Method and system for sleep detection

• US 10,278,639 B2
• Filed: 04/23/2015
• Issued: 05/07/2019
• Est. Priority Date: 05/02/2014
• Status: Active Grant

• Current Assignee: University Health Network
• Original Assignee: University Health Network
• Inventors: Alshaer, Hisham, Bradley, T. Douglas
• Primary Examiner(s): Natnithithadha, Navin
• Application Number:
  US14/694,198
• Publication Number:
  US 20150313535A1
• Time in Patent Office: 1,475 Days
• Field of Search: 600529-543
• CPC Class Codes: A61B 5/0816 Measuring devices for examining respiratory frequency measuring frequency of electric signals G01R23/00
• CPC Class Codes: A61B 5/4809 Sleep detection, i.e. determining whether a subject is asleep or not
• CPC Class Codes: A61B 5/4812 Detecting sleep stages or cycles
• CPC Class Codes: A61B 5/4818 Sleep apnoea
• CPC Class Codes: A61B 5/6803 Head-worn items, e.g. helmets, masks, headphones or goggles
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 7/003 Detecting lung or respiration noise

Abstract

Described herein are various embodiments of a method and system for sleep detection. For example, in one embodiment, a method is described for automatically characterizing digitized breath sounds recorded from a subject over time as indicative of the subject being one of asleep and awake. This method comprises identify individual breathing cycles in a given segment of the recorded breath sounds; calculating one or more preset breathing cycle characteristics from the identified breathing cycles; evaluating a relative regularity of the calculated characteristics for the given segment; and upon the relative regularity satisfying a preset high regularity condition, outputting a sleep status indicator that the subject was likely asleep during the segment, otherwise outputting a wake indicator that the subject was likely awake during the segment.

Claims

1. A computer-implemented method, automatically implemented by one or more processors of a computing device, for automatically characterizing digitized breath sounds recorded from a subject over time as indicative of the subject being one of asleep and awake to accurately assess a potential sleep disorder of the candidate, the method comprising:

identifying, by the one or more processors, individual breathing cycles in a given segment of said recorded breath sounds;

calculating, by the one or more processors, one or more preset breathing cycle characteristics from said identified breathing cycles;

evaluating, by the one or more processors, a relative regularity of said calculated preset breathing cycle characteristics for said given segment; and

upon said relative regularity satisfying a preset high regularity condition, outputting, by the one or more processors, a sleep status indicator indicating that the subject was likely asleep during said segment;

otherwise or upon said relative regularity satisfying a preset low regularity condition, outputting, by the one or more processors, a wake status indicator indicating that the subject was likely awake during said segment; and

compiling, by the one or more processors, each said wake status indicator or said sleep status indicator for each said given segment to compute and output an actual sleep time to be associated with said breath sound recording during which the subject is asleep in accurately assessing the potential sleep disorder of the candidate.

2. The computer-implemented method of claim 1, wherein said sleep status indicator comprises a NREM sleep indicator or a REM sleep indicator, wherein said preset high regularity condition encompasses two distinct high regularity conditions, and wherein satisfying a higher one of said two distinct high regularity conditions triggers said NREM sleep indicator whereas satisfying a lower one of said two distinct high regularity conditions triggers said REM sleep indicator.

3. The computer-implemented method of claim 1, wherein said one or more preset breathing cycle characteristics comprise at least one of an inter-breath interval and an inter-breath cycle energy.

4. The computer-implemented method of claim 3, further comprising extracting one or more designated features from said one or more preset breathing cycle characteristics, and wherein said evaluating step comprises evaluating said regularity as a function of said extracted one or more designated features.

5. The computer-implemented method of claim 4, wherein said extracted one or more designated features comprise at least one of entropy, dispersion or variance.

6. The computer-implemented method of claim 1, wherein the method further comprises automatically determining a severity of the potential disorder by automatically identifying and compiling a number of sleep disorder events from said breath sound recording and computing a severity index therefor as a function of said number and said actual sleep time.

7. The computer-implemented method of claim 6, wherein the potential disorder comprises sleep apnea and wherein said severity index comprises an apnea-hypopnea index (AHI).

8. A computer-implemented method, automatically implemented by one or more processors of a computing device, for automatically characterizing digitized breath sounds recorded from a subject over time as indicative of one of a sleep status or a wake status to accurately assess a potential sleep disorder of the candidate, the method comprising:

identifying, by the one or more processors, individual breathing cycles in a given segment of said recorded breath sounds;

calculating, by the one or more processors, a relative regularity of said identified breathing cycles;

comparing, by the one or more processors, said relative regularity with one or more preset regularity conditions to output at least one of a high regularity and a low regularity indicator indicative of the sleep status and the wake status, respectively, for said given segment; and

confirming, by the one or more processors, in response to said high regularity indicator, the wake status indicated by said high regularity indicator for said given segment, by at least one of;

identifying an instance of snoring during said given segment by extracting one or more designated snore-related features from said recorded breath sounds in said given segment previously determined to distinguish snoring sounds from non-snoring sounds;

or

identifying an instance of relatively high upper airway relaxation during said given segment by extracting one or more designated upper airway relaxation-related features from said recorded breath sounds in said given segment previously determined to distinguish relatively high upper airway narrowing instances from relatively low upper airway narrowing instances; and

compiling, by the one or more processors, each said wake status or said sleep status for each said given segment to compute an actual sleep time to be associated with said breath sound recording during which the subject is asleep in accurately assessing the potential sleep disorder of the candidate.

9. The computer-implemented method of claim 8, wherein said sleep status comprises either of a NREM sleep status and a REM sleep status, wherein said one or more preset regularity conditions encompasse two distinct high regularity conditions, and wherein satisfying a higher one of said two distinct high regularity conditions triggers said NREM sleep status whereas satisfying a lower one of said two distinct high regularity conditions triggers said REM sleep status.

10. The computer-implemented method of claim 9, wherein said NREM sleep status is output irrespective of whether either of said instance of relatively high upper airway narrowing and said instance of snoring is identified.

11. The computer-implemented method of claim 9, wherein said REM sleep status is output only upon the sleep status being confirmed by said instance of relatively high upper airway narrowing and/or said instance of snoring being identified.

12. The computer-implemented method of claim 8, wherein the method further comprises automatically determining a severity of the potential disorder by automatically identifying and compiling a number of sleep disorder events from said breath sound recording and computing a severity index therefor as a function of said number and said actual sleep time.

13. The computer-implemented method of claim 12, wherein the potential disorder comprises sleep apnea and wherein said severity index comprises an apnea-hypopnea index (AHI).