JP7288334B2 - Sleep state estimation device, method of operating sleep state estimation device, and sleep state estimation program - Google Patents

Description

The present invention relates to a sleep state estimation device, a sleep state estimation method, and a sleep state estimation program.

Sleep apnea syndrome (SAS), which is a disease that causes respiratory arrest or hypopnea during sleep, is known. As a method for diagnosing sleep apnea syndrome, there is a polysomnography test in which biological activities such as electroencephalogram, eye movement, electrocardiogram, electromyogram, breathing curve, snoring, and arterial blood oxygen saturation are measured overnight. However, the apparatus is large-scale, and the subject needs to be hospitalized for a wide variety of examinations. In addition, comprehensive judgment by a doctor or clinical laboratory technologist is required. Therefore, there is a high need for a sleep state estimation device that can easily diagnose sleep apnea syndrome.

It is known that respiratory arrest or hypopnea during sleep is associated with heart rate variability. Observation of heart rate variability during sleep reveals that, compared to heart rate variability during normal sleep, when respiratory arrest or hypopnea occurs, a characteristic heart rate variability that alternates between bradycardia and tachycardia, the so-called cycle cyclic variation of heart rate (CVHR) occurs.

Japanese Patent Laid-Open No. 2002-200001 discloses a method for detecting periodic heart rate variability to determine a respiratory arrest or hypopnea state. Further disclosed is the use of changes in the activity ratio of the sympathetic and parasympathetic nerves as an index reflecting the individual health of the subject to improve the accuracy of respiratory arrest or hypopnea.

Further, Patent Document 2 discloses a sleep state determination method for estimating a sleep state such as REM sleep or non-REM sleep by obtaining an activity ratio of sympathetic nerves and parasympathetic nerves from heartbeat intervals.

Furthermore, in patent document 3, when determining the sleep state of a subject using pulse interval data during sleep, if the body movement indicating the degree of movement of the subject's body is within a predetermined range, the corresponding pulse interval A sleep state determination method that excludes data is disclosed.

JP 2010-051387 A JP-A-7-143972 JP-A-2005-279113

However, in the detection method described in Patent Literature 1, the detected periodic heart rate fluctuation may include REM sleep and the like, and there is a possibility that the detection accuracy of respiratory arrest or hypopnea during sleep will be low.

Further, the estimation method described in Patent Document 2 can determine REM sleep and non-REM sleep, but does not disclose determination of respiratory arrest or hypopnea during sleep.

Furthermore, the subject's body movement is caused by various factors other than when the subject's breathing stops or becomes hypopnea. Hard to judge.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above-described problems, and to provide a simple sleep state estimation device with high accuracy in estimating respiratory arrest or hypopnea during sleep.

In order to achieve the above object, a sleep state estimation apparatus according to the present invention includes an acquisition unit that acquires beat interval data of a subject in a predetermined period including a sleep period, and an autonomous a neural index output unit, a respiratory estimator for estimating an apnea or hypopnea sleep state of a subject based on the beat interval data and the autonomic nerve index data, and an output unit for outputting an apnea index signal indicating the estimation result. , is characterized by having

Furthermore, the sleep state estimation apparatus according to the present invention further includes a body motion data output unit that acquires body motion signals of the subject and outputs body motion data. Preferably, the apnea or hypopnea sleep state of the subject is estimated based on the index data and the body motion data.

Further, in the sleep state estimating device according to the present invention, the respiration estimating unit includes an event detecting unit that detects periodic heart rate fluctuation event data from the beat interval data, Excludes periodic heart rate variability event data for a period within, and excludes periodic heart rate variability event data in which the autonomic nerve index data is less than a predetermined autonomic nerve index threshold from the periodic heart rate variability event data, and and an event processor for excluding periodic heart rate variability event data whose motion data is less than a predetermined body motion threshold.

Further, in the sleep state estimation device according to the present invention, it is preferable that the respiration estimation unit estimates the subject's apnea or hypopnea sleep state based on data of a plurality of subjects.

A method for estimating a sleep state according to the present invention includes acquiring beat interval data of a subject in a predetermined period including a sleep period, outputting autonomic nerve index data from the beat interval data, The method includes estimating an apnea or hypopnea sleep state of the subject based on the neural index data, and outputting an apnea index signal indicating the estimation result.

A sleep state estimation program according to the present invention is a program that causes a computer to acquire beat interval data of a subject in a predetermined period including a sleep period, and outputs autonomic nerve index data from the beat interval data. estimating an apnea or hypopnea sleep state of the subject based on the beat interval data and the autonomic nerve index data; and outputting an apnea index signal indicating the estimation result. A sleep state estimation program that causes a computer to perform the method.

According to the sleep state estimation device according to the present invention, it is possible to easily improve the accuracy of estimating respiratory arrest or hypopnea during sleep.

FIG. 2 is a diagram for explaining an electrocardiogram showing heartbeats and a heartbeat interval, which is a time interval between heartbeats. (a) is an electrocardiogram waveform, and (b) is a graph of heart beat interval (RRI). FIG. 4 is a diagram for explaining a conventional method of estimating periodic heart rate variability from pulse wave interval variability. FIG. 10 is a diagram showing an example of a conventional algorithm for estimating a periodic heart rate variability flag from pulse wave intervals; FIG. 10 is a graph depicting the relationship between smoothed pulse intervals and periodic heart rate variability events; It is a figure which shows an example of an autonomic nerve index (LF/HF). It is a figure which shows an example of the process outline|summary of 1st Embodiment of a sleep state estimation apparatus. BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows an example of a structure of the sleep state estimation apparatus of 1st Embodiment. It is a figure which shows an example of the flowchart of a sleep state estimation process. It is a figure which shows an example of the process outline|summary of 2nd Embodiment of a sleep state estimation apparatus. It is a figure which shows an example of a structure of the sleep state estimation apparatus of 2nd Embodiment. 6 is a diagram showing an example of a flowchart of sleep measurement processing performed by the sleep state estimation device 7. FIG. It is an external perspective view showing an example of a sleep state estimation device of a third embodiment. (a) is an external perspective view from the front side, and (b) is an external perspective view from the back side. It is a figure which shows an example of a structure of the main-body part of the sleep state estimation apparatus of 3rd Embodiment. FIG. 10 illustrates the effect of periodic heart rate variability detection according to the present invention;

A sleep state estimation device according to one aspect of the present disclosure will be described below with reference to the drawings. However, it should be noted that the technical scope of the present disclosure is not limited to those embodiments, but extends to the invention described in the claims and equivalents thereof. In the following description and drawings, components having the same functional configuration are denoted by the same reference numerals, thereby omitting redundant description.

[Heart beat and heartbeat interval]
FIG. 1 is a diagram for explaining an electrocardiogram showing heartbeats and a heartbeat interval, which is a time interval between heartbeats. (a) is an electrocardiogram waveform, and (b) is a heartbeat interval graph.

It is well known that the heart beats "slower" when a person is at rest, such as lying down, and "fastens" when the person is exercising or tense. The phenomenon that the heart beats "fast" or "slow" can be expressed, for example, by a heartbeat interval, which is the time between one beat and the next beat. When the heart beats fast, the beat-to-beat interval time decreases, and conversely, when the heart beats slowly, the beat-to-beat interval time increases.

FIG. 1(a) is a diagram showing an electrical signal generated when the heart beats, called an electrocardiogram. The sharpest peak of an electrocardiogram, R, is an electrical signal generated when the ventricles of the heart contract rapidly to pump blood out of the heart. ; R-R Interval).

FIG. 1(b) is a graph plotting the heart rate interval (RRI). It can be seen that the heartbeat interval (RRI) is not always constant but fluctuates. Not only does exercise increase heart rate and change heart rate interval (RRI), variations in heart rate interval (RRI) are also observed during rest and sleep. In the upper figure, the heartbeat interval (RRI) fluctuates between 950 ms and 900 ms. Such variation in heartbeat interval is called heart rate variability.

It is known that when a subject is in an apnea or hypopnea sleep state, a characteristic heart rate variability that alternates between bradycardia and tachycardia, so-called periodic heart rate variability, occurs. However, there are various factors in periodic heart rate fluctuations, and it is necessary to simply and accurately estimate whether the subject is in an apnea or hypopnea sleep state. In this embodiment, the heart rate interval (RRI) and the pulse interval (PPI) are collectively referred to as the beat interval. Pulse interval (PPI) refers to the time interval between the sharpest peaks of the pulse wave. The electrocardiogram captures the potential changes that occur with the activity of the heart muscle, while the pulse wave measures changes in the blood flow from the heart's pumped blood through the arteries to the periphery. PPI is information acquired from each, and it is known that they show almost the same value when relatively at rest such as during sleep. The following embodiments will be described based on pulse wave intervals.

[Conventional method for detecting periodic heart rate variability from interpulse interval variability]
FIG. 2 is a diagram for explaining a conventional method of estimating periodic heart rate variability from pulse wave interval variability.

The upper graph in FIG. 2 is a graph in which the pulse wave interval value is plotted on the vertical axis and the sleep time is plotted on the horizontal axis. A solid line indicates an actual pulse wave interval, and a dashed line indicates a pulse wave interval smoothed by, for example, a moving average. A drop in the smoothed pulse wave interval is called a dip. The lower graph in FIG. 2 is a graph showing the periodic heart rate variability flag estimated from the smoothed pulse wave interval, with the sleep time as the horizontal axis. A periodic heart rate variability flag is a flag indicating that the subject is in an apnea state.

FIG. 3 is a diagram showing an example of a conventional algorithm for estimating a periodic heart rate variability flag from pulse wave intervals. First, a shape near the minimum value of the smoothed pulse wave interval is extracted as a dip candidate (ST401). The extracted dip shape is determined, for example, it is determined whether the dip shape resembles a predetermined parabolic shape (ST402). Further, the depth of each dip is calculated as the difference between the depth of the dip, for example, the value at the center time of the minimum value and the average value of the maximum values of the moving averages before and after, and the threshold determination is performed (ST403). Further, the width of the dip is determined, for example, it is determined whether or not the adjacent dip has a dip width within a predetermined range (ST404).

Furthermore, the similarity of dips is determined (ST405). For example, a dip having a dip width and a dip height within a predetermined range with respect to adjacent dips, and a dip width to dip height ratio within a predetermined range with respect to adjacent dips. are identified as dips with similarity. Furthermore, the periodicity of dips is determined (ST406). For example, for all two adjacent dips out of four consecutive dips, the time difference between the two adjacent dips is within a predetermined time range, and the variation in magnitude of the time difference is within a predetermined range. , the four dips are periodic. A dip that satisfies all of the above conditions is acquired as a periodic heart rate variability flag (ST407). The cyclic heart rate variability flag indicates that the subject is in an apnea state and a cyclic heart rate variability event has occurred at the time corresponding to the cyclic heart rate variability flag.

FIG. 4 is a graph depicting the relationship between smoothed pulse intervals and periodic heart rate variability events.

It can be seen that the subject's smoothed pulse wave interval is stable in the left portion A, but fluctuates greatly in the right portion B. FIG. The left part A obtains the cyclic heart rate variability flags c(1, t1), c(2, t2), c(3, t3) by a conventional algorithm for estimating cyclic heart rate variability from pulse interval variability. Here, c(n, tn) is the n-th periodic heart rate variability flag and indicates detection at time tn (n: natural number). Furthermore, in the right part B, periodic heart rate variability flags c(4, t4), c(5, t5), c(6, t6), c(7, t7), c(8, t8) are acquired.

In the right part B, where the smoothed pulse wave interval fluctuates greatly, the subject is not in an apnea state, but in another sleep state such as REM sleep. It was obtained by comparing with the detected periodic heart rate variability flag. The present invention excludes the periodic heart rate variability flags indicating that the subject is not in an apnea state from the periodic heart rate variability flags obtained by a conventional algorithm for obtaining periodic heart rate variability from pulse interval variability. By excluding the periodic heart rate variability flag indicating that the subject is not in an apnea state, it is accurately detected that the subject is in an apnea state and a periodic heart rate variability event has occurred.

[Example of autonomic nerve index]
It is known that the sympathetic activity of the subject is activated when the subject enters an apnea sleep state. It is known that the activity level of sympathetic nerve activity can be grasped by analyzing heart rate variability, for example.

FIG. 5 is a diagram showing an example of an autonomic nerve index (LF/HF). An example of results of frequency analysis of heart rate variability is shown. For example, it is a graph showing the relationship between the frequency (horizontal axis) and the power (vertical axis) obtained by frequency analysis of the heartbeat interval (RRI) over a predetermined period of time, for example, 5 minutes. Let LF be the peak value in the low frequency section (eg, 0.05 Hz-0.15 Hz), and let HF be the peak value in the high frequency section (eg, 0.15 Hz-0.4 Hz).

LF mainly indicates the activity of the sympathetic nerve function, and HF indicates the activity of the parasympathetic nerve function. However, since LF is involved not only in sympathetic nerves but also in parasympathetic nerves, the ratio of LF to HF (LF/HF) is often used as an index of sympathetic nerve activity. Used as an index of neural activity. Also, instead of the peak value, the ratio of the power spectral densities of the low frequency section and the high frequency section may be used as the autonomic nerve index (LF/HF). Using the autonomic nerve index (LF/HF), which indicates the degree of sympathetic nerve activity, it is possible to increase the accuracy of estimation when the subject is in an apnea sleep state.

[Outline of processing of the first embodiment of the sleep state estimation device]
FIG. 6 is a diagram showing an example of a processing outline of the sleep state estimation device according to the first embodiment of the present invention.

A wristband 1 having an optical pulse wave sensor 11 for detecting the beating state of the subject's heart, for example, the pulse, is attached to the subject's wrist. The optical pulse wave sensor 11 has, for example, an optical pulse wave sensor IC equipped with an LED driver and a green detection photodiode. A pulse wave signal can be obtained by measuring, with a photodiode, the intensity of the reflected light when the LED is directed toward the living body. The pulse wave sensor may be, for example, a pulse pressure detection type pulse wave sensor as well as an optical type.

In this embodiment, the wristband 1 is attached to the subject's wrist, but it may be attached to other parts of the subject.

A pulse signal detected by the pulse wave sensor of the wristband 1 is transmitted to the sleep state estimation device 2 by wireless communication such as WiFi or Bluetooth (registered trademark). The pulse signal may not be transmitted in real time, and for example, after the subject wakes up from sleep, the pulse signal may be transmitted to the sleep state estimation device 2 all at once. Further, it is not necessary to directly wirelessly transmit the pulse signal from the wristband 1 to the sleep state estimation device 2, and the pulse signal may be transmitted to the sleep state estimation device 2 via, for example, a smart phone or an access point.

A pulsation signal transmitted from the wristband 1 is received by the communication interface 22 via the antenna 21 of the sleep state estimation device 2 .

The acquiring unit 221 acquires beat interval data indicating intervals between beats, for example, time intervals between adjacent peak values of pulse waves, from the beat signal indicating the beating state of the heart of the subject. The beat interval data is input to respiration estimator 222 . Also, the beat interval data is output to the autonomic nerve index output unit 226 .

The event detector 223 of the respiration estimator 222 smoothes the beat interval data. Further, the event detector 223 detects cyclical heart rate variability events using the algorithm for estimating the cyclical heart rate variability flag from the conventional pulse interval shown in FIG. The detected periodic heart rate variability event data is input to the event processing section 224 .

The autonomic nerve index output unit 226 outputs autonomic nerve index (LF/HF) data based on the beat interval data. The autonomic nerve index data output from the autonomic nerve index output unit 226 is input to the event processing unit 224 .

The event processing unit 224 excludes periodic heart rate variability event data corresponding to a period in which the beat interval is within a predetermined range from the detected periodic heart rate variability event data. In periodic heart rate variability associated with apnea, the sympathetic nervous system is dominant, and the average beat interval tends to be short. However, the average beating interval tends to be longer than in the case of apnea, so the predetermined range is set around the average beating interval during REM sleep, and this range is excluded. Select only cyclic heart rate variability event data that show .

The event processing unit 224 further filters the excluded periodic heart rate variability event data indicating apnea based on the autonomic nerve index data. The filtering determines, for example, the periodic heart rate variability event data when the autonomic nerve index is greater than or equal to a predetermined autonomic nerve index threshold as the periodic heart rate variability event data indicating apnea. When the autonomic index (LF/HF) is high, it is estimated that the subject is likely to be in an apnea or hypopnea sleep state.

Based on the detection result by the event processing unit 224, the output unit 225 outputs periodic heart rate variability event data as an apnea index. Instead of, or in conjunction with, the cyclic heart rate variability event data, an index of respiration per detected unit time, e.g. Apnea Hypopnea Index) may be output.

The apnea index is input to a display unit (not shown) built into the sleep state estimation device 2, and a value corresponding to the apnea hypopnea index is displayed on the display screen, for example. The apnea index may also be written to a hard disk (not shown) or a removable memory card built into the sleep state estimation device 2 . Furthermore, the value corresponding to the apnea-hypopnea index may be output to a printer, information terminal device, or server device connected to the sleep state estimation device 2 . It may also be transmitted over a network, for example to a server device having a subject database.

According to the sleep state estimation device 2 of the present embodiment, it is possible to easily improve the detection accuracy of respiratory arrest or hypopnea during sleep.
[Configuration example of the sleep state estimation device of the first embodiment]
FIG. 7 is a diagram showing an example of the configuration of the sleep state estimation device of the first embodiment.

The sleep state estimation device 2 has a communication interface 22 that communicates with the outside via an antenna 21 and a control unit 24 that connects to the communication interface 22 . Furthermore, it may have a storage medium interface 23 for connecting with an external storage medium. For example, the storage medium interface 23 allows the sleep state estimation device 2 to read detection data stored in the storage medium of the wristband 1 .

The control unit 24 has a control storage unit 210 and a control processing unit 220 . The control storage unit 210 is composed of one or more semiconductor memories. For example, it has at least one of RAM and non-volatile memory such as flash memory, EPROM, and EEPROM. The control storage unit 210 stores driver programs, operating system programs, application programs, data, etc. used for processing by the control processing unit 220 .

For example, the control storage unit 210 stores a device driver program that controls the communication interface 22 as a driver program. The computer program may be installed in the control storage unit 210 from a computer-readable portable recording medium such as CD-ROM, DVD-ROM, etc. using a known setup program or the like. Alternatively, it may be downloaded from a program server or the like and installed.

Furthermore, the control storage unit 210 may temporarily store temporary data related to predetermined processing. The control storage unit 210 includes a threshold table 211 such as a beat threshold value, a beat interval average range, an autonomic nerve index threshold value, and a body movement threshold value, a time table 212 such as an estimated time, a judgment time, an autonomic nerve index calculation time, and the like, and statistical processing. Statistical master 213 and the like for the purpose are stored.

A threshold table 211 such as a beat threshold value, a beat interval average range, an autonomic nerve index threshold value, and a body movement threshold value, and a time table 212 such as an estimated time, a determination time, an autonomic nerve index output time, etc., stored in the control storage unit 210 are: A unique set value may be used. Also, the threshold table 211 and the time table 212 may be statistically set based on the average, variance, etc. of statistical values obtained by taking statistics of apnea or hypopnea sleep states of multiple subjects. Furthermore, biometric information of the subject, such as age, sex, and set values according to weight, may be stored in the statistics master 213 and used to estimate the apnea or hypopnea sleep state of the subject.

The control processing unit 220 has one or more processors and their peripheral circuits. The control processing unit 220 comprehensively controls the overall operation of the sleep state estimation device 2, and is, for example, an MCU (Micro Controller Unit).

The control processing unit 220 executes processing based on programs (operating system program, driver program, application program, etc.) stored in the control storage unit 210 . Also, the control processing unit 220 may execute a plurality of programs (application programs, etc.) in parallel. The control processing unit 220 includes an acquisition unit 221, a respiration estimation unit 222, an event detection unit 223, an event processing unit 224, an output unit 225, an autonomic nerve index output unit 226, and the like.

These units included in the control processing unit 220 may be implemented in the control unit 24 as independent integrated circuits, circuit modules, microprocessors, or firmware.

[Sleep state estimation processing flow of the first embodiment]
FIG. 8 is a diagram showing an example of a flowchart of sleep state estimation processing performed by the sleep state estimation device 2. As shown in FIG.

In the sleep state estimation processing flow of this example, the pulsation signals for one sleep time of the subject, for example, 7 hours have already been sent from the subject's wristband 1 to the sleep state estimation device 2, and are stored in the control storage unit 210. The sleep estimation process, which is the process when stored, can be performed collectively. Alternatively, for example, a pulse signal may be obtained in units of one hour, and estimation processing may be performed every hour.

Acquisition section 221 acquires beat interval data indicating the beat interval using the beat signal received from subject's wristband 1 (ST801).

Autonomic nerve index output section 226 outputs autonomic nerve index data based on the beat interval data (ST802).

Event detection section 223 smoothes the beat interval data (ST803). Furthermore, event detection section 223 detects a periodic heart rate fluctuation event using the conventional algorithm for estimating a periodic heart rate fluctuation flag from the pulse wave interval shown in FIG. 3 (ST804). Event processing section 224 calculates the moving average (mv) of the beat interval data (ST805). Event processing section 224 calculates the average value (An) of beat interval data for a predetermined period, for example, every hour or for one sleep period (ST806).

Event processing section 224 determines whether the moving average (mv) of the beat interval data is equal to or less than the average value (An) at time tk when the periodic heart rate variability flag is detected (ST807). When the moving average (mv) is equal to or less than the average value (An) (ST807: YES), event processing section 224 determines whether the autonomic nerve index at time tk is equal to or greater than a predetermined autonomic nerve index threshold Th (ST808). . When the moving average (mv) exceeds the average value (An) (ST807: NO), the event processing section 224 returns to ST807 and calculates the moving average (mv) and the average value (An) of the periodic heart rate variability flag at the next time. ). Here, it is assumed that the moving average (mv) is equal to or less than the average value (An), but this is not limitative, and the average value (An) may be multiplied by a constant (for example, 1.1). In addition, although An is the average value in one sleep period, it may be the average value within the ultradian rhythm of sleep, for example, every 1.5 hours.

When the autonomic nerve index at time tk is equal to or greater than the predetermined autonomic nerve index threshold Th (ST808: YES), event processing section 224 changes the periodic heart rate variability flag c(k, tk) at time tk to the periodic heart rate variability flag. (ST809). When the autonomic nerve index at time tk is less than the predetermined autonomic nerve index threshold value Th (ST808: NO), event processing section 224 returns to ST807, and the moving average (mv) of the periodic heart rate variability data at the next time and the average Compare the values (An).

Event processing section 224 determines whether time tk has exceeded the sleep time (ST810). When the time tk has not exceeded the sleep time (ST810: NO), the process returns to ST807. When time tk exceeds the sleep time (ST810; YES), output section 225 outputs an apnea index based on the accumulated periodic heart rate variability flag (ST811). The process flow ends.

[Outline of processing of the second embodiment of the sleep state estimation device]
In the first embodiment, the sleeping state of the subject is estimated based on the beat interval data and the autonomic nerve index data of the sleeping subject. In the second embodiment, the sleeping state of the subject is further estimated based on the body motion data.

FIG. 9 is a diagram showing an example of a processing outline of the second embodiment of the sleep state estimation device according to the present invention.

A wrist band 6 having an optical pulse wave sensor 61 and a body motion sensor 62 for detecting the beating state of the subject's heart, for example, the pulse, is attached to the subject's wrist. The optical pulse wave sensor 61 has, for example, an optical pulse wave sensor IC equipped with an LED driver and a green detection photodiode. A pulse wave signal can be obtained by measuring, with a photodiode, the intensity of the reflected light when the LED is directed toward the living body. The pulse wave sensor may be, for example, a pulse pressure detection type pulse wave sensor as well as an optical type.

The body motion sensor 62 is, for example, a motion sensor having an acceleration sensor. The motion sensor may have a gyro. A plurality of acceleration sensors can detect a three-dimensional motion vector amount, and furthermore, can calculate the amount of movement of the subject's wrist.

In this embodiment, the wristband 6 is attached to the subject's wrist, but it may be attached to other parts of the subject.

The pulse signal detected by the optical pulse wave sensor 61 of the wristband 6 and the body motion signal detected by the body motion sensor 62 are transmitted to the sleep state estimation device 7 by wireless communication, for example, WiFi, Bluetooth (registered trademark). be done. The pulse signal and the body motion signal may not be transmitted in real time. For example, after the subject wakes up from sleep, they may be transmitted to the sleep state estimation device 7 all at once. Moreover, it is not necessary to directly wirelessly transmit the pulse signal and the body motion signal from the wristband 6 to the sleep state estimation device 7, and they may be transmitted to the sleep state estimation device 7 via, for example, a smartphone or an access point.

The communication interface 72 receives the pulse signal and body motion signal transmitted from the wristband 6 via the antenna 71 of the sleep state estimation device 7 .

The acquiring unit 721 acquires beat interval data indicating intervals between beats, for example, time intervals between adjacent peak values of pulse waves, from the beat signal indicating the beating state of the subject's heart. The beat-to-beat data is input to respiration estimator 722 . Also, the beat interval data is input to the autonomic nerve index output unit 726 .

The autonomic nerve index output unit 726 calculates autonomic nerve index (LF/HF) data based on the beat interval data. The autonomic nerve index data output from the autonomic nerve index output unit 726 is input to the respiration estimation unit 722 .

The body motion data output unit 727 acquires the body motion signal of the subject and outputs body motion data indicating the movement amount of the wrist of the subject. It is known that when a subject stops breathing or falls into a state of hypopnea during sleep, when respiration is resumed beyond the limit, an arousal reaction occurs instantaneously with body movement and breathing is resumed. By measuring body movements during sleep, the possibility of occurrence of apnea or hypopnea can be determined.

The respiration estimator 722 estimates the apnea or hypopnea sleep state of the subject using the beat-to-beat data, the autonomic nerve index data, and the body motion data.

Event detector 723 of respiration estimator 722 smoothes the beat-to-beat data. In addition, the event detector 723 detects cyclic heart rate variability events using the algorithm of estimating the cyclic heart rate variability flag from the conventional pulse interval shown in FIG. The detected periodic heart rate variability event data is input to the event processing section 724 .

The autonomic nerve index output unit 726 outputs autonomic nerve index (LF/HF) data based on the beat interval data. The autonomic nerve index data output from the autonomic nerve index output unit 726 is input to the event processing unit 724 .

The event processing unit 724 excludes periodic heart rate variability event data corresponding to a period in which the beat interval is within a predetermined range from the detected periodic heart rate variability event data. In periodic heart rate variability associated with apnea, the sympathetic nervous system is dominant, and the average beat interval tends to be short. However, the average beating interval tends to be longer than in the case of apnea, so the predetermined range is set around the average beating interval during REM sleep, and this range is excluded. Select only cyclic heart rate variability event data that show .

The event processor 724 further filters the excluded periodic heart rate variability event data indicating apnea based on the autonomic nerve index data. The filtering determines, for example, the periodic heart rate variability event data when the autonomic nerve index is greater than or equal to a predetermined autonomic nerve index threshold as the periodic heart rate variability event data indicating apnea. When the autonomic nerve index (LF/HF) is high, ie, when the sympathetic nervous system is activated, it is estimated that the subject is likely to be in an apnea or hypopnea sleep state.

The event processing unit 724 further filters the periodic heart rate variability event data filtered based on the autonomic nerve index data based on the body movement data. In filtering, for example, periodic heart rate variability event data when the amount of body movement is greater than or equal to a predetermined body movement index threshold is determined as periodic heart rate variability event data indicating apnea. When the amount of body movement is large, it is presumed that the subject has caused a momentary arousal response due to an apnea response, and it is highly likely that the subject is in an apnea or hypopnea sleep state.

Based on the detection result by the event processing unit 724, the output unit 725 outputs periodic heart rate variability event data as an apnea index. Instead of, or in conjunction with, the cyclic heart rate variability event data, an index of respiration per detected unit time, e.g. Apnea Hypopnea Index) may be output.

The apnea index is input to a display unit (not shown) built into the sleep state estimation device 2, and a value corresponding to the apnea hypopnea index is displayed on the display screen, for example. The apnea index may also be written to a hard disk (not shown) or a removable memory card built into the sleep state estimation device 2 . Furthermore, the value corresponding to the apnea-hypopnea index may be output to a printer, information terminal device, or server device connected to the sleep state estimation device 2 . It may also be transmitted over a network, for example to a server device having a subject database.

The apnea index is input to a display unit (not shown) built into the sleep state estimation device 7, and the value of the apnea hypopnea index, for example, is displayed on the display screen. The apnea index may also be written to a hard disk (not shown) or a removable memory card built into the sleep state estimation device 2 . Furthermore, the apnea index may be output to a printer, information terminal device, or server device connected to the sleep state estimation device 7 . It may also be transmitted over a network, for example to a server device having a subject database.

According to the sleep state estimation device 7 of the present embodiment, it is possible to easily improve the detection accuracy of respiratory arrest or hypopnea during sleep.

[Configuration example of sleep state estimation device of second embodiment]
FIG. 10 is a diagram showing an example of the configuration of the sleep state estimation device of the second embodiment.

The sleep state estimation device 7 has a communication interface 72 that communicates with the outside via an antenna 71 and a control unit 74 that connects to the communication interface 72 . Furthermore, it may have a storage medium interface 73 for connecting with an external storage medium. For example, the storage medium interface 73 allows the sleep state estimation device 2 to read detection data stored in the storage medium of the wristband 6 .

The control unit 74 has a control storage unit 710 and a control processing unit 720 . The control memory unit 710 is composed of one or more semiconductor memories. For example, it has at least one of RAM and non-volatile memory such as flash memory, EPROM, and EEPROM. The control storage unit 710 stores driver programs, operating system programs, application programs, data, etc. used for processing by the control processing unit 720 .

For example, the control storage unit 710 stores a device driver program that controls the communication interface 22 as a driver program. The computer program may be installed in the control storage unit 710 from a computer-readable portable recording medium such as a CD-ROM, DVD-ROM, etc. using a known setup program or the like. Alternatively, it may be downloaded from a program server or the like and installed.

Furthermore, the control storage unit 710 may temporarily store temporary data related to predetermined processing. The control storage unit 710 includes a threshold table 711 such as a pulse threshold, an autonomic nerve index threshold, an average range of beat intervals, and a body motion threshold, a time table 712 such as an estimated unit time and an autonomic nerve index output unit time, and a table 712 for statistical processing. statistic master 713 and the like.

A threshold table 711 such as a beat threshold value, an autonomic nerve index threshold value, an average range of beat intervals, and a body motion threshold value, and a time table 712 such as an estimated time, a judgment time, an autonomic nerve index calculation time, etc., stored in the control storage unit 710 are: A unique set value may be used. Also, the threshold table 711 and the time table 712 may be statistically set based on the average, variance, etc. of statistical values obtained by taking statistics of apnea or hypopnea sleep states of multiple subjects. Furthermore, biometric information of the subject, such as age, sex, and set values according to weight, may be stored in the statistics master 713 and used to estimate the apnea or hypopnea sleep state of the subject.

The control processing unit 720 has one or more processors and their peripheral circuits. The control processing unit 720 comprehensively controls the overall operation of the sleep state estimation device 2, and is, for example, an MCU.

The control processing unit 720 executes processing based on programs (operating system program, driver program, application program, etc.) stored in the control storage unit 710 . Also, the control processing unit 720 may execute a plurality of programs (application programs, etc.) in parallel. The control processing unit 720 includes an acquisition unit 721, a respiration estimation unit 722, an event detection unit 723, an event processing unit 724, an output unit 725, an autonomic nerve index output unit 726, a body motion data output unit 727, and the like.

Each of these parts of control processing unit 720 may be implemented in control unit 74 as an independent integrated circuit, circuit module, microprocessor, or firmware.

[Sleep state estimation processing flow of the second embodiment]
FIG. 11 is a diagram showing an example of a flowchart of sleep measurement processing performed by the sleep state estimation device 7. As shown in FIG.

In the sleep state estimation processing flow of this example, the pulse signal and body motion signal for one sleep time of the subject, for example, seven hours have already been transmitted from the subject's wristband 6 to the sleep state estimation device 7, and control The sleep estimation process, which is the process when stored in the storage unit 710, can be performed collectively. Alternatively, for example, a pulse signal may be obtained in units of one hour, and estimation processing may be performed every hour.

Obtaining section 721 obtains beat interval data indicating the beat interval using the beat signal stored in control storage section 710 (ST901).

Autonomic nerve index output section 726 outputs autonomic nerve index data based on the beat interval data (ST902).

Body motion data output section 727 outputs data based on the body motion signal stored in control storage section 710 (ST903).

Event detection section 723 smoothes the beat interval data (ST904). Furthermore, event detection section 723 detects a periodic heart rate variability event using an algorithm for estimating a cyclic heart rate variability flag from the conventional pulse wave interval shown in FIG. 3 (ST905). Event processing section 724 calculates the moving average (mv) of the beat interval data (ST906). The event processing section 724 calculates an average value (An) for a predetermined period, for example, every hour or for one sleep period (ST907).

Event processing section 724 determines whether moving average (mv) is equal to or less than average value (An) at time tk when the periodic heart rate variability flag is detected (ST908). When the moving average (mv) exceeds the average value (An) (ST908: NO), the event processing section 724 returns to ST908 and calculates the moving average (mv) and the average value (An) of the periodic heart rate variability flag at the next time. ). When the moving average (mv) is equal to or less than the average value (An) (ST908: YES), event processing section 224 determines whether the autonomic nerve index at time tk is equal to or greater than a predetermined autonomic nerve index threshold Th (ST909). .

When the autonomic nerve index falls below the predetermined autonomic nerve index threshold Th (ST909: NO), event processing section 724 returns to ST908. When the autonomic nerve index at time tk is equal to or greater than the predetermined autonomic nerve index threshold Th (ST909: YES), event processing section 224 determines whether the amount of body movement at time tk is equal to or greater than body movement threshold Tb (ST910). . When the amount of body movement is less than body movement threshold Tb (ST910: NO), event processing section 724 returns to ST908. When the amount of body movement is equal to or greater than body movement threshold Tb (ST910: YES), event processing section 724 accumulates periodic heart rate variability flag c(k, tk) at time tk as a periodic heart rate variability flag (ST911). .

Event processing section 724 determines whether time tk has exceeded the sleep time (ST912). When the time tk has not exceeded the sleep time (ST810: NO), the process returns to ST908. When time tk exceeds the sleep time (ST912; YES), output section 725 outputs an apnea index based on the accumulated periodic heart rate variability flag (ST913). The process flow ends.

[Overview of third embodiment of sleep state estimation device]
In the first and second embodiments, the sleep state estimating device receives a sensor signal from a wristband attached to the subject's wrist and estimates the subject's sleep state. In the third embodiment, for example, a smartwatch-type sleep state estimation device in which the sleep state estimation device is incorporated in a wristband will be described.

FIG. 12 is an external perspective view showing an example of the sleep state estimation device of the third embodiment. (a) is an external perspective view from the front side, and (b) is an external perspective view from the back side.

The sleep state estimation device 9 has a body portion 91 and a wristband portion 92 . The body portion 91 has a display portion 93 on its front side and an operation button 94 on its side surface. An optical pulse wave sensor 95 is provided on the back side of the body portion 91 . An acceleration sensor 96 (not shown) is built in the body portion 91 .

[Configuration example of the sleep state estimation device of the third embodiment]
FIG. 13 is a diagram showing an example of the configuration of the main body of the sleep state estimation device 9 of the third embodiment.

The body portion 91 has a display portion 93 , an operation button 94 , an optical pulse wave sensor 95 , an acceleration sensor 96 , a control portion 97 , an input/output interface 98 and a sensor interface 99 . The display unit 93 and the operation buttons 94 are connected to the control unit 97 via the input/output interface 98 , and the optical pulse wave sensor 95 and the acceleration sensor 96 are connected to the control unit 97 via the sensor interface 99 .

The pulsation signal from the optical pulse wave sensor 95 and the body motion signal from the acceleration sensor 96 are input to the control unit 97 via the sensor interface 99, and the control unit 97 uses the pulsation signal and the body motion signal , the sleep state of the subject who wears the sleep state estimation device 9 is estimated. The apnea index output from the control unit 97 is sent via the input/output interface 98 and the respiratory index value is displayed on the display unit 93 . Further, the operation button 94 operates the operation of the control section 97 via the input/output interface 98 .

[Effect of periodic heart rate variability estimation according to the present invention]
FIG. 14 is a diagram showing the effect of periodic heart rate variability estimation according to the present invention.

The upper portion of FIG. 14 shows the pulse interval (PPI) during one sleep period of the subject, for example, 7 hours of sleep. (i) AHI at the bottom shows the actual state of apnea or hypopnea measured by a polysomnograph in a strip shape. (ii) CVHR (original) on the second row from the bottom shows a belt-like state when it is estimated that a periodic heart rate variability event has occurred using the conventional algorithm. When the pulse interval (PPI) is greater than a predetermined pulse interval threshold, the states when it is estimated that periodic heart rate fluctuations have occurred are shown in strips. Comparing (i) and (ii), it can be seen that (ii) includes many states that are not actually apnea or hypopnea.

(iii) CVHR (modified) in the third row from the bottom is the period when excluding the case where the DC component, which is the pulse wave interval average value of the pulse interval (PPI), is longer than the beat interval average threshold from (ii) heart rate variability is shown in strips. It can be seen that the estimation has been improved by excluding the portion of the period where the pulse wave interval average value is long.

In the (iv) LF/HF filter on the fourth row from the bottom, the gaps in the horizontal thick lines indicate that the body motion data is equal to or greater than the predetermined body motion threshold and the autonomic nerve index data is equal to or higher than the predetermined autonomic nerve index threshold. It is a part corresponding to the autonomic nerve index of the above part.

(v) CVHR (modified) + LF/HF filter on the fifth row from the bottom shows the effect of estimating the sleeping state of the subject according to the first embodiment of the present application. (iii) shows the periodic heart rate variability in consideration of (iv) in a belt shape. It can be seen that (v) is closer to the actual apnea or hypopnea condition of (i).

In the above description, the pulse is used as an example of the beating state of the subject's heart, but the heartbeat may be used as the beating state of the subject's heart.

The sleep state estimation device according to the present invention may be embodied by cooperation between a server device and a client device.

It should be understood by those skilled in the art that various changes, substitutions and modifications can be made thereto without departing from the spirit and scope of the present invention.

1, 6 Wristband 2, 7, 9 Sleep state estimation device 21, 71 Antenna 22, 72 Communication interface 23, 73 Storage medium interface 24, 74 Control unit 210, 710 Control storage unit 220, 720 Control processing unit 221, 721 Acquisition Units 222, 722 Respiration estimation units 223, 723 Event detection units 224, 724 Event processing units 225, 725 Output units 226, 726 Autonomic nerve index output units 727 Body motion data output units

Claims (6)

an acquisition unit that acquires beat-to-beat data of a subject in a predetermined period including a sleep period;
an autonomic nerve index output unit that outputs autonomic nerve index data indicating the activity level of the sympathetic nerve from the beat interval data;
A respiration estimator for estimating an apnea or hypopnea sleep state of the subject based on the beat interval data and the autonomic nerve index data ,
The respiration estimation unit
an event detector that detects periodic heart rate variability event data from the beat interval data;
From the periodic heart rate variability event data, a periodic heart rate corresponding to a period within a predetermined range longer than an average beat interval corresponding to the beat interval in REM sleep of the subject. Excluding fluctuation event data, and excluding periodic heart rate variability event data during a period in which the autonomic nerve index data indicating the degree of sympathetic nerve activity is less than a predetermined autonomic nerve index threshold, from the periodic heart rate variability event data. an event processing unit that
has
estimating the apnea or hypopnea sleep state of the subject based on the periodic heart rate variability event data indicating the occurrence of the apnea or hypopnea sleep state of the subject obtained by the event processing unit; , the respiration estimator;
an output unit for outputting an apnea index signal indicating an estimation result of the respiration estimation unit ;
A sleep state estimation device having further comprising a body motion data output unit that acquires the subject's body motion signal and outputs body motion data;
The sleep state estimation device according to claim 1, wherein the respiration estimation unit estimates an apnea or hypopnea sleep state of the subject based on the beat interval data, the autonomic nerve index data, and the body movement data. . The respiration estimation unit
an event detector that detects periodic heart rate variability event data from the beat interval data;
Excluding periodic heart rate variability event data of a period in which the beat interval is within a predetermined range longer than the average beat interval corresponding to the beat interval in REM sleep from the periodic heart rate variability event data. and excluding periodic heart rate variability event data in a period in which the autonomic nerve index data is less than a predetermined autonomic nerve index threshold from the periodic heart rate variability event data, and wherein the body movement data is a predetermined body movement threshold. an event processing unit that excludes periodic heart rate variability event data for a period less than
The sleep state estimation device according to claim 2, further comprising: 3. The sleep state estimation device according to claim 1, wherein the respiration estimation unit estimates an apnea or hypopnea sleep state of the subject based on data of a plurality of subjects. A method of operating a sleep state estimation device having a computer, comprising:
an acquisition step in which the computer acquires beat-to-beat data of the subject for a predetermined period including sleep periods;
an output step in which the computer outputs autonomic nerve index data indicating sympathetic nerve activity from the beat interval data;
an estimation step in which the computer estimates an apnea or hypopnea sleep state of the subject based on the beat interval data and the autonomic nerve index data,
The estimation step includes
an event detection step of detecting periodic heart rate variability event data from the beat interval data;
From the periodic heart rate variability event data, a periodic heart rate corresponding to a period within a predetermined range longer than an average beat interval corresponding to the beat interval in REM sleep of the subject. Excluding fluctuation event data, and excluding periodic heart rate variability event data during a period in which the autonomic nerve index data indicating the degree of sympathetic nerve activity is less than a predetermined autonomic nerve index threshold, from the periodic heart rate variability event data. an event processing step for
has
estimating the apnea or hypopnea sleep state of the subject based on the periodic heart rate variability event data indicating the occurrence of the apnea or hypopnea sleep state of the subject obtained by the event processing step; ,
the estimation step;
an output step in which the computer outputs an apnea index signal indicating an estimation result in the estimation step ;
A sleep state estimation method comprising: A program that operates a computer,
an acquiring step of acquiring beat-to-beat data of the subject for a predetermined period including a sleeping period;
an output step of outputting autonomic nerve index data indicating the degree of sympathetic nerve activity from the beat interval data;
an estimation step of estimating an apnea or hypopnea sleep state of the subject based on the beat interval data and the autonomic nerve index data,
The estimation step includes
an event detection step of detecting periodic heart rate variability event data from the beat interval data;
From the periodic heart rate variability event data, a periodic heart rate corresponding to a period within a predetermined range longer than an average beat interval corresponding to the beat interval in REM sleep of the subject. Excluding fluctuation event data, and excluding periodic heart rate variability event data during a period in which the autonomic nerve index data indicating the degree of sympathetic nerve activity is less than a predetermined autonomic nerve index threshold, from the periodic heart rate variability event data. an event processing step for
has
estimating the apnea or hypopnea sleep state of the subject based on the periodic heart rate variability event data indicating the occurrence of the apnea or hypopnea sleep state of the subject obtained by the event processing step; ,
the estimation step;
an output step of outputting an apnea index signal indicating an estimation result in the estimation step ;
A sleep state estimation method comprising:
A sleep state estimation program that causes a computer to execute

Images