Physiological state estimation system, physiological state estimation method, and physiological state estimation program product

Abstract

Provided are a physiological state estimation system, a physiological state estimation method, and a physiological state estimation program product for estimating a physiological state with high accuracy. A physiological state estimation system is provided with: a waveform information acquisition unit that acquires brain waves and pulse waves; a filter unit that filters the acquired brain waves and pulse waves in at least one predetermined frequency band; a conversion unit that performs Hilbert conversion on the filtered brain waves and pulse waves in the frequency band; an instantaneous value calculation unit that calculates an instantaneous value including an instantaneous logarithmic amplitude corresponding to a logarithm of an amplitude term of the complex waveform formula after Hilbert transformation and an instantaneous frequency corresponding to a time differential value of a phase term of the complex waveform formula; a distribution calculation unit that calculates a probability density distribution of the instantaneous value, and calculates a feature quantity including an average value and a variance value of the probability density distribution; an extraction unit that extracts an instantaneous value when a predetermined condition is satisfied; and a physiological state estimation unit that estimates a physiological state on the basis of the extracted feature amount of the instantaneous value.

Description

Technical Field

[0001] The present invention relates to a physiological state estimation system, a physiological state estimation method, and a physiological state estimation program. Background Art

[0002] In Patent Document 1, there is described a biological signal analysis method that calculates a feature quantity by assuming chaos based on biological signals such as electroencephalogram, pulse wave, and myoelectricity and using a Lyapunov exponent to obtain quantitative information of an organism. In the biological signal analysis method of Patent Document 1, the feature quantity is calculated at high speed by using a high-frequency region.

[0003] Patent Document 1: Japanese Patent Application Laid-Open No. 2002-017687

[0004] Non-Patent Document 1: W Zong, T Heldt, GB Moody and RG Mark, “An Open-source Algorithm to Detect Onset of Arterial Blood Pressure Pulses”, Computers in Cardiology 2003; 30:259-262. [online], [searched on December 14, 2021], Internet <https: / / lcp.mit.edu / pdf / Zong03a.pdf> Summary of the Invention

[0005] There is a desire for a method for more accurately estimating a physiological state based on biological signals.

[0006] The present invention has been completed to solve such a problem, and provides a physiological state estimation system, a physiological state estimation method, and a physiological state estimation program for accurately estimating a physiological state.

[0007] The physiological state estimation system according to this embodiment includes: a waveform information acquisition unit that acquires electroencephalogram and pulse wave; a filtering unit that filters the acquired electroencephalogram and pulse wave in at least one predetermined frequency band; a transformation unit that performs Hilbert transformation on the electroencephalogram and pulse wave in the filtered frequency band; an instantaneous value calculation unit that calculates instantaneous values including instantaneous logarithmic amplitude and instantaneous frequency, where the instantaneous logarithmic amplitude corresponds to the logarithm of the amplitude term of the complex waveform formula after Hilbert transformation, and the instantaneous frequency corresponds to the time derivative of the phase term of the complex waveform formula; a distribution calculation unit that calculates the probability density distribution of the instantaneous values and calculates characteristic quantities including the average value and variance value of the probability density distribution; an extraction unit that sets the average value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram as μ e and sets the variance value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram as σ e and sets the average value of the instantaneous frequency in the predetermined frequency band of the pulse wave as μ b and sets the variance value of the instantaneous frequency in the predetermined frequency band of the pulse wave as σ b and when setting a parameter of an integer of 3 or more as η, extracts the instantaneous values that satisfy |μ e - μ b | < (σ e + σ b ) / η; a physiological state estimation unit that estimates the physiological state based on the characteristic quantities of the extracted instantaneous values. With such a structure, the physiological state can be estimated with high accuracy.

[0008] In the above physiological state estimation system, the parameter η can also be 3. With such a structure, the physiological state can be estimated with higher accuracy.

[0009] The physiological state estimation method according to this embodiment includes: a waveform information acquisition step of acquiring an electroencephalogram and a pulse wave; a filtering step of filtering the acquired electroencephalogram and pulse wave in at least one predetermined frequency band; a transformation step of performing a Hilbert transform on the electroencephalogram and pulse wave in the filtered frequency band; an instantaneous value calculation step of calculating an instantaneous value including an instantaneous logarithmic amplitude and an instantaneous frequency, where the instantaneous logarithmic amplitude corresponds to the logarithm of the amplitude term of the complex waveform formula after the Hilbert transform, and the instantaneous frequency corresponds to the time derivative of the phase term of the complex waveform formula; a distribution calculation step of calculating the probability density distribution of the instantaneous value and calculating a feature quantity including the average value and variance value of the probability density distribution; an extraction step of setting the average value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram as μ e and setting the variance value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram as σ e and setting the average value of the instantaneous frequency in the predetermined frequency band of the pulse wave as μ b and setting the variance value of the instantaneous frequency in the predetermined frequency band of the pulse wave as σ b and when setting a parameter of an integer of 3 or more as η, extracting the instantaneous value that satisfies |μ e - μ b | < (σ e + σ b ) / η; a physiological state estimation step of estimating the physiological state based on the feature quantity of the extracted instantaneous value. With such a structure, the physiological state can be estimated with high accuracy.

[0010] The physiological state estimation program according to this embodiment causes a computer to execute the following steps, namely: a waveform information acquisition step of acquiring an electroencephalogram and a pulse wave; a filtering step of filtering the acquired electroencephalogram and pulse wave in at least one predetermined frequency band; a transformation step of performing a Hilbert transform on the electroencephalogram and pulse wave in the filtered frequency band; an instantaneous value calculation step of calculating an instantaneous value including an instantaneous logarithmic amplitude and an instantaneous frequency, where the instantaneous logarithmic amplitude corresponds to the logarithm of the amplitude term of the complex waveform formula after the Hilbert transform, and the instantaneous frequency corresponds to the time derivative of the phase term of the complex waveform formula; a distribution calculation step of calculating the probability density distribution of the instantaneous value and calculating a feature quantity including the average value and variance value of the probability density distribution; an extraction step of setting the average value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram as μ e and setting the variance value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram as σe Set the average value of the instantaneous frequency in the predetermined frequency band of the pulse wave as μ b Set the variance value of the instantaneous frequency in the predetermined frequency band of the pulse wave as σ b When setting a parameter of an integer of 3 or more as η, extract those satisfying |μ e - μ b | < (σ e + σ b ) / η of the instantaneous values; a physiological state estimation step that estimates the physiological state based on the feature amount of the extracted instantaneous values. With such a configuration, the physiological state can be estimated with high accuracy.

[0011] According to the present embodiment, it is possible to provide a physiological state estimation system, a physiological state estimation method, and a physiological state estimation program that can estimate the physiological state with high accuracy.

[0012] The above and other objects, features, and advantages of the present disclosure will be more fully understood from the following detailed description and the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 A diagram illustrating a method for measuring a biological wave according to Embodiment 1.

[0014] Figure 2 A graph illustrating a biological wave according to Embodiment 1, where the horizontal axis represents time and the vertical axis represents the pulse interval and the pulse amplitude.

[0015] Figure 3 A graph illustrating a biological wave according to Embodiment 1, where the horizontal axis represents time and the vertical axis represents intensity.

[0016] Figure 4 A graph illustrating a biological wave according to Embodiment 1, where the horizontal axis represents time and the vertical axis represents intensity.

[0017] Figure 5 A graph illustrating a biological wave according to Embodiment 1, where the horizontal axis represents time and the vertical axis represents intensity.

[0018] Figure 6 A graph illustrating waveforms obtained by performing Fourier transform on the electroencephalogram and the pulse wave according to Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents intensity.

[0019] Figure 7 A graph illustrating waveforms obtained by performing Fourier transform on the electroencephalogram and the pulse wave according to Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents intensity.

[0020] Figure 8 A graph showing waveforms obtained by filtering the bio-waves related to Embodiment 1 in multiple frequency bands, where the horizontal axis represents time and the vertical axis represents intensity.

[0021] Figure 9 A graph showing waveforms obtained by filtering the bio-waves related to Embodiment 1 in multiple frequency bands, where the horizontal axis represents time and the vertical axis represents intensity.

[0022] Figure 10 A graph showing waveforms obtained by filtering the bio-waves related to Embodiment 1 in multiple frequency bands, where the horizontal axis represents time and the vertical axis represents intensity.

[0023] Figure 11 A graph representing the complex waveform formula obtained by performing a Hilbert transform on the bio-waves related to Embodiment 1 in the complex plane.

[0024] Figure 12 A graph showing the probability density distribution of the instantaneous frequency of electroencephalogram in the α frequency band when the eyes are open in Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density.

[0025] Figure 13 A graph showing the probability density distribution of the instantaneous frequency of the pulse wave in the α frequency band when the eyes are open in Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density.

[0026] Figure 14 A graph showing the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α frequency band when the eyes are open in Embodiment 1, where the horizontal axis represents the phase difference and the vertical axis represents density.

[0027] Figure 15 A graph showing the probability density distribution of the instantaneous frequency of electroencephalogram in the α frequency band when the eyes are open in Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density.

[0028] Figure 16 A graph showing the probability density distribution of the instantaneous frequency of the pulse wave in the α frequency band when the eyes are open in Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density.

[0029] Figure 17 A graph showing the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α frequency band when the eyes are open in Embodiment 1, where the horizontal axis represents the phase difference and the vertical axis represents density.

[0030] Figure 18 A graph showing the probability density distribution of the instantaneous frequency of electroencephalogram in the α frequency band when the eyes are closed in Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density.

[0031] Figure 19 A graph showing the probability density distribution of the instantaneous frequency of the pulse wave in the α band during eye closure according to Embodiment 1. The horizontal axis represents the frequency and the vertical axis represents the density.

[0032] Figure 20 A graph showing the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α band during eye closure according to Embodiment 1. The horizontal axis represents the phase difference and the vertical axis represents the density.

[0033] Figure 21 A graph showing the probability density distribution of the instantaneous frequency of the electroencephalogram in the α band during eye closure according to Embodiment 1. The horizontal axis represents the frequency and the vertical axis represents the density.

[0034] Figure 22 A graph showing the probability density distribution of the instantaneous frequency of the pulse wave in the α band during eye closure according to Embodiment 1. The horizontal axis represents the frequency and the vertical axis represents the density.

[0035] Figure 23 A graph showing the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α band during eye closure according to Embodiment 1. The horizontal axis represents the phase difference and the vertical axis represents the density.

[0036] Figure 24 A graph showing the characteristic quantities during eye opening and eye closure according to Embodiment 1. The horizontal axis represents time, and the vertical axis represents the characteristic quantity of the instantaneous frequency, the characteristic quantity of the instantaneous logarithmic amplitude, and the concentration degree.

[0037] Figure 25 A three-dimensional graph illustrating the complex waveform formula in each frequency band according to Embodiment 1.

[0038] Figure 26 A three-dimensional graph illustrating the complex waveform formula in each frequency band according to Embodiment 1.

[0039] Figure 27 A block diagram illustrating the physiological state estimation device according to Embodiment 1.

[0040] Figure 28 A flowchart illustrating the physiological state estimation method according to Embodiment 1. Detailed implementation manners

[0041] Hereinafter, the present invention will be described by way of embodiments of the invention, but the present invention is not limited to the following embodiments. In addition, the structures described in the embodiments are not all necessary for solving the problems. For the sake of clear description, the following description and drawings are appropriately omitted or simplified. In each drawing, the same reference numerals are assigned to the same elements, and repeated descriptions are omitted as needed.

[0042] (Embodiment 1)

[0043] First, waveform information related to a living body used in the present embodiment will be described. The living body is, for example, a subject. Based on the waveform information related to the living body, the physiological state of the subject is estimated. The waveform information related to the living body includes bio-waves. The bio-waves include, for example, pulse-related waves and brain-related waves. The pulse-related waves may include pulse waves and pulse interval waves. The brain-related waves may include electroencephalograms, carotid artery pulsation waves, and cerebral blood flow waves. In addition, as long as the bio-waves are waveform information related to the living body, they may also include waveform information other than pulse-related waves and brain-related waves. As long as the pulse-related waves are waveform information related to the pulse, they may also include waveform information other than pulse waves and pulse interval waves. As long as the brain-related waves are waveform information related to the brain, they may also include waveform information other than electroencephalograms, carotid artery pulsation waves, and cerebral blood flow waves. Hereinafter, electroencephalograms, pulse waves, and pulse interval waves will be described as bio-waves.

[0044] <Measurement of Bio-Waves>

[0045] A method for measuring, for example, electroencephalograms, pulse waves, and pulse interval waves in bio-waves will be described. Figure 1 is a diagram illustrating a method for measuring bio-waves according to Embodiment 1. As Figure 1 shown, as a measuring device for measuring bio-waves, for example, an electroencephalogram measuring device 10 and a pulse wave measuring device 20 can be cited.

[0046] The electroencephalogram measuring device 10 acquires the electroencephalogram of the subject as waveform information related to the living body. The electroencephalogram measuring device 10 includes a sensor 11 and a main body 12. The sensor 11 is arranged at a plurality of positions according to the 10-20 method or the like and is simultaneously measured by the main body 12. The sensor 11 is, for example, pasted on the scalp of the subject's head to sense the information of the subject's electroencephalogram from the outside of the living body. The information of the subject's electroencephalogram is, for example, voltage. In addition, as the information of the subject's electroencephalogram, in addition to voltage, the sensor 11 can also sense current, magnetic field, etc. The sensor 11 is worn on the living body in a non-invasive manner.

[0047] The sensor 11 outputs the information of the sensed brain waves to the main body 12 of the electroencephalograph 10. The main body 12 of the electroencephalograph 10 measures the time variation of voltage or the like output from the sensor 11. The sensor 11 is connected to the main body 12 through a wired or wireless communication line.

[0048] As waveform information related to a living body, the pulse wave detector 20 measures the pulse wave of the subject. The pulse wave is waveform information of a living body formed by the pulse interval, the blood output volume, and the physical characteristics of blood vessels. The pulse wave detector 20 includes a sensor 21 and a main body 22. The pulse wave detector 20 can be a photoelectric detector such as a photoplethysmograph, or a piezoelectric detector. In the case of a photoelectric detector, it is preferable to use near-infrared light with a wavelength of 800 to 1000 nm for light. The sensor 21 is, for example, attached to the skin on the neck (carotid artery bifurcation) or the cervical vertebra (vertebral artery) of the subject, and measures the information of the pulse wave flowing into the subject's brain from the outside of the living body.

[0049] Specifically, the sensor 21 can be disposed at least at any one of the vicinity of the left and right carotid artery bifurcations and the vicinity of the left and right vertebral arteries. The information of the pulse wave is, for example, the pulse pressure. In addition, as the information of the pulse wave, in addition to the pulse pressure, the sensor 21 can also sense the blood flow volume and the like. The sensor 21 is worn on the living body in a non-invasive manner.

[0050] The sensor 21 outputs the sensed pulse wave information to the main body 22 of the pulse wave detector 20. The main body 22 of the pulse wave detector 20 measures the time variation of the pulse pressure or the like output from the sensor 21. The sensor 21 is connected to the main body 22 through a wired or wireless communication line.

[0051] <Time-series data of biological waves>

[0052] Next, the time-series data of biological waves will be described. Figure 2 For the graph showing the biological waves according to Embodiment 1, the horizontal axis represents time, and the vertical axis represents the pulse interval and the pulse amplitude. Figure 3 For the graph showing the biological waves according to Embodiment 1, the horizontal axis represents time, and the vertical axis represents the intensity. In Figure 2 , as an example of biological waves, the pulse interval and the pulse amplitude of the pulse waves obtained from the left and right carotid artery bifurcations are shown. In Figure 3 , as an example of biological waves, the electroencephalogram obtained from the frontal lobe is shown. As shown in Figure 2 and Figure 3 , the electroencephalogram and the pulse wave form time-series data. In Figure 2 and Figure 3Among them, 0 seconds to 300 seconds represent the measurement results with eyes open. For example, the measurement results in the case of having an intellectual competition with eyes open. 300 seconds to 600 seconds represent the measurement results with eyes closed. For example, the measurement results in the case of listening to music with eyes closed.

[0053] Figure 4 and Figure 5 is a graph showing the bio-waves related to Embodiment 1. The horizontal axis represents time and the vertical axis represents intensity. In Figure 4 represents the pulse wave used to define the pulse interval wave based on the pulse wave and to define the pulse amplitude based on the pulse wave. In Figure 5 as an example of the bio-wave, represents the pulse interval wave. The pulse interval wave is obtained through the following steps. First, for the time series data of the pulse wave shown in Figure 4 , using the pulse wave rising edge position detection algorithm proposed by Non-Patent Document 1, the pulse interval (PPI: Peak-Peak Interval) is calculated. Next, by plotting the rising edge time of the pulse wave and the pulse interval and performing spline interpolation, a pulse interval wave as shown in Figure 5 is obtained. As shown in Figure 5 , the pulse interval wave also forms time series data. Not limited to electroencephalogram, pulse wave, and pulse interval wave, other bio-waves also form time series data, but only electroencephalogram, pulse wave, and pulse interval wave are shown here. In this embodiment, the physiological state of the organism is inferred based on the time series data of the bio-waves.

[0054] Figure 6 and Figure 7 is a graph showing the waveforms obtained by performing Fourier transform on the electroencephalogram and pulse wave related to Embodiment 1. The horizontal axis represents frequency and the vertical axis represents intensity. In Figure 6 represents Figure 2 and Figure 3 the electroencephalogram and pulse wave with eyes open in Figure 7 represents Figure 2 and Figure 3 the electroencephalogram and pulse wave with eyes closed in

[0055] <Band>

[0056] In this embodiment, the bio-waves are filtered in multiple frequency bands during a predetermined period. The predetermined period is, for example, a period when the physiological state is regarded as a constant state. The frequency bands are, for example, VLF2, VLF1, LF, HF, δ1, δ2, θ, α, β, γ, etc. Figures 8 - 10 is a graph showing the waveforms obtained by filtering the bio-waves related to Embodiment 1 in multiple frequency bands. The horizontal axis represents time and the vertical axis represents intensity. In Figure 8Among them, for example, it represents the brain waves filtered in the frequency bands of δ1 and δ2. In Figure 9 Among them, for example, it represents the brain waves filtered in the frequency bands of θ and α. In Figure 10 Among them, for example, it represents the brain waves filtered in the frequency bands of β and γ.

[0057] As Figures 8 - 10 shown, the bio-waves filtered in each frequency band form time series data. In addition to the above, the brain waves can also be filtered in the frequency bands of VLF2, VLF1, LF, and HF. Not limited to brain waves, bio-waves can be filtered in multiple frequency bands, but only brain waves are shown here. In Figures 8 - 10 it, 0 seconds to 300 seconds represents the measurement result with eyes open, and 300 seconds to 600 represents the measurement result with eyes closed. Hereinafter, each frequency band will be described.

[0058] VLF2 is, for example, a frequency band of 0.004 to 0.015 Hz. The center frequency of VLF2 is, for example, 0.01 Hz. VLF2 is a frequency band showing the functions of the autonomic nerves such as body temperature regulation, digestion, excretion, reproduction, and immunity.

[0059] VLF1 is, for example, a frequency band of 0.015 to 0.04 Hz. The center frequency of VLF1 is, for example, 0.0300 Hz. VLF1 is also a frequency band showing the functions of the autonomic nerves such as body temperature regulation, digestion, excretion, reproduction, and immunity.

[0060] LF is, for example, a frequency band of 0.04 to 0.15 Hz. The center frequency of LF is, for example, 0.1 Hz. LF is a frequency band of the blood pressure fluctuation band showing the function of blood pressure.

[0061] HF is, for example, a frequency band of 0.15 to 0.40 Hz. The center frequency of HF is, for example, 0.30 Hz. HF is a frequency band of the respiratory fluctuation band showing the function of respiration.

[0062] δ1 is, for example, a frequency band of 0.4 to 1.5 Hz. The center frequency of δ1 is, for example, 1.0 Hz.

[0063] δ2 is, for example, a frequency band of 1.5 to 4 Hz. The center frequency of δ2 is, for example, 3.0 Hz.

[0064] θ is, for example, a frequency band of 4.0 to 8.0 Hz. The center frequency of θ is, for example, 5.0 Hz.

[0065] α is, for example, a frequency band of 8.0 to 12.0 Hz. The center frequency of α is, for example, 10.0 Hz.

[0066] β is, for example, a frequency band of 12.0 to 30.0 Hz. The center frequency of β is, for example, 20.0 Hz.

[0067] γ is a frequency band greater than 30 Hz.

[0068] <Complex transformation>

[0069] The biological waves in each filtered frequency band are transformed into a complex waveform formula, for example, by Hilbert transform. Specifically, the time series data φ k (n) (t) of the biological waves in each frequency band is Hilbert-transformed into a complex waveform formula as shown in the following formula (1).

[0070] φ k (n) (t) = exp(a k (n) (t) + iψ k (n) (t)) (1)

[0071] Here, k represents each frequency band. That is, k = 1, 2,... represents frequency bands such as VLF2, VLF1, LF, HF, δ1, δ2, θ, α, β, γ,.... n represents each biological wave. That is, n = 1, 2,... represents any one of the pulse-related wave and the brain-related wave. In particular, in the case of electroencephalogram, pulse wave, and pulse interval wave, n = e, n = b, and n = r are used respectively, and are represented by φ k (e) (t), φ k (b) (t), and φ k (r) (t) as shown in formulas (2) to (4).

[0072] φ k (e) (t) = exp(a k (e) (t) + iψ k (e) (t)) (2)

[0073] φ k (b) (t) = exp(a k (b) (t) + iψ k (b) (t)) (3)

[0074] φ k (r) (t) = exp(a k (r) (t) + iψ k (r) (t)) (4)

[0075] According to equation (1), the following equations (5) to (7) can be obtained.

[0076] a k (n) (t)(5)

[0077] ψ k (n) (t)(6)

[0078] ω k (n) (t) = dψ k (n) (t) / dt (7)

[0079] In addition, as shown in equation (8), the biological wave before filtering in each frequency band can be expressed as a linear combination of frequency bands k.

[0080] φ (n) (t) = Σexp(a k (n) (t) + iψ k (n) (t)) (8)

[0081] Figure 11 It is a diagram showing the complex waveform formula obtained by performing the Hilbert transform on the biological wave according to Embodiment 1 in the complex plane. As Figure 11 shown, the biological wave φ k (n) (t) filtered in the k frequency band and represented by the complex waveform formula can be expressed on the complex plane. When the horizontal axis in the complex plane is set to Re(lnφ k (n) (t)) and the vertical axis is set to Im(lnφ k (n) (t)), a k (n) (t) represents the radius centered at the origin, and ψ k (n) (t) represents the angle with the horizontal axis. Here, k = 1, 2, 3, 4... represent each frequency band.

[0082] <Definition of instantaneous value>

[0083] Next, the instantaneous value will be described. The logarithm a k (n) (t) of the amplitude term of the complex waveform formula as shown in equation (5) is called the instantaneous logarithmic amplitude of the biological wave in the k frequency band, and ψ k (n) (t) in equation (6) is called the instantaneous phase of the biological wave in the k frequency band. The time differential value ω k(n) (t) is called the instantaneous frequency of the bio-wave in the k frequency band. The instantaneous logarithmic amplitude is equivalent to the logarithm of the amplitude term of the complex waveform formula after Hilbert transform. The instantaneous phase is equivalent to the phase term of the complex waveform formula after Hilbert transform. The instantaneous frequency is equivalent to the time differential value of the phase term of the complex waveform formula.

[0084] In addition, as shown in Equation (9), the difference between the instantaneous phases of two different bio-waves is called the instantaneous phase difference θ k (x，y) (t). Here, (x, y) represents the case where the instantaneous phases of two different bio-waves (n = x) and (n = y) are used.

[0085] θ k (x，y) (t) = ψ k (x) (t) - ψ k (y) (t) (9)

[0086] Here, two different bio-waves are represented by Equations (10) and (11).

[0087] φ k (x) (t) = exp(a k (x) (t) + iψ k (x) (t)) (10)

[0088] φ k (y) (t) = exp(a k (y) (t) + iψ k (y) (t)) (11)

[0089] In particular, as shown in (12) - (14), the instantaneous phase difference between the instantaneous phase of the pulse wave and the instantaneous phase of the pulse interval wave is called θ k (b，r) (t), the instantaneous phase difference between the instantaneous phase of the electroencephalogram wave and the instantaneous phase of the pulse wave is called θ k (e，b) (t), and the instantaneous phase difference between the instantaneous phase of the electroencephalogram wave and the instantaneous phase of the pulse interval wave is called θ k (e，r) (t).

[0090] θ k (b，r) (t) = ψ k (b)(t) - ψ k (r) (t) (12)

[0091] θ k (e，b) (t) = ψ k (e) (t) - ψ k (b) (t) (13)

[0092] θ k (e，r) (t) = ψ k (e) (t) - ψ k (r) (t) (14)

[0093] The instantaneous value includes the instantaneous logarithmic amplitude a k (n) (t), the instantaneous phase ψ k (n) (t), the instantaneous frequency ω k (n) (t) and the instantaneous phase difference θ k (x，y) (t).

[0094] <Distribution of instantaneous values>

[0095] By measuring the instantaneous values at predetermined time intervals, the probability density distribution of the instantaneous values can be obtained. Therefore, the probability density distribution of the instantaneous amplitude a k (n) (t), the probability density distribution of the instantaneous phase ψ k (n) (t), the probability density distribution of the instantaneous frequency ω k (n) (t), the probability density distribution of the instantaneous phase difference θ k (x，y) (t) can be obtained.

[0096] Figure 12 is a graph showing the probability density distribution of the instantaneous frequency of the electroencephalogram in the α band during eye opening according to Embodiment 1. The horizontal axis represents the frequency and the vertical axis represents the density. Figure 13 is a graph showing the probability density distribution of the instantaneous frequency of the pulse wave in the α band during eye opening according to Embodiment 1. The horizontal axis represents the frequency and the vertical axis represents the density. Figure 14 is a graph showing the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α band during eye opening according to Embodiment 1. The horizontal axis represents the phase difference and the vertical axis represents the density.

[0097] Figure 15 A graph showing the probability density distribution of the instantaneous frequency of the electroencephalogram in the α band during eye opening according to Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density. Figure 16 A graph showing the probability density distribution of the instantaneous frequency of the pulse wave in the α band during eye opening according to Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density. Figure 17 A graph showing the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α band during eye opening according to Embodiment 1, where the horizontal axis represents the phase difference and the vertical axis represents density.

[0098] Figure 18 A graph showing the probability density distribution of the instantaneous frequency of the electroencephalogram in the α band during eye closure according to Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density. Figure 19 A graph showing the probability density distribution of the instantaneous frequency of the pulse wave in the α band during eye closure according to Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density. Figure 20 A graph showing the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α band during eye closure according to Embodiment 1, where the horizontal axis represents the phase difference and the vertical axis represents density.

[0099] Figure 21 A graph showing the probability density distribution of the instantaneous frequency of the electroencephalogram in the α band during eye closure according to Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density. Figure 22 A graph showing the probability density distribution of the instantaneous frequency of the pulse wave in the α band during eye closure according to Embodiment 1, where the horizontal axis represents frequency and the vertical axis represents density. Figure 23 A graph showing the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α band during eye closure according to Embodiment 1, where the horizontal axis represents the phase difference and the vertical axis represents density.

[0100] As Figures 12 - 23 shown, the instantaneous values measured at a predetermined time interval form a probability density distribution. In addition, the probability density distribution is not limited to the instantaneous values of the electroencephalogram and the pulse wave, and can also be formed by the instantaneous values of other pulse-related waves and brain-related waves. However, the probability density distributions of the instantaneous frequency of the electroencephalogram, the instantaneous frequency of the pulse wave, and the instantaneous phase difference between the electroencephalogram and the pulse wave are shown here.

[0101] <Fitting>

[0102] By fitting the probability density distribution of each instantaneous value with a predetermined function, characteristic quantities of the probability density distribution of the instantaneous value can be calculated. For example, by fitting the probability density distribution of the instantaneous frequency with a Gaussian distribution, characteristic quantities such as the average value and the variance value can be obtained.

[0103] In addition, by fitting the probability density distribution of the instantaneous phase difference with a von Mises distribution, characteristic quantities such as the average value and the concentration degree can be obtained. Since the value of the instantaneous phase difference is limited to the range of -π to +π, the von Mises distribution is applied. The von Mises distribution is the following formula (15). Here, I j (κ) is the Bessel function of the first kind of order j of formula (16).

[0104] f(θ) = exp(κcos(θ - μ)) / 2πI0(κ) (15)

[0105] I j (κ) = (κ / 2) j Σ(κ 2 / 4) i / i!Γ(j + i + 1) (16)

[0106] As Figure 12 shown, in the case where the probability density distribution of the instantaneous frequency of the electroencephalogram in the α frequency band during eye opening is fitted with a Gaussian distribution, the average value μ is 10.41 Hz and the variance value is 1.28. The measurement period is 70 to 90 seconds. As Figure 13 shown, in the case where the probability density distribution of the instantaneous frequency of the pulse wave in the α frequency band during eye opening is fitted with a Gaussian distribution, the average value μ is 9.71 Hz and the variance value is 1.07. The measurement period is 70 to 90 seconds. As Figure 14 shown, in the case where the probability density distribution of the instantaneous phase difference of the electroencephalogram and the pulse wave in the α frequency band during eye opening is fitted with a von Mises distribution, the average value is 0.13 Hz and the concentration degree is 0. The measurement period is 70 to 90 seconds.

[0107] As Figure 15 shown, in the case where the probability density distribution of the instantaneous frequency of the electroencephalogram in the α frequency band during eye opening is fitted with a Gaussian distribution, the average value is 10.26 Hz and the variance value is 1.19. The measurement period is 210 to 230 seconds. As Figure 16 shown, in the case where the probability density distribution of the instantaneous frequency of the pulse wave in the α frequency band during eye opening is fitted with a Gaussian distribution, the average value is 9.5 Hz and the variance value is 0.92. The measurement period is 210 to 230 seconds. As Figure 17As shown, when the probability density distribution of the instantaneous phase difference between the brain wave and the pulse wave in the α frequency band during eye opening is fitted with the von Mises distribution, the average value is 0.78 Hz and the concentration is 0.13. The measurement period is 210 to 230 seconds.

[0108] As Figure 18 shown, when the probability density distribution of the instantaneous frequency of the brain wave in the α frequency band during eye closing is fitted with the Gaussian distribution, the average value is 9.1 Hz and the variance value is 0.96. The measurement period is 340 to 360 seconds. As Figure 19 shown, when the probability density distribution of the instantaneous frequency of the pulse wave in the α frequency band during eye closing is fitted with the Gaussian distribution, the average value is 8.99 Hz and the variance value is 1.07. The measurement period is 340 to 360 seconds. As Figure 20 shown, when the probability density distribution of the instantaneous phase difference between the brain wave and the pulse wave in the α frequency band during eye closing is fitted with the von Mises distribution, the average value is 0.28 Hz and the concentration is 0.29. The measurement period is 340 to 360 seconds.

[0109] As Figure 21 shown, when the probability density distribution of the instantaneous frequency of the brain wave in the α frequency band during eye closing is fitted with the Gaussian distribution, the average value is 9.39 Hz and the variance value is 0.97. The measurement period is 410 to 430 seconds. As Figure 22 shown, when the probability density distribution of the instantaneous frequency of the pulse wave in the α frequency band during eye closing is fitted with the Gaussian distribution, the average value is 9.03 Hz and the variance value is 0.98. The measurement period is 410 to 430 seconds. As Figure 23 shown, when the probability density distribution of the instantaneous phase difference between the brain wave and the pulse wave in the α frequency band during eye closing is fitted with the von Mises distribution, the average value is 0.2 and the concentration is 0.33. The measurement period is 410 to 430 seconds.

[0110] <Tuning>

[0111] Next, the tuning process will be described. In the tuning process, characteristic quantities that satisfy a predetermined condition are extracted from the relationship between characteristic quantities of two different biological waves. Hereinafter, the brain wave and the pulse wave will be used as examples for description. The tuning process extracts, for example, the case where the distribution of the instantaneous frequency of the brain wave is similar to the distribution of the instantaneous frequency of the pulse wave. The case where the distribution of the instantaneous frequency of the brain wave is similar to the distribution of the instantaneous frequency of the pulse wave means the case that satisfies the following formula (17).

[0112] (|μ e - μ b | < (σ e + σ b)(1 / η)(17)

[0113] Here, η is, for example, given by the following equation (18).

[0114] η = 3 (18)

[0115] In Figure 12 and Figure 13 cases, the left side of equation (17) is |10.41 - 9.71| = 0.7, and the right side of equation (17) is (1.28 + 1.07) / 3 = 0.783, which holds. In Figure 15 and Figure 16 cases, the left side of equation (17) is |10.26 - 9.5| = 0.76, and the right side of equation (17) is (1.19 + 0.92) / 3 = 0.703, which does not hold.

[0116] In Figure 18 and Figure 19 cases, the left side of equation (17) is |9.1 - 8.99| = 0.11, and the right side of equation (17) is (0.96 + 1.07) / 3 = 0.676, which holds. In Figure 21 and Figure 22 cases, the left side of equation (15) is |9.39 - 9.03| = 0.36, and the right side of equation (17) is (0.97 + 0.98) / 3 = 0.65, which holds. In this way, the case of extracting the instantaneous values of the electroencephalogram and the pulse wave for which equation (17) holds is called the tuning process.

[0117] In the tuning process, characteristic quantities in the case where the electroencephalogram and the pulse wave are related are extracted. For example, the pulse wave measured from blood vessels such as the carotid artery bifurcation and the vertebral artery that flow into the brain is closely related to the activity of the electroencephalogram. Therefore, by extracting the characteristic quantities in the case where equation (17) holds, the electroencephalogram and the pulse wave that are related to each other can be extracted.

[0118] <Estimation of Physiological State>

[0119] Next, the estimation of the physiological state will be described. Here, a method for estimating the physiological state using the electroencephalogram and the pulse wave will be described. The physiological state is estimated based on the characteristic quantities of the instantaneous values of the electroencephalogram and the pulse wave extracted by the tuning process. Hereinafter, several examples of estimating the physiological state will be shown.

[0120] Figures 12 - 14 and Figures 18 - 20 The characteristic quantities shown are characteristic quantities that satisfy equation (17) and are extracted by the tuning process. As Figure 12 and Figure 13As shown, in the probability density distribution of the instantaneous frequency in the α band during eye opening, the average value of the electroencephalogram is 10.41, and the average value of the pulse wave is 9.71, both being approximately 10 Hz. On the other hand, as Figure 18 and Figure 19 shown, in the probability density distribution of the instantaneous frequency in the α band during eye closing, the average value of the electroencephalogram is 9.1, and the average value of the pulse wave is 8.99, both being approximately 9 Hz. Thus, by changing from eye opening to eye closing, the average values of the electroencephalogram and the pulse wave change from approximately 10 Hz to approximately 9 Hz.

[0121] In addition, as Figure 14 shown, in the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α band during eye opening, the concentration is 0. On the other hand, as Figure 20 shown, in the probability density distribution of the instantaneous phase difference between the electroencephalogram and the pulse wave in the α band during eye closing, the concentration is 0.29. Thus, by changing from eye opening to eye closing, the concentration of the instantaneous phase difference between the electroencephalogram and the pulse wave increases from 0 to 0.29.

[0122] Figure 24 is a graph showing the characteristic quantities during eye opening and eye closing according to Embodiment 1. The horizontal axis represents time, and the vertical axis represents the characteristic quantity of the instantaneous frequency, the characteristic quantity of the instantaneous logarithmic amplitude, and the concentration. As described above, by changing from eye opening to eye closing, the average value of the instantaneous frequency of the electroencephalogram and the average value of the instantaneous frequency of the pulse wave change from approximately 10 Hz to approximately 9 Hz. As a result, the probability density distributions of the instantaneous frequencies of the electroencephalogram and the pulse wave overlap. Therefore, as Figure 24 shown, the concentration κ is greater than 0. This shows the following situation: Since multiple centers are active during eye opening, the pulse wave measured from the carotid bifurcation is not synchronized with the electroencephalogram, but the number of active centers decreases during eye closing. Therefore, the pulse wave measured from the carotid bifurcation is synchronized with the electroencephalogram. Thus, the state of the centers active in the brain can be inferred from the change in the characteristic quantity.

[0123] Figure 25 and Figure 26 are diagrams three-dimensionally illustrating the complex waveform formula in each frequency band according to Embodiment 1. The horizontal axis and the vertical axis on the complex plane represent Re(lnφ k (n) (t)) and Im(lnφ k (n)(t)), the axis orthogonal to the complex plane represents the frequency band (energy axis). For example, let k = 4 for the frequency band longer than the pulse interval be HF (respiration), k = 3 be LF (blood pressure), k = 2 be VLF1 (possibly related to the autonomic nervous system), k = 1 be VLF2 (possibly related to the autonomic nervous system) for representation. Let k = 5 for the electroencephalogram frequency band be δ1, k = 6 be δ2, k = 7 be θ, k = 8 be α, etc. Then, the radius vector on the complex plane becomes the instantaneous logarithmic amplitude of the biological wave of each frequency band. The rotation direction centered on the energy axis becomes the instantaneous phase of the biological wave of each frequency band. The energy axis becomes the frequency band axis. Moreover, it can be seen that while being approximately Gaussian distributed on their respective axes, they rotate around the energy axis. In this way, an image can be obtained as if the electrons of each energy level rotate around the atomic nucleus as an electron cloud expressed by the probability of existence. Moreover, the approximately Gaussian distribution of each frequency band similar to the electron cloud is formed in a manner along the constant energy surface of 1 / F.

[0124] Therefore, in the case where there is an instantaneous value (characteristic quantity) of a frequency band that deviates from the Gaussian distribution such as along the constant energy surface of 1 / F, it can be inferred that an abnormality has occurred in the physiological state characteristic of that frequency band. For example, in the case where the instantaneous values (characteristic quantities) of VLF1 and VLF2 deviate from the Gaussian distribution such as along the constant energy surface of 1 / F, it can be inferred that an abnormality has occurred in the physiological state related to the autonomic nervous system.

[0125] <Physiological State Inference System>

[0126] Next, the physiological state inference system of the present embodiment will be described. The physiological state inference system of the present embodiment includes a physiological state inference device. Hereinafter, as the physiological state inference system, the physiological state inference device will be described.

[0127] Figure 27 is a block diagram illustrating the physiological state inference device according to Embodiment 1. As Figure 27 shown, the physiological state inference device 50 includes a control unit 50a, a communication unit 50b, a storage unit 50c, an interface unit 50d, a waveform information acquisition unit 51, a filtering unit 52, a transformation unit 53, an instantaneous value calculation unit 54, a distribution calculation unit 55, an extraction unit 56, and a physiological state inference unit 57. The control unit 50a, the communication unit 50b, the storage unit 50c, the interface unit 50d, the waveform information acquisition unit 51, the filtering unit 52, the transformation unit 53, the instantaneous value calculation unit 54, the distribution calculation unit 55, the extraction unit 56, and the physiological state inference unit 57 respectively have functions as a control unit, a communication unit, a storage unit, an interface unit, a waveform information acquisition unit, a filtering unit, a transformation unit, an instantaneous value calculation unit, a distribution calculation unit, an extraction unit, and a physiological state inference unit.

[0128] The physiological state estimation device 50 is an information processing device including a computer. The control unit 50a includes, for example, a processor such as a CPU (Central Processing Unit), an MPU (Micro Processing Unit), an ECU (Electronic Control Unit), an FPGA (Field-Programmable Gate Array), or an ASIC (Application Specific Integrated Circuit). The control unit 50a functions as an arithmetic unit that performs control processing, arithmetic processing, and the like. In addition, the control unit 50a controls the operations of the respective structural elements of the communication unit 50b, the storage unit 50c, the interface unit 50d, the waveform information acquisition unit 51, the filtering unit 52, the conversion unit 53, the instantaneous value calculation unit 54, the distribution calculation unit 55, the extraction unit 56, and the physiological state estimation unit 57.

[0129] Each structural element of the physiological state estimation device 50 can be implemented, for example, by executing a program under the control of the control unit 50a. More specifically, each structural element can be implemented by the control unit 50a executing a program stored in the storage unit 50c. In addition, it can also be configured that by pre-recording the required program in an arbitrary non-volatile recording medium and installing it as needed, each structural element can be implemented. In addition, each structural element is not limited to being implemented by software based on a program, and can also be implemented by any combination of hardware, firmware, and software.

[0130] The communication unit 50b receives the waveform information of the bio-wave measured by the waveform measuring device from waveform measuring devices such as the electroencephalogram measuring device 10 and the pulse wave measuring device 20. The waveform information of the bio-wave includes, for example, the time series data of the bio-wave.

[0131] The storage unit 50c can also have a storage device such as a memory or a hard disk, for example. The storage device is, for example, a ROM (ReadOnly Memory) or a RAM (Random Access Memory). The storage unit 50c has a function of storing control programs, arithmetic programs, and the like executed by the control unit 50a. In addition, the storage unit 50c has a function of temporarily storing processing data and the like. Moreover, the storage unit 50c stores the correspondence relationship between the characteristic quantities in the instantaneous values of the bio-waves of the subject pre-measured and the physiological state of the subject. The storage unit 50c can also store the waveform information of the bio-wave received by the communication unit 50b.

[0132] The interface unit 50d is, for example, a User Interface. The interface unit 50d has input devices such as a keyboard, a touch screen, or a mouse, and output devices such as a display or a speaker. The interface unit 50d receives the operation of inputting data performed by a user (such as an operator), and outputs information to the user.

[0133] The waveform information acquisition unit 51 acquires the waveform information obtained by the communication unit 50b from the waveform measuring device. The waveform information is, for example, time series data of biological waves. The waveform information acquisition unit 51 acquires, for example, electroencephalogram and pulse wave. It may also include a pulse correlation wave acquisition unit that acquires pulse correlation waves and a brain correlation wave acquisition unit that acquires brain correlation waves.

[0134] The filtering unit 52 filters the acquired waveform information in at least one predetermined frequency band.

[0135] The transformation unit 53 performs a transformation that complexifies the waveform information in the filtered frequency band. The complexifying transformation is, for example, the Hilbert transform.

[0136] The instantaneous value calculation unit 54 calculates instantaneous values. As described above, the instantaneous values include the instantaneous logarithmic amplitude a k (n) (t), the instantaneous phase ψ k (n) (t), the instantaneous frequency ω k (n) (t), and the instantaneous phase difference θ k (x，y) (t). The instantaneous value calculation unit 54 calculates, for example, the instantaneous logarithmic amplitude a that includes the logarithm of the amplitude term equivalent to the complex waveform formula after the Hilbert transform k (n) (t) and the instantaneous frequency ω that is the time differential value of the phase term equivalent to the complex waveform formula k (n) (t) among the instantaneous values.

[0137] The distribution calculation unit 55 calculates the probability density distribution of the instantaneous values. Moreover, the distribution calculation unit 55 calculates characteristic quantities including the average value and the variance value of the probability density distribution of the instantaneous values. In addition, the distribution calculation unit 55 may fit a Gaussian distribution to the probability density distribution by the instantaneous values. In this case, the average value and the variance value of the probability density distribution are the average value and the variance value when the Gaussian distribution is fitted to the probability density distribution. Furthermore, the distribution calculation unit 55 may fit a von Mises distribution to the probability density distribution by the instantaneous values. In this case, the average value of the probability density distribution is the average value when the von Mises distribution is fitted to the probability density distribution.

[0138] The extraction unit 56 extracts the instantaneous value when a predetermined condition is satisfied. The predetermined condition is, for example, the condition shown below. That is, when the average value of the instantaneous frequency in a predetermined frequency band of the electroencephalogram is set to μ e and the variance value of the instantaneous frequency in a predetermined frequency band of the electroencephalogram is set to σ e and the average value of the instantaneous frequency in a predetermined frequency band of the pulse wave is set to μ b and the variance value of the instantaneous frequency in a predetermined frequency band of the pulse wave is set to σ b and when a parameter of an integer of 3 or more is set to η, the extraction unit 56 extracts the instantaneous value when the above formula (17) is satisfied. The parameter η is, for example, 3.

[0139] The physiological state estimation unit 58 estimates the physiological state based on the characteristic quantity of the extracted instantaneous value. For example, the characteristic quantity calculated for the biological wave of the subject is previously associated with the physiological states such as the subjective evaluation RAS and the stress scale PSS of the subject. Then, based on the characteristic quantity of the extracted instantaneous value, the physiological state of the subject is estimated.

[0140] <Physiological State Estimation Method>

[0141] Next, the physiological state estimation method will be described. Figure 28 FIG. is a flowchart illustrating the physiological state estimation method according to Embodiment 1. As Figure 28 shown, the physiological state estimation method includes a waveform information acquisition step (step S11) of acquiring waveform information related to a living body, a filtering step (step S12) of filtering the acquired waveform information in at least one frequency band, a transformation step (step S13) of transforming the waveform information of the filtered frequency band into a complex waveform formula, an instantaneous value calculation step (step S14) of calculating an instantaneous value including at least any one of an instantaneous logarithmic amplitude, an instantaneous frequency, an instantaneous phase, and an instantaneous phase difference, a distribution calculation step (step S15) of calculating the distribution of the instantaneous value, an extraction step (step S16) of extracting the instantaneous value when a predetermined condition is satisfied, and a physiological state estimation step (step S17) of estimating the physiological state.

[0142] <Waveform Information Acquisition Step>

[0143] First, in the waveform information acquisition step (step S11), the waveform information acquisition unit 51 acquires waveform information of biological waves such as electroencephalograms and pulse waves. Specifically, the waveform information acquisition unit 51 acquires the electroencephalogram measured by the electroencephalogram detector 10 and the pulse wave measured by the pulse wave detector 20.

[0144] <Filtering Step>

[0145] Next, in the filtering step (step S12), the filtering unit 52 filters the acquired waveform information in at least one predetermined frequency band.

[0146] <Transformation Step>

[0147] Next, in the transformation step (step S13), the transformation unit 53 Hilbert-transforms the time-series data φ k (n) (t) of the filtered frequency band into a complex waveform formula.

[0148] <Instantaneous Value Calculation Step>

[0149] Next, in the instantaneous value calculation step (step S14), the instantaneous value calculation unit 54 calculates the above-mentioned instantaneous values. Specifically, the instantaneous value calculation unit 54 calculates at least any one of the instantaneous logarithmic amplitude including the logarithm of the amplitude term of the complex waveform formula equivalent to the Hilbert-transformed one, the instantaneous phase equivalent to the phase term of the complex waveform formula, the instantaneous frequency equivalent to the time differential value of the phase term of the complex waveform formula, and the instantaneous phase difference equivalent to the difference between the instantaneous phases of two different biological waves.

[0150] <Instantaneous Value Distribution Calculation Step>

[0151] Next, in the instantaneous value distribution calculation step (step S15), the distribution calculation unit 55 calculates the distribution of the instantaneous values. Specifically, the distribution calculation unit 55 calculates the probability density distributions of the instantaneous logarithmic amplitude, the instantaneous phase, the instantaneous frequency, and the instantaneous phase difference in a desired time interval. Moreover, as the characteristic quantities of the above-mentioned probability density distributions, the distribution calculation unit 55 calculates, for example, the average value and the variance value.

[0152] <Extraction Step>

[0153] Next, in the extraction step (step S16), the extraction unit 56 sets the average value of the instantaneous frequency in a predetermined frequency band of the electroencephalogram as μ e , sets the variance value of the instantaneous frequency in a predetermined frequency band of the electroencephalogram as σ e , sets the average value of the instantaneous frequency in a predetermined frequency band of the pulse wave as μ b , sets the variance value of the instantaneous frequency in a predetermined frequency band of the pulse wave as σ b , and when setting a parameter of an integer of 3 or more as η, extracts the instantaneous values in the case where the formula (17) is satisfied.

[0154] <Physiological State Deduction Step>

[0155] Next, in the physiological state estimation step (step S17), the physiological state estimation unit 57 estimates the physiological state based on the feature amounts of the extracted instantaneous values.

[0156] Next, the effects of the present embodiment will be described. In the present embodiment, instantaneous values in which the electroencephalogram and the pulse wave satisfy a predetermined condition of correlation are extracted, and the physiological state is estimated based on the feature amounts of the extracted instantaneous values. Therefore, the physiological state can be estimated with high accuracy. In addition, since the electroencephalogram and the pulse wave are filtered in a specific frequency band related to the physiological state, the physiological state related to the specific frequency band can be estimated with high accuracy. Since the probability density distributions of the instantaneous log amplitude, the instantaneous phase, and the instantaneous frequency are fitted to a Gaussian distribution, the average value and the variance value can be obtained as feature amounts. In addition, since the probability density distribution of the instantaneous phase difference is fitted to a von Mises distribution, the average value and the concentration degree can be obtained as feature amounts. Thereby, the physiological state can be estimated with high accuracy.

[0157] In addition, the present invention is not limited to the above-described embodiment, and can be appropriately changed without departing from the gist. For example, a physiological state estimation program for causing a computer to execute the physiological state estimation method is also included in the technical idea of the embodiment.

[0158] The program can be stored using any type of non-transitory computer-readable medium and provided to the computer. Non-transitory computer-readable media include any type of tangible storage medium. Examples of non-transitory computer-readable media include magnetic storage media (such as floppy disks, magnetic tapes, hard disk drives, etc.), magneto-optical storage media (such as magneto-optical discs), CD-ROM (compact disc read-only memory), CD-R (compact disc recordable), CD-R / W (compact disc rewritable), semiconductor memories (such as mask ROM, PROM (programmable ROM), EPROM (erasable PROM), flash ROM, RAM (random access memory), etc.). The program can be provided to the computer using any type of transitory computer-readable medium. Examples of transitory computer-readable media include electrical signals, optical signals, and electromagnetic waves. The transitory computer-readable medium can provide the program to the computer via a wired communication line (such as wires and optical fibers) or a wireless communication line.

[0159] It is apparent from the present disclosure thus described that the embodiments of the present disclosure can vary in many ways. Such variations should not be regarded as departing from the spirit and scope of the present disclosure, and all such modifications that are obvious to those skilled in the art are included within the scope of the technical solution.

Claims

1. A physiological state estimation system, comprising: a waveform information acquisition unit that acquires brain waves and pulse waves; a filtering unit configured to filter the acquired brain waves and pulse waves in at least one predetermined frequency band; a transforming unit that performs Hilbert transform on the filtered brain wave and pulse wave in the frequency band; an instantaneous value calculation unit that calculates an instantaneous value including an instantaneous logarithmic amplitude and an instantaneous frequency, wherein the instantaneous logarithmic amplitude is equivalent to the logarithm of the amplitude term of the complex waveform formula subjected to the Hilbert transform, and the instantaneous frequency is equivalent to the time differential value of the phase term of the complex waveform formula; a distribution calculation unit that calculates the probability density distribution of the instantaneous value and calculates a feature quantity including an average value and a variance value of the probability density distribution; an extracting unit, which sets the average value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram to μ e , setting the variance value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram to σ e , setting the average value of the instantaneous frequency in the predetermined frequency band of the pulse wave to μ b , setting the variance value of the instantaneous frequency in the predetermined frequency band of the pulse wave to σ b , when the parameter is an integer greater than 3 and is set to η, extract the e -μ b |<(σ e +σ b ) / η; A physiological state estimation unit estimates a physiological state based on the extracted feature amount of the instantaneous value.

2. The physiological state estimation system according to claim 1, wherein: The parameter n is 3.

3. A physiological state estimation method, comprising: A waveform information acquisition step, which acquires brain waves and pulse waves; A filtering step of filtering the acquired brain waves and pulse waves in at least one predetermined frequency band; a transforming step of performing Hilbert transform on the brain wave and the pulse wave of the filtered frequency band; An instantaneous value calculation step, which calculates an instantaneous value including an instantaneous logarithmic amplitude and an instantaneous frequency, wherein the instantaneous logarithmic amplitude is equivalent to the logarithm of the amplitude term of the complex waveform formula subjected to Hilbert transformation, and the instantaneous frequency is equivalent to the time differential value of the phase term of the complex waveform formula; a distribution calculation step, which calculates the probability density distribution of the instantaneous value and calculates the feature quantity including the mean value and variance value of the probability density distribution; The extraction step includes setting the average value of the instantaneous frequency in the predetermined frequency band of the brain wave to μ e , setting the variance value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram to σ e , setting the average value of the instantaneous frequency in the predetermined frequency band of the pulse wave to μ b , setting the variance value of the instantaneous frequency in the predetermined frequency band of the pulse wave to σ b , when the parameter is an integer greater than 3 and is set to η, extract the e -μ b |<(σ e +σ b ) / η; A physiological state estimation step is performed to estimate the physiological state based on the extracted feature quantity of the instantaneous value.

4. A physiological state estimation program product, which enables a computer to execute the following steps, namely: A waveform information acquisition step, which acquires brain waves and pulse waves; A filtering step of filtering the acquired brain waves and pulse waves in at least one predetermined frequency band; a transforming step of performing Hilbert transform on the brain wave and the pulse wave of the filtered frequency band; An instantaneous value calculation step, which calculates an instantaneous value including an instantaneous logarithmic amplitude and an instantaneous frequency, wherein the instantaneous logarithmic amplitude is equivalent to the logarithm of the amplitude term of the complex waveform formula subjected to Hilbert transformation, and the instantaneous frequency is equivalent to the time differential value of the phase term of the complex waveform formula; a distribution calculation step, which calculates the probability density distribution of the instantaneous value and calculates the feature quantity including the mean value and variance value of the probability density distribution; The extraction step includes setting the average value of the instantaneous frequency in the predetermined frequency band of the brain wave to μ e , setting the variance value of the instantaneous frequency in the predetermined frequency band of the electroencephalogram to σ e , setting the average value of the instantaneous frequency in the predetermined frequency band of the pulse wave to μ b , setting the variance value of the instantaneous frequency in the predetermined frequency band of the pulse wave to σ b , when the parameter is an integer greater than 3 and is set to η, extract the e -μ b |<(σ e +σ b ) / η; A physiological state estimation step is performed to estimate the physiological state based on the extracted feature quantity of the instantaneous value.

Images