JP2014193263A - Biological state estimation device, and computer program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for more accurately comprehending a biological state.SOLUTION: The present invention estimates a biological state, by using a predetermined central nervous system index value obtained from a central nervous system biological signal information and a predetermined peripheral nervous system index value obtained from a peripheral nervous system biological signal information, from a positional relationship between coordinate points on a coordinate system in which those indexes are set as coordinate axes. When performing state estimation of a person, a state of the person can be more accurately estimated by using both indexes as the present invention to have a configuration for determining the state of the person on a single coordinate system, so that each of variations of the indexes with respect to the state change of the person is emphasized, than using any one of indexes.

Description

  The present invention relates to a biological state estimation device and a computer program that estimate a human state using a biological signal.

  The present applicant, in Patent Document 1, obtains a time series waveform of a frequency from a time series waveform of a biological signal that is mainly a wave of the cardiovascular system collected from a human upper body, and further, a time series waveform of a frequency gradient, An apparatus having a procedure for obtaining time-series waveforms of frequency fluctuations and performing frequency analysis on these waveforms is disclosed. In the frequency analysis, the power spectrum of each frequency corresponding to a predetermined function adjustment signal, fatigue acceptance signal, and activity adjustment signal is obtained. And a person's state is judged from the time series change of each power spectrum.

  In addition, in the patent document 2, the present applicant uses a time-series waveform of a biological signal collected from a human upper body, obtains a regression line in the relationship between logarithmic power spectral density and logarithmic frequency obtained by frequency analysis, A technique for more accurately determining the state of a person from the shape of the regression line is disclosed.

JP 2011-167362 A JP 2012-179202 A

  All of the above-described techniques detect and analyze vibration generated on the body surface of the back in the upper body of a person via an air cushion. A pulse wave (Aortic Pulse Wave (APW)), which is a vibration generated on the body surface of the back, is a pressure vibration generated from the motion of the heart and the aorta, and information on the ventricular systole and diastole. And the elasticity information of the blood vessel wall which becomes an auxiliary pump for circulation. The signal waveform associated with heart rate variability includes sympathetic and parasympathetic nervous system activity information (parasympathetic activity information including the compensation of sympathetic nerves), and the signal waveform associated with aortic oscillation is sympathetic. Contains information on neural activity.

  On the other hand, for example, as shown in Patent Document 2 described above, not only the back body surface pulse wave is analyzed, but also the fingertip volume pulse wave is subjected to frequency analysis, and each of the back body surface pulse wave and the fingertip volume pulse wave is analyzed. There is also known a method for determining the state of a person by comparing analysis results. It is normal for the dorsal body surface pulse wave, which is the biological signal information of the central system that captures the characteristics of the heart movement, to correspond to the volume pulse wave of the fingertip that captures the biological signal information of the peripheral system with a predetermined time delay. However, in the case of a heart disease patient or the like, since a characteristic characteristic of heart disease may be captured only by the back body surface pulse wave, the degree of divergence between the two in the analysis result is compared.

  However, the analysis results of the back body surface pulse wave and fingertip volume pulse wave are compared with each other, and the analysis results of both are similar when the subject is in a healthy state, and the subject has some disease near the heart. Discloses that divergence occurs. In other words, regardless of the basic health status of a person, whether or not they are in a state of low alertness or attention (a low consciousness state) When estimating the state change, all of the conventional methods are performed using only the back body surface pulse wave or only the fingertip volume pulse wave, and the both are not estimated together.

  The present invention has been made in view of the above-described points, and is capable of estimating a hypoxia state and a sleep onset symptom more accurately than before by considering the central biological signal information and the peripheral biological signal information together. It is an object of the present invention to provide a biological state estimation device and a computer program capable of performing the above.

  In order to solve the above problems, the biological state estimating apparatus of the present invention includes a central biological signal processing means for obtaining a predetermined central system index value from the central biological signal information, and a predetermined peripheral system index value from the peripheral biological signal information. A biological signal processing means for obtaining a peripheral system, and a living body from a positional relationship of a plurality of coordinate points plotted on a coordinate system using the central system index value and the index value obtained from the peripheral system index value as at least one coordinate axis. And a state estimating means for estimating the state.

  The state estimation means determines whether each coordinate point is in a predetermined dispersion state, or is aligned substantially on a straight line, or has a singular point as a positional relationship between the plurality of coordinate points. It is preferable that the biological state is estimated in consideration of the above. The state estimating means uses, as the central system index value and the peripheral system index value, each frequency at the same time in the original waveform of the central system biosignal information and the peripheral formation body signal information, and the central system index value and the peripheral system index value A configuration in which the coordinate points of corresponding index values are plotted on a coordinate system on a coordinate system in which peripheral index values are taken on the respective coordinate axes, and the biological state is estimated from the positional relationship of the plurality of plotted coordinate points It is preferable that

  In the inclination time series waveform obtained by performing slide calculation on the original waveform of the central system biosignal information and the peripheral formation body signal information under a predetermined condition as the central system index value and the peripheral system index value. Using each amplitude value at the same time, plotting the central system index value and the peripheral system index value on the coordinate system on each coordinate axis, plotting the coordinate points between the corresponding index values on the coordinate system, and plotting It is preferable that the biological state is estimated from the positional relationship between the plurality of coordinate points. In the inclination time series waveform obtained by performing slide calculation on the original waveform of the central system biosignal information and the peripheral formation body signal information under a predetermined condition as the central system index value and the peripheral system index value. The corresponding index value is plotted on the coordinate system using the ratio of each amplitude value at the same time on one axis and the time axis on the other axis, and the biological state is determined from the positional relationship between the plotted coordinate points. A configuration to be estimated is preferable. The central biological signal information is preferably a back body surface pulse wave collected from the back, and the peripheral biological signal information is preferably a fingertip volume pulse wave.

  The computer program of the present invention provides a computer as a biological state estimation device for estimating a biological state to a central biological signal processing procedure for obtaining a predetermined central system index value from central biological signal information and a predetermined value from peripheral biological signal information. A plurality of coordinate points plotted on a coordinate system using a peripheral biomedical signal processing procedure for obtaining a peripheral system index value and an index value obtained from the central system index value and the peripheral system index value as at least one coordinate axis And a state estimation procedure for estimating the biological state from the positional relationship.

  In the state estimation procedure, as the positional relationship of the plurality of coordinate points, whether each coordinate point is in a predetermined dispersion state, is oriented in a substantially straight line, or has a singular point. It is preferable to estimate the biological state in consideration of. As the central system index value and the peripheral system index value, it is preferable to use respective frequencies at the same time in the original waveforms of the central system biosignal information and the peripheral formation body signal information. As the central system index value and the peripheral system index value, each amplitude value at the same time in the tilt time series waveform obtained by sliding calculation of the original waveform of the central system biological signal information and the peripheral formation body signal information under a predetermined condition Is preferably used. As the central system index value and the peripheral system index value, each amplitude value at the same time in the tilt time series waveform obtained by sliding calculation of the original waveform of the central system biological signal information and the peripheral formation body signal information under a predetermined condition It is preferable to use this ratio.

  The present invention uses a predetermined central system index value obtained from the central system biosignal information and a predetermined peripheral system index value obtained from the peripheral system biosignal information, and uses these as coordinate axes on the coordinate system on the coordinate system. The biological state is estimated from the positional relationship. Negative feedback between a human sleep center and awake center is controlled by the functions of the autonomic nervous system and endocrine system. As described above, fingertip plethysmogram is conventionally known as an index for capturing the state of these controls. On the other hand, as an index that can be detected without restraining force, the back body surface pulse wave (APW) proposed by the present applicant is proposed. )It has been known. However, there is a difference in sensitivity between the central biological signal information represented by the back body surface pulse wave and the peripheral biological signal information represented by the fingertip volume pulse wave. That is, when compared with the back body surface pulse wave, the fingertip volume pulse wave is accompanied by a time delay corresponding to the pulse wave velocity, and the influence of the reflected wave in the bloodstream that reaches the fingertip Is big. Thereby, in the back body surface pulse wave (APW), in the frequency analysis of the original waveform, there is a characteristic frequency component of the 0.5th order component that is not found in the fingertip volume pulse wave. In other words, when comparing the detected waveforms of the central biological signal information and the peripheral biological signal information, the original waveform and the tilt time series waveform processed there are not exactly the same, there is a shift between them, Responsiveness of those changes to human state changes is different. Therefore, when estimating the state of a person, rather than using either one, it is possible to change the state of the person by using both indicators and determining them on one coordinate system as in the present invention. Each variation with respect to is emphasized, and the state of the person can be estimated more accurately.

FIG. 1 is a perspective view showing an example of a biological signal measuring apparatus for measuring a back body surface pulse wave used in an embodiment of the present invention. 2 is an exploded perspective view of the biological signal measuring apparatus shown in FIG. 3 is a cross-sectional view of a main part of the biological signal measuring apparatus shown in FIG. FIG. 4 is a diagram schematically showing the configuration of the biological state estimation apparatus according to one embodiment of the present invention. FIG. 5 is a diagram showing the load-deflection characteristics of the biological signal measuring apparatus shown in FIG. FIG. 6 is a diagram in which the vertical axis of the graph of FIG. 5 is converted into a spring constant. FIG. 7 shows the relationship between the electrocardiogram, heart sound diagram, APW (original waveform), APW second-order differential waveform, fingertip volume pulse wave (original waveform), and second-order differential waveform of fingertip volume pulse wave obtained in Experimental Example 1. It is the figure which compared the time phase of the event. FIG. 8A is a diagram comparing related events in the arousal state (stable state (1)) in the data of the subject A obtained in Experimental Example 1, and FIG. 8B is an experimental example. It is the figure which compared the related event in the sleep state (stable state (2)) in the data of the subject A obtained in. FIG. 9A is a diagram comparing the related events in the sleep onset sign phenomenon expression state (transition state (1)) in the data of the subject A obtained in Experimental Example 1, and FIG. It is the figure which compared the related event in a hypothetical state (transition state (2)) in the data of the subject A obtained in the experimental example. FIG. 10A is a diagram comparing the related events in the sympathetic nerve enhanced state in the data of the subject A obtained in Experimental Example 1, and FIG. 10B is obtained in the experimental example. It is the figure which compared the related event in the parasympathetic nerve enhancement state in the data of the subject A. FIGS. 11A to 11F show the six states of the subject H obtained by adopting the frequency fluctuation values as the central system index value obtained from the APW and the peripheral system index value obtained from the fingertip volume pulse wave. It is the figure which showed another analysis result. FIGS. 12A to 12F show the subject H obtained by adopting the amplitude value of the time series waveform of the frequency gradient as the central system index value obtained from the APW and the peripheral system index value obtained from the fingertip volume pulse wave. It is the figure which showed the analysis result according to these six states. FIGS. 13A to 13F are diagrams showing the analysis results for each of the six states of the subject H obtained using the ratio of the amplitude values in FIG. FIGS. 14A to 14D are diagrams showing statistical processing of the determination using FIGS. 11 and 12 and the determination of a sleep onset symptom phenomenon, a wakefulness state, and a relaxed wakefulness state based on a medical index. . FIG. 15A is a diagram illustrating the analysis result of the subject H who has accepted sleep desire in the sleep experiment and has reached sleep, and FIG. 15B illustrates the analysis result of the hypothetical state of the subject B. FIG. 15C is a diagram showing an analysis result of a case where the subject C resisted sleep. FIG. 16A is a diagram showing the frequency analysis results of the fingertip plethysmogram and the APW original waveforms of 11 subjects in Experimental Example 1, and FIGS. It is the figure which showed the frequency analysis result of the fingertip volume pulse wave according to the presence or absence of sleepiness and the original waveform of APW. FIG. 17 is a diagram showing a power value gradient time series waveform of the fingertip volume pulse wave of the subject F in Experimental Example 2. FIG. 18 is a diagram showing a frequency gradient time-series waveform by the zero cross detection method of the APW of subject F in Experimental Example 2. FIG. 19A is a diagram showing a tilt time series waveform of the fingertip volume pulse wave for each state of the subject F, and FIG. 19B is a slope time series of the APW for each state of the subject F. It is the figure which showed the waveform. FIG. 20 is a diagram illustrating the analysis results of the fingertip volume pulse wave and APW frequency of the subject F. FIGS. 21A to 21F are diagrams showing analysis results of the subject F using the central system index value obtained from the APW and the peripheral system index value obtained from the fingertip volume pulse wave. 22 (a) to 22 (f) are diagrams showing the analysis results of subjects who had a high feeling of fatigue immediately after the start of the experiment, slept for a moment in the middle of the experiment, then drowsed and shifted to a re-wake state. 23 (a) to (d) show how the fingertip plethysmogram and APW slope time-series waveforms in FIG. 22 change, and the states of wakefulness, emergence of sleep onset, and wakefulness. It is the figure which showed the correlation of these. FIG. 24A shows an example of the original waveform of the APW of the subject D in the time zone in which no noise is on, and FIG. 24B shows the original waveform of the APW of the subject A in the time zone on which the noise is superimposed. It is the figure which showed the example. FIG. 25 is a diagram showing a related event of the subject A. FIG. 26 is a diagram illustrating a related event of the subject B. 27 (a) to (c) are diagrams comparing the frequency fluctuations of the fingertip volume pulse wave and the original waveform of APW according to the state of wakefulness, sleep onset phenomenon, and hypoxia of subject A, FIGS. 27D to 27F are diagrams comparing frequency fluctuations of the fingertip volume pulse wave and the original waveform of the APW for each state of the awake state of the subject B, the sleep onset phenomenon, and the state of the hypoxia. 28 (a) to 28 (c) are diagrams showing data on the wakefulness state, the sleep onset phenomenon, and the hypoactive state of the subject A, and FIGS. 28 (d) to 28 (f) are the wakefulness states of the subject B, It is the figure which showed the data in the sleep onset symptom phenomenon and a consciousness low state. FIGS. 29A and 29B are diagrams showing a 2 × 2 cross table in which the correct answer rate is calculated for each valid data ratio. FIG. 30 is a diagram showing the results of Experimental Example 4.

  Hereinafter, the present invention will be described in more detail based on the embodiments of the present invention shown in the drawings. There are two types of biological signal information collected in the present invention: central biological signal information and peripheral biological signal information. The peripheral biological signal information is typically a fingertip volume pulse wave, and is collected by a commercially available fingertip volume pulse wave meter. The central biological signal information is typically a back body surface pulse wave (APW), which is a biological signal measuring device (hereinafter referred to as “back body” proposed by the applicant in the above-mentioned Patent Document 2). Table pulse wave measuring device ") 1 is measured. The back body surface wave (APW) is a pressure vibration generated from the motion of the heart and aorta detected from the upper back of a person, and information on ventricular systole and diastole and the blood vessel wall serving as an auxiliary pump for circulation Elasticity information and blood pressure elasticity information. The signal waveform associated with heart rate variability includes sympathetic and parasympathetic nervous system activity information (parasympathetic activity information including the compensation of sympathetic nerves), and the signal waveform associated with aortic oscillation is sympathetic. Contains information on neural activity.

  Here, a schematic configuration of the back body surface pulse wave measuring device 1 will be described with reference to FIGS. 2 and 3. As shown in these drawings, the back body surface pulse wave measuring apparatus 1 includes a core pad 11, a spacer pad 12, a sensor 13, a front film 14, and a rear film 15.

  The core pad 11 is formed in, for example, a plate shape, and two vertically long through holes 11a and 11a are formed at symmetrical positions across a portion corresponding to the spinal column. The core pad 11 is preferably composed of a polypropylene bead foam formed in a plate shape. When the core pad 11 is composed of a bead foam, the foaming ratio is preferably in the range of 25 to 50 times, and the thickness is preferably less than the average diameter of the beads. For example, when the average diameter of 30 times expanded beads is about 4 to 6 mm, the core pad 11 is sliced to have a thickness of about 3 to 5 mm.

  The spacer pad 12 is loaded into the through holes 11 a and 11 a of the core pad 11. The spacer pad 12 is preferably formed from a three-dimensional solid knitted fabric. The three-dimensional solid knitted fabric includes, for example, a pair of ground knitted fabrics spaced apart from each other and the pair of grounds as disclosed in JP 2002-331603 A, JP 2003-182427 A, and the like. The knitted fabric has a three-dimensional three-dimensional structure having a large number of connecting yarns that reciprocate between the knitted fabrics to couple them together. When the 3D solid knitted fabric is pressed by the person's back, the connecting yarn of the 3D solid knitted fabric is compressed, tension is generated in the connecting yarn, and the vibration of the body surface through the human muscle accompanying the biological signal is propagated. The In addition, it is preferable to use a spacer pad 12 made of a three-dimensional solid knitted fabric that is thicker than the core pad 11. Accordingly, when the peripheral portions of the front film 14 and the rear film 15 are attached to the peripheral portions of the through holes 11a and 11a, the spacer pad 12 made of a three-dimensional solid knitted fabric is pressed in the thickness direction. Tension due to the reaction force of the film 15 is generated, and solid vibration (membrane vibration) is likely to occur in the front film 14 and the rear film 15. On the other hand, the spacer pad 12 made of a three-dimensional solid knitted fabric is also pre-compressed, and the connecting yarn that holds the shape of the three-dimensional solid knitted fabric in the thickness direction is also subjected to tension due to a reaction force, and string vibration is likely to occur.

  The sensor 13 is fixedly disposed on one of the spacer pads 12 before the front film 14 and the rear film 15 are laminated. As described above, the three-dimensional solid knitted fabric constituting the spacer pad 12 is composed of a pair of ground knitted fabrics and connecting yarns, and the string vibration of each connecting yarn is connected to the front film 14 and the ground via the nodes of the ground knitted fabric. In order to be transmitted to the rear film 15, the sensor 13 is preferably fixed to the surface of the spacer pad 12 (the surface of the ground knitted fabric). As the sensor 13, it is preferable to use a microphone sensor, in particular, a condenser microphone sensor.

  Next, the configuration of the biological state estimating device 60 of the present embodiment will be described based on FIG. The biological state estimation device 60 includes a central biological signal processing means 61, a peripheral biological signal processing means 62, and a state estimation means 63. The biological state estimation device 60 is constituted by a computer, and causes the computer to execute a central biological signal processing procedure that functions as the central biological signal processing means 61, and a peripheral biological signal processing that functions as the peripheral biological signal processing means 62. A computer program for executing a procedure and executing a state estimation procedure that functions as the state estimation means 63 is set. The computer program can be provided by being stored in a recording medium such as a flexible disk, a hard disk, a CD-ROM, an MO (magneto-optical disk), a DVD-ROM, or a memory card, or transmitted through a communication line. Is possible.

  The central biological signal processing means 61 is a time series waveform of the back body surface pulse wave (APW) collected from the back by the back body surface pulse wave measuring device 1, that is, a time series output signal obtained from the sensor 13 (preferably Has a means for receiving a filtering process (for example, a filtering process for removing a frequency component caused by body movement) and a time series output signal (original waveform)), and further, means for processing the original waveform Have central system index value.

  The processing means here includes means for obtaining an APW inclination time series waveform. This means first obtains a time-series waveform of frequency (hereinafter referred to as “zero cross detection method”) using a point where the APW original waveform is switched from positive to negative (hereinafter referred to as “zero cross point”). That is, when the zero cross point is obtained, it is divided every 5 seconds, for example, and the reciprocal of the time interval between the zero cross points of the time-series waveform included in the 5 second is obtained as the individual frequency f. The average value of f is adopted as the value of frequency F for 5 seconds. Then, by plotting the frequency F obtained every 5 seconds in time series, a time series waveform of frequency fluctuation is obtained. Next, a time window of a predetermined time width is set with a predetermined overlap time from a time-series waveform of frequency fluctuations, and the frequency slope is obtained by the least square method for each time window. Output the waveform. This calculation (movement calculation) is sequentially repeated to output a time-series change in the APW frequency gradient as a frequency gradient time-series waveform.

  Note that the time series waveform of the frequency fluctuation uses a maximum value (peak) by smoothing and differentiating the time series waveform of the back body surface pulse wave (APW) obtained from the sensor 13 of the back body surface pulse wave measuring device 1. (Hereinafter, referred to as “peak detection method”). For example, the maximum value is obtained by a smoothing differential method using Savitzky and Golay. Next, for example, the local maximum value is divided every 5 seconds, and the reciprocal of the time interval between the local maximum values of the time series waveform included in the 5 seconds (the peak on the peak side of the waveform) is obtained as the individual frequency f. The average value of f is adopted as the value of frequency F for 5 seconds. Then, by plotting the frequency F obtained every 5 seconds in a time series, a time series waveform of the frequency is obtained, and a time series waveform of the frequency gradient is obtained by using the same.

  The dorsal body surface pulse wave (APW) is a biological signal mainly including the state of control of the heart, which is the central system, that is, the state of sympathetic innervation of the arteries, and the appearance of the sympathetic and parasympathetic nervous systems of the autonomic nervous system A biological signal including information, and a waveform obtained by performing absolute value processing on a tilt time-series waveform of the biological signal by the zero-cross detection method is more related to the state of control of the heart and reflects the appearance state of the sympathetic nerve. The method based on the peak detection method is related to heart rate variability, and captures the dynamics of the parasympathetic nervous system in consideration of the compensation effect of the sympathetic nerve. Note that the absolute value processing of the slope time series waveform by the peak detection method is based on the parasympathetic nerve dynamics by wavelet analysis of the fingertip plethysmogram (this parasympathetic nerve dynamics take into account the sympathetic compensation effect). It is relatively approximate. Therefore, it is considered that the zero-cross detection method can be used as an index representing a physical condition resulting from adaptation to stress that is dealt with by control of the autonomic nervous system. Since the zero-crossing detection method is highly related to the state of control of the heart, it also includes information on notches of heart rate variability, and it is 0 which is the 0.5th-order component of the heart rate component that cannot be obtained with the fingertip volume pulse wave. A frequency component in the vicinity of .5 Hz is also obtained as information. Therefore, it is preferable to mainly use data obtained by the zero cross detection method when determining the biological state using the APW.

  The peripheral biological signal processing means 62 includes means for receiving a time series waveform (original waveform) of the fingertip volume pulse wave obtained from the fingertip volume pulse wave meter, and includes means for processing the original waveform, Obtain system index values. The processing means includes means for obtaining the inclination time series waveform of the power value of the fingertip volume pulse wave, similar to the means for obtaining the APW inclination time series waveform. Specifically, the means for obtaining the power time gradient time series waveform is obtained by calculating the maximum value and the minimum value of the time series waveform (original waveform) of the fingertip plethysmogram by the smoothing differentiation method using Savitzky and Golay, respectively. Ask. Then, the maximum value and the minimum value are divided every 5 seconds, and the average value of each is obtained. The square of the difference between the average values of the obtained local maximum and local minimum is used as a power value, and this power value is plotted every 5 seconds to create a time series waveform of the power value. From this time-series waveform, the slope of the power value is obtained by the least square method for a certain time window Tw (180 seconds). Next, the next time window Tw is similarly calculated at the overlap time Tl (162 seconds), and the result is plotted. This calculation (movement calculation) is sequentially repeated to obtain a time series waveform of the gradient of the power value.

  The state estimation means 63 corresponds to a coordinate system in which the central system index value obtained by the central biological signal processing means 61 and the peripheral system index value obtained by the peripheral biological signal processing means 62 are taken on each coordinate axis. The coordinate points between the index values are plotted on the coordinate system, and the biological state is estimated from the positional relationship between the plotted coordinate points. As the coordinate system, for example, an orthogonal coordinate system is used, and the central system index value is formed on one of the vertical axis and the horizontal axis, and the peripheral system index value is formed on the other side.

  As the central system index value and the peripheral system index value used in the state estimation means 63, various values can be used as long as they are obtained with respect to the central system biosignal information (APW) and the peripheral body formation signal information (finger plethysmogram). Although an index can be used, it is preferable to use an index value as in an experimental example described later as an index that causes a significant difference depending on the biological state.

  That is, (1) means for using each frequency at the same time in the original waveform of central biological signal information (APW) and peripheral formation signal information (finger plethysmogram) as an index value, and (2) from the original waveform of APW The frequency slope time series waveform obtained by slide calculation using the zero cross detection method and the power value slope time series waveform of the power value obtained by sliding the fingertip plethysmograph power value time series waveform are the same. Means using each amplitude value of time as an index value, (3) The ratio of each amplitude value at the same time in the slope time series waveform of APW frequency and the slope time series waveform of the power value of fingertip volume pulse wave It is preferable to employ at least one of the means used.

  Further, the state estimating means 63 determines whether each coordinate point is in a predetermined dispersion state, or is oriented substantially in a straight line, or a singular point as the positional relationship between the plotted coordinate points. It is determined in consideration of whether or not it has. Details will be described in an experimental example to be described later. For example, when comparing the frequency of the fingertip plethysmogram and the APW original waveforms in (1) above, in the arousal state, it is plotted on almost the same straight line, When the phenomenon is manifested, it shows a characteristic tendency that the coordinate points are dispersed within a predetermined range. Also, for example, when the amplitude values at the same time in the fingertip volume pulse wave and the APW slope time-series waveform of (3) are used, there is no correlation when the sleep onset symptom is expressed. It shows a characteristic tendency that singular points are plotted.

(Experimental example 1)
The back body surface pulse wave (APW) is measured by the back body surface pulse wave measuring device 1 according to the above embodiment, the finger tip volume pulse wave is measured by the finger tip volume pulse wave meter, and the central system index value and the peripheral system are measured. The state was determined by the coordinate system using the index value.

  The back body surface pulse wave measuring apparatus 1 used in the experiment has the configuration shown in FIGS. 1 to 3 and was measured in contact with the back of the subject. Moreover, the physical characteristics of this back body surface pulse wave measuring apparatus 1 were as follows. That is, when a 98 mm diameter wooden disk is mounted on an autograph manufactured by Shimadzu Corporation and a load of up to 200 N is applied in the Z direction in FIG. 3 at a moving speed of 50 mm / min, the load-deflection characteristics shown in FIG. It is obtained. FIG. 6 is a diagram in which the vertical axis of FIG. 5 is converted into a spring constant, and the back body surface pulse wave measuring apparatus 1 shows a constant value of the spring constant during the compression margin of 1 to 4.5 mm. The value was k = 19400 N / m.

  The fingertip plethysmogram was measured using an optical fingertip plethysmograph sensor (Amco's finger clip probe SR-5C) on the left index finger of the subject. In addition, the subject's condition was verified by simultaneously measuring heart sound, electrocardiogram, and electroencephalogram. The heart sound was measured using TA701T and AS101D manufactured by Nippon Koden Kogyo Co., Ltd. as a heart sound sensor / heart sound pulse wave amplifier, and the electrocardiogram was measured using BSM-2301 manufactured by Nihon Koden Kogyo Co., Ltd. as an electrocardiograph. The heart sound and the fingertip plethysmogram were obtained by synchronizing the analog signal with the APW and using A / D-converted data by Contec's USB 2.0 compatible analog input terminal AI-1608AY-USB. . The electroencephalogram measurement was performed by using an electroencephalograph EEG-9100 manufactured by Nihon Kohden Kogyo Co., Ltd., with electrodes attached based on the Ten-twenty electrocode system of the International Electroencephalographic Society and measured by monopolar induction. . The sleep stage was determined by the sleep stage determination method standardized by Rechtschanffen & Kales.

The subjects were 11 healthy men in their 20s (average age 26.78 years), and measured for 1 hour in a resting / supposed posture. The subject was instructed to remain awake as much as possible for the first 30 minutes and to sleep in the relaxed state for the latter 30 minutes.
FIG. 7 shows related events of the electrocardiogram, heart sound diagram, APW (original waveform), second-order differential waveform of APW, fingertip volume pulse wave (original waveform), and second-order differential waveform of fingertip volume pulse wave obtained by this experiment. The time phases are compared, and one waveform of the cardiac cycle is shown. The cardiac cycle was roughly divided into three phases: atrial systole, ventricular systole, and ventricular diastole, and the time phases of related events were compared. As a result, the R wave of the electrocardiogram at the beginning of the ventricular systole coincides with the I sound of the electrocardiogram and the bottom (α point) of the APW, with a delay of about 0.2 seconds related to the pulse wave velocity. It was recognized that the bottom of the fingertip plethysmogram was opposite. Next, the coincidence of ECG T wave, ECG II sound, APW top (β point), and fingertip plethysmogram notches in the early ventricular dilation period was recognized.

  That is, the APW and fingertip plethysmogram original waveforms, second-order differential waveforms and time phases in the cardiac cycle are very similar, and both are considered to be born in the same hemodynamics. Further, when a change occurs in the elasticity of the blood vessel, the waveform is disturbed due to the influence of the elastic resistance and the reflected wave. Therefore, the APW shows hemodynamics flowing in the same blood vessel, further shows the dynamics of the aorta and the peripheral circulatory system, and can be said to pick up vibrations of the same quality as the fingertip volume pulse wave.

  FIGS. 8 to 10 show that among the data obtained in this experiment, two stable states (FIGS. 8A and 8B), two transition states (FIGS. 9A and 9B), Various related events are compared according to six states of two types of autonomic nervous system hyperfunction states (FIGS. 10A and 10B). Related events are wavelet analysis of fingertip plethysmogram original waveform and second-order differential waveform, APW original waveform and second-order differential waveform α and β points, heart sounds I and II sounds, and fingertip plethysmogram The degree of appearance of sympathy / subsympathy obtained in this way, and the time difference between these various related events was also obtained. When the sympathetic nerve or parasympathetic nerve is enhanced, as shown in FIGS. 9A, 9B, and 10A, the time difference between the related events of fingertip volume pulse wave, APW, and heart sound diagram is disturbed. Was observed to occur.

  The two stable states are the arousal state (stable state (1)) in which the sympathetic / parasympathetic nerves appear in a balanced manner in FIG. 8A and the parasympathetic dominant sleep state in FIG. 8B ( It is a stable state (2)). Two types of transition states are states in which a burst wave representing temporary and rapid enhancement of the sympathetic nerve of FIG. 9A appears: a sleep onset sign phenomenon (transition state (1)), and FIG. 9B. This is a state of low wakefulness (transition state (2)) in which the degree of arousal and attention are reduced, in which temporary and rapid enhancement of sympathetic / parasympathetic nerves is observed. The two types of the autonomic nervous system hyperfunction state are the state where the sympathetic nerve is enhanced in FIG. 10A and the state where the parasympathetic nerve is enhanced in FIG. By capturing these states, it was possible to capture changes in the state of the living body, in particular, drowsiness and drowsiness during driving.

  The following can be said from FIGS. First, when transition state sympathetic nerve enhancement, so-called burst wave enhancement, is shown, the difference between fingertip volume pulse wave and APW and the difference between APW and heart sound diagram are shown in FIG. , (B) and FIG. 10 (a) show the tendency of the linear correlation shown by the regression line, as shown in the plots of the time differences in the bottom diagram. It shows that each axis of a-α and α-I corresponding to the ventricular systole tends to be shifted from each axis of e-β and β-II corresponding to the ventricular diastole. From the original waveform, it can be seen that the fingertip plethysmogram has a fluctuation in frequency while exhibiting baseline fluctuations. This indicates that the power value varies while the maximum Lyapunov exponent obtained from the fingertip plethysmogram changes. On the other hand, in the APW, the waveform is represented by the box in FIGS. 9A, 9B, and 10A, and the frequency variation occurs at the zero cross point that is the average value of the α point and the β point. Therefore, at the time of transition state sympathetic nerve enhancement, a correlation between the frequency of the fingertip volume pulse wave and the frequency of the APW zero-cross point is recognized.

  Next, in the transition state parasympathetic nerve enhancement, there is no frequency fluctuation of the fingertip volume pulse wave, and the baseline fluctuation occurs. Therefore, a change has occurred mainly in the maximum Lyapunov exponent. On the other hand, in the APW, amplitude fluctuation occurs, and a low frequency component and a high frequency component are superimposed. This suggests that APW may be affected by changes in vascular elasticity or reflected waves due to parasympathetic enhancement. The finger plethysmogram and the APW time phase show a tendency that the axes of a-α and α-I corresponding to the ventricular systole and the axes of e-β and β-II corresponding to the ventricular diastolic phase are shifted together. It becomes the same way of fluctuation as in the sympathetic nerve enhancement in the transition state.

  Next, the increase in sympathetic nerve appeared as a combination of baseline fluctuation in which no frequency fluctuation occurred in fingertip plethysmogram and frequency fluctuation with little baseline fluctuation. On the other hand, the APW has a subharmonic resonance waveform represented by a box in FIGS. 9A, 9B, and 10A. In this case, the difference value between the related events of fingertip plethysmogram, APW, and phonocardiogram is in the convergence direction having two regression curves. It shows that the axis of a-α and α-I corresponding to the ventricular systole fluctuates and the APW zero cross point moves. Correlation between the fluctuation of the frequency of the fingertip volume pulse wave and the fluctuation of the frequency of the APW zero cross point Is suggested. Moreover, in the stable state (both wakefulness and sleep state) and the enhancement of the parasympathetic nerve, there is little fluctuation. Although the fluctuation of the frequency occurs, a harmonious waveform is continuous, and the change in the fluctuation of the fingertip volume pulse wave and the APW frequency has a one-to-one relationship.

  Next, on the premise of the above, the analysis results by the central biological signal processing means 61, the peripheral biological signal processing means 62, and the state estimating means 63 will be described.

  FIG. 11 adopts the frequency of the APW original waveform as the central system index value obtained by the central biological signal processing means 61, and the fingertip volume pulse wave as the peripheral system index value obtained by the peripheral biological signal processing means 62. The frequency of the original waveform is adopted, and the state estimation means 63 takes the frequency of the APW original waveform on the vertical axis and the frequency of the fingertip volume pulse wave on the horizontal axis, and is a coordinate point obtained from each frequency at the same time , And the analysis result of subject H is shown.

  From FIG. 11, the steady state of the awake state (a), the sleep state (b), and the hypoactive state (d) show a one-to-one relationship between the fingertip volume pulse wave and the APW frequency variation, and the state change It can be seen that the frequency varies corresponding to. On the other hand, at the onset of sleep onset, which is thought to be negative feedback to the sleep center (c), when the sympathetic nerve is elevated (e), when the parasympathetic nerve is elevated (f) It can be seen that the correlation between the finger plethysmogram and the APW frequency component is low, and the frequency variation is random and dispersed. From these results, it can be said that when negative feedback is acting on the autonomic nervous system, the sympathetic and parasympathetic nerves are temporarily and suddenly increased, and the fingertip volume pulse wave and the APW frequency are disturbed. The expression of fluctuations due to these frequency fluctuations and dispersions indicates that both the APW and fingertip plethysmogram frequencies are fluctuating greatly. Therefore, the enhancement of both nervous systems means that the APW β-α point becomes small and changes occur in the down cross point, and the enhancement of the sympathetic nerve is considered to affect the periodic characteristics by the zero cross detection method. . It should be noted that although the sympathetic nerve is less prominent and the hypotension state, in which the function tends to be suppressed, the frequency fluctuation of the original waveform is large, but because the correlation between the fingertip volume pulse wave and the APW frequency fluctuation is high, there is little disturbance. It is difficult to distinguish from a stable state.

  For these reasons, the APW and fingertip plethysmogram waveform frequencies are used as central system index values and peripheral system index values and plotted on the coordinate system, and the obtained coordinate points correspond one-to-one. Whether or not the sympathetic / parasympathetic nerves are in an enhanced state by determining whether or not they are oriented (substantially in a straight line) and whether or not they are in a predetermined dispersion state by setting an arbitrary threshold value. Such state determination can be clearly performed. The threshold value can be set by statistically processing a plurality of data.

  FIG. 12 shows that the state estimation means 63 uses an absolute value processed inclination time series waveform calculated using the APW zero cross detection method as the central system index value, and the fingertip volume pulse wave as the peripheral system index value. Each amplitude value at the same time is plotted on a coordinate system using a power value gradient time series waveform obtained by absolute value processing.

  The sleep onset symptom, sympathetic state, and sympathetic nerve enhancement shown in FIG. 12, that is, sleepiness before going to sleep and a state that resists drowsiness are based on the power value of finger plethysmogram and APW zero cross detection method. It can be seen that the amplitude variation of each tilt time series waveform increases. Further, a waveform such as a subharmonic resonance with a high frequency component shows an aspect of two beats, and the amplitude increases and becomes a long cycle. In addition, the point that becomes a sudden change part in the original waveform is reflected in the time-series waveform of the gradient as a long-period waveform with the time width in which the waveform of the subharmonic resonance and the fluctuation of the heartbeat fluctuation fluctuate in the original waveform. In other words, capture accuracy can be expected to improve by capturing changes in the amplitude of the tilt time-series waveform rather than heartbeat fluctuations of each original waveform, which is difficult to capture in the hypoxia state.

  More specifically, in the two stable states of FIGS. 12A and 12B, the coordinate points are plotted within the range of the L line in these drawings. Therefore, the state determination means 63 is set so as to determine the degree of dispersion of each coordinate point with reference to the range surrounded by the L line. The L line indicates the boundary between a stable state such as when awakening and an autonomic nervous system hyperfunction state in a transition state such as a sleep symptom phenomenon or a wakefulness state. The L value is obtained from the average value of the maximum value group of the slope values in the stable state, and the state change with the passage of time is taken as a relative change with respect to the reference value. Note that this L value is a unique value for each subject.

  That is, the L value is set based on FIG. In FIG. 12B, each coordinate point shows a tendency of convergence within the range of the L value, which means that the amplitudes of the two time-series waveforms become smaller in the parasympathetic dominant state. 12 (c), (e), and (f) are all in a state in which the sympathetic / parasympathetic nerves are enhanced tend to be dispersed in the direction along the axis of the peripheral index value obtained from the fingertip volume pulse wave. On the other hand, in the hypoxic state of FIG. 12D, there is a tendency to disperse in the direction along the axis of the central system index value obtained from the APW. From this, it can be seen that the state determination means 63 can distinguish the hypoactive state from the enhanced state of the sympathetic nerve and the enhanced state of the parasympathetic nerve by setting a threshold value in the dispersion direction. That is, in the case of using the tilt time series waveform of FIG. 12, in addition to the distinction between the two stable states and the two sympathetic / parasympathetic enhancement states, it is possible to make a state determination that also distinguishes the conscious state.

  FIG. 13 shows that the state estimation means 63 uses the APW zero cross detection method as the central system index value and the amplitude value of the gradient time series waveform by the APW zero cross detection method and the amplitude of the gradient time series waveform of the power value of the fingertip plethysmogram as the peripheral system index value. It is the figure which showed the coordinate which employ | adopted ratio as a vertical axis | shaft and time was plotted as a horizontal axis. When the vertical axis of this figure is 1, it indicates that the fingertip volume pulse waveform tilt time-series waveform and APW tilt time-series waveform have a one-to-one relationship. Accordingly, the correlation between the fingertip volume pulse wave where the burst wave does not appear and the time series waveform of both slopes of the APW becomes high except for a time zone in which the sympathetic nerve is enhanced, such as a sleep symptom predictive phenomenon. Specifically, the amplitude ratio in FIG. 13 is obtained as follows. In other words, the gradient time series waveform calculated using the APW zero cross detection method and the gradient time series waveform of the power value of the fingertip plethysmogram are respectively subjected to differential processing and absolute value processing. Next, as shown in FIG. 12, the L value is set from the differential processing of the time series waveform indicating the stable state, each point of each time series waveform subjected to the absolute value processing is divided by the L value, and the APW Using the obtained value as the denominator, the value obtained from the fingertip volume pulse wave as the numerator is plotted against the time on the horizontal axis.

  In FIG. 13 (c), it can be seen that the coordinate points are dispersed and there are singular points greatly deviating from other coordinate points. This is a point with low correlation between APW and finger plethysmogram, and this corresponds to a so-called sympathetic burst wave. Therefore, from FIG. 13C, when such a singular point exists, it can be determined that a sleep onset symptom in which a sympathetic burst wave appears is occurring.

  From the above, it can be seen that the state estimation means 63 can estimate the state of the person by capturing the tendency of the coordinate points in FIGS. Moreover, in order to perform more accurate state estimation, it is preferable to use the estimation means in FIGS. 11, 12, and 13 in combination.

  FIGS. 14A to 14D show statistical processing of the determination using FIGS. 11 and 12 and the determination of a sleep onset symptom phenomenon, a wakefulness state, and a relaxed wakefulness state by a medical index. The p-value by the chi-square test was 0.05 or less, indicating a significant difference. Here, it can be seen that the detection sensitivity is improved by the homogeneity of the finger plethysmogram and the APW, and the analysis based on each inclination time series waveform.

  FIG. 15 (a) shows an example in which sleep desires were accepted in sleep experiments and resulted in sleep. FIG. 15A shows, in order from the top, determination of sleep stage based on brain waves, brain wave distribution rate, slope time series waveform of power value of fingertip plethysmogram, slope calculated using zero cross detection method of APW original waveform. A time series waveform, a coordinate point in which the amplitude ratio calculated in the calculation procedure of FIG. 13 is plotted with respect to the time axis (the one that appears apart as a singular point in the coordinate point corresponds to a burst wave of a sympathetic nerve), a finger The distribution map of sympathetic / parasympathetic nerve appearance of autonomic nervous system function using wavelet analysis of apical volume pulse wave is shown. FIG. 15B is an example of the hypothetical state of the subject B, and FIG. 15C shows an example of the subject C resisting sleep. Here, the degree of appearance of the sympathetic nerve obtained by wavelet analysis of the fingertip volume pulse wave is calculated using the APW zero cross detection method, the amplitude value obtained from the slope of the power value of the fingertip volume pulse wave using the calculation procedure of FIG. It can be seen that it approximates the simulated sympathetic burst wave.

  FIG. 16A shows the frequency analysis results of the original waveform of fingertip volume pulse wave and APW of 11 subjects (total 11 hours) of this experiment. Differences were observed between finger plethysmogram and APW at 0.5 Hz, 1.0 Hz, 1.3 Hz, 2.1 Hz, and 2.6 Hz. FIGS. 16B and 16C show the frequency analysis results of the fingertip volume pulse wave and the APW original waveform for about 10 minutes for each subject H with or without sleepiness. Depending on the presence or absence of drowsiness, a difference between the heart rate component and 2.5 Hz was observed in the fingertip volume pulse wave, and a significant difference appeared in the 0.5 Hz component in APW. Therefore, the fluctuation of the heart rate component caused by the presence or absence of sleepiness of the fingertip volume pulse wave appeared as a 0.5 Hz component in the APW. Here, in the frequency analysis results for 11 subjects in this experiment, the fingertip volume pulse wave and APW have peaks at 1 Hz and 1.3 Hz, which are heart rate variability components, and the fingertip volume at 0.5 Hz in APW. Since the power spectrum was larger than the pulse wave, it was suggested that many subjects were in the transition state.

(Experimental example 2)
An experiment was conducted to capture a hypoxic state and a sleep symptom phenomenon in a sitting position in which the sympathetic nerve is more likely to be enhanced than in the supine position in Experimental Example 1.

  The subject performed the sleep transition experiment while sitting in the driver's seat and the engine was idling or idling stopped. APW was performed by setting the back body surface pulse wave measuring device 1 used in Experimental Example 1 on the seat back of the driver's seat. The same fingertip plethysmogram sensor as in Experimental Example 1 was used for measurement of the fingertip plethysmogram. The condition of the subject was observed with a camera, and sleepiness was determined every 5 minutes using VAS (Visual Analog Scale) as a psychological index. The experiment time was 60 minutes, and the subjects were told that they would remain awake as much as possible until the end of the experiment. The test subjects were 10 healthy men in their 20s and 30s and 2 women.

  FIG. 17 shows a power value gradient time series waveform of the fingertip volume pulse wave, and FIG. 18 shows a frequency gradient time series waveform by the APW zero cross detection method. The time zone in which drowsiness increased most in the sensory evaluation is 300 seconds from 2400 to 2700 seconds indicated by hatching in FIG. In addition, before the time period when sleepiness increased, the amplitude increases (increases from a to b) in FIG. 17 and FIG. 18, and the period becomes longer (f becomes longer). It is suggested that this timing is a predictive phenomenon of falling asleep. It was suggested that the time when sleepiness increased was in the middle of increasing or decreasing the amplitude, and that the consciousness state was mixed with increasing or decreasing amplitude. The timing of the appearance time of these phenomena coincides in both FIG. 17 and FIG.

  More specifically, first, from the psychological index shown in FIG. 18, subject F has a relatively high arousal level for 900 seconds from the start of the experiment, and once the arousal level tends to decrease for a period of 900 to 2100 seconds. Shows a certain level of wakefulness, and the degree of wakefulness further decreases between 2100 and 2700 seconds, and after 3000 seconds it returns to the start of the experiment.

This will be considered with the time-series waveforms shown in FIGS. Both the fingertip plethysmogram and the APW have a long period in each inclination time-series waveform, and the amplitude tends to increase during 900 to 2100 seconds when the sleep onset sign phenomenon appears. During 2100 to 3000 seconds, when the arousal level rises and rises, the amplitude increases and decreases with almost a long period, and when the arousal level returns to the original level, a relatively short period and small amplitude tend to remain It became.
FIG. 19 shows a state in which there is some sleepiness but is relaxed and awake (0 to 900 seconds), a sympathetic burst wave is generated and a sleep onset symptom is expressed (900 to 2100 seconds), The frequency analysis result of the inclination time series waveform according to each state of the fall of arousal level, generation | occurrence | production of sleepiness, a raise of arousal level (2100-2700 second), and the state which relaxed and re-awakened (2700-3600 second) is shown.

  As seen from the frequency analysis results of each tilt time series waveform in FIG. 19, the peak time peak waveform of the fingertip volume pulse wave is observed at 0.003 Hz, 0.007 Hz, and 0.011 Hz at the time of waking. At the time of re-wakening, it was observed at 0.002 Hz, 0.005 Hz, 0.007 Hz, and 0.011 Hz. A high peak appeared at 0.003 Hz or less when the sleep onset symptom occurred. When the arousal level was lowered, the PSD of the power spectrum was significantly lowered, and the peak frequency was 0.005 Hz. These results suggest that the fingertip plethysmogram tends to have a large amplitude / long period when the sleep onset symptom appears, and a small amplitude / high period when the hypotension appears.

  On the other hand, in APW, the peak is observed at 0.003 Hz and 0.0065 Hz at the time of arousal accompanied by sleepiness, and the peak is observed at 0.0045 Hz at the time of re-awakening. The peak of the power spectrum shifted to the high frequency side at the last re-awakening. Peaks were observed at 0.003 Hz and 0.005 Hz at the onset of the onset of sleep phenomenon, and peaks were observed at 0.003 Hz in the conscious state. The power spectrum tended to be larger at the onset of sleep onset than at the time of awakening with a high level of arousal but drowsiness. From these facts, it can be said that the frequencies representing the boundary between the arousal state and the sleep symptom predictive phenomenon and the boundary between the arousal state and the low wake state exist in the vicinity of 0.003 Hz.

  FIG. 20 shows the frequency analysis results of the fingertip volume pulse wave and the APW original waveform for 1040 to 1060 seconds in which the occurrence of sleepiness is recognized in the sensory evaluation in FIG. From FIG. 13, it can be seen that even in the sitting position in the idling state, the same results as those of the subject H experimented in the supine position in FIG. 16 are shown.

  On the other hand, FIG. 21 shows a sleep onset symptom (1400 to 2100 seconds) and a wakefulness state (2100 to 2800 seconds) by using the methods shown in FIGS. 11, 12, and 13 which are analysis methods of the present invention. The time zone is illustrated. Specifically, FIGS. 21A to 21F are sequentially shown in FIGS. 11C, 12C, 13C, 11D, 12D, and 13 respectively. Corresponds to (d). When comparing each of FIGS. 11 to 13 corresponding to FIGS. 21A to 21F, in particular, the coordinate point of the frequency variation of the original waveform in the hypothetical state shown in FIG. As shown in FIG. 11 (d), there is no one-to-one correspondence, and there is a significant difference in that it is dispersed. In the supine state in the supine position in FIG. 11D, as described above, there is little increase in sympathetic nerves and functions tend to be suppressed. However, in the sitting position in the idling state, a hypoxia state occurs in the sympathetic nerve dominant. This means that the sympathetic nerve function is enhanced only by sitting in the driver's seat, and when the determination of the present invention is applied at the time of driving, the coordinate point is higher than that of the supine position in the determination of the conscious state. It can be seen that it is necessary to set a threshold value in consideration of the degree of variance of.

  FIG. 22 shows an analysis result of a subject who has a high feeling of fatigue immediately after the start of the experiment, sleeps for a moment in the middle of the experiment, then becomes sleepy and then shifts to a re-awake state. Among these, FIGS. 22A to 22C are FIGS. 11A and 11B in which frequency fluctuations of the original waveforms of APW and fingertip plethysmogram in an arousal state, a sleep onset phenomenon, and a hypoxia state are plotted as index values. A similar coordinate system is shown. 22D to 22F are the same as FIG. 12 in which the amplitude values of the time series waveforms of the APW and fingertip plethysmogram in the awake state, the sleep onset phenomenon, and the hypoxia state are plotted as index values. The coordinate system of is shown.

  From these figures, it is difficult to distinguish the degree of dispersion of coordinate points in the original waveform, and it is difficult to determine the state. This is because the sympathetic nerve function is superior to the supine position. On the other hand, FIGS. 22D to 22F using the gradient time-series waveform are shown in FIG. 22D at the time of awakening (at the time of awakening in a stable state) with the APW frequency around 0.003 Hz as a boundary. On the other hand, the degree of dispersion is clearly different between the appearance of the sleep onset symptom and the state of wakefulness. Therefore, by setting the threshold value at a frequency of 0.003 Hz obtained from the APW slope time series waveform, the state estimation unit 63 can determine whether or not the user is in the awake state.

  Also, in the hypoactive state of FIG. 22 (f), the distribution of coordinate points is along the vertical axis compared to the appearance of the sleep onset sign phenomenon of FIG. 22 (e). This is the same when comparing the hypoactive state of FIG. 12D in the supine position with the appearance of the sleep onset symptom in FIG. 12C, but the hypoactive state in the sitting position of FIG. 22F. The state has a larger spread in the horizontal axis direction than in the hypothetical state in the supine position of FIG. As mentioned above, this is because the sympathetic nerve function is enhanced by sitting in the driver's seat. Set the threshold.

  FIG. 23 is a chi-square test for the correlation between the fingertip volume pulse wave and APW slope time-series waveform changes of FIG. 22 and the states of awakening, the appearance of a sleep symptom, and a wakefulness state. It is the result calculated | required by. Compare the number of times of “constant amplitude” time period and the number of times “increase of amplitude” occurred in the comparison of awakening and sleep onset symptom phenomenon. The number of occurrences of “increase / decrease in amplitude” as shown in FIG. 18 was compared. As a result, the pp value was 0.005 or less, which was found to be statistically significant.

(Experimental example 3)
In Experiment 2, the effectiveness of state estimation using a time-series waveform of inclination in an environment affected by noise was examined using data for six male subjects. All subjects measured in both idling and idling stop states, and set the maximum amplitude value of the APW original waveform excluding the influence of body movement as a reference value in the measurement results in idling stop state. Data exceeding this reference value was defined as noise in both the measurement result in the idling state and the measurement result in the idling stop state.

  FIG. 24A shows an example of the original waveform of the APW of the subject D in the time zone when no noise is on, and FIG. 24B shows the original waveform of the APW of the subject A in the time zone when the noise is superimposed. An example is shown. The APW original waveform in FIG. 24 (a) is a waveform in a region of amplitude ± 0.1 (V) or less. In FIG. 24 (b), the influence of the reflected wave in the APW original waveform is There is a mixture of things that cannot be determined from the effects of vibration. In the following analysis, the state of the sleep onset symptom is estimated using the waveform in this state.

  First, the ratio of valid data that was not affected by noise was determined from the waveform data of all time. The percentage of valid data for each subject is: 41.3% for subject A, 70.5% for subject B, 67.0% for subject C, 71.5% for subject D E was 68.4% and subject F was 62.2%.

  FIG. 25 shows a related event for subject A with a percentage of valid data of 41.3%, and FIG. 26 shows a related event for subject B with a percentage of valid data of 70.5%. The related events here are, from the top, the sleepiness determination by VAS, the distribution of the appearance of sympathetic / parasympathetic nerves of autonomic nervous system function using wavelet analysis of fingertip volume pulse wave, fingertip volume pulse wave The slope time series waveform of the power value and the maximum Lyapunov exponent, and the slope time series waveform calculated using the zero cross detection method of the APW original waveform.

  FIGS. 27 (a) to (c) are comparisons of the frequency fluctuations of the fingertip volume pulse wave and the original waveform of the APW for each state of the subject A's arousal state, sleep onset symptom phenomenon, and hypoxia state, FIGS. 27D to 27F compare the frequency fluctuations of the fingertip volume pulse wave and the original waveform of the APW for each state of the subject B's arousal state, sleep onset symptom phenomenon, and hypoxia state.

  FIG. 28 shows the correlation between the slopes at the same time by performing absolute value processing on the slope time series waveform of the power value of the fingertip plethysmogram and the slope time series waveform using the APW zero cross detection method. It is. 28 (a) to (c) are data on the awakening state, sleep onset phenomenon, and low onset state of the subject A, and FIGS. 28 (d) to (f) are data on the awakening state of the subject B, the sleep onset sign. This is data in a phenomenon, hypoxia. The meaning of the L line and L value in the figure is the same as in the above experimental example.

  FIGS. 29A and 29B are 2 × 2 cross tables in which the correct answer rate is calculated for each valid data ratio. FIG. 29A is obtained from data of five subjects B to F, and FIG. 29B is obtained from data of six subjects A to F. Comparing FIGS. 29 (a) and 29 (b), it can be seen that when subject A has a valid data ratio of 41.3%, the correct answer rate has decreased from 78.0% to 73.5%. . Therefore, according to this experiment, if the proportion of valid data excluding subject A is 62.2% or more of subject D, which is the lowest, accurate state estimation is possible even in noisy environments. It can be said that there is.

  In addition, as shown in FIG. 27, it is difficult to estimate the state using the fingertip volume pulse wave and the original APW waveform under the idling condition with external vibration noise, whereas FIG. When using the analysis results of fingertip plethysmogram and APW slope time-series waveform as shown above, both subjects A and B fall below the L value in the wakefulness state, and exceed the L value in the sleep onset symptom phenomenon and the wakefulness state. Many values are detected, and it can be seen that the influence of noise is reduced.

  From these facts, it can be seen that the determination using the gradient time-series waveform has higher robustness than the determination using the original waveform, and can be determined with high accuracy even in a noisy environment.

  In addition, when the related events shown in FIGS. 25 and 26 are compared with each other, it can be seen that both the subjects A and B appear to alternate from the early sleep onset phenomenon and the hypoxia state to maintain the arousal state. Comparing the determination result of the subject's state by the VAS and the fingertip volume pulse wave with the determination result by the APW inclination time series analysis, it tends to be relatively close. However, in the case of the subject A having a lot of noise and a valid data ratio of 41.3%, the judgment of the sleep onset symptom for 2100 to 3000 seconds and the state of hypoxia use the time series waveform of the fingertip volume pulse wave inclination. It can be seen that there is a slightly different tendency between the one using the APW slope time series waveform. This is considered to be due to a difference in determination accuracy because the proportion of valid data is low.

(Experimental example 4)
The back body surface pulse wave measuring device 1 was mounted on the driver's seat of the truck, and the body surface pulse wave of the driver who was operating normally was measured and evaluated. The test subjects were professional drivers (9 men) in their 20s to 50s engaged in the transportation business, and the total number of operations for all subjects was 91 times.

  FIG. 30 shows an autonomic nerve in a transition state such as a sleep symptom and a hypoxia state obtained from an inclination time-series waveform using the APW zero-cross detection method of actual driving history for 91 occupational drivers (9 men). The number of occurrences of system hyperactivity state is plotted on the vertical axis against the operation time on the horizontal axis.

  From this graph, it can be seen that in work over 10 hours, the operation time is from 2 hours to 10 hours is a time zone in which warnings frequently occur, and also a time zone in which variations due to business differences and individual differences occur. However, there are extremely few warnings at the end of work over 10 hours after the start of work. This is a good sense of tension that lasts for more than 8 hours from the start of work for more than 2 hours, but this rebound of tension may create a relaxed state at the end of work. That is, there is a time zone in which the sympathetic nerve compensation effect occurs from more than 2 hours after the start of business, and once it does not function after 10 hours after the business starts. This relaxed state may create an environment in which human error is likely to occur due to a combined action with the progress of fatigue. According to the operating business manager who manages the subjects who cooperated in this experiment, the rule of thumb is that accidents are most likely to occur during the 30 minutes from the start of the operation and 30 minutes before the end of the operation. In the experiment, the result according to this rule of thumb was obtained as described above.

DESCRIPTION OF SYMBOLS 1 Back body surface pulse wave measuring apparatus 11 Core pad 12 Spacer pad 13 Sensor 14 Front film 15 Rear film 60 Living body state estimation apparatus 61 Central body biological signal processing means 62 Peripheral body biological signal processing means 63 State estimation means

Claims (12)

A central biological signal processing means for obtaining a predetermined central system index value from the central biological signal information;
Peripheral biosignal processing means for obtaining a predetermined peripheral index value from peripheral biosignal information;
State estimation means for estimating a biological state from a positional relationship of a plurality of coordinate points plotted on a coordinate system using an index value obtained from the central system index value and the peripheral system index value as at least one coordinate axis The biological state estimation apparatus characterized by the above-mentioned.   The state estimation means determines whether each coordinate point is in a predetermined dispersion state, or is aligned substantially on a straight line, or has a singular point as a positional relationship between the plurality of coordinate points. The biological state estimation apparatus according to claim 1, wherein the biological state is estimated in consideration of   The state estimating means uses, as the central system index value and the peripheral system index value, each frequency at the same time in the original waveform of the central system biosignal information and the peripheral formation body signal information, and the central system index value and the peripheral system index value Claims: On a coordinate system in which peripheral index values are taken on respective coordinate axes, coordinate points between corresponding index values are plotted on a coordinate system, and a biological state is estimated from the positional relationship between the plotted coordinate points. Item 3. The biological state estimation apparatus according to Item 1 or 2.   In the inclination time series waveform obtained by performing slide calculation on the original waveform of the central system biosignal information and the peripheral formation body signal information under a predetermined condition as the central system index value and the peripheral system index value. Using each amplitude value at the same time, plotting the central system index value and the peripheral system index value on the coordinate system on each coordinate axis, plotting the coordinate points between the corresponding index values on the coordinate system, and plotting The biological state estimation apparatus according to claim 1, wherein the biological state is estimated from a positional relationship between the plurality of coordinate points.   In the inclination time series waveform obtained by performing slide calculation on the original waveform of the central system biosignal information and the peripheral formation body signal information under a predetermined condition as the central system index value and the peripheral system index value. The corresponding index value is plotted on the coordinate system using the ratio of each amplitude value at the same time on one axis and the time axis on the other axis, and the biological state is determined from the positional relationship between the plotted coordinate points. The biological state estimation apparatus according to claim 1 or 2 for estimation.   The biological state estimating apparatus according to any one of claims 1 to 5, wherein the central biological signal information is a back body surface pulse wave collected from the back.   The biological state estimation device according to any one of claims 1 to 6, wherein the peripheral biological signal information is a fingertip volume pulse wave. To a computer as a biological state estimating device for estimating a biological state,
A central biological signal processing procedure for obtaining a predetermined central system index value from the central biological signal information;
Peripheral biosignal processing procedure for obtaining a predetermined peripheral index value from peripheral biosignal information;
A state estimation procedure for estimating a biological state from a positional relationship of a plurality of coordinate points plotted on a coordinate system using an index value obtained from the central system index value and the peripheral system index value as at least one coordinate axis Computer program to make.   In the state estimation procedure, as the positional relationship of the plurality of coordinate points, whether each coordinate point is in a predetermined dispersion state, is oriented in a substantially straight line, or has a singular point. The computer program according to claim 8, wherein the biological state is estimated in consideration of.   The computer program according to claim 8 or 9, wherein each frequency at the same time in an original waveform of the central biological signal information and the peripheral formed body signal information is used as the central system index value and the peripheral system index value.   As the central system index value and the peripheral system index value, each amplitude value at the same time in the tilt time series waveform obtained by sliding calculation of the original waveform of the central system biological signal information and the peripheral formation body signal information under a predetermined condition 10. The computer program according to claim 8 or 9, wherein:   As the central system index value and the peripheral system index value, each amplitude value at the same time in the tilt time series waveform obtained by sliding calculation of the original waveform of the central system biological signal information and the peripheral formation body signal information under a predetermined condition The computer program according to claim 8 or 9, wherein a ratio of