JP2018093997A - Biological state estimation device, biological state estimation method, computer program, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To determine tiredness in which a possibility of a feeling of the tiredness being masked is high, and which is hardly realized.SOLUTION: A biological state estimation device 100 includes: biological adjustment function element determination means 200 for analyzing a biological signal and obtaining a plurality of indexes involved in a biological adjustment function including indexes with high relevance to an autonomic nervous function, and physical and mental tiredness or feelings; and feeling-of-tiredness estimation means 300 for combining the plurality of indexes obtained by the biological adjustment function element determination means 200, and estimating whether or not the biological state is the state in which masking of a feeling of tiredness is generated.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a technique for estimating a state of a living body using a biological signal obtained from a human back.

  In the patent documents 1 to 5 and the like, the present inventors have proposed a technique for detecting a vibration generated on the body surface of the back in a human upper body with a biological signal measuring device and analyzing a human state. Sound / vibration information detected from the heart and aorta motion detected from the upper back of a person is pressure vibration generated from the heart and aorta motion. Information on ventricular systole and diastole, circulation auxiliary pump, It includes elasticity information of the blood vessel wall and reflected wave information. That is, a back body surface pulse wave (including Aortic Pulse Wave (APW)) around 1 Hz generated on the back surface from the motion of the heart and aorta, and a sound transmitted to the back side with a heartbeat (“pseudo heart sound” (this specification) The book includes information on the heart sound, which is a heart sound collected from the chest side, as a “pseudo heart sound”). The signal waveform associated with heart rate variability includes sympathetic nervous system and parasympathetic nervous system nerve activity information, and the signal waveform associated with aortic oscillation includes sympathetic nerve activity information.

  In Patent Document 1, slide calculation is performed by applying a predetermined time width to a time series waveform of a back body surface pulse wave (APW) near 1 Hz extracted from a collected biological signal (sound / vibration information), and a frequency slope is calculated. A time-series waveform is obtained, and the biological state is estimated from the tendency of the change depending on, for example, whether the amplitude tends to be amplified or attenuated. In addition, the biological signal is subjected to frequency analysis, and the power spectrum of each frequency corresponding to the function adjustment signal, fatigue acceptance signal, and activity adjustment signal belonging to a predetermined ULF band (very low frequency band) to VLF band (very low frequency band). And determining the state of a person from the time series change of each power spectrum. Since the fatigue acceptance signal indicates the degree of progress of fatigue in the normal activity state, in addition to this, by comparing the degree of dominance of the power spectrum of the function adjustment signal and the activity adjustment signal, Status, parasympathetic dominant status, etc.). In addition, the distribution rate of each frequency component is obtained in a time series when the sum of the power spectrum values of the frequency components corresponding to these three signals is set to 100, and the human condition is obtained using the time series change of the distribution rate. Is also disclosed.

  Patent Document 2 proposes a technique for displaying a biological state as a physical condition map and a sensory map as a quantification method of the biological state. This is a frequency analysis of the APW described above, and for the analysis section to be analyzed, the analysis waveform is shown in a logarithmic axis display, the analysis waveform is divided into a low frequency band, a middle frequency band, and a high frequency band, and the divided analysis waveforms are divided. The analysis waveform is scored based on a certain standard from the slope and the shape of the entire analysis waveform, and plotted on the coordinate axes. The physical condition map is a state in which the state of control of the autonomic nervous system is viewed as a balance between the sympathetic nerve and the parasympathetic nerve, and the sensory map is obtained by superimposing the state of change of heartbeat variability on the physical condition map.

  Patent Documents 3 to 5 disclose means for determining the homeostasis maintenance function level. Consistency maintenance function level judgment means uses positive / negative of differential waveform of frequency gradient time series waveform, positive / negative of integral waveform integrating frequency gradient time series waveform, frequency gradient time series waveform using zero cross method and peak detection method The determination is made using at least one of the absolute values of the frequency gradient time series waveforms obtained by performing absolute value processing on the frequency gradient time series waveforms. Based on these combinations, the level of the homeostasis maintenance function is determined. For example, using a frequency gradient and an integral value, if the value is equal to or greater than a predetermined value, it is determined as “constancy maintaining function level 1”, or the differential value is equal to or less than a predetermined position and “ In the case of “peak advantage”, it can be set to be determined as “constancy maintaining function level 4”. These combinations, threshold values for determination, and the like are determined by statistically processing data of a large number of subjects. Moreover, the level of the homeostasis maintenance function is divided into, for example, 5 to 7 stages, from the case where the homeostasis maintenance function is excellent (when the degree of concentration is high, etc.) to the case where the homeostasis maintenance function is inferior (in the case of an overtension state, Decrease of concentration due to driving aside) When displaying on the monitor, the level of 5 to 7 levels is displayed with characters, or if it is higher than the intermediate level (normal state), it is judged that the homeostasis maintenance function is excellent, and that In other cases, it is determined that the homeostasis maintenance function is inferior collectively, and the monitor is set so that different color displays are made.

  In Non-Patent Document 1, regarding fingertip plethysmogram information, a frequency gradient time series waveform of a power value reflecting sympathetic nerve information is obtained, and an integral value obtained by processing the absolute value is plotted as a fatigue level in time series. Thus, a technique for drawing a fatigue curve and capturing muscle fatigue is disclosed. Non-Patent Document 2 discloses a technique for capturing muscle fatigue by computing a biological signal obtained from a human back using an air pack sensor and drawing a fatigue curve using a similar method. That is, the state of muscle fatigue can be grasped by using a frequency gradient time-series waveform of power values reflecting sympathetic information (in the case of APW, a frequency gradient time-series waveform by the zero cross method).

JP 2011-167362 A JP 2012-239480 A WO2011 / 046178 JP 2014-117425 A JP 2014-223271 A

Original "Development of a simple evaluation method for long-term sitting fatigue using fingertip volume pulse wave information" Yasunori Fujita (8 others), Ergonomics Vol. 40, no. 5 (’04) "Application of biological fluctuation signals measured by non-invasive sensors to fatigue and sleep prediction", Naoki Ochiai (6 others), 39th Annual Meeting of the Japan Ergonomics Society, Chugoku-Shikoku Branch, 2006 Issued on May 25, Publisher: Japan Ergonomics Society Chugoku-Shikoku Branch Office

  Each of the above-described techniques is to determine each person's state by analyzing each element caused by fluctuations related to the bioregulatory function. The timing of the sign is determined, the degree of fatigue is determined, the change in the homeostasis maintenance function level is determined, and the purpose is determined. However, these are all output separately. In Patent Literature 4, time-series changes of a plurality of indicators related to each element caused by fluctuations in the biological regulation function, such as a sleep onset symptom phenomenon, an imminent sleep phenomenon, a low running state, a homeostasis maintenance function level, an initial fatigue state, and mood determination However, in any case, the time-series change of each index is individually determined.

  By the way, when a person gets tired, he / she feels tiredness. However, there is a case where a person feels fatigued even though the body is actually fatigued, such as when the field of view concentrates on things and the field of view is narrow. That is, the feeling of fatigue that a person feels does not necessarily coincide with the fatigue that occurs in the body. The areas where fatigue occurs are mainly the hypothalamus and frontal gyrus, which are the centers of the autonomic nerves, and the area that recognizes the fatigue is the orbital frontal cortex, but the function of the frontal lobe of the brain, which controls motivation and achievement, It is known that the feeling of fatigue generated in the frontal area is masked. For this reason, there is a case where fatigue accumulates without being aware of the feeling of fatigue, and it is desired that a state in which the feeling of fatigue is masked can be estimated.

  This invention is made | formed in view of the above, and makes it a subject to provide the technique which can connect accumulation | storage of unconsciousness fatigue by making it possible to estimate the state where masking of a feeling of fatigue occurs.

  In order to solve the above-mentioned problems, the present inventor diligently studied and came to complete the present invention paying attention to the following points. That is, sound / vibration information generated from the motion of the heart and aorta detected from the upper back of a person, in particular, the back body surface pulse wave (APW) in the vicinity of 1 Hz among them is the elasticity information of the blood vessels and the reflected wave. Contains information. For this reason, by analyzing the back body surface pulse wave (APW) and obtaining a frequency gradient time-series waveform described later, it is possible to obtain information on the overall fluctuation of the living body which is a bioregulatory functional element. In this fluctuation information, if the state of fluctuation for each frequency is captured, the distribution rate for each frequency takes the fluctuation method reflecting the fluctuation element of the brain wave (θ wave, α wave, β wave). By analyzing using the series waveform, fluctuation information such as which frequency band of the electroencephalogram is the dominant fluctuation can be captured.

  In addition, while the frequency-slope time-series waveform by the zero-cross method is under the control of the autonomic nervous system, the frequency-sequential time-series waveform by the zero-cross method is frequency-analyzed. The function adjustment signal typified by 0.0017 Hz, the fatigue acceptance signal typified by 0.0035 Hz, and the activity spectrum signal typified by 0.0053 Hz are represented by power spectrum ratios of three frequency components. In other words, the distribution rate of the three frequency components representing the shape of the power spectrum obtained by frequency analysis can be considered as a site showing the expression of endocrine function rather than an autonomic nervous reaction.

  The frequency gradient time series waveform obtained by the zero cross method indicates the degree of sympathetic nerve expression by performing absolute value processing, and the frequency gradient time series waveform obtained by the peak detection method indicates the degree of parasympathetic nerve expression. Therefore, by using these, it is possible to capture the expression of the biological regulation function in more detail. For example, an absolute value process is performed on a frequency gradient time series waveform obtained by the zero cross method, and by integrating this, a fatigue curve indicating the degree of human fatigue is obtained, and the state of muscle fatigue can be grasped.

  Further, by using at least one or more of positive / negative of the differential waveform of each frequency gradient time series waveform using the zero cross method, the absolute value of each frequency gradient time series waveform of the zero cross method or the peak detection method, etc. It is possible to capture changes in the maintenance function level. The time series waveform of the fatigue curve and the homeostasis maintenance function level is also derived from the frequency gradient time series waveform, and is an index reflecting information on fluctuations in the autonomic nervous system, brain function, and the like. In addition to this, indicators that are highly relevant to physical and mental fatigue (physical condition map, sensory map), and indicators that are highly relevant to sensation (attention level of homeostasis function level, frequency of the level equivalent to warning) The frequency at which a sense of fatigue or fatigue that is required as fatigue is required is also obtained. These capture the fluctuations of each function of the brain, autonomic nervous system, and endocrine system from multiple viewpoints. Combining these multiple types of fluctuation indicators can cause fatigue masking. We thought that it was possible to estimate the presence or absence of gender.

That is, the biological state estimation apparatus of the present invention is
A biological state estimation device that estimates a biological state using a biological signal obtained from a biological signal measurement device that is in contact with a human back,
A bioregulatory function element determination unit that analyzes the biosignal and obtains a plurality of indices related to the bioregulatory function, including an index that is highly related to autonomic nervous function, physical / mental fatigue, or sensation;
Fatigue sensation estimating means for estimating whether or not the biological condition is a state where fatigue masking is occurring by combining a plurality of indices obtained by the biological adjustment functional element determination means. And

The fatigue estimation means is
First estimation means that estimates that the state of masking of fatigue occurs when a time-series change of an index highly relevant to the autonomic nervous function of the biological signal to be analyzed satisfies a predetermined criterion When,
With respect to the biological signal to be analyzed that is not estimated to be a state in which the first estimator masks the feeling of fatigue, the time series change of the index highly relevant to the physical / mental fatigue satisfies a predetermined criterion A second estimating means for estimating that no fatigue masking has occurred,
The biological signal to be analyzed that is not estimated to be in a state in which no fatigue masking has occurred in the second estimating means, based on a predetermined criterion, at least the state in which the fatigue feeling masking has occurred and the fatigue feeling It is preferable to have the 3rd estimation means classify | categorized into either of the states which masking has not arisen.

The biological state estimation method of the present invention is a biological state estimation method for estimating a biological state using a biological signal obtained from a biological signal measurement device that is in contact with a human back,
Analyzing the biological signal, to obtain a plurality of indicators involved in the bioregulatory function, including indicators highly related to autonomic nervous function, physical / mental fatigue or sensation,
By combining the plurality of indices, it is estimated whether or not the biological state is a state in which fatigue masking occurs.

In the biological state estimation method of the present invention, when the time series change of the index highly relevant to the autonomic nerve function for the biological signal to be analyzed satisfies a predetermined criterion, the fatigue feeling masking occurs. Performing a first estimation procedure to estimate that
With respect to the biological signal to be analyzed that is not estimated to be in a state where the masking of the feeling of fatigue has occurred by the first estimation procedure, the time series change of the index highly related to the physical / mental fatigue satisfies a predetermined criterion The second estimation procedure for estimating that no fatigue masking has occurred,
Based on a predetermined standard, at least the state of the fatigue feeling masking and the fatigue feeling of the biological signal to be analyzed that is not estimated as the state where the fatigue feeling masking does not occur in the second estimation procedure Perform a third estimation procedure to classify any of the states where masking has not occurred,
It is preferable to make a configuration for estimating whether or not the fatigue masking is occurring.

The computer program of the present invention is a computer program that causes a computer as a biological state estimation device to analyze a biological signal obtained from a biological signal measurement device that is in contact with a human back and to execute a procedure for estimating a biological state. And
Analyzing the biological signal, and obtaining a plurality of indicators involved in the biological regulation function, including an indicator highly related to autonomic nervous function, physical / mental fatigue or sensation;
Combining the plurality of indices, a procedure for performing masking of a feeling of fatigue and estimating whether the feeling of fatigue is masked as the biological state is executed.

In the computer program of the present invention, the procedure for estimating whether or not the fatigue masking is in a state occurs,
A first estimation procedure for estimating a state in which the masking of the feeling of fatigue occurs when a time-series change of an index highly relevant to the autonomic nervous function of the biological signal to be analyzed satisfies a predetermined criterion And execute
With respect to the biological signal to be analyzed that is not estimated to be in a state where the masking of the feeling of fatigue has occurred by the first estimation procedure, the time series change of the index highly related to the physical / mental fatigue satisfies a predetermined criterion A second estimation procedure for estimating that no fatigue masking has occurred,
Based on a predetermined standard, at least the state of the fatigue feeling masking and the fatigue feeling of the biological signal to be analyzed that is not estimated as the state where the fatigue feeling masking does not occur in the second estimation procedure It is preferable to adopt a configuration in which the third estimation procedure for classifying into any state where masking has not occurred is executed.

  According to another aspect of the present invention, there is provided a computer program that causes a computer as a biological state estimation device to analyze a biological signal obtained from a biological signal measurement device that is in contact with a human back and to execute a procedure for estimating a biological state. A recorded computer-readable recording medium is provided.

  According to the present invention, the combination of a plurality of indicators indicating fluctuations in the state of the bioregulatory functional element, including indicators caused by fluctuations highly related to autonomic nervous function, physical / mental fatigue, or sense, The possibility of feeling masking can be captured.

FIG. 1A is an exploded view showing an example of a biological signal measuring apparatus for measuring a back body surface pulse wave used in one embodiment of the present invention, and FIG. FIG. FIG. 2 is a diagram schematically showing the configuration of the biological state estimation apparatus according to one embodiment of the present invention. FIG. 3A is a diagram showing an example of a frequency slope time series waveform obtained by the zero cross method and the peak detection method obtained by the frequency slope time series waveform calculating means, and FIG. FIG. 3A shows a waveform obtained by performing the smoothing process. FIG. 4 is a diagram showing an example of a time-series waveform of the distribution rate obtained by the distribution rate calculating means. FIG. 5 is a diagram showing an example of a fatigue curve obtained by the fatigue curve calculation means. FIG. 6 is a diagram showing an example of a time series waveform of the homeostatic function level obtained by the homeostatic function level calculating means. FIG. 7A is a diagram showing an example of a physical condition map obtained by the physical condition map calculating means, and FIG. 7B is a diagram showing an example of a sensory map obtained by the sensory map calculating means. FIG. 8 is a flowchart for explaining an estimation step by the fatigue estimator. FIGS. 9A and 9B are diagrams showing the RR interval time series data of the electrocardiogram and the inclination angle of the approximate line of the APW fluctuation waveform in Experimental Example 1. FIG. FIG. 10 is a diagram illustrating the inclination angle of the approximate line of the APW fluctuation waveform in the actual vehicle running experiment of Experimental Example 1. FIG. 11 is a diagram showing an estimation result in an actual vehicle running experiment of Experimental Example 1. FIG. 12 is a diagram showing a transition of estimation results in Experimental Example 2. FIG. 13 is a distribution diagram showing the travel time (time during which the car was driven) and the number of determinations for each operation in Experimental Example 2. FIG. 14 is a diagram showing the variation in travel time in Experimental Example 2 in units of two weeks. FIG. 15 is a diagram illustrating the number of warnings for each day of the week by counting the determination results on a monthly basis in Experimental Example 2.

  Hereinafter, the present invention will be described in more detail based on the embodiments of the present invention shown in the drawings. The biological signal collected in the present invention is sound / vibration information collected from the back (hereinafter, “back sound / vibration information”). As described above, the back sound / vibration information is sound / vibration information generated from the motion of the heart and aorta detected from the upper back of a person, and includes information on ventricular systole and diastole, blood circulation It includes elasticity information of the blood vessel wall serving as an auxiliary pump, elasticity information based on blood pressure, and reflected wave information, that is, back body surface wave (APW) and pseudo heart sound 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. Since it contains information on neural activity and information on the endocrine system, it is suitable for determining bioregulatory functional elements from different viewpoints.

  For example, a pressure sensor can be used as a biological signal measuring device for collecting a biological signal, but preferably, it is used in a drowsy driving warning device (Sleep Buster (registered trademark)) manufactured by Delta Touring Co., Ltd. The biological signal measuring apparatus 1 is used. FIG. 1 shows a schematic configuration of the biological signal measuring apparatus 1. This biological signal measuring apparatus 1 can be used by being incorporated in a driver's seat of a vehicle, and can collect a biological signal without restraining fingers.

  The biological signal measuring apparatus 1 will be briefly described. As shown in FIGS. 1A and 1B, the first layer 11, the second layer 12, and the third layer 13 are stacked in order from the upper layer side. The first layer 11 made of a layer structure and made of a three-dimensional solid knitted fabric or the like is used by being positioned on the human body side that is a detection target of a biological signal. Therefore, sound / vibration information of the heart / vascular system including biological signals from the back of the trunk of the human body, in particular, biological sounds generated by vibrations of the ventricle, the atrium, and the large blood vessels (direct trunk sound or bioacoustic signal). (Back body surface pulse wave (including APW)) is first propagated to the first layer 11 which is a biological signal input system. The second layer 12 functions as a resonance layer that emphasizes the biological signal propagated from the first layer 11, particularly the sound / vibration of the heart / vascular system by a resonance phenomenon or a beat phenomenon, and a casing 121 made of a bead foam or the like. And a three-dimensional solid knitted fabric 122 that functions as a natural vibrator, and a film 123 that generates membrane vibration. A microphone sensor 14 is provided in the second layer 12 to detect sound / vibration information. The third layer 13 is laminated on the opposite side of the first layer 11 via the second layer 12 to reduce external sound / vibration input.

  Next, the configuration of the biological state estimation apparatus 100 of the present embodiment will be described based on FIG. The biological state estimation device 100 includes a biological adjustment functional element determination unit 200 and a fatigue estimation unit 300. The biological state estimation device 100 is configured by a computer (including a microcomputer), and a computer program that executes a procedure for causing the computer to function as the biological adjustment functional element determination unit 200 and the fatigue estimation unit 300 is stored in the storage unit. Has been. In addition, the biological state estimation device 100 uses the biological adjustment function element determination unit 200, the fatigue feeling estimation unit 300, and the like as a biological adjustment element determination circuit, a fatigue feeling estimation circuit, and the like that are electronic circuits that operate according to a predetermined procedure using a computer program. It can also be configured. In the following description, it is obvious that a configuration expressed by adding “means” other than the biological adjustment functional element determination unit 200 and the fatigue estimation unit 300 can also be configured as an electronic circuit component. is there.

  The computer program may be stored in a recording medium. If this recording medium is used, the program can be installed in the computer, for example. Here, the recording medium storing the program may be a non-transitory recording medium. The non-transitory recording medium is not particularly limited, and examples thereof include a recording medium such as a flexible disk, a hard disk, a CD-ROM, an MO (magneto-optical disk), a DVD-ROM, and a memory card. It is also possible to install the program by transmitting it to the computer through a communication line.

  In the present embodiment, the biological adjustment functional element determination unit 200 analyzes the back sound / vibration information, which is a biological signal measured by the biological signal measuring device 1 described above, and is used for estimation of a basic physical condition of a person. A plurality of types of indices related to the functional elements are calculated at predetermined time intervals that are set in advance, and their time series changes are obtained.

  A plurality of types of bioregulatory function elements determined by the bioregulatory function element determining means 200 are not limited, but at least an index highly relevant to brain function and autonomic nerve function, and relation to physical / mental fatigue It is preferable that the index includes a highly sensitive index and an index highly related to sense. These are indicators that show fluctuations in EEG fluctuations that affect human homeostasis functions, or how bioregulatory functions represented by body temperature control functions work. This is because the influence is large.

  Examples of indices highly relevant to brain function and autonomic nerve function include time-series waveforms of frequency gradients obtained by processing collected biological signals, and the distribution ratios of the three signals described in the above section of the prior art. Examples include time-series waveforms, fatigue curves, and time-series fluctuations in the determination of the homeostasis maintenance function level. The time-series waveform of the frequency gradient shows fluctuations that are at the base of the regulation of the homeostatic function, and the fluctuations are well-balanced to balance the function of anti-trait, Relevance is particularly high. This is supported by statistical methods. The time-series waveform of each frequency band obtained from the distribution rate corresponds to the types of brain waves (θ waves, α waves, β waves) that are indirectly related to fluctuation rhythms, and functions in human brain function and autonomic nerve function. In addition, it is highly related to the regulatory function of the endocrine system. Since homeostasis is maintained by various regulatory systems such as the endocrine system and the autonomic nervous system, fluctuations in the level are deeply related to the regulatory performance due to fluctuations in the brain, autonomic nervous system and endocrine system.

  The bioregulatory function element determining means 200 is a frequency slope time series waveform computing means for obtaining a time series waveform of a frequency slope as a computing means for obtaining an index that captures how the fluctuations of the brain function, autonomic nervous function and endocrine system fluctuate. 210, a distribution rate calculating means 220 for obtaining a distribution rate, a fatigue curve calculating means 230 for obtaining a fatigue curve, and a constancy maintaining function level calculating means 240 for obtaining a constancy maintaining function level.

  The frequency gradient time series waveform calculation means 210 is a time series of frequencies from the time series waveform of the back body surface pulse wave (APW) near 1 Hz obtained by filtering the back sound / vibration information obtained from the sensor 14 of the biological signal measuring apparatus 1. After obtaining the waveform, the time-series waveform of the frequency is slid to calculate the frequency-gradient time-series waveform (see FIGS. 3A and 3B). The frequency gradient time series waveform calculation means 210 is a point (zero cross) that switches from positive to negative in the time series waveform of the back body surface pulse wave (APW) as disclosed in the above-mentioned Patent Document 1 by the present inventors. Point method) (zero-cross method) and a method of obtaining a time-series waveform using a maximum value (peak point) by smoothing and differentiating the time-series waveform of the back body surface pulse wave (APW) (peak detection method) There are two ways.

  In the zero cross method, 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 the individual frequency f is adopted as the value of the 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.

  In the peak detection method, the maximum value of the time series waveform of the back body surface pulse wave (APW) is obtained by, for example, the smoothing differential method using Savitzky and Golay. Next, for example, the local maximum value is divided every 5 seconds, the reciprocal of the time interval between the local maximum values of the time-series waveform included in the 5 seconds is obtained as the individual frequency f, and the average value of the individual frequency f in the 5 seconds is calculated This is adopted as the value of the 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.

  The frequency gradient time-series waveform computing means 210 has a predetermined overlap time (for example, 18 seconds) and a predetermined time width (for example, 180 seconds) from a time-series waveform of frequency fluctuations obtained by the zero cross method or the peak detection method. A time window is set, a frequency gradient is obtained for each time window by the method of least squares, and a time series waveform of the gradient is output. This slide calculation is sequentially repeated to output the APW frequency gradient time-series change as a frequency gradient time-series waveform.

  The dorsal body surface 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 artery, and the appearance information of the sympathetic nervous system and the parasympathetic nervous system. The frequency-gradient time-series waveform obtained by the zero-cross method (the waveform indicated as “0x” in FIGS. 3A and 3B) is more related to the state of control of the heart, and the appearance of the sympathetic nerve Although reflecting the state, the time-series waveform of the frequency gradient obtained by the peak detection method (the waveform indicated as “Peak” in FIGS. 3A and 3B) is more related to the heartbeat variability. Therefore, in order to grasp the state of the autonomic nervous function more clearly, it is preferable to use a frequency gradient time series waveform obtained by using the zero cross method.

Sympathetic nerve activity affects vascular elasticity and diameter, and the influence of reflected waves from the vascular wall is simulated heart sound information (detected from the back) included in sound and vibration information detected from the human back. Therefore, between a pseudo I sound (corresponding to a heart sound I sound) and a pseudo II sound (corresponding to a heart sound II sound) detected by a muscle, skin, etc. between the heart and the back surface. It is superimposed on the waveform component. This is the reason why the width between the zero cross points in the zero cross method differs from the width between the peak points in the peak detection method, and the zero cross method has a period affected by the reflected wave. Therefore, it is possible to capture sympathetic nerve information by looking at the frequency gradient time-series waveform by the zero cross method.
On the other hand, the peak value reflects the information of heart rate variability as described above, but the heart rate variability is mainly controlled by the parasympathetic nerve. Therefore, parasympathetic information can be captured by looking at the peak value.

  The frequency gradient time series waveform obtained by the zero-cross method obtained by the frequency gradient time series waveform calculation means 210 has an amplitude that increases with a temporary increase in sympathetic nerve activity that occurs as resistance to sleepiness at a predetermined timing before sleep. It is known that when the tendency to change is shown, it can be regarded as an index of the predictive sleep phenomenon (see Patent Document 4). In addition, after a waveform showing a sleep onset symptom appears, the waveform shows a tendency to converge, and after that, when the fluctuation fluctuation of a long period is larger, the point at which the fluctuation of the long period starts to show the imminent sleep immediately before falling asleep. It is known that it can be regarded as an index indicating a phenomenon.

  The distribution rate calculating means 220 first analyzes the frequency inclination time series waveforms obtained from the frequency inclination time series waveform calculating means 210, respectively, and from the above 0.0033 Hz, which is the frequency at which the fluctuation characteristics of the cardiovascular system are switched. Each frequency component belonging to the VLF band is extracted from the ULF band corresponding to the lower frequency function adjustment signal, the fatigue acceptance signal having a higher frequency than the function adjustment signal, and the activity adjustment signal having a higher frequency than the fatigue acceptance signal. Next, the distribution ratios of these frequency components are obtained in time series. That is, the ratio of each frequency component when the sum of the power spectrum values of the three frequency components is 1 is obtained as a distribution rate in time series (see FIG. 4).

  In the present embodiment, as shown in FIG. 4, a frequency component of 0.0017 Hz is used as the function adjustment signal, a frequency component of 0.0035 Hz is used as the fatigue acceptance signal, and a frequency component of 0.0053 Hz is used as the activity adjustment signal. Used. In atrial fibrillation, which is one of the heart diseases, the frequency at which the fluctuation characteristics of the heart and circulatory system are switched is said to be 0.0033 Hz. By capturing fluctuations in the vicinity of 0.0033 Hz, Information on activities and homeostasis can be obtained. In addition, the frequency band near 0.0033 Hz or less and the band near 0.0053 Hz are mainly related to body temperature regulation, and the frequency band of 0.01 to 0.04 Hz is said to be related to autonomic nerve control. Yes. And when the inventors actually obtained a frequency gradient time-series waveform for calculating these low-frequency fluctuations inherent in the biological signal and analyzed it, 0.0017 Hz, which is a lower frequency than 0.0033 Hz, It was confirmed that there were fluctuations in the frequency band centered at 0.0035 Hz in the vicinity of 0.0033 Hz and fluctuations in the frequency band centered at 0.0053 Hz in addition to these two. However, the frequency component of each signal can be adjusted by individual differences, etc., the function adjustment signal is less than 0.0033 Hz, preferably 0.001 to 0.0027 Hz, and the fatigue acceptance signal is 0. In the range of 002 to 0.0052 Hz, the activity adjustment signal can be adjusted and used in the range of 0.004 to 0.007 Hz.

  As shown in Patent Document 2, the time series change of the distribution ratio obtained by the distribution ratio calculation means 220 is, for example, a rapid decrease in the distribution ratio of 0.0017 Hz and a rapid increase in the distribution ratio of 0.0053 Hz. The time point showing the change can be regarded as the present time point of the impending sleep phenomenon.

  The fatigue curve calculation means 230 is a means disclosed in the Japanese Patent Application Laid-Open No. 2009-22610 by the present inventors, calculates an integral value by performing absolute value processing on a frequency gradient time series waveform obtained by the zero cross method, This integrated value is a means for obtaining the fatigue curve as shown in FIG. 5 by obtaining the integrated value as the degree of fatigue at every predetermined judgment time and plotting it corresponding to the time. The muscle activity is muscle contraction or relaxation, and the fatigue curve, which is the integrated information of the frequency gradient time series waveform by the zero cross method reflecting the information of the sympathetic nerve, is highly correlated with the muscle activity (Non-patent Document 1). reference). Therefore, in the fatigue curve, the point where the slope fluctuates more than a predetermined value indicates a singular point, and each singular point corresponds to the increasing fatigue and the point indicating that muscle activity has occurred and the blood flow increased. Shows the point.

  The homeostasis maintenance function level determination unit 240 is based on the technique disclosed in Patent Document 3, and the positive / negative of the differential waveform of each frequency gradient time series waveform obtained by the frequency gradient time series waveform calculation unit 210 using the zero cross method. At the time of each frequency slope obtained by absolute value processing of the positive and negative of the integral waveform obtained by integrating the frequency slope time series waveform, the frequency slope time series waveform using the zero cross method and the frequency slope time series waveform using the peak detection method. The determination is made using at least one of the absolute values of the series waveform. Based on these combinations, the level of the homeostasis maintenance function is determined. For example, using a frequency gradient and an integral value, if the value is equal to or greater than a predetermined value, it is determined as “constancy maintaining function level 1”, or the differential value is equal to or less than a predetermined position and “ In the case of “peak advantage”, it can be set to be determined as “constancy maintaining function level 4”. And, for example, when the above conditions are variously combined and correlated with the state of the person and determined to be level 1 to 3, attention should be paid to the case of normal to good state, level 4 to 6. Is determined to be necessary. Further, when it is determined that a sign of sleep onset or a sign of imminent sleep has occurred, an index such as levels 7 to 11 is assigned as a level that requires immediate warning depending on the state. In the product name “Sleep Buster” manufactured by Delta Touring Co., Ltd., the determination result by the homeostasis maintenance function level determination means 240 is set to be displayed as shown in FIG. 6, for example.

  As an index highly relevant to physical / mental fatigue, a physical condition map and a sensory map, which are indices disclosed in Patent Document 2, can be used. These are graphs showing how fluctuations fluctuate, and show high relevance to human physical and mental fatigue.

  Therefore, the living body adjustment functional element determination unit 200 of the present embodiment further includes a physical condition map calculation unit 250 and a sensory map calculation unit 260. The back body surface pulse wave (APW) obtained from the back sound / vibration information acquired from the biological signal measuring apparatus 1 is frequency-analyzed, and the analysis waveform is displayed on the logarithmic axis display for the target analysis section. Divided into low frequency band, medium frequency band, and high frequency band, the analysis waveform is scored based on a certain standard from the slope of the analysis waveform and the shape of the entire analysis waveform, and plotted on the coordinate axis It is. The physical condition map is a state in which the state of control of the autonomic nervous system is viewed as a balance between the sympathetic nerve and the parasympathetic nerve, and the sensory map is obtained by superimposing the state of change of heartbeat variability on the physical condition map.

  Specifically, the physical condition map calculation unit 250 first obtains a regression line for each predetermined period region with respect to an analysis waveform obtained by frequency analysis of the back body surface pulse wave. Next, each regression line obtained for each periodic area is given a region score based on its slope, and the difference in power spectral density value between regression lines in the adjacent frequency domain and the difference in slope between regression lines Based on the above, the number of break points indicating the branching phenomenon in each regression line is obtained, a shape score based on the number of break points is given, and at least one of the area score and the shape score is used to determine a determination reference point for each analysis waveform. Ask. As the area score, the slope of each regression line in each area is divided into three, approximately horizontal state, upward and downward, for example, increasing or decreasing the score in the upward direction and downward direction on the basis of the score in the approximately horizontal state To give a score. As the shape score, the smaller the number of break points, the higher the score.

  When obtaining the determination reference point, the first determination reference point is obtained using the frequency slope time series waveform obtained by the zero cross method, and the second judgment reference is obtained using the frequency slope time series waveform obtained by the peak detection method. Find a point. Then, the coordinate point is plotted with the index based on the first determination reference point on one axis and the index based on the second determination reference point on the other axis, and the physical condition map as illustrated in FIG. Is created. In the physical condition map, the coordinate time-series change line connecting the coordinate points is determined to be comfortable when it is determined that the change trend approximates the 1 / f slope, and is determined to be changing vertically. If it is determined that it is uncomfortable. In FIG. 7A, a plurality of coordinate points are connected without being aligned with the coordinate origin, but when the change tendency of two points that are temporally different is viewed, the first point is aligned with the coordinate origin and the second point is When plotted in the fourth quadrant, this bioregulatory functional element is “good”, and determination becomes easier.

  The sensory map calculation means 260 performs a movement calculation to obtain an average value of the frequency for each predetermined time window set with a predetermined overlap time in the time series waveform of the frequency using the peak detection method related to the heartbeat fluctuation, The time series change of the average value of the frequency obtained for each time window is obtained as a frequency fluctuation time series waveform, and an index corresponding to the functional point obtained from the time series waveform of the frequency using the zero cross method is taken on one axis. In addition, an index corresponding to the amount of change in the predetermined time width of the frequency fluctuation time series waveform obtained by the peak detection method is taken on the other axis, and the time series change of the coordinates obtained from the functional point and the amount of change is obtained. It is a way to go. FIG. 7B is an example of the sensory map obtained in this way. In FIG. 7B, a plurality of coordinate points are connected without being aligned with the coordinate origin. However, when the change tendency of two points that are different in time is observed, the first point is aligned with the coordinate origin and 2 points are obtained. When the eyes are plotted, it becomes easy to determine the separation distance and the separation direction between the two.

In addition, the functional point is between the determination reference points of the analysis waveform in the two time ranges before and after the comparison target:
Function point = Judgment reference point of the later time range + (Judgment reference point of the later time range-Judgment reference point of the previous time range) x n (where n is a correction coefficient)
Is required.

  As an index highly relevant to the sense, among the time-series changes of the homeostatic function level determined by the homeostatic function level determination unit 240, for example, normal to good using a frequency slope and an integral value It can be determined using the frequency of the frequency (levels 1 to 3 in the above example) and the level of attention (levels 4 to 6 in the above example). The homeostasis maintenance function level is highly related to the state of autonomic nervous function as described above, but when sympathetic decompensation acts on fatigue due to physical condition, basic physical strength, or motivation, fatigue There are times when you feel a feeling and sometimes you don't feel it. Therefore, the sympathetic compensation effect on fatigue and the basic physical condition are highly related to the sense of feeling it as fatigue. In addition, the sense here is a sensation similar to feeling of loss accompanied by fatigue or hypoxia.

  The fatigue sensation estimation means 300 calculates the basic physical condition of the person to be analyzed (based on a predetermined standard) from each time series change regarding the fluctuation performance of each biological adjustment functional element obtained by the biological adjustment functional element determination means 200 described above. It is a means to estimate the basic physical condition). In the biological adjustment functional element determination unit 200, as described above, a plurality of types of time-series changes regarding the fluctuation performance of the biological adjustment functional elements are obtained. It is obtained every predetermined determination time. For example, the frequency gradient time-series waveform calculation unit 210 takes several minutes after obtaining the data from the biological signal measuring apparatus 1 until the first calculation result is output, but thereafter, for example, is obtained every 18 seconds. As a result, a time series change is required. The distribution rate obtained by the distribution rate computation means 220, the fatigue degree obtained by the fatigue curve computation means 230, and the homeostasis maintenance function level obtained by the homeostasis maintenance function level computation means 240 are also numbers until the first computation result is output. It takes minutes, and is obtained every 18 seconds, for example, and a time series change is obtained for each. The calculation results obtained by the physical condition map calculation means 250 and the sensation map calculation means 260 each take 20 to 30 minutes at first, but the second point is obtained about ten minutes later, and the third and subsequent points are obtained every few minutes. . On the other hand, the basic physical condition estimation means 300 estimates the basic physical condition for a time (basic physical condition estimation time) that is longer than each determination time in each bioregulatory function element.

  The fatigue sensation estimation means 300 combines the determination results of the respective bioregulatory functional elements to estimate whether or not the state of fatigue feeling masking has occurred. The fatigue estimation means 300 includes first estimation means 310, second estimation means 320, and third estimation means 330. FIG. 8 is a flowchart showing an estimation procedure by the fatigue feeling estimation means 300. As shown in this figure, the first estimating means 310 masks the feeling of fatigue when the time-series change of the index highly relevant to the autonomic nervous function for the biological signal to be analyzed satisfies a predetermined criterion. Is estimated (category 1) (S110). The second estimation means 320 is a time series of indicators highly relevant to physical / mental fatigue for the biological signal to be analyzed that is not estimated as a state (category 1) in which the first estimation means 310 has masked fatigue. When the change satisfies a predetermined standard, it is estimated that the state (category 3) in which the masking of fatigue does not occur (S120). The third estimation unit 330 generates at least masking of fatigue based on a predetermined criterion for a biological signal to be analyzed that is not estimated that the fatigue estimation masking has not occurred in the second estimation unit 320 (category 3). It is classified into either the state (category 1) in which the masking of fatigue is not occurring (category 3) (S130).

  That is, a state in which a biological signal to be analyzed is discriminated using at least three types of bioregulatory functional elements, and a feeling of fatigue that makes it difficult to recognize fatigue due to a narrow field of view is performed (category 1). The state (category 3) in which the masking of the feeling of fatigue does not occur in a resting state is estimated. In addition, it is also preferable to estimate a blurred state (category 2) at an intermediate level between the two.

  More specifically, in the first estimation means 310, the above-described frequency gradient time series waveform, distribution rate, fatigue, which is an index relating to a homeostasis maintenance function based on fluctuations in brain function / autonomic nerve function or endocrine system regulation function A curve (fatigue level) is used. The homeostasis maintenance function level is an index that makes it easy to discriminate signs resulting from the accumulation of fatigue, such as a sign of sleep onset, imminent sleep, and low-altitude driving. In addition, fluctuations in the homeostatic function that is regulated by brain function vary depending on the frequency band, and the regulated system governed by the endocrine system, etc., but the distribution rate among these is the regulation system. It is an index that reflects well when there is sudden change, decay and increase. Therefore, when these are used, it is possible to grasp the situation that requires attention and warning when performing operations such as driving, and when the frequency of sudden change in the distribution rate exceeds a predetermined value, the field of view is narrow for things. It is presumed that the state (category 1) in which the masking of the feeling of fatigue that makes it difficult to recognize the fatigue is occurring.

  The physical condition map / sensory map, which is an index highly related to physical / mental fatigue, can easily distinguish the index when feeling well and comfortable. Therefore, the second estimation means 320 uses this index to extract a state (category 3) in which the user does not mask his / her fatigue feeling in a resting state in the case of a condition indicating goodness and comfort. Whether to use an index highly relevant to brain function, autonomic nervous function, or endocrine regulatory function, or an index highly relevant to physical / mental fatigue, the modulation of physical / mental fatigue Since there is basically a relationship with the autonomic nerve, the first estimation unit 310 performs the estimation using the index highly relevant to the brain function, the autonomic nerve function, and the like as in the present embodiment. In addition, it is preferable that the second estimation unit 320 performs estimation using an index highly relevant to physical / mental fatigue.

  In the present embodiment, as described above, the brain function, the autonomic nervous function, and the endocrine function are highly relevant to the hormone secretion regulation function. As described above, the frequency gradient time series waveform, distribution rate, fatigue curve (fatigue degree), and constant There are four sex maintenance function levels. Of these, it is possible to set to estimate “category 1” when a sign such as a sleep onset sign is detected more than a predetermined value, but when a predetermined sign is detected in a plurality of indicators, , “Category 1” is preferable because it increases the reliability.

  Therefore, for example, the first estimating means 310 obtains the frequency slope time series waveform obtained from the frequency slope time series waveform computing means 210, the time series waveform of the distribution rate obtained from the distribution rate computing means 220, and the fatigue curve computing means 230. When three or more indices satisfy a predetermined standard among the fatigue curve (time series waveform of fatigue level) and the homeostasis maintenance function level obtained from the homeostasis maintenance function level calculation means 240 (in S110 of FIG. 8) When “Yes” is determined, it can be set to be estimated as “category 1” (S111 in FIG. 8).

In the present embodiment, the predetermined standard estimated as “category 1” is set as follows.
(A) Index obtained from frequency gradient time-series waveform calculation means 210 In the frequency gradient time-series waveform using the zero cross method, amplitude changes are compared, and 1 is obtained continuously several times (usually set in a range of 2 to 4 times). When a convergence point that changes to less than 90 to 60% of the previous amplitude occurs (an index that estimates that sympathetic nerve activity has declined and the patient cannot fall asleep)
(B) Index obtained from the distribution rate calculation means 220 In the time series change of the distribution rate of the frequency gradient time series waveform using the zero cross method, 0 in a predetermined time range (usually set in the range of 60 to 120 seconds). When the distribution rate at .0017 Hz suddenly decreases (usually set at a reduction rate of 15% or more) and the distribution rate at 0.0053 Hz suddenly increases (usually set at an increase rate of 15% or more) during this period ( Estimated metrics)
(C) Index obtained from fatigue curve calculation means 230 Fatigue curve using peak detection method (frequency gradient time series waveform using peak detection method) during a predetermined time (usually set in a range of 3 to 10 minutes) Where the slope of the absolute value integration time series waveform) changes more significantly than the slope of the fatigue curve using the zero-cross method (frequency slope time series waveform integration of the frequency slope time-series waveform using the zero-cross method) When the fatigue curve using the peak detection method reaches a predetermined value or more when a predetermined time elapses (an index that estimates that the parasympathetic nerve activity is extremely dominant)
(D) Index obtained from the homeostasis maintenance function level calculating means 240 Of the homeostasis maintenance function levels obtained every 18 seconds, a level lower than the normal level, in the above example, several times of caution determination of levels 4-6 Appearing more than a dozen times (an indicator that appears when parasympathetic activity is presumed to be dominant), or a level that requires a warning, in the above example, a warning judgment of level 7 to 11 appears several times (An index that appears when a sudden increase or a drastic decrease in sympathetic nerve activity is estimated)

  The second estimation unit 320 relates to data that is not estimated as “category 1” in three or more of the indicators (a) to (d) (data determined as “No” in S110 of FIG. 8). Then, it is determined whether or not a predetermined standard is satisfied by using the indices of the physical condition map calculating unit 250 and the sensory map calculating unit 260 (S120 in FIG. 8), and if the predetermined standard is satisfied, it is estimated that the category corresponds to “Category 3”. (S121 in FIG. 8).

(E) Index when it is determined as “Category 3” In this embodiment, the time series change obtained from the physical condition map calculation means 250 is a point at which the previous calculation result is output (as described above, 1 When the second and second points are output after a predetermined time has passed, but the third and subsequent points are output every few minutes), the next point is plotted in the fourth quadrant, In addition, when the time-series change obtained from the sensory map calculation means 260 is plotted at a predetermined distance or more in the X-axis direction when the previous point is also set to the coordinate origin, “Category 3” It is set to be estimated (in the case of “Yes” in S120 of FIG. 8, S121).

  Note that the determination criteria (a) to (d) estimated as “Category 1” and the determination criterion (e) estimated as “Category 3” are based on statistical analysis of a large number of cases. However, the present invention is not limited to this. For example, data may be accumulated for each individual, and the condition may be set statistically for each individual.

  The third estimation unit 330 determines that the data to be estimated does not satisfy any of the criteria of “category 1” in the first estimation unit 310 and “category 3” in the second estimation unit (“No” in S110 of FIG. 8). And when it is determined “No” in S120) (S130 in FIG. 8). The third estimation means 330 is an index of a level that can be said to be good from normal to good with sympathetic nerve activity among time-series changes in the homeostasis maintenance function level obtained by the homeostasis maintenance function level determination means 240 (in the above example, Levels 1 to 3) are compared with the frequency of appearance of the level near the boundary between the indicators of parasympathetic nerve activity and the level requiring attention (in the above example, levels 4 to 6). However, since the difference in state is small at one level difference, it is preferable to compare at two or more different levels. In the present embodiment, the ratio of the appearance frequency of the level 2 index in the middle of the levels of normal to good level and the level 4 index that needs to be noted is compared. Basically, it can be estimated as “Category 3” when the appearance frequency of level 2 showing good state with superior sympathetic nerve activity is high and the appearance frequency of level 4 showing attention state with low parasympathetic nerve activity is low. When the appearance frequency is reversed, it can be estimated as “category 1”, but the data to be analyzed by the third estimating means 330 is clearly “category 1” in the first estimating means 310 and the second estimating means 320. ”And“ Category 3 ”are not estimated, and data that is difficult to classify is assumed. Therefore, in the present invention, a large number of cases are analyzed, and a criterion for classification into “category 1” and “category 3” and “category 2” which is an intermediate state thereof is set by a Bayesian estimation method (see FIG. 8 S131).

  According to the present embodiment, for example, when analyzing the state of a person, the fatigue estimation means 300 analyzes a combination of a plurality of biological regulation functional elements. Due to the use of multiple element criteria, there is a narrow field of view for things that cause fatigue masking that makes it difficult to recognize fatigue (category 1), and that there is masking of your own fatigue in a resting state No state (category 3), and a blurry state (category 2) at an intermediate level between them can be estimated appropriately.

  Further, the fatigue estimator 300 of the present embodiment first extracts the “category 1” state by the first estimator 310 for the biological signal data to be analyzed, and only the data that has not been extracted by the first estimator 310. The second estimation unit 320 extracts the state of “category 3”, and the third estimation unit 330 determines only the data that has not been extracted by either the first estimation unit 310 or the second estimation unit 320. It becomes a target. That is, the first estimation means 310 and the second estimation means 320 first extract only those clearly corresponding to “category 1” and “category 3”, and then only the remaining analysis target data is the third estimation means. 330 is processing. Since a plurality of bioregulatory functional elements are used as the determination elements, the calculation means of the computer that processes each data by narrowing down the extracted data in each of the estimation means 310 to 330 and determining in stages in this way, The exchange and calculation of data with the storage unit for storing the data received from the biological signal measuring device can be simplified, and the processing of the arithmetic processing device can be made more efficient. When each of the estimation means 310 to 330 analyzes at one time based on the respective criteria, further processing is required depending on the difference in the analysis result, and there is a concern that the calculation processing becomes complicated and the calculation speed is slowed down. .

(Experimental example 1)
A: Electroencephalogram measurement experiment (1) Experimental method The frontal polar electroencephalogram is measured in order to confirm the action of the prefrontal cortex that forms feelings such as achievement. Measurement sites are Fp1, Fp2, A1, A2 and eye movement based on the International 10-20 method. This was measured under two conditions: a state in which the subject watched a movie that the subject wanted to watch again in the past, and a seated seat without stress. All were seated on a car seat in the laboratory and measured for 60 minutes. The measurement items include an electrocardiogram, a back body surface pulse wave (APW: a waveform in the vicinity of 1 Hz obtained by filtering back sound / vibration information obtained from the sensor 14 of the biological signal measuring device 1), and a fingertip volume pulse wave in addition to the electroencephalogram. . The test subjects were three healthy Japanese men aged 24, 25 and 30.

(2) Experimental result Frequency analysis is performed on the time series data of the RR interval of the electrocardiogram and the frequency gradient time series waveform of the APW to obtain a fluctuation waveform indicating the relationship between the logarithmic power spectrum density and the logarithmic frequency. The angle of inclination was determined. The results are shown in FIGS. 9 (a) and 9 (b). From this figure, it can be seen that the fluctuation of the RR interval of the electrocardiogram and the fluctuation tendency of the APW frequency gradient time series waveform coincide. In other words, the slopes of both the electrocardiogram and the APW are close to -1 at rest, and the slope is tighter than -1 when watching a movie.
Therefore, from this assumption, it can be understood that the measurement using the APW is effective for estimating the state of the person. In addition, the frequency band used for the frequency analysis was set to a range of 0.01 to 0.03 Hz. The degree of appearance of sympathetic nerve activity and parasympathetic nerve activity is said to appear in the frequency band of 0.01 to 0.04 Hz, but in order to avoid noise as much as possible and eliminate variation, data up to 0.03 Hz is used. It is preferable to use it.

  Next, the biological state is estimated by the biological state estimation device 100. With respect to the waveform of the APW analysis target section of each subject, the above-described category 1, category 2, and category 3 were estimated. Moreover, the content rate of the beta wave of the electroencephalogram at the timing of the analysis target section where each category was estimated was measured. In the electroencephalogram analysis, in order to remove artifacts such as blinking and body movement, the section where the amplitude of Fp1 or Fp2 is equal to or greater than a certain value is excluded from the analysis target. The frequency band of 14 to 30 Hz of Fp1 and Fp2 is obtained for each analysis target section as a prefrontal β wave, the average value of the prefrontal β wave content of Fp1 and Fp2 is calculated, and the number of analysis target sections higher than the average value The number of analysis target sections below the average value was obtained by category. The results are shown in the following table. According to Fisher's exact test, p = 0.017, indicating a correlation.

  From Table 1, it can be seen that when the β wave content rate is high, the prefrontal cortex is activated, and this has a high correlation with category 1.

B: Actual vehicle running experiment (1) Experimental method Fabric carrying astaxanthin known to show physiologically active effects such as blood flow improvement and muscle fatigue improvement (AX carrying fabric), fabric not carrying astaxanthin ( AX unsupported fabric) is prepared in advance. The actual vehicle running experiment is performed with respect to three types of cases where the AX carrying fabric is stuck on both shoulders of the subject, the non-AX carrying fabric is stuck on both shoulders of the subject, and nothing is stuck. However, the experiment was conducted by telling the test subject that both were AX-supported fabric and AX-unsupported fabric, both of which were supported by astaxanthin. The test subject was a truck driver and was a 35-year-old, 43-year-old, and 45-year-old Japanese male. The measurement item is the back body surface pulse wave (APW: waveform in the vicinity of 1 Hz obtained by filtering the back sound / vibration information obtained from the sensor 14 of the biological signal measuring device 1).

(2) Experimental results Frequency analysis was performed on the APW frequency gradient time series waveform to obtain a fluctuation waveform indicating the relationship between the logarithmic power spectrum density and the logarithmic frequency, and an approximation line was drawn for the fluctuation waveform to obtain the inclination angle. Note that the frequency band of the APW frequency gradient time series waveform used for frequency analysis was in the range of 0.01 to 0.03 Hz, as described above. The result is shown in FIG. As shown in FIG. 10, the average value of the slope when the AX carrying fabric is affixed is −1.61, and the average value of the slope when the fabric without the AX is affixed is −0.97. The average value of the slope when not sticking was -1.00. When the AX-supported fabric is affixed, it can be said that due to the effect of astaxanthin, the concentration tends to increase and the field of view tends to narrow, as in the case of the above-mentioned movie viewing. On the other hand, when the fabric without AX carrying was affixed and when nothing was affixed, it can be considered that the subject was relaxed without feeling great stress because the test subject was a professional driver. Therefore, it is thought that astaxanthin has an effect of enhancing the action of masking the feeling of fatigue.

  FIG. 11 shows the number of times of category 1 determination by the biological state estimation device 100 of the present invention when the AX-supported fabric is affixed, when the fabric without AX is affixed, and when nothing is affixed. .

  In all cases, the number of determinations of category 1 increased as the operating time became longer, but when AX carrying fabric was affixed in the order of the smaller number of determinations, when AX carrying no fabric was affixed, and nothing was affixed It has become a case. That is, the change in the number of determinations in category 1 indicates the degree of accumulation in the actual fatigue process, but as shown in Table 1, the determination in category 1 correlates with the content of the prefrontal β wave. Therefore, it is considered that as the degree of increase with the passage of time is smaller, the degree of masking so as not to feel fatigue is increased by the action of the prefrontal cortex. Therefore, the degree of masking when the AX carrying fabric is affixed is the highest. It is considered that the effect of the placebo effect by the fact that it was determined that the fabric with astaxanthin supported in advance was determined that the degree of masking was higher when the fabric without AX was affixed than when no fabric was affixed It is done. Therefore, it can be seen that the presence / absence of fatigue masking can be estimated by the fatigue estimator 300 by the biological state estimating apparatus 100 of the present invention.

(Experimental example 2)
The biological signal measuring device 1 of the above-described embodiment (sleeping buster (registered trademark) manufactured by Delta Touring Co., Ltd.) was mounted on a truck seat, and biological signals were measured for a plurality of occupation drivers. Among them, the data of the biosignal measuring device 1 was extracted from the vehicle of a male driver in the 40s who caused the accident for about 2 years, a total of 539 operations.The data was 21 months from the 2 months before the accident occurred. The driver is working at night, and the operation route is almost the same every time.From the extracted data, the fatigue condition estimation means 300 uses the biological state estimation device 100 of the above embodiment. 3 categories (the state in which fatigue is difficult to recognize due to the masking effect of fatigue (category 1), the state in which fatigue is not masked and the fatigue can be recognized (category 1) Li 3), and were calculated estimation results of those intermediate states (Category 2)) every 5 minutes.

  FIG. 12 is a matrix of estimation results from 45 days before the accident occurrence date (“0 day” in the figure, occurrence time “6 o'clock”) to 8 days after the occurrence. In the normal state before the accident occurred, there was a moderate change in judgment, and it was found that there was fluctuation in the physical condition change. On the other hand, in the week of the accident occurrence day, it was found that there were many states (category 1) in which it was difficult to notice fatigue, the change in judgment was small, and fluctuation was reduced. The week after the accident occurred again with fluctuations. Therefore, it was shown that the body was not able to cope with changes in physical condition such as fatigue and sleepiness on the day of the accident. Therefore, in the case of the above pattern, it was suggested that the physical condition might cause an accident.

  Fig. 13 shows the travel time for each service (time during which the car was driven) and the number of each judgment in the distribution chart, which summarizes the data of 2 months before the accident and 2 weeks immediately before the accident. is there. Two months before the accident, the running time during one operation remained unchanged, and there was fluctuation in the number of judgments for each operation. Although there are many cases, the number of judgments tends to be close to a positive correlation. In other words, the driver immediately before the accident was in a state where he could not cope with the change as indicated by the rapid increase in the number of judgments. Also, it can be seen from FIG. 14 that the variation in travel time during one operation was on the increase immediately before the accident. In FIG. 12, fluctuations tended to decrease when the accident occurred, suggesting the possibility that some sort of poor physical condition had occurred immediately before the accident due to variations in travel time.

  For reference, the result of giving advice to the driver once a week will be described. In the product name “Sleep Buster” manufactured by Delta Touring Co., Ltd., the determination result by the homeostasis maintenance function level determination means 240 is set to be displayed as shown in FIG. The result determined by the homeostasis maintenance function level determination means 240 was used. FIG. 15 summarizes the determination results in units of months and calculates an average value of the number of frequent warning determinations for each day of the week. In the figure, the bar graph on the left shows the average value of the frequency of warnings, the line graph shows the operating days, and the right shows the actual values for each operation. In the month following the accident, the number of warnings on Sunday or Monday after the holiday was high, and it tended to be the same or less from the middle to the second half of the week. From these things, it is thought that the way of spending holidays is important in the night work type. The night shift at the end of the holiday seems to have increased the number of warnings at the end of the holiday because the body is less likely to switch from the holiday mode to the work mode. Therefore, advice was given to pay attention to how to spend holidays. As a result, after 13 months and 21 months, there was a change in the tendency that the days after the holidays were less than other days of the week. Through this interview, it seems that he was ready for his day off work. As a result, it is considered that an accident prevention reduction effect is obtained by a weekly interview, and as a result, no accidents can be continued for more than two years.

  The physical condition estimation method calculated from the back body surface pulse wave was applied to professional drivers, and the tendency of physical condition change by interviews immediately before and after the accident for 2 years was observed and the effect was verified. As a result, it is speculated that the change of vacation and work, such as the effects of travel time variations and night work after holidays, will affect the physical condition, and attention and improvement through interviews are useful for reducing accidents. I can say that.

DESCRIPTION OF SYMBOLS 1 Biosignal measuring apparatus 11 Core pad 12 Spacer pad 14 Sensor 100 Living body state estimation apparatus 200 Biological regulation functional element determination means 210 Frequency inclination time series waveform calculation means 220 Distribution rate calculation means 230 Fatigue curve calculation means 240 Constancy maintenance function level calculation means 250 physical condition map calculation means 260 sensory map calculation means 300 fatigue estimation means 310 first estimation means 320 second estimation means 330 third estimation means

Claims (7)

A biological state estimation device that estimates a biological state using a biological signal obtained from a biological signal measurement device that is in contact with a human back,
A bioregulatory function element determination unit that analyzes the biosignal and obtains a plurality of indices related to the bioregulatory function, including an index that is highly related to autonomic nervous function, physical / mental fatigue, or sensation;
Fatigue sensation estimating means for estimating whether or not the biological condition is a state where fatigue masking is occurring by combining a plurality of indices obtained by the biological adjustment functional element determination means. A biological state estimating device. The fatigue estimation means is
First estimation means that estimates that the state of masking of fatigue occurs when a time-series change of an index highly relevant to the autonomic nervous function of the biological signal to be analyzed satisfies a predetermined criterion When,
With respect to the biological signal to be analyzed that is not estimated to be a state in which the first estimator masks the feeling of fatigue, the time series change of the index highly relevant to the physical / mental fatigue satisfies a predetermined criterion A second estimating means for estimating that no fatigue masking has occurred,
The biological signal to be analyzed that is not estimated to be in a state in which no fatigue masking has occurred in the second estimating means, based on a predetermined criterion, at least the state in which the fatigue feeling masking has occurred and the fatigue feeling The biological state estimation apparatus according to claim 1, further comprising: a third estimation unit that classifies the mask into any state where masking has not occurred. A biological state estimation method for estimating a biological state using a biological signal obtained from a biological signal measurement device in contact with a human back,
Analyzing the biological signal, to obtain a plurality of indicators involved in the bioregulatory function, including indicators highly related to autonomic nervous function, physical / mental fatigue or sensation,
A biological state estimation method, comprising: combining the plurality of indices to estimate whether or not the biological state is a state in which fatigue masking has occurred. A first estimation procedure for estimating a state in which the masking of the feeling of fatigue occurs when a time-series change of an index highly relevant to the autonomic nervous function of the biological signal to be analyzed satisfies a predetermined criterion Carried out
With respect to the biological signal to be analyzed that is not estimated to be in a state where the masking of the feeling of fatigue has occurred by the first estimation procedure, the time series change of the index highly related to the physical / mental fatigue satisfies a predetermined criterion The second estimation procedure for estimating that no fatigue masking has occurred,
Based on a predetermined standard, at least the state of the fatigue feeling masking and the fatigue feeling of the biological signal to be analyzed that is not estimated as the state where the fatigue feeling masking does not occur in the second estimation procedure Perform a third estimation procedure to classify any of the states where masking has not occurred,
The biological state estimation method according to claim 3, wherein it is estimated whether or not the fatigue feeling is masked. A computer program for causing a computer as a biological state estimation device to analyze a biological signal obtained from a biological signal measurement device in contact with a human back and to execute a procedure for estimating a biological state,
Analyzing the biological signal, and obtaining a plurality of indicators involved in the biological regulation function, including an indicator highly related to autonomic nervous function, physical / mental fatigue or sensation;
A computer program that executes a procedure of combining the plurality of indices to estimate whether or not the biological state is a state in which the feeling of fatigue is masked and the state of masking of the feeling of fatigue has occurred. In the procedure for estimating whether or not the fatigue masking is occurring,
A first estimation procedure for estimating a state in which the masking of the feeling of fatigue occurs when a time-series change of an index highly relevant to the autonomic nervous function of the biological signal to be analyzed satisfies a predetermined criterion And execute
With respect to the biological signal to be analyzed that is not estimated to be in a state where the masking of the feeling of fatigue has occurred by the first estimation procedure, the time series change of the index highly related to the physical / mental fatigue satisfies a predetermined criterion A second estimation procedure for estimating that no fatigue masking has occurred,
Based on a predetermined standard, at least the state of the fatigue feeling masking and the fatigue feeling of the biological signal to be analyzed that is not estimated as the state where the fatigue feeling masking does not occur in the second estimation procedure The computer program according to claim 5, wherein a third estimation procedure for classifying the state into any state where masking has not occurred is executed.   The computer program according to claim 5 or 6, which causes a computer as a biological state estimation device to analyze a biological signal obtained from a biological signal measurement device that is in contact with a human back and to perform a procedure for estimating a biological state. A recorded computer-readable recording medium.