WO2015194833A1

WO2015194833A1 - During-sleep breath sound analysis apparatus and method

Info

Publication number: WO2015194833A1
Application number: PCT/KR2015/006095
Authority: WO
Inventors: 고쇼이
Original assignee: 고쇼이; 가부시키가이샤 이료후쿠시고가쿠겐큐쇼; 주식회사 마스테코
Priority date: 2014-06-16
Filing date: 2015-06-16
Publication date: 2015-12-23
Also published as: JP6315576B2; JP2016002189A

Abstract

(Technical Problem) The purpose of the present invention is to detect the during-sleep breath sound of a person who is sleeping in a natural state, and to determine, with excellent precision, an oxygen uptake state, etc. in a sleeping state or during sleep on the basis of a detected during-sleep signal. (Technical Solution) The during-sleep signal of a subject (2), which has been detected by a during-sleep breath sound detection means (3), is transmitted to a during-sleep breath sound analysis apparatus (6) through a portable terminal (5). The during-sleep breath sound analysis apparatus (6) comprises: a breath characteristic waveform acquisition means for processing the during-sleep signal and thus acquiring a breath characteristic waveform; a breath waveform discrimination means for discriminating between an inhalation waveform and an exhalation waveform in the breath characteristic waveform; a characteristic value extraction means for extracting one or a plurality of characteristic values selected from the inhalation waveform, the exhalation waveform, each time interval from the exhalation waveform to the exhalation waveform or from the exhalation waveform to the inhalation waveform, the time width of the inhalation waveform, the time width of the exhalation waveform, the amplitude value of the inhalation waveform, and the amplitude value of the exhalation waveform; and a state determination means for, on the basis of the extracted characteristic values, determining the sleep state of the subject (2), determining an apnea state or a hypopnea state or estimating an oxygen uptake state.

Description

Erosion breathing sound analysis device and method

The present invention relates to an erosion respiratory sound analysis device and method for analyzing the measured erosion breathing sound (sleeping sound when sleeping) to determine apnea and low respiration state during sleep or sleep. will be.

In the conventional sleep measurement apparatus, a sensor or the like must be attached to the body or anesthetized by the subject, which not only consumes a lot of time and quality, but also interferes with sleep.

In addition, in the commercially available simple water gauge, the precision was not high, such as a large deviation in the measurement result depending on the operation of the device or the sleeping mode.

In addition, although some can be measured only by attaching the sensor mat under the bedding to be used, there are problems such as high price or difficulty in use in a business trip.

For example, Patent Document 1 (Japanese Patent Laid-Open No. 2014-64675) discloses that an exhalation sound generator is attached to cover a subject's nose and / or mouth to convert a respiratory sound into a high frequency sound. An apnea detection system that can detect apnea with no restraint and low invasiveness by receiving and interpreting is described.

In addition, Patent Document 2 (Japanese Patent Laid-Open No. 2006-320641) discloses quantitatively determining the presence or absence of apnea during sleep by causing an artificial short-time sleep caused by anesthesia to the subject, and analyzing the breathing sound of the subject during sleep. A system is described.

[Preceding technical literature]

Patent Document 1: Japanese Patent Application Laid-Open No. 2014-64675

Patent Document 2: Japanese Patent Application Laid-Open No. 2006-320641

In order to solve various problems in various conventional systems, the present invention detects the erosion breathing sound of a person sleeping in a natural state and collects a microphone or a microphone installed near the face, and detects the erosion breathing sound. The purpose of the present invention is to provide an erosion respiratory sound analyzing apparatus and method capable of accurately determining the apnea and low breathing in the sleep state or the sleep state and estimating the oxygen intake state by accurately analyzing the data.

The invention according to Claim 1 includes: erosion breathing sound detecting means for detecting erosion breathing sound of the subject; Respiratory feature waveform acquisition means for obtaining respiratory feature waveforms on the basis of the erosion signal detected by the erosion respiratory sound detection means; Breathing waveform discriminating means for discriminating an intake waveform and an exhalation waveform in said breathing characteristic waveform; The time interval of the inhalation waveform, the time interval of the aerobic waveform, the time interval of the aerobic waveform to the aerobic waveform, the time interval of the aerobic waveform to the aerobic waveform, the time interval of the intake waveform, the time width of the aerobic waveform, Feature value extracting means for extracting one or a plurality of feature values selected from the amplitude value of the waveform and the amplitude value of the expiratory waveform; State determination means for performing the determination of the subject's sleep state, the apnea state or low breath state, or the estimation of oxygen intake state in the predetermined period based on one or a plurality of feature values in a predetermined period; Erosion breathing sound analysis device having a.

In the invention according to claim 2, in the erosion respiratory sound analysis device according to claim 1, the feature value extracting means extracts at least an amplitude value of the intake waveform and an amplitude value of the exhalation waveform, and the state determining means includes: The apnea of the subject on the basis of a change in the proportion of the time width of the portion where the amplitude value of the intake waveform and the amplitude value of the exhalation waveform exceed the threshold set for each period in the predetermined period. It is characterized in that the determination of the state or hyporespiratory state or the estimation of the oxygen intake state.

In the invention according to claim 3, in the erosion respiratory sound analysis device according to claim 1, the feature value extracting means extracts at least the time interval of the aerobic waveform from the intake waveform and the time interval of the intake waveform from the exhalation waveform, The state determining means is based on the number of times of extraction or the sum of the time intervals in which the time intervals from the intake waveform to the exhalation waveform or the time interval from the exhalation waveform to the intake waveform exceed a threshold for each period. It is characterized in that the determination of the apnea state or low breathing state or the estimation of the oxygen intake state.

In the invention according to claim 4, in the erosion respiratory sound analysis device according to claim 1, the feature value extracting means extracts at least a time interval of the intake waveform or a time interval of the aerobic waveform, and the state determining means includes: Whether the subject's sleep state is REM (rapid eye movement) sleep or non-REM sleep based on a change in the time interval of the intake waveform or the time interval of the expiratory waveform in the predetermined period. Characterized in that the determination of.

In accordance with a fifth aspect of the present invention, in the erosion respiratory sound analyzer according to claim 1, the feature value extracting means includes at least the time width of the intake waveform and the time width of the exhalation waveform or the amplitude value of the intake waveform and The amplitude value of the exhalation waveform is extracted, and the state determining means has a length relationship between the time width of the intake waveform and the time width of the exhalation waveform or the amplitude value of the intake waveform and the amplitude of the exhalation waveform in the predetermined period. Based on the magnitude relationship of the values, it is characterized in that the subject determines whether or not the subject is hypoxic.

The invention according to claim 6, further comprising: detecting an erosion breathing sound of the subject to obtain an erosion signal; Obtaining a respiratory characteristic waveform based on the erosion signal; Determining an intake waveform and an exhalation waveform in the respiratory characteristic waveform; The time interval of the inhalation waveform, the time interval of the aerobic waveform, the time interval of the aerobic waveform to the aerobic waveform, the time interval of the aerobic waveform to the aerobic waveform, the time interval of the intake waveform, the time width of the aerobic waveform, Extracting one or a plurality of feature values selected from amplitude values of the waveforms and amplitude values of the expiratory waveforms; Performing the determination of the subject's sleep state, the apnea state or low breath state, or the estimation of the oxygen intake state, based on the one or a plurality of feature values in a predetermined period. Erosion breathing sound analysis method.

According to the erosion respiratory sound analyzer according to claim 1, the respiratory characteristic waveform acquiring means for obtaining the respiratory characteristic waveform based on the erosion signal detected by the erosive respiratory sound detection means, and the intake waveform and the aerobic waveform in the respiratory characteristic waveform Since a respiratory waveform discriminating means for discriminating is provided, the feature value extracting means makes a time interval for the intake waveform, a time interval for the aerobic waveform, a time interval for the aerobic waveform to the aerobic waveform, a time interval for the intake waveform to the intake waveform, and an intake waveform. One or a plurality of feature values selected from the time width, the time width of the exhalation waveform, the amplitude value of the intake waveform, and the amplitude value of the exhalation waveform can be extracted.

Then, by the state determining means, based on the total value, the average value, or the degree of change in the predetermined period of the extracted feature value, the determination of the subject's sleep state or the estimation of the oxygen intake state in the predetermined period is performed. Although the system directly detects sound unconstrained, it is possible to obtain an effect that the determination or estimation accuracy is greatly improved as compared with the conventional homogeneous system.

In addition, since there is no need to attach a sensor or the like to the body or to anesthetize the subject, it takes no time and arms and does not disturb the subject's sleep.

Therefore, it is very useful for early detection of sleep disturbances, confirmation of sleep status of a person engaged in work where drowsy driving of drivers in subways, buses and taxis may cause an accident.

According to the erosion respiratory sound analyzer according to claim 2, in addition to the effect of the invention according to claim 1, the amplitude of the intake waveform and the amplitude of the expiratory waveform exceed the limits set for each period. Based on the change in the proportion of the total in a certain period, the effect of determining the subject's apnea state or low breathing state or estimating the oxygen intake state can be obtained with high accuracy.

According to the erosion respiratory sound analyzer according to claim 3, in addition to the effect of the invention according to claim 1, the time interval of the intake waveform to the exhalation waveform or the time interval of the exhalation waveform to the intake waveform for each fixed period exceeds the limit. Based on the number of extractions or the total value of the time intervals, the effect of determining the apnea state or the low breathing state or the estimation of the oxygen intake state of the subject can be accurately performed.

According to the erosion respiratory sound analyzer according to claim 4, in addition to the effects of the invention according to claim 1, the subject's sleeping state is based on the degree of change in the time interval of the intake waveform or the time interval of the aerobic waveform in a predetermined period. The effect of determining whether or not sleep is non-REM sleep can be obtained with high precision.

According to the erosion respiratory sound analyzer according to claim 5, in addition to the effect of the invention according to claim 1, the length relationship between the time width of the intake waveform and the time width of the expiratory waveform or the amplitude value of the intake waveform and the exhalation in a predetermined period Based on the magnitude relationship between the amplitude values of the waveforms, it is possible to obtain an effect of accurately determining whether or not the test subject is in a hypoxic state.

According to the erosion respiratory sound analysis method of the invention according to claim 6, the step of obtaining the erosion signal by detecting the erosion respiration sound of the subject, the step of obtaining the respiratory characteristic waveform based on the erosion signal, the inspiration waveform and aerobic waveform in the respiratory characteristic waveform Since it has a step of discriminating, the time interval of the intake waveform, the time interval of the aerobic waveform, the time interval of the intake waveform to the aerobic waveform, the time interval of the intake waveform to the aerobic waveform, the time width of the intake waveform, the time width of the aerobic waveform, One or more feature values selected from the amplitude value of the intake waveform and the amplitude value of the exhalation waveform can be extracted.

Then, similarly to the invention of claim 1, based on the sum value, the average value, or the degree of change in the predetermined period of the extracted feature value, the determination of the subject's sleep state or the estimation of the oxygen intake state in the predetermined period is performed. As a result, it is possible to obtain an effect that the judgment or estimation accuracy is greatly improved as compared with the conventional erosion respiratory sound detection method.

BRIEF DESCRIPTION OF THE DRAWINGS The figure which shows the outline | summary of the system containing the erosion respiratory sound analyzer concerning an Example.

2 is a diagram showing an example of an erosion signal and a graph of the result of processing the erosion signal.

3 is an explanatory diagram of a first method of determining whether or not apnea is present;

4 is an explanatory diagram of a second method of determining whether or not an apnea state is present;

5 is an explanatory diagram of a method of determining whether REM sleep or non-REM sleep

6 is an explanatory diagram for a method of determining whether or not a hypoxic state is present;

Before explaining an Example, the outline | summary of the erosion measurement / analysis system containing the erosion respiratory sound analyzer of this invention is demonstrated.

As shown in FIG. 1, erosion breathing sound detecting means 3 (collecting microphone, etc.) for detecting erosion breathing sound is installed around the 'nose' portion of the subject 2 who is sleeping in a private house, nursing home facility, etc. (1). do.

The erosion signal (audio data) detected by the erosion breath sound detecting means 3 is carried by the subject through a wireless communication means 4 such as Bluetooth (registered trademark) included in the erosion breath sound detection means 3. It is once transmitted to the terminal 5 (tablet type terminal or smart phone), and is transmitted from this portable terminal 5 to the erosion respiratory sound analyzer 6 which has a data server function via a wireless communication line, such as the Internet.

The erosion breathing sound analyzing device 6 stores the erosion signal received in the built-in storage device or the cloud memory and analyzes the data by performing various signal processing. The result of the analysis is sent to the terminal 8 (PC, etc.) provided in the portable terminal 5 or the hospital 2 where the subject 2 is hospitalized, the hospital or the like, and the subject 2 or the attending physician. 9) can confirm the result of the analysis.

The present invention also relates to an apparatus and method for analyzing erosion breath sounds in such an erosion measurement and analysis system.

EMBODIMENT OF THE INVENTION Hereinafter, embodiment of this invention is described based on an Example.

EXAMPLE

2 is a diagram showing an example of an erosion signal and a graph of the result of processing the erosion signal. The graph describes how the erosion breathing sound analyzing apparatus according to the embodiment processes the erosion signal.

The erosion respiratory sound analyzer first sets the threshold value Hv for every predetermined length (for example, 10 seconds) in the erosion signal, and then the time δi of the portion where the erosion signal exceeds the threshold value Hv ( The total of i = 1 to n divided by the predetermined length, that is, the total number of data exceeding the threshold value Hv for the total number of data (for example, when the sampling frequency is 11025 kHz) for 10 seconds The total number of data is regarded as a change in respiration as a change in the value divided by 110250, and is used as a parameter for apnea determination.

Then, the amplification process is performed so that a signal having a smaller value, either inspiratory or exhalation, appears remarkably with respect to the erosion signal, and then subjected to band pass filter processing and normalization of 50 to 2000 Hz to obtain a post-processed signal (Fig. Graph of signal after 2).

Then, the envelope of the signal after processing is obtained to obtain a respiratory characteristic waveform (graph of the respiratory characteristic waveform in FIG. 2).

This process is performed equally for every predetermined length of data.

Thereafter, various treatments are performed on the respiratory characteristic waveforms obtained in the appropriately selected predetermined period to extract one or a plurality of characteristic values, and the determination of the sleep state of the subject 2 in the predetermined period, the apnea state or Determination of low respiration or estimation of oxygen uptake is performed. Examples of the processing include the time interval of the intake waveform, that is, the process of obtaining the respiration cycle Td, the process of extracting the time interval Te of the aerobic waveform from the intake waveform, the process of extracting the time width T1 of the intake waveform, and the exhalation. The process of extracting the time width T2 of a waveform, and the process of extracting the respiratory stop time interval Ts between inhalation and exhalation are mentioned.

In addition, a frequency analysis is performed on the respiratory characteristic waveforms during the appropriately set period, and as shown in the graph of the spectrum of FIG. 2, the frequency fd (0.25 Hz) or the second peak corresponding to the first peak is shown. The frequency fe (0.5 kHz) corresponding to and the average value of Td (≒ 4 seconds) or the average value of Te (≒ 2 seconds) are derived from the relationship of T / 1 / f. From Td) and the intake waveform, the average value of the time interval Te of the expiratory waveform can be obtained.

3 is an explanatory diagram of a first method for determining whether or not apnea state or low breathing state, and a method of determining whether or not the test subject 2 is apnea state or low breathing state using this graph.

The graph of respiration fluctuations shows the variation of the above-mentioned apnea determination parameters obtained based on the erosion signal.

The graph of smoothed breathing fluctuations shows time series data obtained by performing a smoothing process on the respiratory fluctuation data of a predetermined period, and the apnea graph obtains the difference between the respiratory fluctuation data and its moving average value, and the change of the difference is a predetermined threshold value. The smaller case is Low and the larger case is High.

In the graph of apnea, it shows that the part which is low is in the state of apnea or low breathing, and when the last low part of four low parts more than 50 minutes was examined precisely, 57 minute 7 second-57 minute 42 The apnea state continued for about 35 seconds over a second.

FIG. 4 is an explanatory diagram of a second method for determining whether or not apnea state or a low breathing state, and a method of determining whether or not the test subject 2 is in an apnea state or a low breathing state using this graph.

In the respiratory characteristic waveform for a predetermined period, after performing a process of extracting the respiratory pause time interval Ts between the intake and exhalation, the portion Tn whose Ts exceeded the limit value (for example, 10 seconds) is extracted. . The number of extractions and the interval length of Tn are parameters indicating sleep apnea or sleep apnea. In addition, the change of the time interval Tn exceeding a predetermined limit value is used for the estimation of the oxygen intake state.

FIG. 5 is an explanatory diagram of a method of determining whether REM sleep or non-REM sleep is used, and a method of determining whether the test subject 2 is in REM sleep state or non-REM sleep state using this graph. Explain.

The graph of the respiratory cycle shows the fluctuation of the respiratory cycle (Td) obtained from the respiratory characteristic waveform.Td fluctuates severely, and the respiratory cycle is in an irregular REM sleep state where the respiratory cycle variation is large, and Td is stable. Therefore, in the place where the deviation of the breathing cycle is small, the breath is in a regular nonremed state of sleep.

The graph of the moving average shows time series data obtained by performing moving average processing on the respiratory cycle data of the measurement target period, where the REM sleep state is at a short period and the non-REM sleep state is at a long period. .

The graph of the sleep state calculates the difference between the respiratory cycle data and the moving average value, and sets the case where the change of the difference is greater than the predetermined limit value as Low and the case where the change is smaller.

In the graph of the sleep state, the low portion indicates that the wake state or the REM sleep state is shown, and the high portion indicates the non-remme sleep state.

FIG. 6 is an explanatory diagram for determining whether or not a hypoxic state is described, and a method for determining whether or not the test subject 2 is in a hypoxic state using this graph and the graph of the respiratory characteristic waveform of FIG. 2 will be described.

In the graph of the respiratory characteristic waveform of FIG. 2, when the time width T1 of the intake waveform is longer than the time width T2 of the exhalation waveform, the airway becomes narrow and the time width T1 of the intake waveform becomes long due to difficulty in breathing. It can be thought that it is lost, which is determined to be a hypoxic state.

Also, when the amplitude value of the intake waveform is larger than the amplitude value of the exhalation waveform, it is determined as in the low oxygen state as described above.

That is, in the graph of the respiratory characteristic waveform of FIG. 6, the amplitude value of the intake waveform is larger than the amplitude value of the immediately exhaled waveform, and the time width of the intake waveform is larger than the time width of the immediately exhaled waveform ( Although there are individual differences, it is suspected that the blood oxygen concentration is lowered due to the possibility that the test subject 2 has not sufficiently inhaled air.

Examples of the erosion measurement and analysis system to which the embodiment is applied and the apparatus and method for analyzing the erosion breathing sound of the embodiment are listed.

(1) In the erosion measurement and analysis system to which the embodiment is applied, the erosion respiratory sound detection means 3 (collecting microphone, etc.) is installed around the 'nose' portion of the subject 2 who is sleeping, but the It may be provided in a bed head, an upper side (ceiling, etc.), a bookshelf or a lighting fixture in the vicinity of the head of the subject 2, or a pajama worn by the subject 2.

(2) In the erosion measurement and analysis system to which the embodiment is applied, the erosion signal detected by the erosion breathing sound detection means 3 is once transmitted to the portable terminal 5 owned by the subject, but to a personal computer or a dedicated terminal. The erosion breathing sound detection means 3 itself may have a transmission function, and the erosion breathing sound analysis device 6 may be directly transmitted.

(3) In the embodiment, only the determination of the sleep state of the subject 2, the determination of the apnea state or the low breathing state, or the estimation of the oxygen intake state are described. Drawings, drawings, documents, etc., which are easily explained, are displayed on the mobile terminal 5 or the terminal 8 so that the subject 2 and the physician's back 9 can sleep, apnea, hypopnea or oxygen uptake. It's better to make it easier to grasp.

(4) In the description of FIGS. 2 and 5 of the embodiment, the time interval of the intake waveform is regarded as the respiration cycle, and it is determined whether the subject 2 is in the REM sleep state or the non-REM sleep state according to the magnitude of the deviation of the respiration cycle. Although the time interval of the aerobic waveform is a breathing cycle, the test subject 2 may determine whether the subject 2 is in a REM sleep state or a non-REM sleep state as described above.

Claims

Erosion breathing sound detection means for detecting the erosion breathing sound of the subject,

Respiratory feature waveform acquisition means for obtaining a respiratory feature waveform on the basis of the erosion signal detected by the erosion respiratory sound detection means;

Breathing waveform discriminating means for discriminating an intake waveform and an exhalation waveform in said breathing characteristic waveform;

The time interval of the inhalation waveform, the time interval of the aerobic waveform, the time interval of the aerobic waveform to the aerobic waveform, the time interval of the aerobic waveform to the aerobic waveform, the time interval of the intake waveform, the time width of the aerobic waveform, Feature value extraction means for extracting one or a plurality of feature values selected from the amplitude value of the waveform and the amplitude value of the expiratory waveform;

Based on one or a plurality of feature values in a predetermined period, state determination means for performing the determination of the subject's sleep state, the apnea or low breath state, or the estimation of oxygen intake state in the predetermined period; Erosion breathing sound analysis device provided.
The method according to claim 1,

The feature value extracting means extracts at least an amplitude value of the intake waveform and an amplitude value of the exhalation waveform,

The state determining means is based on the change in the proportion of the time width of the portion where the amplitude value of the intake waveform and the amplitude value of the exhalation waveform exceed the threshold set for each predetermined period, in the predetermined period. Erosion breathing sound analysis device characterized in that for performing the determination of the apnea state or hyporespiratory state of the subject or the oxygen intake state.
The method according to claim 1,

The feature value extracting means extracts at least a time interval of the inhalation waveform from the intake waveform and a time interval of the intake waveform from the exhalation waveform,

The state determining means is based on the number of times of extraction or the sum of the time intervals in which the time intervals from the intake waveform to the exhalation waveform or the time interval from the exhalation waveform to the intake waveform exceed a threshold for each period. Erosion call sound analysis device characterized in that for performing the determination of the apnea state or low breathing state or the estimation of oxygen intake state.
The method according to claim 1,

The feature value extracting means extracts at least a time interval of the intake waveform or a time interval of the exhalation waveform,

The state determining means performs the determination of whether the subject's sleep state is REM sleep or non-REM sleep based on the degree of change of the time interval of the intake waveform or the time interval of the exhalation waveform in the predetermined period. Erosion breathing sound analyzer, characterized in that.
The method according to claim 1,

The feature value extracting means extracts at least a time width of the intake waveform and a time width of the exhalation waveform or at least an amplitude value of the intake waveform and an amplitude value of the exhalation waveform,

The state determining means is based on the length relationship between the time width of the intake waveform and the time width of the exhalation waveform or the magnitude relationship between the amplitude value of the intake waveform and the amplitude value of the exhalation waveform in the predetermined period. Erosion breathing sound analysis device, characterized in that for performing a determination of whether or not the subject is in a hypoxic state.
Detecting the erosion breathing sound of the subject to obtain an erosion signal,

Obtaining a respiratory characteristic waveform based on the erosion signal;

Determining an intake waveform and an exhalation waveform in the respiratory characteristic waveform;

The time interval of the inhalation waveform, the time interval of the aerobic waveform, the time interval of the aerobic waveform to the aerobic waveform, the time interval of the aerobic waveform to the aerobic waveform, the time interval of the intake waveform, the time width of the aerobic waveform, Extracting one or more feature values selected from the amplitude value of the waveform and the amplitude value of the expiratory waveform,

Performing the determination of the subject's sleep state, the apnea state or low breath state, or the estimation of the oxygen intake state, based on the one or a plurality of feature values in a predetermined period. Erosion breathing sound analysis method.