JP6707015B2 - State estimation device, method and program - Google Patents

Description

The present invention relates to a technique for estimating a subject's state based on the subject's biological information.

There are two autonomic nerves, the sympathetic nerve and the vagus nerve. Sympathetic nerves and vagus nerves are widely distributed in various organs, and control involuntary body functions such as circulation and metabolism. In many cases, it is said that the sympathetic nerve and the vagus nerve dominate one organ in an antagonistic manner.

It is known that sympathetic nerve activity, which is one of autonomic nerve activity, is enhanced by stress stimulation such as mental arithmetic load (see Non-Patent Document 1). In this specification, a stress state in which sympathetic nerve activity similar to that at the time of executing mental arithmetic load is considered to be induced is referred to as a “task state”.

On the other hand, vagal nerve activity, which is another autonomic nerve activity, is often understood to be equivalent to parasympathetic nerve activity, since the vagus nerve is mainly responsible for parasympathetic nerve activity in each organ. The vagus nerve is, strictly speaking, the name of the Xth nerve, which is one of the cranial nerves, and refers to all the nerves from the brain to each organ. Therefore, for example, by adding the name of the organ to be controlled, such as the cardiac vagus nerve, the parasympathetic nerve activity in the target organ may be indicated.

The heart is one of the organs controlled by the autonomic nerves. The heart rate is antagonistically controlled by the sympathetic nerve and the vagus nerve, and is said to reflect the static balance between the sympathetic nerve activity and the vagus nerve activity (see Reference Document 2). In particular, it is known that the fluctuation of the instantaneous heartbeat (RRI: R-R interval), which is the interval between two adjacent R waves as shown in FIG. 6, changes due to both sympathetic nerve activity and vagal nerve activity. The R wave is one of the electrocardiographic waveforms obtained by electrocardiographic measurement, and reflects the depolarizing activity of the heart (see Non-Patent Document 3).

Frequency spectrum analysis of instantaneous heartbeat variability is a method for estimating autonomic nervous activity in a real environment. According to this method, low frequency components (hereinafter referred to as HRV _LF ) when frequency spectrum analysis of instantaneous heartbeats having unequal intervals are performed, both sympathetic nerve activity and cardiac vagus nerve activity, and high frequency components (hereafter HRV _LF ). (Described as _HF ) is interpreted as an index that reflects cardiac vagal activity (see Non-Patent Documents 1 and 2).

HRV _LF and HRV _HF are also said to reflect the effects of other organs. According to this interpretation, it is said that HRV _LF contains a Mayer wave due to baroreceptive reflex of blood vessels, and HRV _HF contains a component of respiratory sinus arrhythmia (see Non-Patent Document 4). That is, there is a relationship between the instantaneous heart rate variability and respiration, and at least HRV _HF contains a respiratory component.

In order to perform appropriate stress management in a real environment, it is necessary to determine the task state and to understand the stress load received during the task state. In the conventional method, the task state is determined using HRV _LF/HF , which is the ratio of HRV _LF and HRV _HF (see Non-Patent Document 1 and Non-Patent Document 5). This method assumes that the HRV _LF/HF in the task state is higher than the HRV _LF/HF in the rest state. HRV _LF/HF is calculated from HRV _LF and HRV _HF obtained by performing frequency spectrum analysis on the instantaneous heartbeat variability, and based on the value, state determination is performed by a method such as Support Vector Machine (SVM). The calculation formula of HRV _LF/HF used at this time is shown in Formula (1).

K. Yamaguchi, "Analysis of Mental Stress due to Heart Rate Variability", Bulletin of Faculty of Human Relations, Shigakukan University, Vol.31, No.1, pp.1-10, 2010 Hiroshi Inoue, "Cardiovascular Diseases and Autonomic Nervous Function, 2nd Edition", Medical School, 2001 Jun Okude, “If You Know This! Easy Point ECG 2nd Edition”, Medical Institute, 2011 Yutaka Yoshida, et al., “Estimation of respiratory-related parameters from heartbeat variability time series”, Biomedical engineering, Vol.43, No.3, pp.456-460, 2005 Hiroto Ide, 6 others, “Study of stress estimation method without relying on individual based on multivariate analysis of physiological index”, Proceedings of 50th anniversary of IPSJ (72nd) National Convention, pp.561-562, 2010

Heart rate variability changes not only with autonomic nervous activity but also with breathing. In particular, since HRV _HF changes depending on the respiratory rate and the ventilation volume, it cannot be said that the HRV _LF/HF in the task state is always higher than the HRV _LF/HF in the rest state. Therefore, the task state may not be properly estimated by the conventional method.

As a method for avoiding the above problem, it is possible to intentionally adjust the respiratory rate and the ventilation volume (hereinafter referred to as respiratory control). However, it is extremely difficult to keep the breathing rate and the ventilation volume constant in a real environment at all times, and it is almost impossible to incorporate the breathing control into the daily service.

The present invention has been made in view of the above circumstances, and an object thereof is to provide a state estimation device, a state estimation method, and a program that can more accurately determine the state of a subject without performing respiratory control. It is to be.

A first aspect of the present invention is, in a state estimating device, a calculating means for calculating a high frequency component of heartbeat fluctuation based on a signal output from an electrocardiographic measuring means for measuring an electrocardiogram of a subject, and the calculated heartbeat. An estimating means for estimating a respiratory characteristic amount based on a high frequency component of the fluctuation and a determining means for determining a state of the subject based on the estimated respiratory characteristic amount are provided.

According to a second aspect of the present invention, the estimating means calculates the respiratory characteristic amount from the high frequency component of the calculated heartbeat variability by a single regression analysis.

In a third aspect of the present invention, the calculating means further calculates a low frequency component of the heartbeat variability based on the signal, and the estimating means based on the calculated low frequency component and high frequency component of the heartbeat variability. The respiratory feature amount is estimated by the above.

In a fourth aspect of the present invention, the estimating means calculates the respiratory feature amount from the calculated low frequency component and high frequency component of the heartbeat variability by multiple regression analysis.

The respiratory feature may be based on, for example, respiratory rate, ventilation, or both. The respiratory characteristic amount may be based on, for example, the ratio of the low frequency component in the ventilation amount.

According to the present invention, it becomes possible to more accurately determine the state without performing respiratory control.

The block diagram which shows the structural example of the state estimation apparatus which concerns on 1st Embodiment. FIG. 3 is a block diagram showing an example of a system for generating data stored in a respiratory feature amount estimation data recording unit shown in FIG. 1. FIG. 2 is a block diagram showing an example of a system that generates data stored in a state determination data recording unit shown in FIG. 1. FIG. 6 is a block diagram showing another example of a system for generating data stored in the state determination data recording section shown in FIG. 1. 6 is a flowchart showing an example of a procedure of state determination processing according to the first embodiment. The figure which shows instantaneous heartbeat (RRI).

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[First Embodiment]
The state determination method according to the first embodiment estimates a respiratory feature amount from a low-frequency component HRV _LF and a high-frequency component HRV _HF in heart rate variability (HRV: Heart Rate Variability), and states based on the estimated value of the respiratory feature amount. It is to determine. In the present embodiment, as an example, a case will be described in which the ratio of low-frequency components to the ventilation volume (for example, tidal volume) (hereinafter referred to as RVV _(LF,n)) is used as the respiratory feature amount.

(Constitution)
FIG. 1 shows a configuration example of a state estimation device according to the first embodiment of the present invention. The state estimation device 10 shown in FIG. 1 includes an electrocardiographic measurement unit 11 and a state estimation unit 12. As an example, the state estimation apparatus 10 is a system in which the electrocardiographic measurement unit 11 is a wearable device that can be worn by a subject (user), and the state estimation unit 12 is a computer device such as a smartphone, a tablet terminal, or a personal computer (PC). It is realized by. For example, the computer device includes a processor such as a CPU (Central Processing Unit), a memory connected to the processor, and a communication interface for communicating (for example, wirelessly) with the electrocardiography measurement unit 11. Note that the implementation of the state estimation device 10 is not limited to this example. For example, the state estimation device 10 may be realized as one device. Further, the electrocardiographic measurement unit 11 may be provided outside the state estimation device 10. In other words, the state estimation device 10 may acquire the result of measuring the electrocardiogram of the subject from an external electrocardiographic measurement device corresponding to the electrocardiographic measurement unit 11.

The electrocardiogram measurement unit 11 measures the electrocardiogram of the subject and sends the measurement result to the state estimation unit 12. The electrocardiogram is a biological signal of the circulatory system and includes, for example, a periodic signal synchronized with the contraction of the ventricle. The electrocardiography measurement unit 11 measures electrocardiography with at least two electrodes. The measurement result includes data capable of extracting the electrocardiogram corresponding to the R wave in the electrocardiogram. For example, the measurement result includes electrocardiographic data. It suffices that the electrocardiography measurement unit 11 be capable of measuring electrocardiograms corresponding to R waves, and the implementation thereof does not matter. For example, the electrocardiographic measurement unit 11 includes a Holter electrocardiograph.

The state estimation unit 12 receives the measurement result from the electrocardiographic measurement unit 11 and estimates the state of the subject based on the received measurement result. For example, the state estimation unit 12 includes a heartbeat characteristic amount calculation unit 121, a respiratory characteristic amount estimation unit 122, a state determination unit 123, a respiratory characteristic amount estimation data recording unit 124, and a state determination data recording unit 125. The functions of the heartbeat characteristic amount calculation unit 121, the respiratory characteristic amount estimation unit 122, the state determination unit 123, the respiratory characteristic amount estimation data recording unit 124, and the state determination data recording unit 125 are, for example, stored in a memory by a processor. It is realized by reading and executing the existing program. Note that some or all of these functions may be realized by a circuit such as an application specific integrated circuit (ASIC).

The heartbeat feature amount calculation unit 121 calculates the heartbeat feature amount HRV _LF and HRV _HF based on the measurement result from the electrocardiographic measurement unit 11. For example, the heartbeat characteristic amount calculation unit 121 extracts an R wave from the measurement result, calculates an instantaneous heartbeat (RRI) that is an interval between adjacent R waves, and thereby generates RRI time-series data. Then, the heartbeat characteristic amount calculation unit 121 performs frequency spectrum analysis on the time series data of RRI to calculate HRV _LF and HRV _HF . In order to improve the calculation accuracy of HRV _LF and HRV _HF , processing such as exclusion of abnormal values may be executed before calculation of the heartbeat characteristic amount. Further, as the calculation method used in the frequency spectrum analysis, any method such as a periodogram or an autoregressive model may be used as long as the same analysis method is used for calculating HRV _LF and HRV _HF . As an example, the heartbeat characteristic amount calculation unit 121 calculates the power spectral density using an autoregressive model, calculates the integrated value of the power over the frequency range from 0.05 Hz to 0.15 Hz as HRV _LF , and calculates it as 0.15 Hz. the integrated value of the power over the frequency range of up to 0.40Hz from calculated as _{HRV HF.}

The respiratory feature amount estimation unit 122 estimates the respiratory feature amount based on the heartbeat feature amount calculated by the heartbeat feature amount calculation unit 121. For example, the respiratory feature amount estimation unit 122 receives the data (respiratory feature amount estimation data) necessary for estimating the respiratory feature amount from the respiratory feature amount estimation data recording unit 124, and uses the respiratory feature amount estimation data. Then, the respiratory feature amount is estimated from the heartbeat feature amount. The processing by the respiratory feature amount estimation unit 122 may be any process as long as the respiratory feature amount can be estimated from the heartbeat feature amount. As an example, respiratory characteristic quantity estimating section 122 calculates a respiration characteristic amount _{RVV (LF, n)} using equation (2) generated by multiple regression analysis.
In this example, α, β, and γ correspond to the respiratory feature amount estimation data.

The state determination unit 123 determines the task state of the subject based on the respiratory feature amount calculated by the respiratory feature amount estimation unit 122. For example, the state determination unit 123 receives data necessary for state determination (state determination data) from the state determination data recording unit 125, and uses the state determination data to determine the task state from the respiratory feature amount. The state determination data may be, for example, a pattern recognition model such as SVM. The state determination unit 123 may determine whether the subject is in one of a plurality of types of task states, and determines whether the subject is in a task state (specifically, whether the subject is in a stress state or not). Which of a relaxed state such as a rest state) is determined.

(motion)
Next, the state estimation operation by the state estimation device 10 according to the present embodiment will be described.
The state estimation processing according to this embodiment includes two phases, a learning phase and an operation phase. The learning phase is for generating data to be recorded in the respiratory feature amount estimating data recording unit 124 and the state determining data recording unit 125. In the operation phase, the respiratory feature quantity and the state are estimated using the data generated in the learning phase.

When using the same method as this embodiment, it is necessary to carry out the learning phase before the execution of the operation phase. In the present embodiment, the update frequency of the values held in the respiratory feature amount estimation data recording unit 124 and the state determination data recording unit 125 created in the learning phase is not specified. The value obtained in the initially implemented learning phase may be continuously used, or the learning phase may be re-executed using appropriate learning data after execution of the operation phase.

<Learning phase>
In the learning phase, respiratory feature amount estimation data used when estimating the respiratory feature amount from the heartbeat feature amount and state determination data used when determining the task state from the respiratory feature amount are generated. In this embodiment, as an example, the respiratory feature amount estimating unit 122 uses multiple regression analysis, and the state determining unit 123 uses a support vector machine.

(A) Generation of Respiratory Feature Amount Estimation Data In this process, the coefficient and intercept of the multiple regression equation used when estimating the respiratory feature amount from the heartbeat feature amount are determined. For this reason, in performing this process, a plurality of data pairs of the respiratory feature amount and the heartbeat feature amount that are correct values are required. An example of a system used for this processing is shown in FIG. The respiratory characteristic amount estimation data generation system shown in FIG. 2 includes an electrocardiographic measurement unit 21, a heartbeat characteristic amount calculation unit 22, a respiratory measurement unit 23, a respiratory characteristic amount calculation unit 24, a respiratory characteristic amount estimation data generation unit 25, and A respiratory feature quantity estimation data recording unit 26 is provided.

The electrocardiography measurement unit 21 measures the electrocardiogram of the subject, which is a sample. The heartbeat feature amount calculation unit 22 calculates the heartbeat feature amount HRV _LF and HRV _HF based on the measurement result by the electrocardiographic measurement unit 21. The electrocardiographic measurement unit 21 and the heartbeat characteristic amount calculation unit 22 can perform the same processing as the electrocardiographic measurement unit 11 and the heartbeat characteristic amount calculation unit 121 illustrated in FIG. 1, respectively. Therefore, detailed description of the electrocardiographic measurement unit 21 and the heartbeat characteristic amount calculation unit 22 will be omitted.

The respiration measurement unit 23 measures respiration of the same subject as the subject of electrocardiography measurement, and generates respiration curve data. The respiratory measurement unit 23 only needs to be able to acquire the characteristic amount corresponding to the ventilation amount of inspiration and expiration, and the realization form thereof does not matter. For example, the respiration measurement unit 23 can use a piezo sensor, a heat wire respiration flowmeter, a pneumotachography, or the like.

The respiratory characteristic amount calculation unit 24 analyzes the respiratory curve obtained by the respiratory measurement unit 23 to calculate the respiratory characteristic amount. As an example, a case will be described in which a feature amount related to ventilation is used as a breathing feature amount. The respiratory characteristic amount calculation unit 24 extracts time series data of ventilation volume from the respiratory curve, and calculates the frequency spectrum by the same method as the heartbeat characteristic amount calculation unit 22. Here, the low frequency component of the calculated ventilation is called RRV _LF and the high frequency component is called RRV _HF . RRV _{(LF,n), which} is the objective variable of multiple regression analysis, is calculated according to equation (3).

The respiratory feature amount estimation data creation unit 25 determines the coefficients and intercepts of the multiple regression equation for estimating the respiratory feature amount from the heartbeat feature amount, using HRV _LF and HRV _HF as explanatory variables and RRV _(LF,n) as the objective variable. .. In the present embodiment, the above equation (2) is assumed as a multiple regression equation. A plurality of RRV _(LF,n) and corresponding HRV _LF and HRV _HF data pairs are prepared by using the electrocardiographic measurement unit 21, the heartbeat characteristic amount calculation unit 22, the respiration measurement unit 23, and the respiration characteristic amount calculation unit 24. Then, the respiratory feature amount estimation data creation unit 25 obtains the coefficients α and β and the intercept γ in the equation (2) based on these data pairs, and estimates the obtained coefficients α and β and the intercept γ as the respiratory feature amount. The data is stored in the business data recording unit 26. The same data as that stored in the respiratory characteristic amount estimation data recording unit 26 is stored in the respiratory characteristic amount estimation data recording unit 124 shown in FIG.

(B) Generation of State Discrimination Data In this processing, a discriminator used for discriminating the task state is constructed. For this reason, in order to carry out this processing, it is necessary to have a plurality of data pairs of the state label and the respiratory feature amount that are correct values. An example of a system used for this processing is shown in FIG. The state determination data generation system shown in FIG. 3 includes a respiration measurement unit 31, a respiratory feature amount calculation unit 32, a state determination device creation unit 33, a state label recording unit 34, and a state determination data recording unit 35.

The respiration measurement unit 31 measures the respiration of the subject, which is a sample, and generates respiration curve data. The respiratory characteristic amount calculation unit 24 analyzes the respiratory curve obtained by the respiratory measurement unit 23 and calculates the respiratory characteristic amount RVV _(LF,n) . The respiration measurement unit 31 and the respiration characteristic amount calculation unit 32 can perform the same processing as the respiration measurement unit 23 and the respiration characteristic amount calculation unit 24 illustrated in FIG. 2, respectively. Therefore, detailed description of the breath measuring unit 31 and the breath characteristic amount calculating unit 32 will be omitted.

The state label recording unit 34 holds a state label indicating the task state of the subject at the time when the respiratory feature amount is being measured. The state discriminator creation unit 33 determines a discriminator necessary for estimating the task state from the respiratory feature amount based on the respiratory feature amount calculated by the respiratory feature amount calculation unit 32 and the corresponding task state label. create. Specifically, RVV _(LF,n) is calculated in various task states, and a large number of data labeled with RVV _(LF,n) in task states are prepared. The state discriminator creation unit 33 creates an SVM that discriminates the task state by RVV _(LF,n) and stores it in the state discrimination data recording unit 35. The same data as that stored in the state determination data recording section 35 is stored in the state determination data recording section 125 shown in FIG.

An estimated value may be used for the respiratory feature amount used by the state discriminator creation unit 33. For example, the breathing feature amount may be estimated using the breathing feature amount estimation data created in the process (A) described above. FIG. 4 shows an example of a system using the estimated value of the respiratory feature amount. The state discrimination data generation system shown in FIG. 4 includes an electrocardiographic measurement unit 41, a heartbeat characteristic amount calculation unit 42, a respiratory characteristic amount estimation unit 43, a respiratory characteristic amount estimation data recording unit 44, a state discriminator creation unit 33, and a state. A label recording unit 34 and a state determination data recording unit 35 are provided. 4, parts that are the same as those in FIG. 3 are given the same reference numerals, and descriptions thereof will be omitted.

The respiratory feature amount estimation data recording unit 44 holds the respiratory feature amount estimation data created in the process (A) described above. The electrocardiogram measurement unit 41 measures the electrocardiogram of the subject as a sample. The heartbeat characteristic amount calculation unit 42 calculates the heartbeat characteristic amount based on the measurement result from the electrocardiographic measurement unit 41. The respiratory characteristic amount estimating unit 43 uses the respiratory characteristic amount estimating data held in the respiratory characteristic amount estimating data recording unit 44 to calculate the respiratory characteristic amount from the heartbeat characteristic amount calculated by the heartbeat characteristic amount calculating unit 42. Estimate and output the estimated value of the respiratory feature amount to the state discriminator creation unit 33. The electrocardiographic measurement unit 41, the heartbeat characteristic amount calculation unit 42, and the respiratory characteristic amount estimation unit 43 are similar to the electrocardiographic measurement unit 11, the heartbeat characteristic amount calculation unit 121, and the respiratory characteristic amount estimation unit 122 illustrated in FIG. 1, respectively. Can be performed. Therefore, detailed description of the electrocardiography measurement unit 41, the heartbeat feature amount calculation unit 42, and the respiratory feature amount estimation unit 43 will be omitted.

<Operation phase>
In the operation phase, the state estimation device 10 shown in FIG. 1 estimates the respiratory feature amount from the heartbeat feature amount, and the state is determined using the estimation result.

FIG. 5 shows an example of the procedure of the state determination processing according to this embodiment.
In step S501 of FIG. 5, the electrocardiographic measurement unit 11 measures the electrocardiogram of the subject. In step S502, the heartbeat characteristic amount calculation unit 121 calculates the low frequency component HRV _LF and the high frequency component HRV _HF of the heartbeat fluctuation based on the measurement result by the electrocardiography measurement unit 11. For example, the heartbeat characteristic amount calculation unit 121 extracts an R wave from the electrocardiographic signal output from the electrocardiographic measurement unit 11, calculates an RRI from two adjacent R waves, and calculates the frequency spectrum of the RRI sequence as an autoregressive model. And HRV _LF and HRV _HF are calculated.

In step S503, the respiratory feature amount estimation unit 122 uses the data stored in the respiratory feature amount estimation data recording unit 124 to calculate the estimated respiratory feature amount value RVV _(LF, HRV _LF and HRV _{HF) . n)} is calculated. For example, the respiratory feature amount estimation unit 122 calculates RVV _(LF,n) according to the above equation (2), which is a multiple regression equation with HRV _LF and HRV _HF as explanatory variables.

In step S504, the state determination unit 123 determines the task state of the subject from the estimated value RVV _{(LF,n) of} the respiratory feature amount using the data stored in the state determination data recording unit 125.

(effect)
As described above, in the present embodiment, the relationship between the heartbeat characteristic amount and the respiratory characteristic amount, and the relationship between the respiratory characteristic amount and the task state are learned in advance, and the respiratory characteristic amount is estimated from the heartbeat characteristic amount. , The task state is determined from the estimated respiratory feature amount. This enables state estimation that reflects the influence of breathing on heart rate variability. As a result, it is possible to estimate the state with high accuracy without controlling breathing. Furthermore, since the respiratory feature amount is estimated from the heartbeat feature amount, it is not necessary to introduce a device for measuring the respiratory feature amount in the operation phase.

[Other Embodiments]
The present invention is not limited to the above-described first embodiment. For example, the respiratory characteristic amount is not limited to RVV _(LF,n), and other indices related to respiration may be used. For example, the respiratory characteristic amount may be based on the respiratory rate, or may be based on both the ventilation amount and the respiratory rate.

Further, the respiratory feature amount may be estimated based on the high frequency component HRV _HF of the heartbeat variability without using the low frequency component HRV _LF of the heartbeat variability. This is because the contribution rate of the high-frequency component of the heart rate variability is higher, which is effective when the accuracy reduction can be tolerated. As an example, the respiratory feature amount RVV _(LF,n) can be calculated from the high-frequency component HRV _{HF of} heart rate variability according to the equation (4) generated by simple regression analysis .
In this example, β and γ correspond to the respiratory feature amount estimation data.

In short, the present invention is not limited to the above-described embodiments as they are, and can be embodied by modifying the constituent elements within a range not departing from the gist of the invention in an implementation stage. Further, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the above embodiments. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, the constituent elements of different embodiments may be combined appropriately.

DESCRIPTION OF SYMBOLS 10... State estimation device, 11... Electrocardiography measurement part, 12... State estimation part, 121... Heartbeat feature amount calculation part, 122... Respiration feature amount estimation part, 123... State determination part, 124... Respiration feature amount estimation data record Reference numeral 125: State determination data recording unit, 21... Electrocardiographic measurement unit, 22... Heartbeat feature amount calculation unit, 23... Respiration measurement unit, 24... Respiration feature amount calculation unit, 25... Respiration feature amount estimation data creation unit , 26... Respiratory feature amount estimating data recording unit, 31... Respiratory measuring unit, 32... Respiratory feature amount calculating unit, 33... State discriminator creating unit, 34... State label recording unit, 35... State discriminating data recording unit, 41... Electrocardiography measurement unit, 42... Heartbeat feature amount calculation unit, 43... Respiration feature amount estimation unit, 44... Respiration feature amount estimation data recording unit.

Claims (6)

Calculation means for calculating a high-frequency component of heart rate variability based on a signal output from an electrocardiographic measurement means for measuring the electrocardiogram of the subject,
By applying the high frequency component of heart rate variability in relation, which is calculated to represent the relationship between the respiratory characteristic quantity is a ratio of the low-frequency component of the high frequency component and ventilation of the heart rate variability, the breathing characteristic quantity An estimation means for estimating
Discriminating means for discriminating the state of the subject based on the estimated respiratory feature amount,
A state estimation device including. The state estimation device according to claim 1, wherein the relational expression is generated in advance by a single regression analysis using a high frequency component of the heartbeat variability as an explanatory variable and the respiratory feature amount as an objective variable . The calculating means further calculates a low frequency component of heart rate variability based on the signal,
The relational expression represents a relation between the low-frequency component and the high-frequency component of the heartbeat variability and the respiratory characteristic amount that is the ratio of the low-frequency component of the ventilation volume,
The estimating means, the relational expression by applying a low frequency component and the high frequency component of the calculated the heart rate variability, estimates the respiratory feature amount, the state estimation device according to claim 1. The state estimation device according to claim 3, wherein the relational expression is generated in advance by multiple regression analysis using the low frequency component and the high frequency component of the heartbeat variability as explanatory variables and the respiratory feature amount as an objective variable . Calculating a high-frequency component of heart rate variability based on a signal obtained by measuring the electrocardiogram of the subject,
By applying the high frequency component of heart rate variability in relation, which is calculated to represent the relationship between the respiratory characteristic quantity is a ratio of the low-frequency component of the high frequency component and ventilation of the heart rate variability, the breathing characteristic quantity To estimate
Discriminating the state of the subject based on the estimated respiratory feature amount,
A state estimation method comprising: A program for causing a computer to function as each unit included in the state estimation device according to any one of claims 1 to 4 .