformation type (e.g. the same transformation type, sigmoid, etc., but with different weights). During the processing of the input data, the mathematical operation of each neuron is performed on the input data to produce a digital output, and the output of each layer in the neural network is fed to the next layer in sequence. The last layer provides the output.

[0084] Methods for training machine learning algorithms are well known. Typically, such methods include obtaining a training data set, including training input data entries and corresponding training output data entries. An initialized machine learning algorithm is applied to each input data entry to generate a predicted output data entry. The error between the predicted output data entry and the corresponding training output data entry is used to modify the machine learning algorithm. The process can be repeated until the error converges and the predicted output data entry is sufficiently similar to the training output data entry (e.g., ±1%). This is generally referred to as a supervised learning technique.

[0085] For example, in the case where the machine learning algorithm is formed by a neural network, the mathematical operation (weight) of each neuron can be modified until the error converges. Known methods for modifying neural networks include gradient descent, back propagation algorithms, etc.

[0086] The training input data entries correspond to example breathing features of breathing data during a snoring cycle (or example raw breathing data during a snoring cycle). Preferably, the training input data entries are obtained from a large number of individuals. The training output data entries correspond to identifications of snoring types.

[0087] In some examples, a snoring type may be identified based on subject-specific data acquired during at least one previous sleep session for subject 140. The subject-specific data may include characteristics of one or more breathing features of the subject for each of a plurality of snoring types. The plurality of snoring types may include: a first snoring type exhibited by the subject during a hypopnea event; a second snoring type exhibited by the subject prior to an obstructive apnea event (i.e., immediately prior to an obstructive apnea event); a third snoring type exhibited by the subject after obstruction resolves (i.e., immediately after an obstructive apnea event ends); and / or a fourth snoring type exhibited by the subject during regular (i.e., non-obstructed) breathing.

[0088] Characteristics of one or more breathing features of the subject for each of a plurality of snoring types may be determined using respiratory data responsive to the subject's breathing during at least one prior sleep session and data including times when obstructive apnea and hypopnea events were detected during the same at least one prior sleep session. The data including times when obstructive apnea and hypopnea events were detected may be acquired using any known technique for detecting such events, including polysomnography (PSG), home sleep apnea testing (HSAT), or pulse oximetry.

[0089] Respiratory data for one or more time periods (e.g., periods of 5 to 10 minutes in length) during at least one previous sleep process immediately before a detected obstructive apnea event (i.e., during one or more periods ending at the start of a detected obstructive apnea event, respectively), respiratory data for one or more time periods (e.g., periods of 5 to 10 minutes in length) immediately after the end of a detected obstructive apnea event, respiratory data for one or more time periods corresponding to the duration of a detected hypopnea event, and / or respiratory data for one or more time periods corresponding to the duration of normal breathing may be processed to determine characteristics of one or more respiratory features for each of a plurality of snoring types.

[0090] In some examples, a machine learning algorithm trained using subject-specific data can be used to identify snoring types. Suitable machine learning algorithms for identifying snoring types that can be trained using subject-specific data will be apparent to those skilled in the art. Examples of suitable machine learning algorithms include reinforcement learning algorithms, particularly active reinforcement learning algorithms.

[0091] Active reinforcement learning may be used to maximize the detection of oxygen desaturation events while minimizing the time that the pulse oximetry sensor is activated. The pulse oximetry sensor may initially (i.e., prior to training the machine learning algorithm) run continuously during at least one sleep session (or run continuously during a portion of each sleep session, e.g., until the sensor's battery is depleted). The machine learning algorithm may then learn to detect snoring types that occur before and after obstructive apnea events and during hypopnea events by rewarding when the algorithm correctly identifies a snoring type that indicates an oxygen desaturation event and penalizing when the algorithm identifies a snoring type that indicates an oxygen desaturation event but no oxygen desaturation event has occurred. In some examples, the pulse oximetry sensor may run continuously during each sleep session (or run continuously during a portion of each sleep session, e.g., until the sensor's battery is depleted) until sufficient examples of different snoring types associated with apnea / hypopnea events are found for subject 140 (e.g., until a predetermined number of examples are found for each snoring type, or until the machine learning algorithm has reached a predetermined level of accuracy).

[0092] After identifying the snoring type, processing system 130 may control pulse oximetry sensor 110 according to the identified snoring type.

[0093] For example, in response to determining that the subject's current snoring corresponds to normal (i.e., non-obstructed) snoring, processing system 130 may control pulse oximetry sensor 110 to not collect (or continue to collect) pulse oximetry data. The likelihood of an oxygen desaturation event occurring during normal snoring is extremely low, and therefore, this reduces the likelihood that the pulse oximetry sensor will acquire unnecessary data.

[0094] In response to determining that the subject's current snoring corresponds to a type of snoring that occurs during a hypopnea event, processing system 130 may control pulse oximetry sensor 110 to begin collecting pulse oximetry data 115. In other words, once processing system determines that subject 140 is snoring and that the snoring is of a type of snoring that occurs during a hypopnea event, pulse oximetry sensor 110 may be turned on / activated. This allows pulse oximetry data to be collected during hypopnea events throughout the subject's sleep without the need to continuously collect pulse oximetry data.

[0095] In response to determining that the subject's current snoring corresponds to a type of snoring that occurs prior to an obstructive apnea event, the processing system 130 may process the one or more respiratory features to determine a time since a most recent snoring event. The processing system may then control the pulse oximetry sensor 110 to begin collecting pulse oximetry data 115 in response to determining that the time since the most recent snoring event exceeds a first predetermined time period. If the time since the most recent snoring event does not exceed the first predetermined time period, the processing system may continue to determine the time since the most recent snoring event until the interval exceeds the first predetermined time period.

[0096] In other words, once it is determined that an obstructive apnea event may be beginning, the pulse oximetry sensor 110 begins collecting pulse oximetry data 115. A cessation of snoring patterns that precedes an obstructive apnea event generally indicates the onset of an obstructive apnea event (during which the subject is not breathing, and therefore not snoring).

[0097] The first predetermined time period is a long enough period of time to detect the cessation of snoring (i.e., long enough to distinguish the cessation of snoring from the interval between multiple snorings in the same snoring cycle). The first predetermined time period can be less than a typical breathing cycle length (e.g., based on a population average or a subject-specific average breathing rate). For example, the first predetermined time period can be half the length of a typical breathing cycle (e.g., for a breathing rate of 12 breaths per minute, the first predetermined time period can be 2.5 seconds). If the first predetermined time period is based on a subject-specific average breathing rate, the processing system can process the breathing data 125 to determine the breathing rate of the subject during regular breathing, and determine the first predetermined time period based on the determined breathing rate.

[0098] In response to determining that the subject's current snoring corresponds to a type of snoring that occurs after the obstruction resolves, processing system 130 may control pulse oximetry sensor 110 to begin collecting pulse oximetry data 115. In other words, once the processing system determines that the obstructive apnea event is over, the pulse oximetry sensor may be turned on / activated. Collecting pulse oximetry data immediately after the obstructive apnea event ends should result in the detection of a decrease in oxygen saturation because it may take several breathing cycles (e.g., up to one minute) for the subject's oxygen saturation level to return to normal (i.e., to the oxygen saturation level before any obstructive apnea and / or hypopnea events occurred).

[0099] In some examples, additionally or alternatively, the processing system 130 may control the pulse oximetry sensor 110 based on one or more inter-snore variability features. The one or more inter-snore variability features are features that represent the variability between snoring and may include any suitable statistical parameters (e.g., variance, standard deviation interquartile range, etc.). For example, the one or more inter-snore variability features may include a running variance for one or more snoring characteristics, wherein the one or more snoring characteristics include at least one of the following: snoring energy, snoring duration, time interval between consecutive snorings, fundamental frequency of the signal during snoring, amplitude of the signal envelope, and / or a variation (e.g., slope or range) of any of these features.

[0100] Processing system 130 may, in response to determining that subject 140 is snoring, process one or more respiratory features to determine a running variance for each of one or more snoring characteristics. The running variance(s) may be determined each time a new snoring is detected in the respiratory data. Processing system may then control pulse oximetry sensor 110 to collect pulse oximetry data 115 in response to determining that each running variance exceeds a predetermined running variance threshold, and to not collect pulse oximetry data in response to determining that the running variance of one or more snoring characteristics does not exceed the predetermined threshold. For example, the predetermined energy running variance threshold may be a value between 0.2 and 0.3. Preferably, the value of the predetermined running variance threshold is in the range of 0.24 to 0.26 (e.g., 0.25).

[0101] The running variance of each of the one or more snoring characteristics may be determined as a within-group running variance for the one or more snoring characteristics, wherein a group of snoring is a plurality of snorings in which each time interval between consecutive snorings is below a predetermined time interval threshold (e.g., 60 seconds). If the time interval between a snoring and a previous snoring is greater than the above time interval threshold, the snoring is considered to belong to a different group from the previous snoring. In other words, the running variance for each of the one or more characteristics may be determined based only on the characteristics of the snoring in the current group of snorings (i.e., the group of snorings to which the most recent snoring belongs). If the time interval between the most recent snoring and the previous snoring exceeds the predetermined time interval threshold, the most recent snoring is the first snoring in the current group of snorings.

[0102] As described in Nir Ben-Israel et al., (2012), "Obstructive apnea hypopnea indexestimation by analysis of nocturnal snoring signals in adults", SLEEP, 35(9):1299–1305C, it was found that the running variance is closer to apnea events than during simple snoring phases (i.e., snoring associated with normal breathing). Therefore, controlling the pulse oximetry sensor based on the running variance can effectively classify the subject's current snoring as normal snoring or apneic snoring, and control the pulse oximetry sensor to collect pulse oximetry data only during apneic snoring.

[0103] In some examples, one or more snoring-to-snoring variability characteristics may be used in conjunction with other characteristics to control the pulse oximetry sensor 110. Characteristics used in conjunction with one or more snoring-to-snoring variability characteristics to determine when to control the pulse oximetry sensor to start and / or stop collecting pulse oximetry data may include a temporal relationship between groups of snoring events.

[0104] In some examples, one or more respiratory characteristics may be used to determine when to stop collecting pulse oximetry data (i.e., when to turn off / deactivate the pulse oximetry sensor). For example, the processing system 130 may process the one or more respiratory characteristics to determine a time since a most recent snoring. The processing system may then control the pulse oximetry sensor 110 to stop collecting pulse oximetry data 115 in response to determining that the time since the most recent snoring exceeds a second predetermined time period. If the time since the most recent snoring does not exceed the second predetermined time period, the processing system may control the pulse oximetry sensor to continue collecting pulse oximetry data until the time since the most recent snoring exceeds the second predetermined time period.

[0105] In this manner, the pulse oximetry sensor 110 continues to collect pulse oximetry data 115 during periods when oxygen desaturation may occur, and conserves battery power by ceasing to collect pulse oximetry data once oxygen desaturation is no longer likely to occur.

[0106] The second predetermined time period may be long enough not only to detect the cessation of snoring, but also to ensure that pulse oximetry data is collected in the period immediately following the cessation of snoring. The cessation of snoring may be caused by the onset of an obstructive apnea event, in which case the subject's oxygen saturation level will not begin to drop until the subject stops snoring. For example, a second predetermined time period of 15 to 30 seconds from the last snoring may be long enough to collect useful pulse oximetry data while reducing the amount of unnecessary pulse oximetry data collected. Preferably, the second predetermined time period has a value between 20 and 25 seconds.

[0107] Other mechanisms for determining when to stop collecting pulse oximetry data based on one or more respiratory characteristics will be apparent to those skilled in the art. In particular, in examples where a type of snoring is identified, the time after which oxygen saturation will no longer occur can be more accurately determined. For example, if the subject's snoring is of the type that occurs after an obstructive apnea event, the subject's oxygen saturation level may have returned to normal before the subject stops snoring. In this case, the processing system may instead control the pulse oximetry data to stop collecting pulse oximetry data once the time of the period since the subject began snoring exceeds a predetermined time period.

[0108] In some examples, the acquired pulse oximetry data may additionally or alternatively be used to determine when to stop acquiring pulse oximetry data. Processing system 130 may acquire pulse oximetry data 115 from pulse oximetry sensor 110 and process the pulse oximetry data to determine an oxygen saturation measurement. The processing system may then control the pulse oximetry sensor based on the determined oxygen saturation measurement.

[0109] For example, in response to determining that the oxygen saturation measurement exceeds the oxygen saturation threshold throughout the third predetermined time period, the processing system 130 may control the pulse oximetry sensor 110 to stop collecting pulse oximetry data 115. In other words, if the oxygen saturation measurement does not drop below the oxygen saturation threshold during the third predetermined time period (i.e., if no oxygen desaturation event is detected), the pulse oximetry sensor may be turned off / deactivated.

[0110] The third predetermined time period may begin at the time that the pulse oximetry sensor 110 last began collecting pulse oximetry data 115 (i.e., the time that the pulse oximetry sensor was last turned on / activated). The third predetermined time period may be long enough so that an oxygen desaturation event is expected to occur before the end of the third predetermined time period in the event of an obstructive apnea or hypopnea event. Thus, the length of the third predetermined time period may depend on the mechanism used to begin collecting pulse oximetry data, and the type of snoring identified if the pulse oximetry sensor is controlled using a snoring type.

[0111] For example, if the pulse oximetry sensor 110 begins collecting pulse oximetry data 115 in response to a cessation of snoring type that precedes an obstructive apnea event, it would be expected that the obstructive apnea event had occurred shortly before the pulse oximetry sensor began collecting pulse oximetry data. Therefore, a third predetermined time period of a few seconds (preferably less than 15 seconds, such as 10 seconds) would be sufficient: if an obstructive apnea event occurs, it is likely that a decrease in oxygen saturation will be detected by the end of this period.

[0112] The oxygen saturation threshold used to determine whether an oxygen desaturation event has occurred can be a predetermined threshold, such as based on a typical clinical threshold for detecting an oxygen desaturation event. For example, the value of the oxygen saturation threshold can be 95% (because oxygen saturation levels above 95% are generally considered normal).

[0113] Alternatively, the oxygen saturation threshold may be a subject-specific threshold, such as one or more values ​​based on measurements of oxygen saturation taken during a regular breathing cycle, or based on one or more values ​​of measurements of oxygen saturation determined from pulse oximetry data collected at the beginning of a current pulse oximetry data collection period (where the current collection period is the period since the pulse oximetry sensor last began collecting pulse oximetry data).

[0114] For example, the oxygen saturation threshold may be a value corresponding to a 3% decrease in oxygen saturation from a pre-event baseline value (e.g., an oxygen saturation value from a period immediately preceding the decrease in oxygen saturation). A 3% decrease in oxygen saturation is one of the criteria used by the American Academy of Sleep Medicine (AASM) for assessing hypopnea events (see, e.g., Version 2.5: Berry RB, Albertario CL, Harding SM, et al., “The AASM Manual for the Scoring of Sleep and Associated Events”. 2.5. American Academy of Sleep Medicine; 2018). In some examples, the processing system 130 may control the pulse oximetry sensor 110 to begin collecting pulse oximetry data 115 in response to determining that all of the AASM criteria for assessing a hypopnea event have been met (i.e., oxygen saturation has decreased by at least 3% from a pre-event baseline value, peak signal excursion has decreased by at least 30% from a pre-event baseline, and the duration of the decrease in peak signal excursion is at least 10 seconds).

[0115] In some examples, the processing system 130 may control the pulse oximetry sensor 110 to stop collecting pulse oximetry data 115 in response to determining that the measured value of oxygen saturation has dropped below the oxygen saturation threshold since the pulse oximetry sensor last started collecting pulse oximetry data, and the measured value of oxygen saturation at this time exceeds the oxygen saturation threshold. In other words, the pulse oximetry sensor may be controlled to stop collecting pulse oximetry data in response to determining that the oxygen desaturation event has ended.

[0116] Figure 2 A computer-implemented method 200 for controlling a pulse oximetry sensor according to an embodiment of the present invention is shown.

[0117] The method begins at step 210 where respiration data responsive to a subject's respiration during a sleep session of the subject is received from a detector other than a pulse oximetry sensor.

[0118] At step 220, the breathing data is processed to determine one or more breathing characteristics.

[0119] At step 230, a pulse oximetry sensor is controlled in response to one or more respiratory characteristics.

[0120] It should be understood that the disclosed method is a computer-implemented method. Therefore, a concept of a computer program comprising code means is also proposed, which, when the program is run on a processing system, is used to implement any of the above methods.

[0121] Those skilled in the art will be able to easily develop a processor for executing any method described herein. Therefore, each step in the flowchart may represent a different action performed by the processor and may be performed by a corresponding module of the processor.

[0122] As described above, the system utilizes a processor to perform data processing. The processor can be implemented in a variety of ways using software and / or hardware to perform various desired functions. The processor typically employs one or more microprocessors that can be programmed with software (such as microcode) to perform the desired functions. The processor can be implemented as a combination of dedicated hardware that performs some functions and one or more programmed microprocessors and their associated circuits that perform other functions.

[0123] Examples of circuits that may be employed in various embodiments of the present disclosure include, but are not limited to, conventional microprocessors, application specific integrated circuits (ASICs), and field programmable gate arrays (FPGAs).

[0124] In various embodiments, the processor may be associated with one or more storage media, such as volatile and non-volatile computer memory, such as RAM, PROM, EPROM, and EEPROM. The storage media may be encoded with one or more programs that, when executed on one or more processors and / or controllers, perform the desired functions. The various storage media may be fixed in the processor or controller, or may be transportable so that one or more programs stored therein may be loaded into the processor.

[0125] In the process of implementing the claimed invention, those skilled in the art can understand and implement variations of the disclosed embodiments by studying the drawings, the disclosure and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

[0126] The functions implemented by a processor may be implemented by a single processor or multiple separate processing units, which may be considered to collectively constitute a “processor.” In some cases, these processing units may be remote from each other and communicate with each other in a wired or wireless manner.

[0127] The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0128] The computer program may be stored / distributed on an appropriate medium, such as an optical storage medium or solid-state medium provided with the hardware or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[0129] If the term "suitable for" is used in the claims or the specification, it should be noted that the term "suitable for" is intended to be equivalent to the term "configured to". If the term "device" is used in the claims or the specification, it should be noted that the term "device" is intended to be equivalent to the term "system" and vice versa.

[0130] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A processing system (130) for controlling a pulse oximetry sensor (110), the processing system being configured to: acquiring respiration data (125) responsive to the subject's respiration from a detector (120) distinct from the pulse oximetry sensor; processing the breathing data to determine one or more breathing characteristics; and In response to the one or more respiratory characteristics, controlling the pulse oximetry sensor by: processing the one or more breathing characteristics to determine whether the subject is snoring; and, in response to determining that the subject is snoring: processing the one or more respiratory features to identify a snoring type; and The pulse oximetry sensor is controlled based on the identified snoring type.

2. The processing system (130) of claim 1, wherein: The detector (120) includes at least one of the following: a microphone, a resonator, an accelerometer, a gyroscope, an inertial measurement unit, a pressure sensor, a flow sensor, a seismocardiographic sensor, and / or a ballistocardiographic sensor.

3. The processing system (130) according to claim 1 or 2, wherein: The processing system is configured to: in response to the identified snoring type being a first snoring type occurring during a period of hypopnea, The pulse oximetry sensor (110) is controlled to begin collecting pulse oximetry data (115).

4. The processing system (130) according to any one of claims 1 to 3, wherein: The processing system is configured to: in response to the identified snoring type being a second snoring type that occurs prior to an obstructive apnea event, processing the one or more respiratory features to determine a time since a most recent snoring; and In response to determining that the time since the most recent snoring exceeds a first predetermined time period, the pulse oximetry sensor (110) is controlled to begin collecting pulse oximetry data (115).

5. The processing system (130) according to any one of claims 1 to 4, wherein: The processing system is configured to: in response to the identified snoring type being a third snoring type that occurs after obstruction resolves, The pulse oximetry sensor (110) is controlled to begin collecting pulse oximetry data (115).

6. The processing system (130) according to any one of claims 1 to 5, wherein: The processing system is configured to identify a snoring type based on subject-specific data acquired during at least one previous sleep session for the subject (140).

7. The processing system (130) of claim 6, wherein: The subject-specific data includes characteristics of one or more breathing features of the subject (140) for each of a plurality of snoring types, wherein the plurality of snoring types include one or more of: a first snoring type exhibited by the subject during a hypopnea event; a second snoring type exhibited by the subject prior to an obstructive apnea event; a third snoring type exhibited by the subject after resolution of the obstruction; and / or a fourth snoring type exhibited by the subject during regular breathing.

8. The processing system (130) according to claim 6 or 7, wherein: The processing system is configured to identify the snoring type using a machine learning algorithm, wherein the machine learning algorithm has been trained using the subject-specific data acquired during at least one previous sleep session for the subject (140).

9. The processing system according to claim 8, wherein: The machine learning algorithm is an active reinforcement learning algorithm.

10. The processing system (130) according to any one of claims 1 to 9, wherein: The processing system is further configured to, in response to determining that the subject is snoring, control the pulse oximetry sensor (110) by: processing the one or more respiratory features to determine a running variance of one or more snoring characteristics; In response to determining that each operating variance exceeds a predetermined operating variance threshold, controlling the pulse oximetry sensor to collect pulse oximetry data (115); and In response to determining that the operating variance for one or more snoring characteristics does not exceed the predetermined threshold, the pulse oximetry sensor is controlled to not acquire pulse oximetry data.

11. The processing system (130) according to any one of claims 1 to 10, wherein: The processing system is configured to control the pulse oximetry sensor (110) by: processing the one or more respiratory features to determine a time since a most recent snoring; and In response to determining that the time since the most recent snoring exceeds a second predetermined time period, the pulse oximetry sensor is controlled to stop collecting pulse oximetry data (115).

12. The processing system (130) according to any one of claims 1 to 11, wherein: The processing system is configured to control the pulse oximetry sensor (110) by: acquiring pulse oximetry data from the pulse oximetry sensor (115); processing the pulse oximetry data to determine an oxygen saturation measurement; and In response to determining that the oxygen saturation measurement exceeds the oxygen saturation threshold throughout the third predetermined time period, the pulse oximetry sensor is controlled to stop acquiring pulse oximetry data.

13. A system (100) for measuring oxygen saturation, comprising: a pulse oximetry sensor (110); a detector (120) capable of acquiring respiratory data responsive to the subject's respiration; as well as A processing system (130) according to any one of claims 1 to 12.

14. A computer-implemented method (200) for controlling a pulse oximetry sensor (110), the computer-implemented method comprising: acquiring respiration data (125) responsive to the subject's respiration from a detector (120) distinct from the pulse oximetry sensor; processing the breathing data to determine one or more breathing characteristics; as well as In response to the one or more respiratory characteristics, controlling the pulse oximetry sensor by: processing the one or more breathing characteristics to determine whether the subject is snoring; and, in response to determining that the subject is snoring: processing the one or more respiratory features to identify a snoring type; and The pulse oximetry sensor is controlled based on the identified snoring type.

15. A computer program product comprising instructions which, when executed on a computer device having a processing system, cause the processing system to perform the steps of the method (200) according to claim 14.

Images