Detection and treatment of obstructive sleep apnea

1 . A method of detecting an occurrence of disturbance during a sleep period of a user, the method comprising:

receiving data corresponding to a single parameter of the user via a sensor during the sleep period, wherein the single parameter comprises either a nasal air pressure of the user or an audio output of the user;

using a trained machine learning model, determining patterns in the data and deriving an SpO2 level over the sleep period based on the patterns;

based at least on the derived SpO2 level, determining a sleep disturbance severity index (SDSI) over the sleep period; and

in response to detecting the occurrence of disturbance during the sleep period of the user, treating the disturbance, the treating comprising activating a first patch that is affixed externally on a dermis of the user and comprises electrodes, the activating comprising generating an electrical stimuli via the electrodes.

2 . The method of claim 1 , further comprising, deriving at least one additional breath metric comprising: a rising edge rate of each breath, a falling edge rate of each breath, a breathing cadence, a variability in breath cadence or a volume of each breath.

3 . The method of claim 2 , wherein the determining an SpO2 level over the sleep period is based on one or more of any of the breath metrics, a sleep position or a comorbidity in addition to the SpO2 level.

4 . The method of claim 1 , further comprising:

based on the patterns, determining a number of apnea events and a number of hypopnea events over the sleep period.

5 . The method of claim 1 , the determining patterns in the data comprising training the machine learning model using input polysomnography (PSG) recordings.

6 . The method of claim 1 , further comprising:

generating a user interface that displays over a common timeline during the sleep period at least the SDSI and a corresponding Apnea Hypopnea Index (AHI).

7 . The method of claim 1 , further comprising:

generating a user interface using explainable artificial intelligence (XAI) to graphically display the patterns.

8 . The method of claim 1 , wherein the receiving data corresponding to the single parameter of the user during the sleep period comprises affixing a second patch externally on the dermis of the user, the second patch comprising a flexible substrate, an adhesive on a first side adapted to adhere to the dermis of the user, a processor directly coupled to the substrate, and the sensor directly coupled to the substrate.

9 . A sleep system for detecting an occurrence of disturbance during a sleep period of a user, the system comprising:

a sensor configured to receive data corresponding to a single parameter of the user during the sleep period, wherein the single parameter comprises either a nasal air pressure of the user or an audio output of the user;

a first patch that is affixed externally on a dermis of the user and comprises electrodes;

one or more processors configured to, using a trained machine learning model, determine patterns in the data and derive an SpO2 level over the sleep period based on the patterns, and, based at least on the derived SpO2 level, determine a sleep disturbance severity index (SDSI) over the sleep period;

the one or more processors further configured to, in response to detecting the occurrence of disturbance during the sleep period of the user, treating the disturbance, the treating comprising activating the first patch, the activating comprising generating an electrical stimuli via the electrodes.

10 . The sleep system of claim 9 , the one or more processors further configured to: derive at least one additional breath metric comprising: a rising edge rate of each breath, a falling edge rate of each breath, a breathing cadence, a variability in breath cadence or a volume of each breath.

11 . The sleep system of claim 10 , wherein the determining an SpO2 level over the sleep period is based on one or more of any of the breath metrics, a sleep position or a comorbidity in addition to the SpO2 level.

12 . The sleep system of claim 9 , further comprising:

based on the patterns, the processors further configured to determine a number of apnea events and a number of hypopnea events over the sleep period.

13 . The sleep system of claim 9 , the determining patterns in the data comprising training a machine learning model using input polysomnography (PSG) recordings.

14 . The sleep system of claim 9 , further comprising:

the processors generating a user interface that displays over a common timeline during the sleep period at least the SDSI and a corresponding Apnea Hypopnea Index (AHI).

15 . The sleep system of claim 9 , further comprising:

the processors further configured to generate a user interface using explainable artificial intelligence (XAI) to graphically display the patterns.

16 . The sleep system of claim 9 , wherein the sensor and the processors are formed on a second patch that is adapted to be affixed externally on the dermis of the user, the second patch comprising a flexible substrate, an adhesive on a first side adapted to adhere to the dermis of the user, the processors directly coupled to the substrate, and the sensors directly coupled to the substrate.

17 . A non-transitory computer readable medium having instructions stored thereon that, when executed by one or more processors, cause the processors to detect an occurrence of disturbance during a sleep period of a user, the detecting comprising:

receiving data corresponding to a single parameter of the user via a sensor during the sleep period, wherein the single parameter comprises either a nasal air pressure of the user or an audio output of the user;

using a trained machine learning model, determining patterns in the data and deriving an SpO2 level over the sleep period based on the patterns;

based at least on the derived SpO2 level, determining a sleep disturbance severity index (SDSI) over the sleep period; and

in response to detecting the occurrence of disturbance during the sleep period of the user, treating the disturbance, the treating comprising activating a first patch that is affixed externally on a dermis of the user and comprises electrodes, the activating comprising generating an electrical stimuli via the electrodes.

18 . The non-transitory computer readable medium of claim 17 , further comprising, using the trained machine learning model, deriving at least one additional breath metric comprising: a rising edge rate of each breath, a falling edge rate of each breath, a breathing cadence, a variability in breath cadence or a volume of each breath.

19 . The non-transitory computer readable medium of claim 18 , the detecting further comprising:

wherein the determining an SpO2 level over the sleep period is based on one or more of any of the breath metrics, a sleep position or a comorbidity in addition to the SpO2 level.

20 . The non-transitory computer readable medium of claim 17 , the detecting further comprising:

based on the patterns, determining a number of apnea events and a number of hypopnea events over the sleep period.