WO2004107978A1

WO2004107978A1 - Sleep stage judgment method and judgment device

Info

Publication number: WO2004107978A1
Application number: PCT/JP2003/016745
Authority: WO
Inventors: Shin Nemoto
Original assignee: Cb System Co.
Priority date: 2003-06-03
Filing date: 2003-12-25
Publication date: 2004-12-16
Also published as: JP4461388B2; JPWO2004107978A1; AU2003296122A1

Abstract

From a biological signal detected by a nondestructive sensor, a breath signal, a heart beat signal, and a biological signal intensity are detected. A signal extracted from these or a parameter calculated from this signal is used as an index value so as to provide a sleep stage judgment method and a judgment device. The method and the device are used for judging the sleep stage of an examinee who is sleeping. The device includes a nondestructive sensor arranged on a bed for detecting a biological signal, detection means for detecting hear beat, breath, a biological intensity signal, and the like from the output of the nondestructive sensor, and index value calculation means for evaluating the autonomic nerve from a power spectrum density obtained by subjecting the R-R interval signal detected from the heart beat signal to the Fourier transform and calculating the index value of the sleep stage. By using a plurality of parameters as index signals and a threshold value according to the sleep stage is calculated from data of a predetermined time for each of the index signals, thereby judging the sleep stage. Similarly, a sleep stage judgment method and judgment device using at least one of the parameters calculated from the nondestructive sensor signal as an index value are also disclosed.

Description

Sleep stage determination method and device

The present invention relates to a sleep stage judging method and judging device for judging a sleep stage from a biological signal detected by a biological signal detecting means, and is not affected by a difference in the age of a subject or a change in physical condition. The present invention also relates to a sleep stage determination method and a sleep stage determination method for clearly and clearly determining a sleep stage. Technology background

Sleep is often used as a criterion when examining an individual's health, and it is well known that sleep and health are closely related. Health and the depth and quality of sleep at night are closely related to the mood and morale of the next day, while the pattern of sleep depth and sleep stage when mental stress or physical condition is poor Changes occur and comfortable sleep cannot be obtained.

In healthy sleep, the non-REM sleep phase and the REM sleep phase appear repeatedly at regular intervals after falling asleep, but the rhythm may be disturbed when the patient is sick or mentally stressed. It has been known. Therefore, by monitoring the sleep stages during sleep at night and their occurrence patterns, it becomes possible to know the mental stress and poor physical condition of the subject.

In particular, many elderly people complain of poor sleep, such as light sleep, and the quality of sleep is a problem. In order to know the quality of sleep, it is possible to find corrective measures / measures that can be improved by knowing the transition of sleep stages.

As a conventional method of knowing the sleep stage, a method using a sleep polysomnograph (PSG) is generally used. With the method using PSG, the brain during sleep, the activity of the menstrual system can be estimated from EEG, surface EMG, eye movements, etc., and much information about sleep can be obtained. It is difficult to get a natural sleep because the measurement is carried out with the electrodes attached. It costs. Therefore, the physical and physical burden on the subject is very large, and furthermore, this measurement must be performed by a hospital or other specialized facility and an expert in handling. Therefore, the cost required for this is large.

For this reason, it is difficult to use PSG for daily health care, even if it can be an effective treatment for patients with obvious sleep disorders.

Therefore, there is a high demand for a method of easily knowing the sleep stage without using PSG to know the subject's daily health status. In order to determine the sleep stage, several methods have been proposed to change the judgment criteria depending on age, differences in the autonomic nervous component, etc. There is a problem that the accuracy may vary due to variation and has not been put to practical use.

Therefore, an object of the present invention is to provide a sleep stage determination method and a determination device capable of determining a stable sleep stage irrespective of a difference due to a subject's age and a change in physical condition.

In addition, it is easy to handle, can be used on a daily basis in terms of price and maintenance costs, and can be used to determine the subject's sleep stage without placing any physical or mental burden on the subject. The purpose is to provide the equipment. Disclosure of the invention

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned circumstances, and has been created with the object of solving these problems. At least one signal among the parameters derived from is used as an index signal, and a threshold for judging a sleep stage is calculated from data of a predetermined time for each index signal, and the sleep stage is determined using the threshold. This is a characteristic sleep stage determination method.

By doing so, the biological signal detecting means can detect minute vibrations emitted from the living body and determine the sleep stage.

In claim 1, at least one of the signals obtained by adding the signal strength values of the signals detected by the biological signal detecting means can be used as the index signal.

3. The method according to claim 2, wherein the determination of each of the sleep stages includes determining a small number of the index signals. At least it is possible to make a determination by using the logical product of the determination information of two or more index signals, thereby improving the reliability of the determination.

In claim 3, the threshold value of the index signal can be calculated using an average value and a variance value of the index signal at a predetermined time, and the statistical value of the latest index signal is used. It is possible to obtain a threshold value corresponding to the state, and it is possible to avoid the effects of differences in age and physical condition.

In claim 3, the threshold value of the index signal may be obtained by calculating a signal of a moving average of the index signal and using an average value and a variance value of the signal.

In claim 4, the threshold value of the index signal is obtained by obtaining a signal of a difference between a long-time moving average and a short-time moving average of the index signal, and calculating using a mean value and a variance of the difference signal. By calculating the moving average in this way, high-frequency components of the index signal and long-term level fluctuations can be eliminated, and the result of the signal analysis processing becomes stable.

In claim 3, one of the index signals can be a signal obtained as a reciprocal of a coefficient obtained by performing gain control on a signal detected by the biological signal detecting means.

In claim 3, one of the index signals is an RR interval value of a heartbeat signal. In claim 3, one of the index signals is a peak value interval value of a respiratory signal. 4. The signal according to claim 3, wherein one of the index signals is an integrated value (LF value) of a maximum value band appearing in a lower frequency band in the power spectrum density of the heartbeat signal. It can be an integrated value (HF value) signal of the maximum value band.

In claim 3, one of the index signals may be a signal having a ratio between an LF value and an HF value in a power spectrum density of a heartbeat signal.

In claim 3, one of the index signals is a ratio of a ratio of an LF value signal in a power spectrum density signal obtained from an R-R interval value of a heartbeat signal to a signal of a sum of the LF value signal and the HF value signal. It can be a signal.

In claim 3, one of the index signals may be a logarithmic value of a ratio between an LF value and an HF value in the power spectrum density of the heartbeat signal. These index signals are public parameters for judging the sleep stage, and by using them, the reliability of the sleep stage can be improved.

In claim 2, it is desirable that the biological signal detecting means is a non-invasive detecting means. ,

In claim 14, the biological signal detecting means may include a pressure detecting tube, a pressure detecting sensor, and a biological signal extracting means, and may extract a biological signal from a pressure change detected by the pressure detecting sensor.

In claim 2, the biological signal detecting means may be a heart rate signal detecting means such as an electrocardiograph or a pulse meter and a respiratory state detecting means for detecting a respiratory rate or a respiratory state. Further, the present invention provides a heartbeat signal or a respiratory signal extracted from a signal detected by the biological signal detecting means, and at least one of the parameters derived from these signals as an index signal. A sleep stage determining apparatus comprising: calculating a threshold for determining a sleep stage from data of a predetermined period of time; and determining means for determining a sleep stage using the threshold.

In claim 17, at least one of the signals obtained by adding the signal strength values of the signals detected by the biological signal detecting means can be used as the index signal.

In claim 18, in each of the sleep stages, the determination can be made using a logical product of determination information of at least two or more of the index signals.

In claim 19, the threshold value of the index signal can be calculated using an average value and a variance of the index signal for a predetermined time.

In claim 19, the threshold value of the index signal can be calculated using a moving average signal of the index signal and using an average value and a variance value of the signal.

20. The threshold value of the index signal according to claim 20, wherein a signal of a difference between a long-term moving average and a short-time moving average of the index signal is obtained, and the threshold value is calculated using an average value and a variance value of the difference signal. be able to.

In claim 19, one of the index signals is a signal obtained as a reciprocal of a coefficient obtained by performing gain control on a signal detected by the biological signal detecting means. In claim 19, one of the index signals is an RR interval value of a heartbeat signal. In claim 19, one of the index signals is a peak value interval value of a respiratory signal.

20. The method according to claim 19, wherein one of the index signals is an integrated value (LF value) of a maximum value band that appears in a lower frequency region in the power spectrum density of the heartbeat signal and a signal that appears in a higher or higher frequency region. It can be an integrated value (HF value) signal of the maximum value band.

In claim 19, one of the index signals can be a signal having a ratio between an LF value and an HF value in the power spectrum density of the heartbeat signal.

In claim 19, one of the index signals is an LF value signal in a power spectrum density signal obtained from an R_R interval value of a heartbeat signal, and a signal of a sum of the LF value signal and the HF value signal. It can be a ratio signal.

In claim 19, one of the index signals can be a logarithmic value of a ratio between an LF value and an HF value in the power spectrum density of the heartbeat signal.

In claim 18, it is desirable that the biological signal detecting means is a non-invasive detecting means.

In claim 30, the biological signal detecting means comprises a pressure detecting tube, a pressure detecting sensor, and a biological signal extracting means, and can extract a biological signal from a pressure change detected by the pressure detecting sensor.

In claim 18, the biological signal detecting means may be a heart rate signal detecting means such as an electrocardiograph or a pulse meter and a respiratory state detecting means for detecting a respiratory rate or a respiratory state. BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 (A) is a block diagram showing a flow for judging a sleep stage in the sleep stage judging method of the present invention, and FIG. 1 (B) is a sectional view taken along line XX of FIG. 1 (A).

FIG. 2 is an explanatory diagram showing the power spectrum density when the sympathetic nerve is dominant. FIG. 3 is an explanatory diagram showing the power spectrum density when the parasympathetic nerve is dominant.

FIG. 4 is a flowchart for determining a sleep stage using one index signal. FIG. 5 is a flowchart for determining a sleep stage using two index signals. FIG. 6 is a flowchart showing a processing result of a parameter for determining whether to wake up from a REM sleep stage or a non-REM sleep stage.

FIGS. 7 (A) and (B) are output graphs of the index signals used for determining the awakening. REM sleep stage and the non-REM sleep stage.

FIG. 8 is a graph showing the determination results of the awake / REM sleep stage and the non-REM sleep stage.

FIGS. 9 (A) and 9 (B) are output graphs of the index signals used to determine the awake state and the REM sleep stage.

FIG. 10 is a graph showing the determination results of the awake state and the REM sleep stage. FIGS. 11 (A) and (B) are output graphs of the index signals used to determine the light non-REM sleep stage and the deep non-REM sleep stage.

FIG. 12 is a graph showing the determination results of a light non-REM sleep stage and a deep non-REM sleep stage.

FIGS. 13 (A) and (B) are graphs comparing the results of sleep determination according to the present embodiment with the results of a conventional method using a sleep polynomograph (PSG). BEST MODE FOR CARRYING OUT THE INVENTION

Embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 (A) is a block diagram showing a flow of judging a sleep stage according to the embodiment of the present invention, and FIG. 1 (B) is a block diagram of the biological signal detecting means 1 shown in the block diagram. It is the side view seen from the arrow direction shown by a section.

The noninvasive sensor 1 shown in FIG. 1 (A) is a biological signal detecting means for detecting a minute biological signal of a sleeping subject, and from this biological signal, a heartbeat signal detecting means 2 and a call P and signal detecting means. At 7, the respiratory signal and the heartbeat signal are extracted via a filter or the like.

The noninvasive sensor 1 includes a pressure sensor 1a and a pressure detection tube 1b. The pressure sensor 1a is a sensor that detects minute pressure fluctuations. In the present embodiment, a low-frequency condenser microphone type is used. Instead, it just needs to have the appropriate performance and dynamic range. The condenser microphone for low frequency used in the present embodiment is replaced by a general microphone microphone for sound, which is not considered for the low frequency region, and is provided with a chamber behind the pressure receiving surface. This is a type in which the characteristics in the frequency domain are greatly improved, and is suitable for detecting minute pressure fluctuations in the pressure detection tube 1b. It is also excellent for measuring minute differential pressure, has a resolution of 0.2 Pa and a dynamic range of about 50 Pa, and is compared with a micro-differential pressure sensor using ceramic that is usually used. It is suitable for detecting a minute pressure applied to the pressure detection tube 1b when a biological signal passes through the body surface. Or frequency characteristic 0. Ί H _Z 2 0 H showed almost flat output value between the z, it is suitable for detecting minute biosignal beauty respiratory rate, etc. heartbeat Oyo.

The pressure detection tube 1b has an appropriate elasticity so that the internal pressure fluctuates according to the pressure fluctuation range of the biological signal. In addition, it is necessary to appropriately select the volume of the hollow portion of the tube in order to transmit the pressure change to the fine differential pressure sensor 1a at an appropriate response speed. If the pressure detection tube 1b cannot satisfy both the appropriate elasticity and the hollow volume at the same time, load a core wire of appropriate thickness into the hollow portion of the pressure detection tube 1b over the entire length of the tube, and The volume can be appropriately taken.

The pressure detection tube 1b is placed on a hard sheet 16 laid on a bed 15 and an elastic cushion sheet 17 is laid thereon, and a subject is placed on the pressure detection tube 1b. Ka yellowish. In addition, the pressure detection tube 1 b may be configured to be incorporated in the cushion sheet 17 or the like so that the position of the pressure detection tube 1 b is stabilized.

In the present embodiment, two sets of non-invasive sensors 1 are provided, one of which detects a biological signal of the subject's thoracic region, and the other detects a subject's buttock region, so that the subject goes to sleep. It is configured to detect a biological signal regardless of the posture.

The biological signal detected by the non-invasive sensor 1 is a signal in which various vibrations emitted from the human body are mixed, and includes signals such as a heartbeat signal, a respiratory signal, and a turn signal. Therefore, a biological signal such as a heartbeat signal or a respiratory signal is extracted by a biological signal detecting means using a filter or a means such as a gun gauge processing. Needless to turn over Other signals can be detected.

In the present embodiment, the heart rate signal is extracted from the detection signal of the non-invasive sensor 1, but the present invention is not limited to this. For example, the heart rate signal may be obtained by wearing a dedicated heart rate monitor or detecting a pulse. It is possible to obtain information on breathing or turning over by using a microphone or imaging means.

The heart rate is detected by the heart rate detection means 3 from the heart rate signal detected by the biological signal detection means 1 and the interval between adjacent peaks of the R wave of the heart rate signal by the R-R interval signal calculation means 4, ie, the R-R interval Detect signal.

The RR interval signal described above is a signal that uses the interval of the waveform (R wave) near the peak of the intensity of the heartbeat signal as a variable, and is often used for heart rate variability analysis. The RR interval signal detected by the RR interval signal calculating means 4 is sent to a power spectrum density calculating means 5.

Here, the power spectrum density of the RR interval signal will be described.

FIG. 2 shows the power spectrum density when the sympathetic nerve is dominant, and FIG. 3 shows the power spectrum density when the parasympathetic nerve is dominant. As can be seen, the power spectrum density varies depending on the state of the autonomic nervous system.

That is, a remarkable maximum value appears in a band of about 0.05 to 0.15 Hz and a band of about 0.2 to 0.4 Hz. Here, the integrated signal of the local maximum value band appearing in the lower frequency range of approximately 0.05 to 0.15 Hz is called the LF value signal, and the local maximum signal appears in the higher frequency region of approximately 0.2 to 0.4 Hz. The integrated signal in the value band is called the HF value signal. A large LF value and a small HF value indicate that the sympathetic nerve is active and nervous, while a small LF value and a large HF value indicate that the parasympathetic nerve is active. I have.

During sleep, the heart rate decreases, due to a decrease in the sympathetic activity that is active during tension and an increase in the parasympathetic activity that is active during relaxation. That is, the values of HF and LF vary significantly depending on the state of sleep depth.

The HFZLF detecting means 6 is a means for detecting the above HF and LF values from the power spectrum density, and includes the HF value and the HF value detected by the HF / LF detecting means 6. LF values fluctuate according to sleep stages. This data is sent to the judgment parameter generating means 13 for judging the sleep stage.

The respiratory signal detecting means 7 is a means for extracting a respiratory signal from a signal detected from the biological signal detecting means, detects a respiratory rate by the respiratory rate detecting means 8, and calculates a respiratory interval by the respiratory interval signal calculating means 9. The value is used as a breathing interval value signal.

Both respiration and signals are signals that are significantly affected by the autonomic nervous system, the sympathetic nervous system and the parasympathetic nervous system, and are closely correlated with sleep stages.

Next, the processing of the signal strength of the biological signal will be described.

The signal amplification and shaping section 10 sets the characteristics of the amplifier circuit so as to amplify only the main frequency band of the biological signal and reduce the frequency band corresponding to the other noises, and further includes a bandpass filter. Further, the noise may be reduced. The automatic gain control unit 11 is a so-called AGC circuit that automatically performs gain control so that the output of the signal control shaping unit 10 falls within a predetermined signal level range. Output to the strength calculation unit 12. For gain control, for example, set the gain so that the amplitude of the output signal decreases when the peak value of the signal exceeds the upper threshold, and set the gain so that the amplitude increases when the peak value falls below the lower threshold. are doing.

The signal strength calculator 12 calculates the signal strength from the gain control coefficient applied to the biological signal in the automatic gain controller 11. The gain value obtained from the AGC circuit described above is set to be small when the signal size is large, and is set to be large when the signal size is small. In order to indicate, it is better to set a function indicating the signal strength so as to be inversely proportional to the value of the gain. On the other hand, when the output value of the non-invasive sensor 1 continuously exceeds the upper limit of the automatic gain control within a predetermined time, it can be determined that a body movement such as turning over has occurred.

As described above, the intensity of the biological signal including the body motion is considered to be closely related to the sleep state, and thus is used as a parameter for determining the sleep stage. The determination parameter generation means 13 calculates and obtains a parameter used for the determination using the heartbeat signal or the HF value signal and the LF value signal. The determination parameters include, for example, a heart rate signal, an HF value signal and an LF value signal, and an HF value signal and an LF A parameter such as a signal of the value of the ratio of the value signal is generated by calculation.

The sleep stage determination means 14 uses the parameters generated by the determination parameter generation means 13 to wake up, judge between REM sleep and non-REM sleep, judge between waking and REM sleep, and the deeper sleep stages of non-REM sleep The sleep stage is determined by determining that the sleep stage is light.

The non-REM determining means determines whether or not the force is in the non-REM sleep state. That is, when it is confirmed that the subject is not in the non-REM sleep state, it is known that the subject is in the REM sleep state or the awake state. .

The REM sleep determination means is means for determining the REM sleep state or the awake state after confirming that the state is not the non-REM state, that is, after confirming that the state is the REM sleep state or the awake state.

Non-REM sleep stages are usually categorized into four sleep stages, 1st to 4th, with the first non-REM sleep stage being the shallowest, deeper into j, and the fourth sleep stage being the deepest. It is. Here, the first and second sleep stages are assumed to be light sleep stages, and the third and fourth sleep stages are assumed to be deep sleep stages. The non-REM sleep deep / shallow determining means determines whether the sleep state is light or deep after the sleep state is confirmed.

By the above three determinations, it is possible to determine four sleep stages: awake state, REM sleep state, light non-REM sleep state, and deep non-REM sleep state.

Next, an HF value signal and an LF value signal will be described as examples of the generation of the index signal. Based on the heartbeat signal sent from the heartbeat signal detection means 2, the R-R interval signal calculation means 4 detects the R-R interval signal. Fourier expansion is performed on the detected R-R interval signal to determine the power spectral density of the R-R interval signal.

From the power spectrum density signal of the R-interval signal, the HF / LF detection means 6 detects HF and LF momentarily. Using these HF and LF, it is possible to generate sleep stage determination parameters effective for sleep stage determination.

Other parameters that can be used as the index signal will be described below.

The heart rate signal directly uses the heart rate extracted by the heart rate detecting means 3 and is affected by the sympathetic and parasympathetic changes. Respiratory rate Since the signal is also affected by changes in the sympathetic and parasympathetic nerves, it can be adopted as an index signal for determining the sleep stage. In addition, the RN LF signal is a signal that takes in the LF value as it is. RN LFR is the ratio between the LF value and the HF value. Further, RN LOG is the logarithmic value of the value indicated by LF / HF.

Next, the procedure for determining the actual sleep stage will be described.

Fig. 4 is a flowchart for judging the sleep stage using one of the index signals derived from the biological signal. One of the parameters generated by the determination parameter generating means 13 is shown. One is selected as an index function, and the sleep stage determination means 14 performs determination according to this flowchart.

Since the index signal contains many high-frequency components, that is, minute fluctuations, a moving average process for a predetermined time is performed to remove the high-frequency components. The difference between the long-term moving average and the short-term moving average is calculated in consideration of the case where the index signal fluctuates for a long time, and a pure fluctuation value of the index signal is obtained. That is, this is to correct the long-term variation of the parameter signal and extract a pure variation. The number of moving average data used for this operation is 50,000 points for the short-term moving average and 100,000 points for the long-term moving average, but is not limited to this. Selected appropriately. Set a threshold for judging the sleep stage for the adopted index signal. At this time, different thresholds are set for different sleep stages.

FIG. 5 shows an example of determining a sleep stage using two index signals. The combination to be adopted can be selected according to the sleep stage. For example, in determining which of the sleep stages is awake / REM sleep or non-REM sleep, a good determination result can be obtained by using an LF value signal as one of the index signals. However, the present invention is not limited to this, and suboptimal results can be obtained using other parameters.

Combining the two signals is a means to make the determination accuracy more reliable, and depending on the purpose of use, only one parameter may be used as the index signal.

The threshold generation for judging the sleep stage in FIG. 4 and FIG. 5 is performed as follows. Performs moving average processing of the index signal, corrects long-term fluctuations of the index signal, and The average m and the standard deviation s (variance) of the data for a predetermined time are obtained for the signal from which the trend is extracted.

Then, using this value, awakening. To determine the REM sleep state and the non-REM sleep state _{: In} order to determine the difference signal between the index signals, a binary signal is used. The threshold value is, for example, Ask.

a 'm + β

Here, m is the average value, s is the standard deviation, and] 3 is the maximum matching rate between the sleep stage judgment of this embodiment and the sleep stage judgment by PSG using the experimental data of many times. It is determined by calculating the optimum value so that ·

Further, the average plant m is not limited to the arithmetic average value, and a median value or the like may be used. On the other hand, the standard deviation s was used as a parameter indicating the variation, but a value indicating a parameter such as variance can be used as an alternative to this.

The constants for α and in equation (a) differ depending on which sleep stage the parameter used as an index is used for. For example, the values are different depending on whether the wakefulness REM sleep stage and the non-REM sleep stage are used, or the case where it is used for determining light non-REM sleep and deep non-REM sleep. '' In the sleep stage determination method and the determination apparatus of the present invention, the threshold value according to the state of the subject at the time of the test is used because the average value m and the standard deviation s of the parameters are used to determine the threshold value of the parameter used for the determination. As a result, a judgment can be made that is not affected by individual differences, age differences, or the state of the subject at the time of the test.

Figure 6 shows the procedure for determining each sleep stage and determining the sleep stage. The sleep stage is determined by the sleep stage determination unit 14 using the parameters generated by the determination parameter generation unit 13 to wake up. The determination between REM sleep and non-REM sleep, the determination between awakening and REM sleep, and the non-REM sleep The sleep stage is determined by determining the deep sleep stage and the light sleep stage. That is, by performing the above three types of determinations, the power determination belonging to any sleep stage is performed. The procedure first determines whether or not it is in the non-REM sleep stage. At this time, if it is determined that the subject is in the non-REM sleep, it is determined whether the non-REM sleep is a light non-REM sleep stage or a deep non-REM sleep stage. If it is determined that the patient is not in non-REM sleep, it is determined whether the subject is in REM sleep or awake. With the above three determination steps, it is possible to determine the power at which the sleep stage at each time point corresponds to any of the four stages of the awake stage, the REM sleep stage, the light non-REM sleep stage, and the deep sleep stage.

FIG. 7 shows the measurement results of the RNL F signal and the RN LOG signal as signals for determining whether the subject is in the awake / REM sleep state or in the non-REM sleep state. Fig. 7 (A) shows the short-term and long-term moving averages of the RNLF signal, and then the difference processing data. Fig. 7 (B) shows the short-term and long-term moving averages of the NLOG signal. After that, the signal subjected to the difference processing is shown. In addition, the threshold value calculated by the equation (a) is also displayed at the same time, and the threshold value is used to perform the binary threshold. As described with reference to the flow chart of FIG. 6, the power judgment of the awakening or REM sleep state or the non-REM sleep state is performed in the awake and REM sleep states if both signals are equal to or more than the respective thresholds. Judge that there is.

FIG. 8 shows the results of the force determination of awake / REM sleep or non-REM sleep according to the above procedure, together with the results of a conventional method using a sleep polysomnograph (PSG). If the waveform indicates a high position, it indicates awakening. REM sleep state, and if the waveform indicates a low position, it indicates non-REM sleep. FIG. 9 shows the measurement results of the RN signal signal and a parameter for determining whether the subject is in the awake state or in REM sleep. Fig. 9 (A) shows the moving average processed data of the RNLF signal, and Fig. 9 (B) shows the moving average processed signal of the RNLFR signal. In addition, the threshold value calculated by equation (a) is also displayed at the same time, and is binarized by this threshold value. As described with reference to the flowchart of FIG. 4, after binarization using this threshold value, the logical product of the two signals is taken to determine whether the subject is in the awake state or in REM sleep. FIG. 10 shows the results of determining whether the subject is in the awake state or in REM sleep according to the above procedure.

Fig. 11 shows the measurement results of the RNLF signal and the RNLF signal, the parameters for determining whether the sleep is light non-REM sleep or strong non-REM sleep. Fig. 11

(A) shows the short-term and long-term moving averages of the RNL F signal and the difference processing, and Fig. 11 (B) shows the short-term and long-term moving averages of the RNLOG signal. The data subjected to the difference processing later is shown. Also, it is calculated by the formula (A) The displayed threshold is also displayed at the same time. As described with reference to the flowchart of FIG. 6, after binarization using this threshold value, the logical product of the two signals is taken to determine whether the subject is in the awake state or in REM sleep. FIG. 12 shows the result of a force determination that is a strong non-REM sleep that is a shallow non-REM sleep according to the above procedure. FIG. 13 is a graph showing the transition of the sleep stage by integrating the above-described three stages of determination using the sleep stage determination method and the determination device of the present embodiment. At the same time, the results are shown together with the results of the conventional method using sleep polysomnograph (PSG).

By the way, the sleep stage determination method and the determination device of the present invention are a determination method and a determination device based on the behavior of the autonomic nervous system, and the conventional determination method and the determination device using the conventional sleep polysomnograph (PSG) are mainly used for the electroencephalogram. A judgment method and a judgment device using behavior. The electroencephalogram determines the sleep stage by utilizing the characteristic waveform of the neural activity of the cerebral cortex corresponding to various sleep stages according to the degree of synchronization. On the other hand, the sleep stage judging method and the judging device of this method judge the sleep depth from the behavior of the autonomic nervous system, which is the neural activity of the brainstem, and the behavior of the autonomic nervous system has a profound effect on heart rate and respiration. Have given. As a result, there is a slight difference in time between the conventional sleep determination mainly using brain waves and the sleep determination from the autonomic nervous system. Therefore, it should be noted that the coincidence rate of the judgment results described below is not the temporal coincidence rate but the coincidence rate of the ratio of each stage of sleep depth per day.

Fig. 13 (A) shows the results of a comparison between the three stages of awake state, REM sleep stage, and non-REM sleep stage, and the judgment result was 93.3% consistent with the conventional method using sleep polysomnograph (PSG). Shows the rate. Fig. 13 (B) shows the results of a comparison between four stages: awake state, REM sleep stage, light non-REM sleep stage, and deep non-REM sleep stage. And 90.3% match rate. From this, the sleep stage determination device of the present invention has determination accuracy that is practically acceptable.

In the present embodiment, a method of detecting pressure fluctuation by combining a tube and a pressure sensor is employed as a detecting means capable of detecting a biological signal of a subject noninvasively, but the present invention is not limited to this, and a fine biological signal is detected. Any detection means that can be used. For example, even if the detection means is attached to a part of the body such as an electrocardiograph or a pulse rate monitor, If not, it is possible to use. Industrial applicability

It is well known that knowing the state and quality of sleep is useful for knowing the daily health of a subject, but it is easy to use a sleep stage determination method and method that can be easily used for personal health management purposes. At present, no judging device is found.

The sleep stage judging method and judging device of the present invention detects a heartbeat signal by using an appropriate detecting means, and determines the sleep stage of the subject by arithmetically processing the output of the heartbeat signal. In particular, the power spectrum density obtained from the R-R interval signal detected from the heartbeat signal is a good index indicating the state of the autonomic nervous system. Have reliability.

In addition, the sleep stage determination method and determination device of the present invention can determine the sleep stage as long as a heartbeat signal can be detected, and the cost of using and maintaining the device is low. It is possible to provide a sleep stage judging device suitable for daily use. ,

Claims

The scope of the claims

Of the heartbeat signal or respiratory signal extracted from the signal detected by the biological signal detection means, and at least one of the parameters derived from these signals, an at least one signal is used as an index signal. And calculating a sleep stage using the threshold value.

2. The sleep stage determination method according to claim 1, wherein at least one of the signals obtained by adding the signal strength values of the signals detected by the biological signal detection unit is used as an index signal.

The sleep stage determination method according to claim 2, wherein the determination of each stage of the sleep stage is performed using a logical product of determination information of at least two or more of the index signals.

4. The sleep stage judging method according to claim 3, wherein the threshold value of the index signal is calculated using an average value and a variance value of the index signal for a predetermined time.

4. The sleep stage determination method according to claim 3, wherein the threshold value of the index signal is obtained by obtaining a signal of a moving average of the index signal and using an average value and a variance value of the signal.

The threshold value of the index signal is obtained by obtaining a signal of a difference between a long-term moving average and a short-time moving average of the index signal, and calculating using a mean value and a variance of the difference signal.

5. The method for determining a sleep stage according to claim 4, wherein:

4. The sleep stage determination method according to claim 3, wherein one of the index signals is a signal obtained as a reciprocal of a coefficient obtained by performing gain control on a signal detected by a biological signal detection unit. .

4. The sleep stage judging method according to claim 3, wherein one of the index signals is an R-R interval value of a heartbeat signal.

4. The sleep stage determination method according to claim 3, wherein one of the index signals is a peak value interval value of a respiratory signal.

0. One of the index signals is an integral value (LF value) signal of the maximum value band that appears on the lower frequency side in the power spectrum density of the heartbeat signal and Z or high. 4. The sleep stage judging method according to claim 3, wherein the signal is an integrated value (HF value) signal of a maximum value band that appears in a frequency range that is not high.

4. The method for determining a sleep stage according to claim 3, wherein one of the index signals is a signal having a ratio between an LF value and an HF value in a power spectrum density of a heartbeat signal.

One of the index signals is a signal having a ratio of an LF value signal in the power spectrum density signal obtained from the R-R interval value of the heartbeat signal and a signal of the sum of the LF value signal and the HF value signal. 4. The method for determining a sleep stage according to claim 3, wherein: 4. The sleep stage determination method according to claim 3, wherein one of the index signals is a logarithmic value of a ratio between an LF value and an HF value in a power spectrum density of a heartbeat signal. ·

The sleep stage judging method according to claim 2, wherein the biological signal detecting means is a non-invasive detecting means.

The sleep according to claim 14, wherein the biological signal detecting means includes a pressure detecting tube, a pressure detecting sensor, and a biological signal extracting means, and extracts a biological signal from a pressure fluctuation detected by the pressure detecting sensor. Stage determination method.

The sleep stage judging method according to claim 2, wherein said biological signal detecting means is a heartbeat signal detecting means such as an electrocardiograph or a pulse meter and a respiratory state detecting means for detecting a respiratory rate or a respiratory state. .

At least one of the heartbeat signal or respiratory signal extracted from the signals detected by the biological signal detection means and the parameters derived from these signals is used as an index signal, and each index signal is taken from data for a predetermined time to sleep. A sleep stage determination device, comprising: calculating a threshold for determining a stage; and determining means for determining a sleep stage using the threshold.

18. The sleep stage judging device according to claim 17, wherein at least one of the signals obtained by adding the signal strength values of the signals detected by the biological signal detecting means is used as an index signal.

The determination of each of the sleep stages is performed by using a logical product of determination information of at least two or more index signals among the index signals. Item 18. The sleep stage judging device according to Item ¹⁸ .

20. The sleep stage judging device according to claim 19, wherein the threshold value of the index signal is calculated using an average value and a variance value of the index signal for a predetermined time.

21. The sleep stage judging device according to claim 19, wherein the threshold value of the index signal is obtained by calculating a signal of a moving average of the index signal and using an average value and a variance value of the signal.

22. The threshold value of the index signal is obtained by obtaining a signal of a difference between a long-time moving average and a short-time moving average of the index signal, and calculating using a mean value and a variance value of the difference signal. 21. The sleep stage determination device according to claim 20.

23. The sleep according to claim 19, wherein one of the index signals is a signal obtained as a reciprocal of a coefficient obtained by performing gain control on a signal detected by a biological signal detection unit. Stage determination device.

24. The sleep stage judging device according to claim 19, wherein one of the index signals is an R-R interval value of a heartbeat signal.

25. The sleep stage judging device according to claim 19, wherein one of the index signals is a peak value interval value of a respiratory signal.

26. One of the index signals is the approximate power spectral density of the heartbeat signal

Maximum value in the band of 0. 15H _z and (LF value) signal, substantially from 0.2 to 0.4

A maximum value (HF value) signal in a frequency band of Hz.

19. The sleep stage determination device according to 19.

27. The sleep stage judging device according to claim 19, wherein one of the index signals is a signal of a ratio between an LF value and an HF value in a power spectrum density of a heartbeat signal.

28. One of the index signals is a signal having a ratio of an LF value signal in a power spectrum density signal obtained from an R-R interval value of a heartbeat signal and a signal of a sum of the LF value signal and the HF value signal. 20. The sleep stage judging device according to claim 19, wherein there is a sleep stage judging device.

29. The sleep stage judging device according to claim 19, wherein one of the index signals is a logarithmic value of a ratio between an LF value and a HF value in a power spectrum density of a heartbeat signal.

19. The sleep stage judging device according to claim 18, wherein the biological signal detecting means is a non-invasive detecting means.

The biological signal detection means according to claim 30, wherein said biological signal detection means comprises a pressure detection tube, a pressure detection sensor, and a biological signal extraction means, and extracts a biological signal from a pressure fluctuation detected by the pressure detection sensor. Sleep stage determination device.

The sleep stage determination according to claim 18, wherein said biological signal detecting means is a heart rate signal detecting means such as an electrocardiograph or a pulse meter and a respiratory state detecting means for detecting a respiratory rate or a respiratory state. apparatus.