Learning device, information providing device, learning method, information providing method, and program

Abstract

The learning device uses, as a myocardial activity parameter set, a set of parameters representing an approximation function when the approximation function including the difference of the cumulative distribution function or the weighted difference approximates the waveform of the time interval included in the waveform representing the amount of one cycle of the heart cycle of the heart, and the learning data set is composed of a plurality of (Y) pieces of learning data. A heart state estimation device includes a learning unit that learns an estimation model using a learning data group, and estimates heart state information, which is information indicating a state of a heart of a subject of y-th learning data, where each integer of 1 to Y is defined as y, and that estimates the heart state of the subject of the y-th learning data by learning the estimation model using the learning data group, and that estimates the heart state of the subject of the y-th learning data based on the heart state information of the subject of the y-th learning data. The estimation model obtains heart state information, which is information indicating the state of the heart corresponding to a myocardial activity parameter, using the myocardial activity parameter set as an input.

Description

Technical Field

[0001] The present invention relates to a learning device, an information providing device, a learning method, an information providing method and a program. Background Art

[0002] The electrocardiogram is useful information for understanding the state of the heart. For example, using the electrocardiogram, it is possible to determine whether the subject is in a state with a high possibility of developing heart failure (Non-Patent Document 1).

[0003] Prior art literature Non-patent literature Non-patent literature 1: Hiroshi Tanaka, "Methodology of the Electrocardiographic Inverse Problem", Medical Electronics and Bioengineering, 1985, Vol. 23, No. 3, p. 147-158. Summary of the invention

[0004] Problems to be solved by the invention However, the waveform of the electrocardiogram may not be sufficient to understand the state of the heart. For example, even if the waveform of the electrocardiogram is similar, the way in which heart-related diseases develop is sometimes different. In this way, it is sometimes difficult to understand the state of the heart just by observing the waveform of the electrocardiogram itself, depending on the disease. For example, in the case of heart failure, its suppression can be suppressed by observing the state of the heart based on the waveform of the electrocardiogram in daily life. In order to further improve the accuracy of suppressing the onset of the disease, other technologies such as blood collection can be considered to obtain other information, but it is unrealistic to do this in daily life. Therefore, depending on the disease, it is sometimes necessary to essentially grasp the state of the heart based only on the waveform of the electrocardiogram.

[0005] In addition, this situation is not limited to the situation where the state of the heart is understood based on the waveform of the electrocardiogram. This situation is also common in the case where the state of the heart is understood based on biological information of a time series of a channel related to the heart beat obtained by a sensor in contact with the body surface, a sensor close to the body surface, a sensor inserted into the body, a sensor buried in the body, etc. In addition, biological information of a time series related to the heart beat is, for example, a waveform showing changes in cardiac potential, a waveform showing changes in cardiac pressure, a waveform showing changes in blood flow, and a waveform showing changes in heart sounds. In addition, the waveform of the electrocardiogram is also an example of biological information of a time series related to the heart beat.

[0006] In view of the above circumstances, an object of the present invention is to provide a technology for obtaining information useful for understanding the state of the heart from one channel of time-series biological information related to cardiac pulsation.

[0007] Solutions to Solve Problems One embodiment of the present invention is a learning device, which includes a learning unit that performs the following operations: taking a waveform of a time interval of an R wave included in a waveform showing one cycle of a cardiac cycle of a heart as a first target time waveform, taking a waveform of a time interval of a T wave included in the waveform as a second target time waveform, taking a waveform after the time axis of the second target time waveform is reversed as a second target inverse time waveform, taking a cumulative distribution function of a first unimodal distribution as a first cumulative distribution function, taking a cumulative distribution function of a second unimodal distribution as a second cumulative distribution function, and taking the following as a plurality of myocardial activity parameters included in a first group: when the first target time waveform is approximated by a time waveform obtained by a difference or a weighted difference between the first cumulative distribution function and the second cumulative distribution function, that is, a first approximate time waveform, or a time waveform obtained by adding a difference or a weighted difference between the first cumulative distribution function and the second cumulative distribution function and a first level value, that is, a first approximate time waveform, determining parameters of the first unimodal distribution, determining parameters of the first cumulative distribution function, determining parameters of the second unimodal distribution, determining parameters of the second cumulative distribution function, weights of the first cumulative distribution function, weights of the second cumulative distribution function, the ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, the first level value, taking the third unimodal distribution cumulative distribution function as the third cumulative distribution function, taking the fourth unimodal distribution cumulative distribution function as the fourth cumulative distribution function, subtracting the third cumulative distribution function from 1 as the third inverse cumulative distribution function, subtracting the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function, and taking the following as the multiple myocardial activity parameters included in the second group: the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function when the second object inverse time waveform is approximated by a waveform obtained by adding a difference or a weighted difference between the third cumulative distribution function and the fourth cumulative distribution function plus a second level value, namely, a second approximate inverse time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, a weight of the third cumulative distribution function, a weight of the fourth cumulative distribution function, a ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value;Alternatively, when the second object time waveform is approximated by a time waveform obtained by a difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, or a time waveform obtained by adding a second level value to a difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, parameters of the third unimodal distribution are determined, parameters of the third cumulative distribution function are determined, parameters of the fourth unimodal distribution are determined, parameters of the fourth cumulative distribution function are determined, a weight of the third inverse cumulative distribution function, a weight of the fourth inverse cumulative distribution function, a ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value, the first object time waveform is compared with the first approximate any one of the time waveform of the difference between the second object time waveform and the second approximate time waveform, the time waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform, and the time waveform after the time axis of the waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform is inverted as the residual time waveform, the cumulative distribution function of the fifth unimodal distribution is used as the fifth cumulative distribution function, and the cumulative distribution function of the sixth unimodal distribution is used as the sixth cumulative distribution function, and the following contents are used as the multiple myocardial activity parameters included in the third group: when the residual time waveform is approximated by the fifth cumulative distribution function or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, that is, the approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function;Alternatively, when the residual time waveform is approximated by a time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, i.e., an approximate residual time waveform, parameters of the fifth unimodal distribution are determined, parameters of the fifth cumulative distribution function are determined, parameters of the sixth unimodal distribution are determined, parameters of the sixth cumulative distribution function are determined, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function; parameters obtained by calculating multiple myocardial activity parameters included in the first group, the multiple myocardial activity parameters included in the second group, and the multiple myocardial activity parameters included in the third group are used as one or more myocardial activity parameters included in the fourth group, and more than one cycle of the cardiac cycle of the heart is selected. A group formed by the plurality of myocardial activity parameters included in the first group, the plurality of myocardial activity parameters included in the second group, the plurality of myocardial activity parameters included in the third group, and one or more myocardial activity parameters included in the fourth group obtained from the waveform is used as a myocardial activity parameter group, the learning data group is composed of Y (Y is a plurality of) learning data, each integer greater than 1 and less than Y is set to y, each of the Y learning data includes a myocardial activity parameter group of the heart as the object of the yth learning data, and information representing the state of the heart as the object of the yth learning data, namely, heart state information, and the learning data group is used to learn an estimation model, and the estimation model uses the myocardial activity parameter group as an input to obtain information representing the state of the heart corresponding to the myocardial activity parameter, namely, heart state information. ;

[0008] One embodiment of the present invention is an information providing device, comprising a state information generating unit that performs the following operations: taking a waveform of a time interval of an R wave included in a waveform representing one cycle of a cardiac cycle of a heart as a first target time waveform, taking a waveform of a time interval of a T wave included in the waveform as a second target time waveform, taking a waveform obtained by inverting the time axis of the second target time waveform as a second target inverse time waveform, taking a cumulative distribution function of a first unimodal distribution as a first cumulative distribution function, taking a cumulative distribution function of a second unimodal distribution as a second cumulative distribution function, and taking the following as a plurality of myocardial activity parameters included in a first group: when the first target time waveform is approximated by a time waveform that is a first approximate time waveform obtained by a difference or a weighted difference between the first cumulative distribution function and the second cumulative distribution function or a time waveform that is a first approximate time waveform obtained by adding a difference or a weighted difference between the first cumulative distribution function and the second cumulative distribution function to a first level value, determining parameters of the first unimodal distribution, determining parameters of the first cumulative distribution function, determining parameters of the second unimodal distribution, determining parameters of the second cumulative distribution function, weights of the first cumulative distribution function, and weights of the second cumulative distribution function. a weight, a ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, the first level value, taking the third unimodal distribution cumulative distribution function as the third cumulative distribution function, taking the fourth unimodal distribution cumulative distribution function as the fourth cumulative distribution function, subtracting the third cumulative distribution function from 1 as the third inverse cumulative distribution function, subtracting the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function, and taking the following as the multiple myocardial activity parameters included in the second group: a value obtained by the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function when the second object inverse time waveform is approximated by the waveform obtained, i.e., the second approximate inverse time waveform, or the waveform obtained by adding the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function plus the second level value, i.e., the second approximate inverse time waveform, determining the parameters of the third unimodal distribution, determining the parameters of the third cumulative distribution function, determining the parameters of the fourth unimodal distribution, determining the parameters of the fourth cumulative distribution function, the weight of the third cumulative distribution function, the weight of the fourth cumulative distribution function, the ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value;Alternatively, when the second object time waveform is approximated by a time waveform obtained by a difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, or a time waveform obtained by adding a second level value to a difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, parameters of the third unimodal distribution are determined, parameters of the third cumulative distribution function are determined, parameters of the fourth unimodal distribution are determined, parameters of the fourth cumulative distribution function are determined, a weight of the third inverse cumulative distribution function, a weight of the fourth inverse cumulative distribution function, a ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value, the first object time waveform is compared with the first approximate any one of the time waveform of the difference between the second object time waveform and the second approximate time waveform, the time waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform, and the time waveform after the time axis of the waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform is inverted as the residual time waveform, the cumulative distribution function of the fifth unimodal distribution is used as the fifth cumulative distribution function, and the cumulative distribution function of the sixth unimodal distribution is used as the sixth cumulative distribution function, and the following contents are used as the multiple myocardial activity parameters included in the third group: when the residual time waveform is approximated by the fifth cumulative distribution function or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, that is, the approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function;Alternatively, when the residual time waveform is approximated by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, that is, the approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, the parameters of the sixth unimodal distribution are determined, the parameters of the sixth cumulative distribution function, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function, and the parameters obtained by calculating the multiple myocardial activity parameters included in the first group, the multiple myocardial activity parameters included in the second group, and the multiple myocardial activity parameters included in the third group are used as the parameters included in the fourth group. At least one myocardial activity parameter of the first group, at least one myocardial activity parameter of the second group, at least one myocardial activity parameter of the third group, and at least one myocardial activity parameter of the fourth group, obtained from waveforms showing at least one cycle of the cardiac cycle of the heart, is used as a myocardial activity parameter group, an estimation model is pre-stored which uses the myocardial activity parameter group as input to obtain information indicating the state of the heart corresponding to the myocardial activity parameter group, namely, heart state information, and the heart state information of the information providing target heart is obtained by using the estimation model to use the myocardial activity parameter group of the heart as the information providing target heart as input. ;

[0009] One embodiment of the present invention is a learning method executed by a learning device, wherein the learning step includes the following operations: taking a waveform of a time interval of an R wave included in a waveform representing one cycle of a cardiac cycle of a heart as a first target time waveform, taking a waveform of a time interval of a T wave included in the waveform as a second target time waveform, taking a waveform obtained by inverting the time axis of the second target time waveform as a second target inverse time waveform, taking a cumulative distribution function of a first unimodal distribution as a first cumulative distribution function, taking a cumulative distribution function of a second unimodal distribution as a second cumulative distribution function, and taking the following as a plurality of myocardial activity parameters included in a first group: when the first target time waveform is approximated by a time waveform that is a first approximate time waveform obtained by a difference or a weighted difference between the first cumulative distribution function and the second cumulative distribution function or a time waveform that is a first approximate time waveform obtained by adding a difference or a weighted difference between the first cumulative distribution function and the second cumulative distribution function to a first level value, determining parameters of the first unimodal distribution, determining parameters of the first cumulative distribution function, determining parameters of the second unimodal distribution, determining parameters of the second cumulative distribution function, weights of the first cumulative distribution function, and the second cumulative distribution function. the weight of the first cumulative distribution function, the ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, the first level value, taking the third unimodal distribution cumulative distribution function as the third cumulative distribution function, taking the fourth unimodal distribution cumulative distribution function as the fourth cumulative distribution function, subtracting the third cumulative distribution function from 1 as the third inverse cumulative distribution function, subtracting the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function, and taking the following as the multiple myocardial activity parameters included in the second group: the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function When the waveform obtained is the second approximate inverse time waveform or the waveform obtained by adding the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function plus the second level value to approximate the second object inverse time waveform, determine the parameters of the third unimodal distribution, determine the parameters of the third cumulative distribution function, determine the parameters of the fourth unimodal distribution, determine the parameters of the fourth cumulative distribution function, the weight of the third cumulative distribution function, the weight of the fourth cumulative distribution function, the ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value;Alternatively, when the second object time waveform is approximated by a time waveform obtained by a difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, or a time waveform obtained by adding a second level value to a difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, parameters of the third unimodal distribution are determined, parameters of the third cumulative distribution function are determined, parameters of the fourth unimodal distribution are determined, parameters of the fourth cumulative distribution function are determined, a weight of the third inverse cumulative distribution function, a weight of the fourth inverse cumulative distribution function, a ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value, the first object time waveform is compared with the first approximate any one of the time waveform of the difference between the second object time waveform and the second approximate time waveform, the time waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform, and the time waveform after the time axis of the waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform is inverted as the residual time waveform, the cumulative distribution function of the fifth unimodal distribution is used as the fifth cumulative distribution function, and the cumulative distribution function of the sixth unimodal distribution is used as the sixth cumulative distribution function, and the following contents are used as the multiple myocardial activity parameters included in the third group: when the residual time waveform is approximated by the fifth cumulative distribution function or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, that is, the approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function;Alternatively, when the residual time waveform is approximated by a time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, i.e., an approximate residual time waveform, parameters of the fifth unimodal distribution are determined, parameters of the fifth cumulative distribution function are determined, parameters of the sixth unimodal distribution are determined, parameters of the sixth cumulative distribution function are determined, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function; parameters obtained by calculating multiple myocardial activity parameters included in the first group, the multiple myocardial activity parameters included in the second group, and the multiple myocardial activity parameters included in the third group are used as one or more myocardial activity parameters included in the fourth group, and more than one cycle of the cardiac cycle of the heart is selected. A group formed by the plurality of myocardial activity parameters included in the first group, the plurality of myocardial activity parameters included in the second group, the plurality of myocardial activity parameters included in the third group, and one or more myocardial activity parameters included in the fourth group obtained from the waveform is used as a myocardial activity parameter group, the learning data group is composed of Y (Y is a plurality of) learning data, each integer greater than 1 and less than Y is set to y, each of the Y learning data includes a myocardial activity parameter group of the heart as the object of the yth learning data, and information representing the state of the heart as the object of the yth learning data, namely, heart state information, and the learning data group is used to learn an estimation model, and the estimation model uses the myocardial activity parameter group as an input to obtain information representing the state of the heart corresponding to the myocardial activity parameter, namely, heart state information. ;

[0010] One embodiment of the present invention is an information providing method executed by an information providing device, comprising a state information generating step, which performs the following operations: taking a waveform of a time interval of an R wave included in a waveform showing one cycle of a cardiac cycle of a heart as a first target time waveform, taking a waveform of a time interval of a T wave included in the waveform as a second target time waveform, taking a waveform obtained by inverting the time axis of the second target time waveform as a second target inverse time waveform, taking a cumulative distribution function of a first unimodal distribution as a first cumulative distribution function, taking a cumulative distribution function of a second unimodal distribution as a second cumulative distribution function, and taking the following as a plurality of myocardial activity parameters included in a first group: when the first target time waveform is approximated by a time waveform obtained by a difference or a weighted difference between the first cumulative distribution function and the second cumulative distribution function, that is, a first approximate time waveform, or a time waveform obtained by adding a difference or a weighted difference between the first cumulative distribution function and the second cumulative distribution function to a first level value, determining parameters of the first unimodal distribution, determining parameters of the first cumulative distribution function, determining parameters of the second unimodal distribution, determining parameters of the second cumulative distribution function, weights of the first cumulative distribution function, and weights of the second cumulative distribution function; distribution function, the ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, the first level value, taking the third unimodal distribution cumulative distribution function as the third cumulative distribution function, taking the fourth unimodal distribution cumulative distribution function as the fourth cumulative distribution function, subtracting the third cumulative distribution function from 1 as the third inverse cumulative distribution function, subtracting the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function, and taking the following as the multiple myocardial activity parameters included in the second group: the difference or weighted sum of the third cumulative distribution function and the fourth cumulative distribution function when approximating the second object inverse time waveform by a waveform obtained by the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function plus a second level value, namely, a waveform obtained by the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function, namely, a second approximate inverse time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, a weight of the third cumulative distribution function, a weight of the fourth cumulative distribution function, a ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value;Alternatively, when the second object time waveform is approximated by a time waveform obtained by a difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, or a time waveform obtained by adding a second level value to a difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, parameters of the third unimodal distribution are determined, parameters of the third cumulative distribution function are determined, parameters of the fourth unimodal distribution are determined, parameters of the fourth cumulative distribution function are determined, a weight of the third inverse cumulative distribution function, a weight of the fourth inverse cumulative distribution function, a ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value, the first object time waveform is compared with the first approximate any one of the time waveform of the difference between the second object time waveform and the second approximate time waveform, the time waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform, and the time waveform after the time axis of the waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform is inverted as the residual time waveform, the cumulative distribution function of the fifth unimodal distribution is used as the fifth cumulative distribution function, and the cumulative distribution function of the sixth unimodal distribution is used as the sixth cumulative distribution function, and the following contents are used as the multiple myocardial activity parameters included in the third group: when the residual time waveform is approximated by the fifth cumulative distribution function or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, that is, the approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function;Alternatively, when the residual time waveform is approximated by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, that is, the approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, the parameters of the sixth unimodal distribution are determined, the parameters of the sixth cumulative distribution function, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function, and the parameters obtained by calculating the multiple myocardial activity parameters included in the first group, the multiple myocardial activity parameters included in the second group, and the multiple myocardial activity parameters included in the third group are used as the parameters included in the fourth group. At least one myocardial activity parameter of the first group, at least one myocardial activity parameter of the second group, at least one myocardial activity parameter of the third group, and at least one myocardial activity parameter of the fourth group, obtained from waveforms showing at least one cycle of the cardiac cycle of the heart, is used as a myocardial activity parameter group, an estimation model is pre-stored which uses the myocardial activity parameter group as input to obtain information indicating the state of the heart corresponding to the myocardial activity parameter group, namely, heart state information, and the heart state information of the information providing target heart is obtained by using the estimation model to use the myocardial activity parameter group of the heart as the information providing target heart as input. ;

[0011] One embodiment of the present invention is a program for causing a computer to function as the above-mentioned learning device.

[0012] One embodiment of the present invention is a program for causing a computer to function as the above-mentioned information providing device.

[0013] Effects of the Invention According to the present invention, it is possible to provide a technology for obtaining information useful for understanding the state of the heart from one channel of time-series biological information related to cardiac pulsation. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a diagram showing an example of the hardware configuration of the signal analysis device 1 according to the first embodiment. Figure 2 A diagram schematically showing a function of a first cumulative distribution function multiplied by a weight, a function of a second cumulative distribution function multiplied by a weight, and an approximate time waveform which is a weighted difference between the first cumulative distribution function and the second cumulative distribution function, with respect to a first target time waveform. Figure 3A diagram schematically showing a function of the third inverse cumulative distribution function multiplied by a weight, a function of the fourth inverse cumulative distribution function multiplied by a weight, and an approximate time waveform which is a weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, with respect to the second target time waveform. Figure 4 This is a diagram showing an example of the result of fitting the waveform of the electrocardiogram of the target heart in the first embodiment using four cumulative distribution functions. Figure 5 This is an explanatory diagram for explaining that the difference between two cumulative distribution functions in the first embodiment can be fitted to a waveform that is substantially the same as the waveform of the falling T wave. Figure 6 It is a diagram schematically showing, with respect to the time waveform of the first object, a function of multiplying the first cumulative distribution function by a weight and adding a level value, a function of multiplying the second cumulative distribution function by a weight and adding a level value, and an approximate time waveform obtained by adding the weighted difference between the first cumulative distribution function and the second cumulative distribution function and the level value. Figure 7 It is a diagram schematically showing, with respect to the time waveform of the second object, an approximate time waveform obtained by multiplying the third inverse cumulative distribution function by a weight and adding a level value, a function multiplying the fourth inverse cumulative distribution function by a weight and adding a level value, and the weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function plus the level value. Figure 8 is a diagram schematically showing a Δ wave included in a target time waveform. Fig. 9 1 is a diagram showing an example of the functional configuration of the control unit 11 in the first embodiment. Fig.10 This is a flowchart showing an example of a process flow executed by the signal analysis device 1 in the first embodiment. Fig.11 This is a first diagram showing an example of analysis results of the signal analysis device 1 according to the first embodiment. Fig.12 The second diagram shows an example of the analysis result of the signal analysis device 1 according to the first embodiment. Fig.13 The third diagram shows an example of the analysis result of the signal analysis device 1 in the first embodiment. Fig.14 The fourth diagram shows an example of the analysis result of the signal analysis device 1 in the first embodiment. Fig.15 This is a first explanatory diagram of an example of an electrocardiogram of a ventricular extrasystole analyzed by the signal analysis device 1 according to the first embodiment. Fig.16This is a second explanatory diagram showing an example of an electrocardiogram of a ventricular extrasystole analyzed by the signal analysis device 1 according to the first embodiment. Fig.17 This is a third explanatory diagram showing an example of an electrocardiogram of a ventricular extrasystole analyzed by the signal analysis device 1 according to the first embodiment. Fig.18 This is a first explanatory diagram showing an electrocardiogram of a target heart in the depolarization phase of Buddag syndrome type 1 analyzed by the signal analysis device 1 according to the first embodiment. Fig.19 This is a second explanatory diagram showing an electrocardiogram of a target heart in the depolarization phase of Buddag syndrome type 1 analyzed by the signal analysis device 1 according to the first embodiment. Fig. 20 This is a third explanatory diagram showing an electrocardiogram of a target heart in the depolarization phase of Buddag syndrome type 1 analyzed by the signal analysis device 1 according to the first embodiment. Fig.21 This is a diagram showing a first example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig. 22 This is a diagram showing a second example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.23 This is a diagram showing a third example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.24 This is a diagram showing a fourth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.25 This is a diagram showing a fifth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.26 This is a diagram showing a sixth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig. 27 This is a diagram showing a seventh example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.28 This is a diagram showing an eighth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.29 This is a diagram showing a ninth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.30 This is a diagram showing a tenth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.31 This is a diagram showing an eleventh example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.32This is a diagram showing a twelfth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.33 This is a diagram showing a thirteenth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.34 This is a diagram showing a fourteenth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.35 This is a diagram showing a fifteenth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.36 This is a diagram showing a sixteenth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.37 This is a diagram showing a seventeenth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.38 This is a diagram showing an eighteenth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.39 This is a diagram showing a nineteenth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.40 This is a diagram showing the twentieth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.41 This is a diagram showing a twenty-first example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.42 This is a diagram showing a twenty-second example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.43 This is a diagram showing a twenty-third example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.44 This is a diagram showing a twenty-fourth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.45 This is a diagram showing a twenty-fifth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.46 This is a diagram showing a twenty-sixth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.47 This is a diagram showing a twenty-seventh example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.48 This is a diagram showing a twenty-eighth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.49This is a diagram showing a twenty-ninth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.50 This is a diagram showing the thirtieth example of the electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.51 This is a diagram showing a thirty-first example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.52 This is a diagram showing a thirty-second example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.53 This is a diagram showing a thirty-third example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.54 This is a diagram showing a thirty-fourth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.55 This is a diagram showing a thirty-fifth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.56 This is a diagram showing a thirty-sixth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.57 This is a diagram showing a thirty-seventh example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.58 This is a diagram showing a thirty-eighth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.59 This is a diagram showing a thirty-ninth example of an electrocardiogram analyzed by the signal analysis device 1 according to the first embodiment. Fig.60 It is a diagram showing an example of the functional configuration of the information providing system 200 according to the second embodiment. Fig.61 This is a flowchart showing an example of a process flow executed by the learning device 300 according to the second embodiment. Fig.62 This is a flowchart showing an example of the flow of processing executed by the information providing device 400 according to the second embodiment. Fig.63 It is a diagram showing an example of the functional configuration of an information providing system 201 according to the third embodiment. Fig.64 This is a flowchart showing an example of a process flow executed by the learning device 301 according to the third embodiment. Fig.65 This is a flowchart showing an example of the flow of processing executed by the information providing device 401 according to the third embodiment. Fig.66This is a diagram showing an example of the functional configuration of an information providing system 202 according to a modification of the second embodiment and the third embodiment. Fig.67 This is a diagram showing an example of the functional configuration of an information providing system 202 according to a modification of the second embodiment and the third embodiment. DETAILED DESCRIPTION

[0015] <First Embodiment> Figure 1 1 is a diagram showing an example of the hardware structure of the signal analysis device 1 of the first embodiment. In the following, for the sake of simplicity, the signal analysis device 1 is described by taking the case of analyzing the waveform of one channel of the electrocardiogram as an example. However, the signal analysis device 1 is not limited to the waveform of the electrocardiogram, and the same analysis can be performed if it is based on the biological information of the time series related to the heart beat. In addition, the biological information of the time series related to the heart beat is, for example, a waveform showing the change of the cardiac potential, a waveform showing the change of the pressure of the heart, a waveform showing the change of the blood flow, and a waveform showing the change of the heart sound. Therefore, the signal analysis device 1 is not limited to the waveform of the electrical signal obtained from the body surface, and a sensor in contact with the body surface, a sensor close to the body surface, a sensor inserted into the body, a sensor buried in the body, etc., whether it is on the body surface or in the body, etc., as long as it shows the waveform of the cardiac cycle obtained from a certain point, any can be used.

[0016] That is, instead of the waveform of the electrocardiogram, the signal analysis device 1 may use a waveform showing changes in cardiac pressure as biological information of a time series related to cardiac pulsation. In addition, instead of the waveform of the electrocardiogram, the signal analysis device 1 may use a waveform showing changes in blood flow as biological information of a time series related to cardiac pulsation. In addition, instead of the waveform of the electrocardiogram, the signal analysis device 1 may use a waveform showing changes in heart sounds as biological information of a time series related to cardiac pulsation. In addition, the waveform of the electrocardiogram is also an example of biological information of a time series related to cardiac pulsation. In addition, biological information of a time series related to cardiac pulsation may also be biological information of a time series related to the periodic pulsation of the heart.

[0017] The signal analysis device 1 acquires the waveform of an electrocardiogram of the heart to be analyzed (hereinafter referred to as the "target heart"). Based on the acquired electrocardiogram waveform, the signal analysis device 1 acquires a parameter indicating the activity of the myocardium of the target heart (hereinafter referred to as the "myocardial activity parameter"), which includes at least any one of a parameter indicating the activity of the outer layer of the myocardium of the target heart (hereinafter referred to as the "outer myocardial layer parameter") and a parameter indicating the activity of the inner layer of the myocardium of the target heart (hereinafter referred to as the "inner myocardial layer parameter").

[0018] Here, the relationship between the activity of the myocardium and the electrocardiogram waveform will be described. In the medical field, there is a known model called the cardiac electromotive force dipole model (refer to Reference 1) that explains the relationship between myocardial movement and the electrocardiogram. According to the cardiac electromotive force dipole model, the myocardium is modeled by two layers, the outer myocardial layer and the inner myocardial layer.

[0019] Reference 1: Yoshifumi Tanaka, "Understanding Electrocardiogram Waveforms from the Ground Up: Deciphering the Action Potentials of the Myocardium", Gakken Medical Shujunsha (2012).

[0020] In the cardiac electromotive force dipole model, the outer myocardial layer and the inner myocardial layer are respectively modeled as sources of different electromotive forces. According to the cardiac electromotive force dipole model, the composite wave of the epicardial side myocardial action potential and the endocardial side myocardial action potential is roughly consistent with the time change of the potential on the body surface observed on the body surface. The graph representing the time change of the potential on the body surface is the electrocardiogram waveform. The epicardial side myocardial action potential is the result of directly measuring the change in electromotive force generated by the pulsation of the outer myocardial layer by inserting a catheter electrode. The endocardial side myocardial action potential is the result of directly measuring the change in electromotive force generated by the pulsation of the inner myocardial layer by inserting a catheter electrode. The above is a schematic description of the cardiac electromotive force dipole model.

[0021] However, the outer myocardial layer in the cardiac electromotive force dipole model is a collection of cells. Therefore, the timing of the pulsation of the cells in the outer myocardial layer during one pulsation of the outer myocardial layer is not necessarily the same in all cells, and there may be a distribution in the pulsation timing. The same is true for the inner myocardial layer. That is, the timing of the pulsation of the cells in the inner myocardial layer during one pulsation of the inner myocardial layer is not necessarily the same in all cells, and there may be a distribution in the pulsation timing. However, the possibility of such a distribution in the pulsation timing of cells is not considered in the cardiac electromotive force dipole model.

[0022] In addition, there is also a distribution in the distance between each cell and the electrode located on the body surface, and the structure of the body tissue between each cell and the electrode located on the body surface is also different. Therefore, the conversion efficiency of the excitation of the cells in the outer layer of the myocardium reflected in the waveform of the electrocardiogram may not be the same in all cells, and there may be a distribution in the conversion efficiency of the pulsation reflected in the waveform of the electrocardiogram. Similarly, the conversion efficiency of the excitation of the cells in the inner layer of the myocardium reflected in the waveform of the electrocardiogram may not be the same in all cells, and there may be a distribution in the conversion efficiency of the pulsation reflected in the waveform of the electrocardiogram. However, the cardiac electromotive force dipole model does not assume the possibility of a distribution in the conversion efficiency of the pulsation of the cells reflected in the waveform of the electrocardiogram.

[0023] In the signal analysis device 1, taking into account the possibility of distribution in the timing of cell pulsation and the possibility of distribution in the conversion efficiency reflected by the cell pulsation in the waveform of the electrocardiogram, an analysis is performed assuming that the distribution of the timing of the start of the pulsation of each cell in the outer layer of the myocardium as expressed in the waveform of the electrocardiogram, the distribution of the timing of the start of the pulsation of each cell in the inner layer of the myocardium as expressed in the waveform of the electrocardiogram, the distribution of the timing of the end of the pulsation of each cell in the outer layer of the myocardium as expressed in the waveform of the electrocardiogram, and the distribution of the timing of the end of the pulsation of each cell in the inner layer of the myocardium as expressed in the waveform of the electrocardiogram are each Gaussian distributions. That is, in the signal analysis device 1, an analysis is performed assuming the following contents: the start of the activity of the inner myocardium generated by all cells in the inner myocardium is included in the waveform of the electrocardiogram as a cumulative Gaussian distribution, the start of the activity of the outer myocardium generated by all cells in the outer myocardium is included in the waveform of the electrocardiogram as a cumulative Gaussian distribution, the end of the activity of the inner myocardium generated by all cells in the inner myocardium is included in the waveform of the electrocardiogram as a cumulative Gaussian distribution, and the end of the activity of the outer myocardium generated by all cells in the inner myocardium is included in the waveform of the electrocardiogram as a cumulative Gaussian distribution.

[0024] In addition, in the signal analysis device 1, a cumulative Gaussian distribution function can be used, but a sigmoid function, a Gompertz function, a logical function, etc. can also be used instead of the cumulative Gaussian distribution function. That is, in the signal analysis device 1, instead of the cumulative Gaussian distribution function, a cumulative distribution function of a unimodal distribution can also be used, that is, a cumulative distribution function corresponding to a distribution in which the value monotonically increases before the moment when the value becomes the maximum value and the value monotonically decreases after the moment when the value becomes the maximum value. However, the cumulative distribution function used by the signal analysis device 1 needs to be a cumulative distribution function whose shape can be determined by a parameter representing the shape of the cumulative distribution function or a parameter representing the shape of a unimodal distribution as a cumulative source of the cumulative distribution function. Hereinafter, the parameter representing the shape of the cumulative distribution function (i.e., the parameter determining the cumulative distribution function) is referred to as the shape parameter of the cumulative distribution function, and the parameter representing the shape of the unimodal distribution (i.e., the parameter determining the unimodal distribution) is referred to as the shape parameter of the unimodal distribution. Of course, the shape parameter of the cumulative distribution function and the shape parameter of the unimodal distribution are substantially the same. For example, when the cumulative distribution function used by the signal analysis device 1 is a cumulative Gaussian distribution function, the standard deviation (or variance) and mean value of the Gaussian distribution as the cumulative source of the cumulative Gaussian distribution function are shape parameters of the unimodal distribution and are also shape parameters of the cumulative distribution function.

[0025] The signal analysis device 1 takes the waveform of an arbitrary time interval of the R wave and the T wave included in the waveform of one cycle of the electrocardiogram of the target heart obtained as the target time waveform, and obtains the parameters that indicate the characteristics of the target time waveform, i.e., the myocardial activity parameters, by determining the parameters of the first unimodal distribution or the first cumulative distribution function and determining the parameters of the second unimodal distribution or the second cumulative distribution function when approximating the target time waveform by the time waveform obtained by the difference or weighted difference between the first cumulative distribution function as the first unimodal distribution cumulative distribution function and the second cumulative distribution function as the second unimodal distribution cumulative distribution function (hereinafter referred to as the "approximate time waveform"). Hereinafter, approximating the target time waveform by the approximate time waveform, i.e., determining the approximate time waveform, is referred to as "fitting", and the first cumulative distribution function and the second cumulative distribution function included in the approximate time waveform are referred to as "fitting results". In addition, when the signal analysis device 1 uses the weighted difference for approximation, the weight assigned to the first cumulative distribution function and the weight assigned to the second cumulative distribution function can be obtained as parameters representing the characteristics of the object time waveform (i.e., myocardial activity parameters), and the ratio of the weight assigned to the first cumulative distribution function to the weight assigned to the second cumulative distribution function can be obtained as parameters representing the characteristics of the object time waveform (i.e., myocardial activity parameters).

[0026] When the object time waveform is approximated by the approximate time waveform obtained by the difference between the first cumulative distribution function and the second cumulative distribution function, for example, the signal analysis device 1 generates a time waveform obtained by the difference between the first cumulative distribution function and the second cumulative distribution function (hereinafter referred to as "candidate time waveform") using a combination (M×N types) of parameters for determining the cumulative distribution function for each of the plurality (M) candidates of the first cumulative distribution function and a combination (M×N types) of parameters for determining the cumulative distribution function for each of the plurality (N) candidates of the second cumulative distribution function, respectively, determines the candidate time waveform closest to the object time waveform among the generated M×N types of candidate time waveforms as the approximate time waveform, and obtains the parameters for determining the candidate of the first cumulative distribution function and the parameters for determining the candidate of the second cumulative distribution function used in the generation of the determined approximate time waveform as parameters representing the characteristics of the object time waveform. The process of determining the candidate time waveform closest to the object time waveform as the approximate time waveform can be performed, for example, by determining the candidate time waveform with the smallest square error between the candidate time waveform and the object time waveform.

[0027] Alternatively, for example, the signal analysis device 1 determines the candidate time waveform finally obtained as the approximate time waveform by obtaining a time waveform obtained by the difference between a first candidate for the approximate object time waveform and a second candidate for the cumulative distribution function, i.e., the candidate time waveform, and updating at least any one of the parameters for determining each cumulative distribution function in a direction in which the square error between the candidate time waveform and the object time waveform becomes smaller, repeatedly until the square error becomes below a prescribed reference or repeats a prescribed number of times, and obtains the parameters for determining the candidate for the first cumulative distribution function and the parameters for determining the candidate for the second cumulative distribution function used in generating the determined approximate time waveform as parameters representing the characteristics of the object time waveform.

[0028] In the case of approximating the object time waveform using an approximate time waveform obtained by the weighted difference between the first cumulative distribution function and the second cumulative distribution function, for example, the signal analysis device 1 uses a combination (K×L×M×N types) of cumulative distribution function parameters for each of the multiple (M) candidates for the first cumulative distribution function, parameters for each of the multiple (N) candidates for the second cumulative distribution function, multiple (K) candidates for weights assigned to the first cumulative distribution function, and multiple (K) candidates for weights assigned to the second cumulative distribution function, respectively, to generate a time waveform obtained by the weighted difference between the first cumulative distribution function and the second cumulative distribution function, i.e., a candidate time waveform that is closest to the object time waveform among the generated K×L×M×N candidate time waveforms as the approximate time waveform, and obtains the parameters for determining the candidates for the first cumulative distribution function, the parameters for determining the candidates for the second cumulative distribution function, the weights assigned to the first cumulative distribution function, and the weights assigned to the second cumulative distribution function used in the generation of the determined approximate time waveform as parameters representing the characteristics of the object time waveform.

[0029] Alternatively, for example, the signal analysis device 1 obtains a time waveform obtained by obtaining a weighted difference between a candidate for a first cumulative distribution function that approximates an object time waveform and a candidate for a second cumulative distribution function, i.e., a candidate time waveform, and updates at least any one of the parameters for determining each cumulative distribution function and the weight assigned to each cumulative distribution function in a direction in which the square error between the candidate time waveform and the object time waveform becomes smaller, repeatedly until the square error becomes below a prescribed benchmark or a prescribed number of times, thereby determining the candidate time waveform finally obtained as an approximate time waveform, and obtaining the parameters for determining the candidate for the first cumulative distribution function, the parameters for determining the candidate for the second cumulative distribution function, the weight assigned to the first cumulative distribution function, and the weight assigned to the second cumulative distribution function used in generating the determined approximate time waveform as parameters representing the characteristics of the object time waveform.

[0030] Hereinafter, the process of acquiring parameters representing the characteristics of the target time waveform included in the waveform corresponding to one cycle of the acquired electrocardiogram of the target heart is referred to as myocardial activity information parameter acquisition process.

[0031] When the information indicating the time is x, in the case where Gaussian distribution is used as the first unimodal distribution and the second unimodal distribution, the first unimodal distribution is expressed by the following equation (1): the first cumulative distribution function f 1 (x) is expressed by formula (2), the second unimodal distribution is expressed by formula (3), and the second cumulative distribution function f 2 (x) is expressed by formula (4).

[0032] [Formula 1]

[0033] [Formula 2]

[0034] [Formula 3]

[0035] [Formula 4]

[0036] Formula (1) is the average value μ 1 , standard deviation is σ 1 (The variance is σ 1 2 ) is a Gaussian distribution (normal distribution) with a mean value of μ 2 , standard deviation is σ 2 (The variance is σ 2 2 ) is a Gaussian distribution (normal distribution). Formula (2) is the cumulative distribution function of formula (1). Formula (4) is the cumulative distribution function of formula (3). "erf" is a sigmoid function (error function). The unit of the information x representing the time is arbitrary. For example, the sample number or relative time starting from the waveform of one cycle of the electrocardiogram can be used as the information x representing the time.

[0037] The difference between the first cumulative distribution function and the second cumulative distribution function is expressed by, for example, the following equation (5). The function represented by the following equation (5) is a function obtained by subtracting the second cumulative distribution function from the first cumulative distribution function.

[0038] [Formula 5]

[0039] That is, when the target time waveform is approximated by the approximate time waveform which is the difference between the first cumulative distribution function and the second cumulative distribution function, the parameter of the first unimodal distribution or the parameter of the first cumulative distribution function, i.e., the average value μ, is determined. 1 and standard deviation σ 1 , and determining the parameter of the second unimodal distribution or the second cumulative distribution function, namely the mean value μ 2 and standard deviation σ 2 The parameter representing the characteristic of the target time waveform is obtained. In addition, the variance may be obtained as a parameter instead of the standard deviation. This also applies to the subsequent description where the standard deviation is obtained as a parameter.

[0040] Regarding the weighted difference between the first cumulative distribution function and the second cumulative distribution function, the weight of the first cumulative distribution function is set to k 1 , set the weight of the second cumulative distribution function to k2 , for example, is represented by the following equation (6). The function represented by the following equation (6) is obtained by multiplying the first cumulative distribution function by the weight k 1 The second cumulative distribution function is subtracted from the function multiplied by the weight k 2 The function obtained by the function of .

[0041] [Formula 6]

[0042] That is, when the target time waveform is approximated by the approximate time waveform which is the weighted difference between the first cumulative distribution function and the second cumulative distribution function, at least the parameter of the first unimodal distribution or the parameter of the first cumulative distribution function, i.e., the average value μ 1 and standard deviation σ 1 , and determining the parameters of the second unimodal distribution or determining the parameters of the second cumulative distribution function, namely the mean value μ 2 and standard deviation σ 2 The parameters representing the characteristics of the target time waveform can also be obtained. 1 and the weight k of the second cumulative distribution function 2 , or the weight k of the first cumulative distribution function 1 and the weight k of the second cumulative distribution function 2 The ratio (k 1 / k 2 or k 2 / k 1 ) obtains parameters representing the characteristics of the object time waveform.

[0043] When the function of equation (6) is the first cumulative distribution function multiplied by the weight k 1 minus the second cumulative distribution function multiplied by the weight k 2 When the function is obtained by the function of 1 and weight k 2 However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. 1 and weight k 2 Therefore, although the signal analysis device 1 can set the weight k 1 and k 2 are all fitted as positive values, but there is no need to set the weight k 1 and k 2 All fit to positive values.

[0044] In addition, the signal analysis device 1 can use the R wave and the T wave included in the waveform of one cycle of the electrocardiogram of the object heart as the first object time waveform, and the T wave as the second object time waveform, and obtain the above-mentioned parameters representing the characteristics of the object time waveform for the first object time waveform and the second object time waveform, respectively.

[0045] For example, when the first target time waveform (i.e., R wave) is approximated by the difference between the first cumulative distribution function and the second cumulative distribution function, the first target time waveform is approximated by the approximate time waveform of equation (9), which is obtained from the first cumulative distribution function f shown in equation (7). a (x) minus the second cumulative distribution function f shown in equation (8) b (x) will determine the parameter of the first cumulative distribution function, i.e., the mean value μ a and standard deviation σ a , and determine the parameter of the second cumulative distribution function, namely the mean value μ b and standard deviation σ b The parameters representing the characteristics of the first target time waveform (ie, the R wave) are acquired.

[0046] [Formula 7]

[0047] [Formula 8]

[0048] [Formula 9]

[0049] For example, when the first target time waveform (i.e., R wave) is approximated by the weighted difference between the first cumulative distribution function and the second cumulative distribution function, the first target time waveform is approximated by the approximation time waveform of equation (10), which is obtained from the first cumulative distribution function f shown in equation (7). a (x) multiplied by weight k a The function minus the second cumulative distribution function f shown in formula (8) b (x) multiplied by weight k b The function obtained by the function of will at least determine the parameters of the first cumulative distribution function, namely the mean value μ a and standard deviation σ a And determine the parameter of the second cumulative distribution function, namely the mean value μ b and standard deviation σ b The parameters representing the characteristics of the first target time waveform (ie, R wave) are obtained. In addition, the weight k of the first cumulative distribution function may be a and the weight k of the second cumulative distribution function b, or the weight k of the first cumulative distribution function a and the weight k of the second cumulative distribution function b The ratio (k a / k b or k b / k a ) also obtains a parameter that represents the characteristics of the parameter representing the characteristics of the first object time waveform (ie, the R wave).

[0050] [Formula 10]

[0051] The function according to formula (10) is the first cumulative distribution function multiplied by the weight k a minus the second cumulative distribution function multiplied by the weight k b The function obtained by the function of weight k a and weight k b However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. a and weight k b Therefore, although the signal analysis device 1 can set the weight k a and weight k b are all fitted as positive values, but there is no need to set the weight k a and weight k b All fit to positive values.

[0052] Since the R wave corresponds to the excitation of all myocardial cells starting in sequence according to the Gaussian distribution, for the R wave, as described above, it is sufficient to approximate the object time waveform in the positive direction of time by the approximate time waveform of the difference or weighted difference of two cumulative Gaussian distributions. On the other hand, according to the T wave, the excitation of all myocardial cells awakens in sequence according to the Gaussian distribution, which can be explained as a phenomenon that the T wave is in the opposite direction of the R wave on the time axis. That is, for the T wave, it is sufficient to approximate the waveform after the time axis of the object time waveform is reversed by the difference or weighted difference of the cumulative Gaussian distribution. It will be referred to as the first method below. In addition, according to the T wave, all myocardial cells awaken from the excited state according to the Gaussian distribution, for the T wave, it can also be said that the object time waveform in the positive direction of time can be approximated by the difference or weighted difference of two functions (the function of the cumulative Gaussian distribution subtracted from 1). It will be referred to as the second method below. The following describes specific examples of the first method and the second method. In order to avoid confusion between the above-mentioned cumulative distribution function for the R wave and the cumulative distribution function for the T wave described below, the above-mentioned first cumulative distribution function will be referred to as the third cumulative distribution function, and the above-mentioned second cumulative distribution function will be referred to as the fourth cumulative distribution function for the T wave.

[0053] When the first method is used to approximate the second target time waveform (i.e., T wave) by the difference between the third cumulative distribution function and the fourth cumulative distribution function, the information indicating the time in the reverse direction is set to x', and the waveform obtained by inverting the time axis of the second target time waveform is called the second target inverse time waveform. The second target inverse time waveform is approximated by the approximate inverse time waveform of formula (13), which is obtained by converting the third cumulative distribution function f shown in formula (11) into the second target inverse time waveform. e (x') minus the fourth cumulative distribution function f shown in formula (12) g The function obtained by (x') will determine the parameter of the third cumulative distribution function, namely the mean value μ e and standard deviation σ e , and determine the parameter of the fourth cumulative distribution function, namely the mean value μ g and standard deviation σ g The parameters representing the characteristics of the second target time waveform (ie, T wave) are acquired.

[0054] [Formula 11]

[0055] [Formula 12]

[0056] [Formula 13]

[0057] For example, when the second target time waveform (i.e., T wave) is approximated by the weighted difference of the third cumulative distribution function and the fourth cumulative distribution function using the first method, the second target inverse time waveform is approximated by the approximate inverse time waveform of equation (14), which is obtained from the third cumulative distribution function f shown in equation (11): e (x') multiplied by the weight k e The function of minus the fourth cumulative distribution function f shown in formula (12) g (x') multiplied by the weight k g The function obtained by the function of will at least determine the parameters of the third cumulative distribution function, namely the mean value μ e and standard deviation σ e , and determine the parameter of the fourth cumulative distribution function, namely the mean value μ g and standard deviation σ g The parameter representing the characteristic of the second target time waveform (ie, T wave) is obtained. In addition, the weight k of the third cumulative distribution function may be e and the weight k of the fourth cumulative distribution function g , or the weight k of the third cumulative distribution function e With the fourth cumulative distribution function weight k g The ratio (ke / k g or k g / k e ) is also obtained as a parameter representing the characteristics of the second object time waveform (ie, T wave).

[0058] [Formula 14]

[0059] According to formula (14), the function is obtained by multiplying the third cumulative distribution function by the weight k e minus the fourth cumulative distribution function multiplied by the weight k g The function obtained by the function, weight k e and weight k g However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. e and weight k g Therefore, although the signal analysis device 1 can set the weight k e and weight k g are all fitted as positive values, but there is no need to set the weight k e and weight k g All fit to positive values.

[0060] For example, when using the second method by subtracting the third cumulative distribution function f shown in equation (15) from 1 c (x) and the function f' c (x) (hereinafter referred to as the "third inverse cumulative distribution function") is the sum of the fourth cumulative distribution function f shown in equation (16) subtracted from 1 d (x) and the function f' d When the second target time waveform (i.e., T wave) is approximated by the difference between the third inverse cumulative distribution function f' and the second target time waveform (i.e., T wave), the second target time waveform is approximated by the approximation time waveform of formula (17), which is obtained from the third inverse cumulative distribution function f' c (x) minus the fourth inverse cumulative distribution function f' d (x) will determine the parameter of the third cumulative distribution function, namely the mean value μ c and standard deviation σ c , and determine the parameter of the fourth cumulative distribution function, namely the mean value μ d and standard deviation σ d The parameters representing the characteristics of the second target time waveform (ie, T wave) are acquired.

[0061] [Formula 15]

[0062] [Formula 16]

[0063] [Formula 17]

[0064] For example, when the second method is used to approximate the second target time waveform (i.e., T wave) by the weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, the second target time waveform is approximated by the approximate time waveform of formula (18), which is obtained from the third inverse cumulative distribution function f' c (x) multiplied by weight k c minus the fourth inverse cumulative distribution function f' d (x) multiplied by weight k d The function obtained by the function of will determine the parameter of the third cumulative distribution function, namely the mean value μ c and standard deviation σ c , and determine the parameter of the fourth cumulative distribution function, namely the mean value μ d and standard deviation σ d The parameter representing the characteristic of the second target time waveform (ie, T wave) is obtained. In addition, the weight k of the third inverse cumulative distribution function may be c and the weight k of the fourth inverse cumulative distribution function d , or the weight k of the third inverse cumulative distribution function c with the fourth inverse cumulative distribution function with weight k d The ratio (k c / k d or k d / k c ) is obtained as a parameter representing the characteristics of the second object time waveform (ie, T wave).

[0065] [Formula 18]

[0066] According to formula (18), the function is the third inverse cumulative distribution function multiplied by the weight k c minus the fourth inverse cumulative distribution function multiplied by the weight k d The function obtained by the function of weight k c and weight k d However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. c and weight k d Therefore, although the signal analysis device 1 can set the weight k c and weight k d are all fitted as positive values, but there is no need to set the weight k cand weight k d All fit to positive values.

[0067] In addition, since the approximate time waveform of equation (17) is derived from the fourth cumulative distribution function f d (x) minus the third cumulative distribution function f c (x), is the third cumulative distribution function f c (x) and the fourth cumulative distribution function f d In addition, the approximate time waveform of equation (18) is the third cumulative distribution function f c (x) and the fourth cumulative distribution function f d The shape of the curve obtained by adding a constant term to the weighted difference of (x) is similar to the third cumulative distribution function f c (x) and the fourth cumulative distribution function f d In addition, as mentioned above, it can be explained that the T wave is in the opposite direction of the R wave on the time axis. For the T wave, the third cumulative distribution function f e (x) and the fourth cumulative distribution function f g Therefore, in the following description of the T wave, sometimes only the cumulative distribution function is used instead of describing the cumulative distribution function and the inverse cumulative distribution function at the same time.

[0068] Figure 2 is a schematic diagram showing a first cumulative distribution function f with respect to a first target time waveform (ie, R wave). a (x) multiplied by weight k a The function k a f a (x), the second cumulative distribution function f b (x) multiplied by weight k b The function k b f b (x), and an approximate time waveform k which is a weighted difference between the first cumulative distribution function and the second cumulative distribution function a f a (x)-k b f b (x). The single-point dashed line is the first cumulative distribution function f a (x) multiplied by weight k a The function k a f a (x), the double-dotted line is the second cumulative distribution function f b (x) multiplied by weight k b The function k b f b (x), the dotted line is the approximate time waveform ka f a (x)-k b f b (x). The approximate time waveform k a f a (x)-k b f b (x) is a waveform that approximates the first target time waveform (ie, R wave).

[0069] Figure 3 is a schematic diagram showing a third inverse cumulative distribution function f′ with respect to a second object time waveform (ie, T wave). c (x) = 1-f c (x) multiplied by weight k c The function k c f' c (x), the fourth inverse cumulative distribution function f' d (x) = 1-f d (x) multiplied by weight k d The function k d f' d (x), and an approximate time waveform k which is a weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function c f' c (x)-k d f' d (x) is a schematic diagram. The single-point dashed line is the third inverse cumulative distribution function f' c (x) = 1-f c (x) multiplied by weight k c The function k c f' c (x), the double-dotted line is the fourth inverse cumulative distribution function f' d (x) = 1-f d (x) multiplied by weight k d The function k d f' d (x), the dotted line is the approximate time waveform k c f' c (x)-k d f' d (x). The approximate time waveform k c f' c (x)-k d f' d (x) is a waveform that approximates the second target time waveform (ie, T wave).

[0070] Figure 4This is a diagram schematically showing the result of fitting each of the first object time waveform and the second object time waveform by the difference between the two cumulative distribution functions, using the R wave included in the waveform of one cycle of the electrocardiogram of the object heart in the first embodiment as the first object time waveform and the T wave as the second object time waveform. Figure 4 The horizontal axis represents time, and the vertical axis represents potential. The units of the horizontal and vertical axes are both arbitrary units.

[0071] Specifically, Figure 4 This is an example of fitting the first target time waveform (i.e., R wave) by the difference between the first cumulative distribution function and the second cumulative distribution function, and fitting the second target time waveform (i.e., T wave) by the difference between the third cumulative distribution function and the fourth cumulative distribution function. The domain of the first cumulative distribution function is the same as the domain of the second cumulative distribution function, which is from time T1 to time T3, which is the time interval of the first target time waveform (i.e., R wave). The domain of the third cumulative distribution function is the same as the domain of the fourth cumulative distribution function, which is from time T4 to time T6.

[0072] Figure 4 The “first fitting result” and “second fitting result” are the fitting results for the R wave. Figure 4 The “third fitting result” and “fourth fitting result” in FIG. 1 are the fitting results for the T wave.

[0073] exist Figure 4 In , the “first fitting result” shows the first cumulative distribution function in the fitting result of the first object time waveform (ie, R wave) of the electrocardiogram. Figure 4 In , the "second fitting result" shows the second cumulative distribution function in the fitting result of the first object time waveform (ie, R wave) of the electrocardiogram. Figure 4 In FIG. 1 , “third fitting result” shows the third cumulative distribution function in the fitting result of the second object time waveform (ie, T wave) of the electrocardiogram. Figure 4 In FIG. 1 , “the fourth fitting result” shows the fourth cumulative distribution function in the fitting result of the second object time waveform (ie, T wave) of the electrocardiogram. Figure 4 In the figure, “body surface potential” represents the waveform of the electrocardiogram of the object of fitting.

[0074] In addition, the signal analysis device 1 does not perform fitting for the period from time T3 to time T4, which is neither the time period of the first target time waveform (i.e., R wave) nor the time period of the second target time waveform (i.e., T wave). Figure 4In the example, the first fitting result at time T3 and the third fitting result at time T4 are connected, and the second fitting result at time T3 and the fourth fitting result at time T4 are connected. That is, when the signal analysis device 1 displays the fitting results, as shown in FIG. Figure 4 As shown, the signal analysis device 1 only needs to display a line connecting the first fitting result at time T3 and the third fitting result at time T4, and the second fitting result at time T3 and the fourth fitting result at time T4 using a predetermined function such as a constant function or a linear function.

[0075] In addition, when the signal analysis device 1 displays the fitting results, the weight value may be corrected so that the first fitting result at time T3 and the third fitting result at time T4 can be displayed with the same value. That is, the actual first fitting result at time T3 is k a f a (T3), the actual third fitting result at time T4 is k c f' c (T4), but we can also find the solution that satisfies k a f a (T3) = α 1 k c f' c α of (T4) 1 , using α 1 k c Instead of weight k c To display the fitting results, you can also use k a / α 1 Instead of weight k a Similarly, when the signal analysis device 1 displays the fitting results, the weight value can also be corrected so that the second fitting result at time T3 and the fourth fitting result at time T4 can be displayed with the same value. That is, the second fitting result at time T3 is k b f b (T3), the fourth fitting result at time T4 is k d f' d (T4), but we can also find the solution that satisfies k b f b (T3) = α 2 k d f' d α of (T4) 2 , using α 2 k d Instead of weight k d To display the fitting results, you can also use k b / α 2 Instead of weight k b to display the fitting results.

[0076] In addition, the fitting of the first object time waveform and the fitting of the second object time waveform may not be performed separately. That is, the fitting of the first object time waveform and the second object time waveform may be performed together as the fitting of the two. For example, when the signal analysis device 1 performs the fitting of the first object time waveform and the second object time waveform together, the fitting may be performed in consideration of the fact that the difference between the first fitting result at time T3 and the third fitting result at time T4 becomes smaller, and the difference between the second fitting result at time T3 and the fourth fitting result at time T4 becomes smaller.

[0077] Figure 5 This is an explanatory diagram for explaining that typical features of the T wave can be visualized by approximating the T wave with a function obtained by subtracting the fourth inverse cumulative distribution function from the third inverse cumulative distribution function and displaying the third and fourth inverse cumulative distribution functions or displaying parameters that determine each cumulative distribution function. Figure 5 Four images, namely, image G1, image G2, image G3, and image G4, are shown. Each of images G1 to G4 is a graph in which the horizontal axis represents time and the vertical axis represents potential. Figure 5 The units of the horizontal axis and the vertical axis of each of the images G1 to G4 are arbitrary units.

[0078] Figure 5 The "first function" of is an example of the third inverse cumulative distribution function. Figure 5 The “second function” of is an example of the fourth inverse cumulative distribution function. Figure 5 The “third function” of represents a function obtained by subtracting the “second function” from the “first function”, that is, a function obtained by subtracting the fourth inverse cumulative distribution function from the third inverse cumulative distribution function. Figure 5 The shape of the "third function" is roughly the same as the shape of the T wave of a normal heart.

[0079] Figure 5 The "fourth function" of is an example of the third inverse cumulative distribution function. Figure 5 The "fifth function" is an example of the fourth inverse cumulative distribution function. Figure 5 The “sixth function” of represents a function obtained by subtracting the “fifth function” from the “fourth function”, that is, a function obtained by subtracting the fourth inverse cumulative distribution function from the third inverse cumulative distribution function. Figure 5 The shape of the "sixth function" is roughly the same as the shape of a T wave depression, which is one of the three typical patterns of T wave abnormality. Figure 5The interval between the declining part of the third inverse cumulative distribution function and the declining part of the fourth inverse cumulative distribution function in image G2 is narrower than the interval between the declining part of the third inverse cumulative distribution function and the declining part of the fourth inverse cumulative distribution function in image G1 of a normal heart, thereby visualizing that during the T wave decline, the activity of the outer layer of the myocardium is less delayed from the activity of the inner layer of the myocardium.

[0080] Figure 5 The "seventh function" is an example of the third inverse cumulative distribution function. Figure 5 The "eighth function" of is an example of the fourth inverse cumulative distribution function. Figure 5 The “ninth function” of represents a function obtained by subtracting the “eighth function” from the “seventh function”, that is, a function obtained by subtracting the fourth inverse cumulative distribution function from the third inverse cumulative distribution function. Figure 5 The shape of the "ninth function" is roughly the same as the shape of an elevation of the T wave, which is one of the three typical patterns of T wave abnormality. Figure 5 The interval between the declining part of the third inverse cumulative distribution function and the declining part of the fourth inverse cumulative distribution function in image G3 of the normal heart is wider than the interval between the declining part of the third inverse cumulative distribution function and the declining part of the fourth inverse cumulative distribution function in image G1 of the normal heart, thereby visualizing that in the increase of the T wave, the activity of the outer layer of the myocardium is delayed significantly from the activity of the inner layer of the myocardium.

[0081] Figure 5 The "tenth function" of is an example of the third inverse cumulative distribution function. Figure 5 The "eleventh function" is an example of the fourth inverse cumulative distribution function. Figure 5 The “twelfth function” of represents a function obtained by subtracting the “eleventh function” from the “tenth function”, that is, a function obtained by subtracting the fourth inverse cumulative distribution function from the third inverse cumulative distribution function. Figure 5 The shape of the "twelfth function" is roughly the same as the shape of a negative T wave, one of the three typical patterns of T wave abnormalities. Figure 5 The order of the declining part of the third inverse cumulative distribution function and the declining part of the fourth inverse cumulative distribution function in image G4 is opposite to the order of the declining part of the third inverse cumulative distribution function and the declining part of the fourth inverse cumulative distribution function in image G1 of the normal heart, thereby visualizing that in the negative T wave, the outer layer of the myocardium ends the activity earlier than the inner layer of the myocardium.

[0082] The function obtained by subtracting the fourth inverse cumulative distribution function from the third inverse cumulative distribution function can express waves whose widths in the vertical axis direction are different from those in the horizontal axis direction, such as the "third function", "sixth function" and "ninth function". In addition, the function obtained by subtracting the fourth inverse cumulative distribution function from the third inverse cumulative distribution function can express negative waves, such as the "twelfth function". That is, by approximating the T wave with the function obtained by subtracting the fourth inverse cumulative distribution function from the third inverse cumulative distribution function, or by reversing the time axis and approximating the T wave with the function obtained by subtracting the fourth cumulative distribution function from the third cumulative distribution function, it is possible to express the respective activities of the inner and outer myocardial layers included in the T wave and the relationship between the activities of the inner and outer myocardial layers. This is also the same when the R wave is approximated by the function obtained by subtracting the second cumulative distribution function from the first cumulative distribution function.

[0083] Thus, the signal analysis device 1 fits the waveform of the R wave or T wave of the electrocardiogram of the target heart by the difference or weighted difference of the two cumulative distribution functions. Then, the signal analysis device 1 obtains the parameters that determine the approximate time waveform determined by fitting as parameters showing the activity of the myocardium.

[0084] [Approximation by the difference or weighted difference of two cumulative distribution functions plus the value] The signal analysis device 1 may also take the waveform of the time interval of either the R wave or the T wave included in the waveform of one cycle of the electrocardiogram of the target heart obtained as the target time waveform, and take the time waveform obtained by adding a value (hereinafter referred to as a "level value") to the difference or weighted difference between the first cumulative distribution function as the first unimodal cumulative distribution function and the second cumulative distribution function as the second unimodal cumulative distribution function as the approximate time waveform. In this case, in addition to the parameters that determine the first unimodal distribution or the parameters that determine the first cumulative distribution function and the parameters that determine the second unimodal distribution or the parameters that determine the second cumulative distribution function when the target time waveform is approximated by the approximate time waveform, the level value is obtained as a parameter representing the characteristics of the target time waveform. Of course, in the case of approximation by weighted difference, the weight given to the first cumulative distribution function and the weight given to the second cumulative distribution function may also be obtained as parameters representing the characteristics of the target time waveform, and the ratio of the weight given to the first cumulative distribution function and the weight given to the second cumulative distribution function may also be obtained as a parameter representing the characteristics of the target time waveform.

[0085] In the case of using a weighted difference, for example, the signal analysis device 1 uses a combination (J×K×L×M×N types) of the cumulative distribution function parameters for each of the multiple (M) candidates for the first cumulative distribution function, the cumulative distribution function parameters for each of the multiple (N) candidates for the second cumulative distribution function, the multiple (K) candidates for weights assigned to the first cumulative distribution function, the multiple (L) candidates for weights assigned to the second cumulative distribution function, and the multiple (J) candidates for level values, to generate a time waveform obtained by adding the weighted difference between the first cumulative distribution function and the second cumulative distribution function to the level value, i.e., a candidate time waveform, and determines the candidate time waveform that is closest to the object time waveform among the generated J×K×L×M×N types of candidate time waveforms as the approximate time waveform, and obtains the parameters for determining the candidate for the first cumulative distribution function, the parameters for determining the candidate for the second cumulative distribution function, the weight assigned to the first cumulative distribution function, the weight assigned to the second cumulative distribution function, and the level value used in the generation of the determined approximate time waveform as parameters representing the characteristics of the object time waveform.

[0086] Alternatively, for example, the signal analysis device 1 obtains a time waveform obtained by adding a weighted difference between a candidate of a first cumulative distribution function that approximates the object time waveform and a candidate of a second cumulative distribution function plus a level value, i.e., a candidate time waveform, and updates the parameters of each cumulative distribution function in a direction in which the square error between the candidate time waveform and the object time waveform becomes smaller, and repeats at least one of the weights and level values ​​assigned to each cumulative distribution function until the square error becomes below a prescribed benchmark or is repeated a prescribed number of times, thereby determining the final candidate time waveform as an approximate time waveform, and obtaining the parameters for determining the candidate of the first cumulative distribution function, the parameters for determining the candidate of the second cumulative distribution function, the weight assigned to the first cumulative distribution function, the weight assigned to the second cumulative distribution function, and the level value used in the generation of the determined approximate time waveform as parameters representing the characteristics of the object time waveform.

[0087] Alternatively, the level value may be determined before fitting. In this case, the signal analysis device 1 firstly calculates the starting end (equivalent to the R wave) of the target time waveform when the target time waveform is an R wave. Figure 4 The potential at the time T1 in the example is taken as the level value, and when the target time waveform is a T wave, the end end of the target time waveform (equivalent to Figure 4The potential at time T6 in the signal analysis device 1 is obtained as the level value. Furthermore, the signal analysis device 1 uses a combination (K×L×M×N types) of the cumulative distribution function parameters for each of the plurality of (M) candidates for the first cumulative distribution function, the cumulative distribution function parameters for each of the plurality of (N) candidates for the second cumulative distribution function, the plurality of (K) candidates for the weight given to the first cumulative distribution function, and the plurality of (L) candidates for the weight given to the second cumulative distribution function, respectively, to generate a candidate time waveform obtained by adding the level value to the weighted difference between the first cumulative distribution function and the second cumulative distribution function, that is, a candidate time waveform, which is the closest to the target time waveform among the generated K×L×M×N candidate time waveforms, as the approximate time waveform, and obtains the parameters for determining the candidate for the first cumulative distribution function, the parameters for determining the candidate for the second cumulative distribution function, the weight given to the first cumulative distribution function, the weight given to the second cumulative distribution function, and the level value determined in the initial processing as parameters representing the characteristics of the target time waveform.

[0088] Alternatively, for example, when the target time waveform is an R wave, the signal analysis device 1 firstly changes the starting end (equivalent to Figure 4 The potential at the time T1 in the example is taken as the level value, and when the target time waveform is a T wave, the end end of the target time waveform (equivalent to Figure 4 The potential at the time T6 in the signal analysis device 1 is obtained as the level value. Then, the signal analysis device 1 obtains the candidate time waveform obtained by adding the level value to the weighted difference between the candidate of the first cumulative distribution function and the candidate of the second cumulative distribution function that approximates the object time waveform, that is, the candidate time waveform, and updates at least one of the parameters for determining each cumulative distribution function and the weight assigned to each cumulative distribution function in the direction in which the square error between the candidate time waveform and the object time waveform becomes smaller, and repeats until the square error becomes less than a predetermined reference or repeats a predetermined number of times, thereby determining the candidate time waveform finally obtained as the approximate time waveform, and obtaining the parameters for determining the candidate of the first cumulative distribution function, the parameters for determining the candidate of the second cumulative distribution function, the weight assigned to the first cumulative distribution function, the weight assigned to the second cumulative distribution function, and the level value determined in the initial processing as parameters representing the characteristics of the object time waveform.

[0089] When the level value is β, the value obtained by adding the weighted difference between the first cumulative distribution function and the second cumulative distribution function to the level value is expressed by, for example, the following equation (19). The function shown in the following equation (19) is the first cumulative distribution function multiplied by the weight k 1 minus the second cumulative distribution function multiplied by the weight k 2 The function obtained by adding the level value β to the function.

[0090] [Formula 19]

[0091] When the object time waveform is approximated by the approximate time waveform of equation (19) obtained by adding the weighted difference between the first cumulative distribution function and the second cumulative distribution function to the level value, at least the parameters of the first unimodal distribution or the parameters of the first cumulative distribution function, i.e., the average value μ, are determined. 1 and standard deviation σ 1 , determine the parameters of the second unimodal distribution or determine the parameters of the second cumulative distribution function, that is, the mean value μ 2 and standard deviation σ 2 , and the level value β are obtained as parameters representing the characteristics of the target time waveform. In addition, the weight k of the first cumulative distribution function can also be 1 and the weight k of the second cumulative distribution function 2 , or the weight k of the first cumulative distribution function 1 and the weight k of the second cumulative distribution function 2 The ratio (k 1 / k 2 or k 2 / k 1 ) is also obtained as a parameter representing the characteristics of the object time waveform.

[0092] According to formula (19), the function is the first cumulative distribution function multiplied by the weight k 1 minus the second cumulative distribution function multiplied by the weight k 2 The function obtained by adding the level value β, weight k 1 and weight k 2 However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. 1 and weight k 2 Therefore, although the signal analysis device 1 can set the weight k 1 and k 2 are all fitted as positive values, but there is no need to set the weight k 1 and k 2 All fit to positive values.

[0093] The signal analysis device 1 can use the R wave included in the waveform of one cycle of the electrocardiogram of the target heart as the first target time waveform, and the T wave as the second target time waveform, and obtain the above-mentioned parameters representing the characteristics of the target time waveform for the first target time waveform and the second target time waveform, respectively.

[0094] For example, when the first target time waveform (i.e., R wave) is approximated by a function obtained by adding a level value to the weighted difference between the first cumulative distribution function and the second cumulative distribution function, the potential at the starting end of the first target time waveform is taken as the level value β R , the first object time waveform is approximated by the approximate time waveform of equation (20), which is obtained from the first cumulative distribution function f shown in equation (7) a (x) multiplied by weight k a The function minus the second cumulative distribution function f shown in formula (8) b (x) multiplied by weight k b function and add the level value β R The obtained function will at least determine the parameter of the first cumulative distribution function, namely the mean value μ a and standard deviation σ a , determine the parameter of the second cumulative distribution function, i.e., the mean value μ b and standard deviation σ b , and the level value β R The parameters representing the characteristics of the first target time waveform (ie, R wave) are obtained. In addition, the weight k of the first cumulative distribution function may be a and the weight k of the second cumulative distribution function b , or the weight k of the first cumulative distribution function a and the weight k of the second cumulative distribution function b The ratio (k a / k b or k b / k a ) also obtains a parameter that represents the characteristics of the parameter representing the characteristics of the first object time waveform (ie, the R wave).

[0095] [Formula 20]

[0096] According to formula (20), the function is the first cumulative distribution function multiplied by the weight k a minus the second cumulative distribution function multiplied by the weight k b function and add the level value β R The obtained function, weight k a and weight k b However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. a and weight k b Therefore, although the signal analysis device 1 can set the weight k a and weight k b are all fitted as positive values, but there is no need to set the weight k aand weight k b All fit to positive values.

[0097] For example, when the second target inverse time waveform, which is a waveform obtained by inverting the time axis of the second target time waveform (i.e., T wave), is approximated by a function obtained by adding a level value to the weighted difference between the third cumulative distribution function and the fourth cumulative distribution function, the potential at the end of the second target time waveform is taken as the level value β T , the second object inverse time waveform is approximated by the approximate inverse time waveform of equation (21), which is obtained from the third cumulative distribution function f shown in equation (11) e (x') multiplied by the weight k e The function minus the fourth inverse cumulative distribution function f shown in formula (12) g (x') multiplied by the weight k g function and add the level value β T The obtained function will at least determine the parameter of the third cumulative distribution function, namely the mean value μ e and standard deviation σ e , determine the parameter of the fourth cumulative distribution function, that is, the mean value μ g and standard deviation σ g , and the level value β T The parameter representing the characteristic of the second target time waveform (ie, T wave) is obtained. In addition, the weight k of the third cumulative distribution function may be e and the weight k of the fourth cumulative distribution function g , or the weight k of the third cumulative distribution function e With the fourth cumulative distribution function weight k g The ratio (k e / k g or k g / k e ) also obtains a parameter representing the characteristics of a parameter representing the characteristics of a second object time waveform (ie, T wave).

[0098] [Formula 21]

[0099] According to formula (21), the function is obtained by multiplying the third cumulative distribution function by the weight k e minus the fourth cumulative distribution function multiplied by the weight k g function and add the level value β T The obtained function, weight k e and weight k g However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. e and weight k gTherefore, although the signal analysis device 1 can set the weight k e and weight k g are all fitted as positive values, but there is no need to set the weight k e and weight k g All fit to positive values.

[0100] For example, when the second target time waveform (i.e., T wave) is approximated by a function obtained by adding a level value to a weighted difference between a function obtained by subtracting the third cumulative distribution function from 1 (i.e., the third inverse cumulative distribution function) and a function obtained by subtracting the fourth cumulative distribution function from 1 (i.e., the fourth inverse cumulative distribution function), the potential at the end of the second target time waveform is taken as the level value β. T , the second object time waveform is approximated by the approximate time waveform of formula (22), which is obtained from the third inverse cumulative distribution function f' c (x) multiplied by weight k c minus the fourth inverse cumulative distribution function f' d (x) multiplied by weight k c function and add the level value β T The obtained function will determine the parameter of the third cumulative distribution function, namely the mean value μ c and standard deviation σ c , determine the parameter of the fourth cumulative distribution function, that is, the mean value μ d and standard deviation σ d , and the level value β T The parameter representing the characteristic of the second target time waveform (ie, T wave) is obtained. In addition, the weight k of the third inverse cumulative distribution function may be c and the weight k of the fourth inverse cumulative distribution function d , or the weight k of the third inverse cumulative distribution function c with the fourth inverse cumulative distribution function with weight k d The ratio (k c / k d or k d / k c ) is also obtained as a parameter representing the characteristics of the second object time waveform (ie, T wave).

[0101] [Formula 22]

[0102] According to formula (22), the function is obtained by multiplying the third inverse cumulative distribution function by the weight k c minus the fourth inverse cumulative distribution function multiplied by the weight k d function and add the level value β T The obtained function, weight k c and weight kd However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. c and weight k d Therefore, although the signal analysis device 1 can set the weight k c and weight k d are all fitted as positive values, but there is no need to set the weight k c and weight k d All fit to positive values.

[0103] Figure 6 is a schematic diagram showing a first cumulative distribution function f with respect to a first target time waveform (ie, R wave). a (x) multiplied by weight k a And add the level value β R The obtained function k a f a (x)+β R , the second cumulative distribution function f b (x) multiplied by weight k b And add the level value β R The obtained function k b f b (x)+β R , and the weighted difference between the first cumulative distribution function and the second cumulative distribution function plus the level value β R The approximate time waveform k a f a (x)-k b f b (x)+β R The single-point dash line is the first cumulative distribution function f a (x) multiplied by weight k a And add the level value β R The obtained function k a f a (x)+β R , the double-dotted line is the second cumulative distribution function f b (x) multiplied by weight k b And add the level value β R The obtained function k b f b (x)+β R , the dotted line is the approximate time waveform k a f a (x)-k b f b (x)+β R The approximate time waveform k a f a (x)-kb f b (x)+β R This is a waveform that approximates the first target time waveform (ie, the R wave).

[0104] Figure 7 is a schematic diagram showing a third inverse cumulative distribution function f′ with respect to a second object time waveform (ie, T wave). c (x) = 1-f c (x) multiplied by weight k c And add the level value β T The obtained function k c f' c (x)+β T , the fourth inverse cumulative distribution function f' d (x) = 1-f d (x) multiplied by weight k d And add the level value β T The obtained function k d f' d (x)+β T , and the weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function plus the level value β T The approximate time waveform k c f' c (x)-k d f' d (x)+β T The single-point dashed line is the third inverse cumulative distribution function f' c (x) = 1-f c (x) multiplied by weight k c And add the level value β T The obtained function k c f' c (x)+β T , the double-dotted line is the fourth inverse cumulative distribution function f' d (x) = 1-f d (x) multiplied by weight k d And add the level value β T The obtained function k d f' d (x)+β T , the dotted line is the approximate time waveform k c f' c (x)-k d f' d (x)+β T The approximate time waveform k c f' c (x)-k d f' d (x)+βT This is a waveform that approximates the second target time waveform (ie, T wave).

[0105] In addition, the potential at the beginning of the R wave (more precisely, the QRS wave) is the level value β R It is a value indicating the magnitude of the DC component at the beginning of the R wave. When there is an abnormality in the coronary artery, the level value β R Sometimes it drops to the negative side. In addition, the potential at the end of the T wave, i.e., the level value β T It is a value indicating the magnitude of the DC component at the end of the T wave. When there is an abnormality in myocardial repolarization, the level value β T Sometimes it rises to the positive side.

[0106] [Approximate the difference between the target time waveform and the approximate time waveform] When the R wave or T wave in a special state is used as the target time waveform, a portion that cannot approximate the target time waveform (hereinafter referred to as a "residual portion") may remain in the above-mentioned approximate time waveform. For example, when the target heart has early repolarization or conduction disorder (auxiliary conduction pathway, etc.), the target time waveform (R wave) may include such Figure 8 The Δ wave is shown by the dotted line in the middle. The Δ wave portion cannot approximate the target time waveform in the above-mentioned approximate time waveform, and remains as a residual portion. Since the residual portion is also a time waveform caused by a certain activity of the heart, in the signal analysis device 1, the residual portion can be analyzed by assuming a cumulative Gaussian distribution, a function of a cumulative Gaussian distribution multiplied by a weight, a difference of cumulative Gaussian distributions, or a weighted difference of cumulative Gaussian distributions.

[0107] That is, the signal analysis device 1 can also use the cumulative distribution function (conveniently called the "fifth cumulative distribution function") of a certain unimodal distribution (conveniently called the "fifth unimodal distribution") or the fifth cumulative distribution function multiplied by a weight function to approximate the residual time waveform, that is, the difference between the object time waveform and the approximate time waveform, that is, the residual time waveform, to approximate the residual time waveform, and the parameters that determine the fifth unimodal distribution or the parameters that determine the fifth cumulative distribution function are also obtained as parameters representing the characteristics of the object time waveform.

[0108] Alternatively, the signal analysis device 1 may also approximate the residual time waveform by using the difference or weighted difference between the cumulative distribution function (conveniently referred to as the "fifth cumulative distribution function") of a certain unimodal distribution (conveniently referred to as the "fifth cumulative distribution function") and the cumulative distribution function (conveniently referred to as the "sixth cumulative distribution function") of a unimodal distribution different from the fifth unimodal distribution (conveniently referred to as the "sixth cumulative distribution function") for the difference between the object time waveform and the approximate time waveform, i.e., the residual time waveform, and the parameters of the fifth unimodal distribution or the parameters of the fifth cumulative distribution function, as well as the parameters of the sixth unimodal distribution or the parameters of the sixth cumulative distribution function, as parameters representing the characteristics of the object time waveform.

[0109] More specifically, when the residual time waveform is approximated by the fifth cumulative distribution function which is the cumulative distribution function of the fifth unimodal distribution shown in equation (23), the signal analysis device 1 approximates the residual time waveform which is the difference between the target time waveform and the approximated time waveform by the approximate time waveform f of the fifth cumulative distribution function shown in equation (24). 5 (x) to approximate the residual time waveform. In addition to the above parameters representing the characteristics of the target time waveform, the fifth cumulative distribution function parameter, i.e., the average value μ 5 and standard deviation σ 5 Parameters representing characteristics of the target time waveform are also acquired.

[0110] [Formula 23]

[0111] [Formula 24]

[0112] When the residual time waveform is approximated by a function multiplied by the fifth cumulative distribution function and weighted, the signal analysis device 1 uses the fifth cumulative distribution function f shown in equation (24) for the residual time waveform, which is the difference between the target time waveform and the approximated time waveform. 5 (x) multiplied by weight k 5 The approximate time waveform k of the obtained function 5 f 5 (x) to approximate the residual time waveform. In addition to the above parameters representing the characteristics of the target time waveform, the fifth cumulative distribution function parameter, i.e., the average value μ 5 and standard deviation σ 5 The signal analysis device 1 can also obtain the parameter representing the characteristics of the target time waveform. 5 Parameters representing characteristics of the target time waveform are also acquired.

[0113] When the residual time waveform is approximated by the difference between the fifth cumulative distribution function and the sixth cumulative distribution function which is the cumulative distribution function of the sixth unimodal distribution shown in equation (25), for example, the signal analysis device 1 uses the waveform obtained by subtracting the sixth cumulative distribution function shown in equation (26) from the fifth cumulative distribution function shown in equation (24), i.e., the approximate time waveform f, for the difference between the object time waveform and the approximate time waveform, i.e., the residual time waveform. 5 (x)-f 6 (x) to approximate the residual time waveform. In addition to the above parameters representing the characteristics of the target time waveform, the fifth cumulative distribution function parameter, i.e., the average value μ 5 and standard deviation σ 5 , and determine the parameter of the sixth cumulative distribution function, namely the mean value μ 6 and standard deviation σ 6 Parameters representing characteristics of the target time waveform are also acquired.

[0114] [Formula 25]

[0115] [Formula 26]

[0116] When the residual time waveform is approximated by the weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, the signal analysis device 1 approximates the residual time waveform, which is the difference between the target time waveform and the approximated time waveform, by using the fifth cumulative distribution function f 5 (x) multiplied by weight k 5 The function k 5 f 5 (x) minus the sixth cumulative distribution function f 6 (x) multiplied by weight k 6 The function k 6 f 6 (x) and the waveform obtained is the approximate time waveform k 5 f 5 (x)-k 6 f 6 (x) to approximate the residual time waveform. In addition to the above parameters representing the characteristics of the target time waveform, the fifth cumulative distribution function parameter, i.e., the average value μ 5 and standard deviation σ 5 , determine the parameter of the sixth cumulative distribution function, namely the mean value μ 6 and standard deviation σ 6 The signal analysis device 1 can also obtain the parameter representing the characteristics of the target time waveform. 5 and weight k 6 , or weight k 5 With weight k6 The ratio (k 5 / k 6 or k 6 / k 5 ) is also obtained as a parameter representing the characteristics of the object time waveform.

[0117] According to the fifth cumulative distribution function multiplied by the weight k 5 The function of the sixth cumulative distribution function multiplied by the weight k 6 The obtained approximate time waveform is used to approximate the residual time waveform, and the weight k 5 and weight k 6 However, when the subject's heart is in a specific state, the weight k obtained by fitting cannot be denied. 5 and weight k 6 Therefore, although the signal analysis device 1 can set the weight k 5 and weight k 6 are all fitted as positive values, but there is no need to set the weight k 5 and weight k 6 All fit to positive values.

[0118] return Figure 1 The signal analysis device 1 includes a control unit 11 and executes a program. The control unit 11 includes a processor 91 such as a CPU and a memory 92 connected via a bus. The signal analysis device 1 functions as a device including the control unit 11, the input unit 12, the communication unit 13, the storage unit 14, and the output unit 15 by executing the program.

[0119] More specifically, the processor 91 reads out a program stored in the storage unit 14 and stores the read program in the memory 92. As the processor 91 executes the program stored in the memory 92, the signal analysis device 1 functions as a device including the control unit 11, the input unit 12, the communication unit 13, the storage unit 14, and the output unit 15.

[0120] The control unit 11 controls the operation of various functional units included in the signal analysis device 1. For example, the control unit 11 performs a myocardial activity information parameter acquisition process. For example, the control unit 11 controls the operation of the output unit 15 so that the output unit 15 outputs the acquisition result of the myocardial activity information parameter acquisition process. The control unit 11 records various information generated by, for example, performing the myocardial activity information parameter acquisition process in the storage unit 14.

[0121] The input unit 12 includes input devices such as a mouse, a keyboard, and a touch panel. The input unit 12 may be configured as an interface for connecting these input devices to the signal analysis device 1. The input unit 12 receives input of various information to the signal analysis device 1.

[0122] For example, for a plurality of candidates for each cumulative distribution function used for fitting, information indicating the shape of the distribution represented by each cumulative distribution function (hereinafter referred to as “distribution shape designation information”) is input to the input unit 12 .

[0123] In addition, the distribution shape specifying information may be stored in advance in the storage unit 14. In this case, the distribution shape specifying information stored in the storage unit 14 does not need to be input from the input unit 12. In the following, for simplicity of description, the signal analysis device 1 is described by taking the case where the storage unit 14 stores the distribution shape specifying information in advance as an example.

[0124] The communication unit 13 is configured to include a communication interface for connecting the signal analysis device 1 to an external device. The communication unit 13 communicates with the external device via wired or wireless communication. The external device is, for example, a device that is a source of the waveform of the electrocardiogram of the subject's heart. The device that is a source of the waveform of the electrocardiogram of the subject's heart is, for example, an electrocardiogram measuring device. When the external device is, for example, an electrocardiogram measuring device, the communication unit 13 obtains the waveform of the electrocardiogram from the electrocardiogram measuring device through communication. In addition, the waveform of the electrocardiogram may also be input to the input unit 12.

[0125] The storage unit 14 is configured using a non-transitory computer-readable storage medium device such as a magnetic hard disk device or a semiconductor storage device. The storage unit 14 stores various information related to the signal analysis device 1. The storage unit 14 stores, for example, information input via the input unit 12 or the communication unit 13. The storage unit 14 stores, for example, an electrocardiogram input via the input unit 12 or the communication unit 13. The storage unit 14 stores, for example, various information generated by executing a myocardial activity information parameter acquisition process.

[0126] The output unit 15 outputs various information. The output unit 15 is configured to include display devices such as a CRT (Cathode Ray Tube) display, a liquid crystal display, and an organic EL (Electro-Luminescence) display. The output unit 15 can also be configured as an interface for connecting these display devices to the signal analysis device 1. The output unit 15, for example, outputs information input to the input unit 12. The output unit 15 can also display an electrocardiogram input to the input unit 12 or the communication unit 13. The output unit 15 can also display the execution result of the myocardial activity information parameter acquisition process.

[0127] Fig. 9 1 is a diagram showing an example of the functional configuration of the control unit 11 in the first embodiment. The control unit 11 includes an electrocardiogram acquisition unit 110 , a fitting information acquisition unit 120 , an analysis unit 130 , and a recording unit 140 .

[0128] The electrocardiogram acquisition unit 110 acquires a waveform of one cycle from the waveform of the electrocardiogram of the target heart input to the input unit 12 or the communication unit 13, and outputs it to the analysis unit 130. The waveform of the electrocardiogram of the target heart is a waveform in which waveforms of multiple beats (multiple cycles) are arranged in a time series. Even in the case of an abnormality in the target heart, the waveforms of all beats included in the waveform of the electrocardiogram will not be special waveforms, and there are many cases where only the waveforms of any few beats included in the waveform of the electrocardiogram become characteristic waveforms. In the myocardial activity information parameter acquisition process, it is preferred to take this characteristic waveform as an object. Therefore, the electrocardiogram acquisition unit 110 acquires a waveform of one cycle as a characteristic waveform from the waveform of the electrocardiogram of the target heart. For example, the electrocardiogram acquisition unit 110 can use a known technique for determining similarity or specificity to acquire a waveform of one cycle as a characteristic waveform from the waveform of the electrocardiogram of the target heart. In addition, for example, the electrocardiogram acquisition unit 110 can also cause the output unit 15 to display the waveform of the electrocardiogram, cause the input unit 12 to accept the designation of the waveform of one cycle by a user such as a doctor, and obtain the waveform of one cycle corresponding to the designation accepted by the input unit 12 from the electrocardiogram waveform.

[0129] The electrocardiogram acquisition unit 110 outputs the waveform of one cycle as digital time series data sampled at a prescribed sampling frequency. The prescribed sampling frequency refers to the sampling frequency of the signal used in the processing of the fitting information acquisition unit 120, for example, 250 Hz. In the case where the waveform of the input electrocardiogram is sampled at a prescribed sampling frequency, the electrocardiogram acquisition unit 110 only needs to cut out the digital time series data of the waveform of one cycle from the digital time series data of the waveform of the input electrocardiogram and output it. In the case where the waveform of the input electrocardiogram is sampled at a sampling frequency different from the prescribed sampling frequency, the electrocardiogram acquisition unit 110 only needs to cut out the digital time series data of the waveform of one cycle from the digital time series data of the waveform of the input electrocardiogram and convert it to the prescribed sampling frequency before outputting it.

[0130] The electrocardiogram acquisition unit 110 also determines the time interval of the R wave and the time interval of the T wave included in the waveform of one cycle of the electrocardiogram, and also obtains information for determining the time interval of the R wave and information for determining the time interval of the T wave and outputs them to the analysis unit 130. For example, the electrocardiogram acquisition unit 110 only needs to determine the start end of the R wave, the end end of the R wave, the start end of the T wave, and the end end of the T wave by a known technique, and obtain the sample number corresponding to the determined start end of the R wave, the end end of the R wave, the start end of the T wave, and the end end of the T wave, respectively, and the relative time from the start end of the waveform, as information for determining the time interval of the R wave and information for determining the time interval of the T wave. In addition, the R wave in this specification refers to the QRS wave to be precise. Although there are various interpretations about which point in the waveform the start end of the R wave (i.e., the start end of the QRS wave) is, it is true that the electrocardiogram acquisition unit 110 only needs to use the point determined by any known technique as the start end of the R wave.

[0131] In addition, in the waveform of the electrocardiogram of the target heart, a characteristic waveform that changes with the passage of time sometimes appears. Therefore, the electrocardiogram acquisition unit 110 may also acquire waveforms of multiple cycles from the waveform of the electrocardiogram of the target heart as waveforms of the target for myocardial activity information parameter acquisition processing. That is, the electrocardiogram acquisition unit 110 may also acquire a predetermined long-term time series waveform (trend graph) from the waveform of the electrocardiogram of the target heart, and output information about the waveform of each cycle included in the acquired waveform, the waveform and the time interval of the R wave included in the waveform, and the time interval of the T wave included in the waveform to the analysis unit 130.

[0132] The fitting information acquisition unit 120 acquires the distribution shape specifying information. If the distribution shape specifying information is stored in the storage unit 14 , the fitting information acquisition unit 120 reads the distribution shape specifying information from the storage unit 14 .

[0133] The analyzing unit 130 includes a fitting unit 131 and a myocardial activity information parameter acquiring unit 132 .

[0134] The fitting unit 131 fits the target time waveform, which is the waveform of the time section of at least one of the R wave and the T wave included in the waveform of one cycle of the electrocardiogram acquired by the electrocardiogram acquisition unit 110 , using the candidate of the cumulative distribution function indicated by the distribution shape designation information.

[0135] The myocardial activity information parameter acquisition unit 132 acquires parameters representing the characteristics of the target time waveform based on the fitting result performed by the fitting unit 131. For example, the myocardial activity information parameter acquisition unit 132 acquires, as parameters representing the characteristics of the target time waveform, parameters that determine the first unimodal distribution or parameters that determine the first cumulative distribution function, and parameters that determine the second unimodal distribution or parameters that determine the second cumulative distribution function, when the target time waveform is approximated by the difference or weighted difference between the first cumulative distribution function that is the first unimodal distribution cumulative distribution function and the second cumulative distribution function that is the second unimodal distribution cumulative distribution function. The parameters representing the characteristics of the target time waveform acquired by the myocardial activity information parameter acquisition unit 132 are an example of myocardial activity parameters.

[0136] In this way, the analyzing unit 130 acquires the myocardial activity parameter based on the target time waveform, which is the waveform of the time section of at least one of the R wave and the T wave included in the waveform of one cycle of the electrocardiogram of the target heart, and the distribution candidate information.

[0137] The storage unit 14 records various information generated by the processing executed by the control unit 11 in the storage unit 14 .

[0138] Fig.10 1 is a flowchart showing an example of a processing flow performed by the signal analysis device 1 in the first embodiment. The electrocardiogram acquisition unit 110 acquires the waveform of one cycle of the electrocardiogram of the target heart, information for determining the time interval of the R wave, and information for determining the time interval of the T wave via the input unit 12 or the communication unit 13 (step S101). Next, the fitting information acquisition unit 120 acquires the distribution shape designation information (step S102). Next, the fitting unit 131 uses the candidate of the cumulative distribution function indicated by the distribution shape designation information to fit the waveform of the time interval of at least one of the R wave and the T wave included in the waveform of one cycle of the electrocardiogram acquired in step S101, that is, the target time waveform (step S103). Next, the myocardial activity information parameter acquisition unit 132 acquires the myocardial activity parameter based on the fitting result (step S104). The acquired myocardial activity parameter is output to the output unit 15 (step S105).

[0139] In step S105, a graph of the fitting result for each target time waveform may be displayed. In addition, the process of step S102 may be performed before the process of step S103, or before the execution of step S101. The processes of step S103 and step S104 are examples of the processes executed by the analysis unit 130.

[0140] Fig.11 1 is a first diagram showing an example of the analysis result of the signal analysis device 1 in the first embodiment. More specifically, Fig.11This is an example of the analysis result of the waveform of the electrocardiogram of the subject's heart that is operating normally by the signal analysis device 1 . Fig.11 The horizontal axis represents time, and the vertical axis represents potential. The unit of the vertical axis is an arbitrary unit.

[0141] Fig.11 It shows that using k a / α 1 Instead of weight k a In order to make the first fitting result at time T3 and the third fitting result at time T4 the same value, k b / α 2 Instead of weight k b An example of the results of the first and second fittings on the depolarized R wave of the electrocardiogram of the object heart with normal operation and the results of the fourth and third fittings on the repolarized T wave of the electrocardiogram of the object heart with normal operation is displayed so that the second fitting result at time T3 and the fourth fitting result at time T4 have the same value. Hereinafter, the result of connecting the first fitting result and the third fitting result with changed weights via a straight line is referred to as the myocardial inner layer activity approximation function, and the result of connecting the second fitting result and the fourth fitting result with changed weights via a straight line is referred to as the myocardial outer layer activity approximation function.

[0142] Fig.11 Corresponding to the fact that in the depolarization of a normal heart, the activity of the inner myocardium layer (i.e., the activity of the ion channels) starts earlier than that of the outer myocardium layer and progresses rapidly, the activity of the outer myocardium layer starts slightly later than the timing of the start of the activity of the ion channels of the inner myocardium layer, and the difference between the timing of the start of the activity of the ion channels of the inner myocardium layer and the timing of the start of the activity of the outer myocardium layer becomes a positive and sharp R wave. That is, the mean value and standard deviation of the inner myocardium layer activity approximate function of the first fitting result of the inner myocardium layer activity approximate function and the mean value and standard deviation of the second fitting result of the outer myocardium layer activity approximate function are parameters indicating the timing of the activity of the ion channels and the progress of the activity in the depolarization of a normal heart.

[0143] in addition, Fig.11Corresponding to the fact that in the repolarization phase of a normal heart, the outer myocardium starts to inactivate the activity of ion channels earlier than the inner myocardium, the inactivation of the inner myocardium starts later than the inactivation of the outer myocardium, the inactivation of the outer myocardium and the inactivation of the inner myocardium both progress slowly, and the difference between the inactivation of the inner myocardium and the inactivation of the outer myocardium becomes a positive and slow T wave. That is, the average value and standard deviation of the fourth fitting result of the outer myocardial layer activity approximate function and the average value and standard deviation of the third fitting result of the inner myocardial layer activity approximate function are parameters indicating the timing of the inactivation of the activity of ion channels in the repolarization phase of a normal heart and the progress of the inactivation. As described above, the inner myocardial layer activity approximate function and the outer myocardial layer activity approximate function obtained by the first to fourth fittings correspond to the timing and progress of the collective activation of ion channels in the depolarization phase and the collective inactivation of ion channels in the repolarization phase, and the average value and standard deviation of each fitting result of each function are parameters indicating myocardial activity.

[0144] Fig.11 The shapes of the myocardial inner layer activity approximation function and the myocardial outer layer activity approximation function are roughly consistent with the measurement results of the electromotive force of the myocardium directly measured by inserting a catheter electrode into the myocardium of a subject heart with normal operation. This shows that according to the signal analysis device 1, it is possible to obtain information representing myocardial activity only from the electrocardiogram without inserting a catheter electrode.

[0145] Fig.12 2 is a second diagram showing an example of the analysis result of the signal analysis device 1 in the first embodiment. More specifically, Fig.12 This is an example of the analysis result of the waveform of the electrocardiogram of the subject's heart that is operating normally by the signal analysis device 1 .

[0146] Fig.12 The three results of graph G5, graph G6 and result G7 are shown. Fig.12 In the figure, the "inner layer side cumulative distribution function" shows the fitting result of the function representing the collective channel activity timing distribution of ion channels present in the inner layer of the myocardium. Fig.12 In the figure, the "outer layer side cumulative distribution function" shows the fitting result of the function representing the collective channel activity timing distribution of ion channels present in the outer layer of the myocardium. Fig.12 In the figure, the "electrical potential of the body surface" is a function representing the time change of the electric potential of the body surface, which is the waveform of the electrocardiogram. Fig.12 The horizontal axis represents time, and the vertical axis represents potential. The units of the horizontal and vertical axes are both arbitrary units. Figure 12 to Figure 14 The time lengths shown by the intervals between each horizontal axis are the same. Figure 12 to Figure 14 The vertical axis is 1, which represents the maximum value of the cumulative Gaussian distribution.

[0147] Graph G5 shows the entire waveform of the electrocardiogram generated in one beat. Graph G6 shows an enlarged view of the T wave region as a part of graph G5. The T wave region is Fig.12 The result G7 shows the statistics of two Gaussian distributions, namely, the Gaussian distribution as the cumulative source of the inner-side cumulative distribution function and the Gaussian distribution as the cumulative source of the outer-side cumulative distribution function. Each value of the result G7 represents the statistics of the two Gaussian distributions. The statistics of the two Gaussian distributions are specifically the mean value and standard deviation of the Gaussian distributions as the cumulative source of the inner-side cumulative distribution function and the outer-side cumulative distribution function.

[0148] Fig.13 3 is a third diagram showing an example of the analysis result of the signal analysis device 1 in the first embodiment. More specifically, Fig.13 This is an example of the analysis result of the waveform of the electrocardiogram of the target heart whose behavior is T prolongation type 3 by the signal analysis device 1 .

[0149] Fig.13 The three results of graph G8, graph G9 and result G10 are shown. Fig.13 In the figure, the "inner layer side cumulative distribution function" shows the fitting result of the function representing the collective channel activity timing distribution of the channels existing in the inner layer of the myocardium. Fig.13 In the figure, the “outer layer cumulative distribution function” shows the fitting result of the function of the collective channel activity timing distribution of the channels existing in the outer layer of the myocardium. Fig.13 In the figure, the "electrical potential of the body surface" is a function representing the time change of the electric potential of the body surface, which is the waveform of the electrocardiogram. Fig.13 The horizontal axis represents time, and the vertical axis represents potential. The units of the horizontal axis and the vertical axis are both arbitrary units.

[0150] Graph G8 shows the entire waveform of the electrocardiogram generated in one beat. Graph G9 shows an enlarged view of the T wave region as a part of graph G8. The T wave region is Fig.13 The area shown in FIG. 1 is area A2. Result G10 shows the statistics of two Gaussian distributions, namely, the Gaussian distribution as the cumulative source of the inner-side cumulative distribution function and the Gaussian distribution as the cumulative source of the outer-side cumulative distribution function. Each value of result G10 represents the statistics of the two Gaussian distributions, namely, the mean value and standard deviation of the Gaussian distributions as the cumulative source of the inner-side cumulative distribution function and the outer-side cumulative distribution function.

[0151] Fig.13The shapes of the inner layer side cumulative distribution function and the outer layer side cumulative distribution function are substantially consistent with the result of directly measuring the change of the electromotive force generated by the pulsation of the outer layer of the myocardium of the target heart whose motion is T prolongation type 3 by inserting the catheter electrode. This shows that according to the signal analysis device 1, it is possible to obtain information indicating the myocardial activity only from the electrocardiogram without inserting the catheter electrode.

[0152] Fig.14 4 is a fourth diagram showing an example of the analysis result of the signal analysis device 1 in the first embodiment. More specifically, Fig.14 This is an example of the analysis result of the waveform of the electrocardiogram of the target heart whose behavior is QT prolongation type 1 by the signal analysis device 1 .

[0153] Fig.14 The three results of graph G11, graph G12 and result G13 are shown. Fig.14 In the figure, the "inner layer side cumulative distribution function" shows the fitting result of the function representing the collective channel activity timing distribution of the channels existing in the inner layer of the myocardium. Fig.14 In the figure, the “outer layer side cumulative distribution function” shows the fitting result of the function representing the collective channel activity timing distribution of the channels existing in the outer layer of the myocardium. Fig.14 In the figure, the "electrical potential of the body surface" is a function representing the time change of the electric potential of the body surface, which is the waveform of the electrocardiogram. Fig.14 The horizontal axis represents time, and the vertical axis represents potential. The units of the horizontal and vertical axes are both arbitrary units.

[0154] Graph G11 shows the entire waveform of the electrocardiogram generated in one beat. Graph G12 shows an enlarged view of the T wave region as a part of graph G11. The T wave region is Fig.13 The result G13 shows the statistics of two Gaussian distributions, namely, the Gaussian distribution as the cumulative source of the inner-side cumulative distribution function and the Gaussian distribution as the cumulative source of the outer-side cumulative distribution function. Each value of the result G13 represents the statistics of the two Gaussian distributions. The statistics of the two Gaussian distributions are specifically the mean value and the standard deviation of the Gaussian distributions as the cumulative source of the inner-side cumulative distribution function and the outer-side cumulative distribution function.

[0155] Fig.14 The shapes of the inner layer side cumulative distribution function and the outer layer side cumulative distribution function are roughly consistent with the result of directly measuring the change of the electromotive force generated by the pulsation of the outer layer of the myocardium of the target heart whose action is QT prolongation type 3 by inserting the catheter electrode. This shows that according to the signal analysis device 1, it is possible to obtain information representing the myocardial activity only from the electrocardiogram without inserting the catheter electrode.

[0156] in addition, Fig.14This is also an example of the estimation result of the channel current characteristic related to sudden death by the signal analysis device 1 .

[0157] use Figure 15 to Figure 17 , it is explained that even for an electrocardiogram of ventricular extrasystole, the signal analysis device 1 can obtain information indicating myocardial activity only from the electrocardiogram. Figure 15 to Figure 17 The horizontal axis represents time (seconds), and the vertical axis represents potential (mV).

[0158] Fig.15 This is a first explanatory diagram for explaining an example of an electrocardiogram of a ventricular extrasystole analyzed by the signal analysis device 1 according to the first embodiment. Fig.16 This is a second explanatory diagram showing an example of an electrocardiogram of a ventricular extrasystole analyzed by the signal analysis device 1 according to the first embodiment. Fig.17 This is a third explanatory diagram showing an example of an electrocardiogram of a ventricular extrasystole analyzed by the signal analysis device 1 according to the first embodiment.

[0159] More specifically, Fig.15 Shows the cardiac potential on the body surface. That is, Fig.15 A normal heartbeat and two consecutive ventricular extrasystoles are shown in the electrocardiogram. More specifically, Fig.16 The myocardial inner layer activity approximation function including the inner layer cumulative distribution function of the depolarization of the ventricular extrasystole and the inner layer cumulative distribution function of the repolarization phase analyzed by the signal analyzing device 1 and the myocardial outer layer activity approximation function including the outer layer cumulative distribution function of the depolarization of the ventricular extrasystole and the outer layer cumulative distribution function of the repolarization phase analyzed by the signal analyzing device 1 are shown. More specifically, Fig.17 An example of a waveform of a ventricular extrasystole in an electrocardiogram actually measured is shown.

[0160] Fig.16 The inner cumulative distribution function of depolarization precedes the outer cumulative distribution function, and the standard deviation of both is larger than that of a normal heartbeat, indicating that the spread of excitement is slow. This analysis result is consistent with the waveform characteristics of the R wave with a wide downward swing.

[0161] Fig.16 The inner layer cumulative distribution function starts to deactivate earlier than the outer layer cumulative distribution function in the repolarization phase, and the order of deactivation of the inner layer and the outer layer is shown as the magnitude relationship of the average values ​​of the two cumulative distribution functions in the repolarization phase. Fig.16 The function obtained by subtracting the outer cumulative distribution function from the inner cumulative distribution function is consistent with the characteristics of the large negative T wave in the repolarization phase, such as Fig.17 As shown, the waveform obtained by subtracting the outer layer side cumulative distribution function from the inner layer side cumulative distribution function is substantially consistent with the waveform of the ventricular extrasystole in the actually measured electrocardiogram.

[0162] in addition, Figure 15 to Figure 17 The results show that the analysis by the signal analysis device 1 corresponds to an example of a giant wave or negative potential generated by altered conduction, early repolarization or delayed repolarization of the myocardial excitation. Figure 15 to Figure 17 In the depolarized phase, the mean μ of the cumulative distribution function of the inner layer is -1, and the standard deviation σ is 0.32. Figure 15 to Figure 17 In the figure, the mean μ of the cumulative distribution function of the outer layer of depolarization is -0.8 and the standard deviation σ is 0.21. Figure 15 to Figure 17 In the case of the repolarization phase, the mean μ of the cumulative distribution function of the inner layer is 1, and the standard deviation σ is 1. Figure 15 to Figure 17 The average value μ of the cumulative distribution function of the outer layer side of the repolarization phase is 2.99, and the standard deviation σ is 0.7.

[0163] use Figures 18 to 20 , it is explained that, even with respect to an electrocardiogram of a target heart in the depolarization phase of Buddah syndrome type 1, information indicating myocardial activity can be obtained only from the electrocardiogram by the signal analysis device 1. Figures 18 to 20 The vertical axis represents the potential in millivolts.

[0164] Fig.18 This is a first explanatory diagram for explaining an example of an electrocardiogram of a target heart in a depolarization phase of Buddag syndrome type 1 analyzed by the signal analysis device 1 according to the first embodiment. Fig.19 This is a second explanatory diagram showing an example of an electrocardiogram of a target heart in the depolarization phase of Buddag syndrome type 1 analyzed by the signal analysis device 1 according to the first embodiment. Fig. 20 This is a third explanatory diagram showing an example of an electrocardiogram of a target heart in the depolarization phase of Buddag syndrome type 1 analyzed by the signal analysis device 1 according to the first embodiment.

[0165] More specifically, Fig.18 An electrocardiogram showing the second induction of the chest in Buddah syndrome. Fig.18 In the figure, the inner frame W1 shows the depolarization phase, and the inner frame W2 shows the partition of the repolarization phase. Fig.19 The same is true. That is, in Fig.19 In the figure, the inner frame W1 represents the depolarization phase, and the inner frame W2 represents the repolarization phase.

[0166] More specifically, Fig.19 The inner layer side cumulative distribution function and the outer layer side cumulative distribution function in the depolarization and repolarization phases analyzed by the signal analysis device 1 are shown. Fig.19 In the example shown, the repolarization phase of the inner layer side cumulative distribution function starts after the depolarization phase, showing the characteristics of early repolarization. Fig.19In the example of , the potential amplitude of the cumulative distribution function on the outer layer side is different between the depolarization phase and the repolarization phase. Fig.19 The gap and anisotropy between the depolarization phase and the repolarization phase of the outer layer cumulative distribution function are shown. In this way, the signal analysis device 1 can express the anisotropy of early repolarization, depolarization, and repolarization, which are the characteristics of the waveform shown in the electrocardiogram of the subject heart of Buddah syndrome, by the average value and standard deviation of the depolarization phase and the repolarization phase of the inner layer cumulative distribution function and the outer layer cumulative distribution function, and the ratio of the weight given to the outer layer cumulative distribution function to the weight given to the inner layer cumulative distribution function (inner-outer ratio).

[0167] Fig. 20 is the comparison between the analytical result and the measured value. More specifically, Fig. 20 Show Fig.19 The difference between the inner and outer cumulative distribution functions in the depolarization and repolarization phases. Fig. 20 The measured values ​​of the electrocardiogram are also shown. The analysis results are basically consistent with the measured values ​​except for the last one. The last one refers to the potential at the latest time.

[0168] In addition, Figures 18 to 20 In the depolarized phase, the mean μ of the cumulative distribution function of the inner layer is 15, and the standard deviation σ is 0.15. Figures 18 to 20 In the figure, the average value μ of the cumulative distribution function of the outer layer of the depolarized phase is 14, and the standard deviation σ is 0.25. Figures 18 to 20 In the depolarized phase, the ratio of the inner and outer layers is 0.45. Figures 18 to 20 In the figure, the mean μ of the cumulative distribution function of the inner layer of the repolarization phase is 25, and the standard deviation σ is 0.25. Figures 18 to 20 The mean μ of the cumulative distribution function of the outer layer side of the repolarization phase is 20, and the standard deviation σ is 0.5.

[0169] Figures 21 to 59 A library showing the use of publicly available ECG data<https: / / physionet.org / about / database / > The result of analysis by the signal analysis device 1. Figures 21 to 59 The inner layer cumulative distribution function, the outer layer cumulative distribution function, and the fitting result obtained as a result of analyzing the cardiac potential by the signal analyzing device 1 are shown.

[0170] Figures 21 to 59 Each of the figures shows an example of an electrocardiogram analyzed by the signal analysis device 1 of the first embodiment. The determined points recorded in each figure represent the Q point, R point, S point, T start point, and T end point of the cardiac potential in order from the left side of the figure. The determined points are determined by the inflection point detection and peak detection algorithms. Figure 21 to Figure 59The graphs show the inner cumulative distribution function and the outer cumulative distribution function in each interval obtained for the interval of the depolarization phase (QRS wave) and the interval of the repolarization phase (T wave). Figure 21 to Figure 59 It is shown that the signal analysis device 1 can show substantially the same shape for various QRS waves and T waves by adjusting the mean values ​​and standard deviations of the inner layer side cumulative distribution function and the outer layer side cumulative distribution function. Figure 21 to Figure 59 The figure below shows the original waveform of the electrocardiogram and the fitting result. In addition, Figure 21 to Figure 59 The results are the result of 300Hz sampling. Therefore, Figure 21 to Figure 59 The origin of the horizontal axis of each graph represents 0 seconds, and a value of 1 represents 3.33 milliseconds.

[0171] The signal analysis device 1 configured in this way takes as the target time waveform the waveform of an arbitrary time interval of the R wave and the T wave included in the waveform representing one cycle of the cardiac cycle of the target heart, and when the target time waveform is approximated by the difference or weighted difference between the first cumulative distribution function as the first unimodal distribution cumulative distribution function and the second cumulative distribution function as the second unimodal distribution cumulative distribution function, the parameters that determine the first unimodal distribution or the parameters that determine the first cumulative distribution function and the parameters that determine the second unimodal distribution or the parameters that determine the second cumulative distribution function are obtained as parameters that represent the characteristics of the target time waveform. The first cumulative distribution function is information representing the activity of the inner myocardial layer of the target heart, and the second cumulative distribution function is information representing the activity of the outer myocardial layer of the target heart. Therefore, the parameters representing the shape of the first cumulative distribution function are parameters representing the activity of the inner myocardial layer of the target heart (inner myocardial layer parameters), and the parameters representing the shape of the second cumulative distribution function are parameters representing the activity of the outer myocardial layer of the target heart (outer myocardial layer parameters). The conventional analysis of the electrocardiogram waveform cannot obtain information such as these parameters that clearly show the activity characteristics of the inner myocardium layer of the target heart and information that clearly show the activity characteristics of the outer myocardium layer of the target heart. Therefore, according to the signal analysis device 1, it is possible to obtain useful information for understanding the state of the heart from the electrocardiogram waveform.

[0172] (Get only some of the parameters) When only the activity characteristics of the inner myocardium layer of the target heart are to be understood, only the inner myocardium layer parameters may be obtained by the signal analysis device 1. When only the activity characteristics of the outer myocardium layer of the target heart are to be understood, only the outer myocardium layer parameters may be obtained by the signal analysis device 1. In addition, only a part of the parameters representing the shape of the cumulative distribution function may be obtained by the signal analysis device 1 as the inner myocardium layer parameters or the outer myocardium layer parameters.

[0173] For example, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 obtains at least any one of at least a portion of the parameters of the first unimodal distribution, at least a portion of the parameters of the first cumulative distribution function, at least a portion of the parameters of the second unimodal distribution, and at least a portion of the parameters of the second cumulative distribution function as parameters showing the activity of the myocardium of the target heart, namely, myocardial activity parameters, when approximating the first target time waveform as the waveform of the time interval of the R wave of the target heart by the time waveform obtained by the difference or weighted difference between the first cumulative distribution function as the cumulative distribution function of the first unimodal distribution and the second cumulative distribution function as the cumulative distribution function of the second unimodal distribution, or the time waveform obtained by adding the difference or weighted difference between the first cumulative distribution function as the cumulative distribution function of the first unimodal distribution and the second cumulative distribution function as the cumulative distribution function of the second unimodal distribution plus a level value.

[0174] For example, when both the first unimodal distribution and the second unimodal distribution are Gaussian distributions, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the mean value of the first unimodal distribution, the standard deviation or variance of the first unimodal distribution, the mean value of the second unimodal distribution, and the standard deviation or variance of the second unimodal distribution as the myocardial activity parameter. In addition, the mean value of the Gaussian distribution is the time when the frequency value in the unimodal distribution becomes the maximum value, and the time when the slope of the cumulative distribution function of the unimodal distribution becomes the maximum value. Therefore, for example, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the time corresponding to the maximum value of the first unimodal distribution, the time corresponding to the maximum slope of the first cumulative distribution function, the time corresponding to the maximum value of the second unimodal distribution, and the time corresponding to the maximum slope of the second cumulative distribution function as the myocardial activity parameter, regardless of whether the first unimodal distribution and the second unimodal distribution are Gaussian distributions.

[0175] For example, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 obtains at least any one of at least a portion of the parameters of the third unimodal distribution, at least a portion of the parameters of the third cumulative distribution function, at least a portion of the parameters of the fourth unimodal distribution, and at least a portion of the parameters of the fourth cumulative distribution function as a waveform obtained by adding a level value to the difference or weighted difference between the third cumulative distribution function as the third unimodal distribution and the fourth cumulative distribution function as the fourth unimodal distribution as a waveform obtained by adding a level value to the difference or weighted difference between the third cumulative distribution function as the third unimodal distribution and the fourth cumulative distribution function as the fourth unimodal distribution as a waveform obtained by inverting the time axis of the second object time waveform of the time interval of the T wave of the object heart, namely, the second object inverse time waveform, as a parameter showing the activity of the myocardium of the object heart, namely, the myocardial activity parameter.

[0176] For example, when both the third unimodal distribution and the fourth unimodal distribution are Gaussian distributions, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the mean value of the third unimodal distribution, the standard deviation or variance of the third unimodal distribution, and the mean value of the fourth unimodal distribution, and the standard deviation or variance of the fourth unimodal distribution as the myocardial activity parameter. In addition, for example, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the time corresponding to the maximum value of the third unimodal distribution, the time corresponding to the maximum slope of the third cumulative distribution function, the time corresponding to the maximum value of the fourth unimodal distribution, and the time corresponding to the maximum slope of the fourth cumulative distribution function as the myocardial activity parameter, regardless of whether the third unimodal distribution and the fourth unimodal distribution are Gaussian distributions. In addition, when the myocardial activity information parameter acquisition unit 132 acquires the time as the myocardial activity parameter, even when the waveform after the time axis is reversed is approximated, the time (the value of x in the above example) is acquired instead of the information indicating the time in the reverse direction (the value of x' in the above example).

[0177] For example, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 uses the cumulative distribution function of the third unimodal distribution as the third cumulative distribution function, uses the cumulative distribution function of the fourth unimodal distribution as the fourth cumulative distribution function, subtracts the third cumulative distribution function from 1 as the third inverse cumulative distribution function, and subtracts the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function. When the second target time waveform that is the waveform of the time interval of the T wave of the target heart is approximated by the waveform obtained by the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, the second approximate time waveform, or the waveform obtained by adding the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, the level value, at least any one of at least a part of the parameters that determine the third unimodal distribution, at least a part of the parameters that determine the third cumulative distribution function, at least a part of the parameters that determine the fourth unimodal distribution, and at least a part of the parameters that determine the fourth cumulative distribution function is acquired as a parameter representing the activity of the myocardium of the target heart, that is, the myocardial activity parameter.

[0178] For example, when both the third unimodal distribution and the fourth unimodal distribution are Gaussian distributions, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the mean value of the third unimodal distribution, the standard deviation or variance of the third unimodal distribution, and the mean value of the fourth unimodal distribution, and the standard deviation or variance of the fourth unimodal distribution as the myocardial activity parameter. In addition, for example, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the time corresponding to the maximum value of the third unimodal distribution, the time corresponding to the maximum slope of the third cumulative distribution function, the time corresponding to the maximum value of the fourth unimodal distribution, and the time corresponding to the maximum slope of the fourth cumulative distribution function as the myocardial activity parameter, regardless of whether the third unimodal distribution and the fourth unimodal distribution are Gaussian distributions.

[0179] Similarly, for example, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 also acquires at least any one of at least a part of the parameters that determine the fifth unimodal distribution and at least a part of the parameters that determine the fifth cumulative distribution function as a parameter showing the activity of the myocardium of the target heart, i.e., a myocardial activity parameter, when approximating a residual time waveform that is a time waveform of a difference between a first object time waveform and a first approximate time waveform, or a residual time waveform that is a time waveform of a difference between a second object time waveform and a second approximate time waveform, or a residual time waveform that is a time waveform of a difference between a second object time waveform and a waveform obtained by inverting the time axis of the second approximate inverse time waveform, i.e., a second approximate time waveform.

[0180] For example, when the fifth unimodal distribution is a Gaussian distribution, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the mean value of the fifth unimodal distribution, the standard deviation, or the variance of the fifth unimodal distribution as the myocardial activity parameter. In addition, for example, the myocardial activity information parameter acquisition unit 132 only needs to acquire any one of the time corresponding to the maximum value of the fifth unimodal distribution and the time corresponding to the maximum slope of the fifth cumulative distribution function as the myocardial activity parameter regardless of whether the fifth unimodal distribution is a Gaussian distribution.

[0181] In addition, for example, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 also obtains at least any one of at least a portion of the parameters that determine the fifth unimodal distribution, at least a portion of the parameters that determine the fifth cumulative distribution function, at least a portion of the parameters that determine the sixth unimodal distribution, and at least a portion of the parameters that determine the sixth cumulative distribution function as a parameter showing the activity of the myocardium of the object heart, namely, a myocardial activity parameter, when the residual time waveform is approximated by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function as the cumulative distribution function of the fifth unimodal distribution and the sixth cumulative distribution function as the cumulative distribution function of the sixth unimodal distribution.

[0182] For example, when both the fifth unimodal distribution and the sixth unimodal distribution are Gaussian distributions, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the average value of the fifth unimodal distribution, the standard deviation or variance of the fifth unimodal distribution, and the average value of the sixth unimodal distribution, and the standard deviation or variance of the sixth unimodal distribution as the myocardial activity parameter. In addition, for example, the myocardial activity information parameter acquisition unit 132 only needs to acquire at least one of the time corresponding to the maximum value of the fifth unimodal distribution, the time corresponding to the maximum slope of the fifth cumulative distribution function, the time corresponding to the maximum value of the sixth unimodal distribution, and the time corresponding to the maximum slope of the sixth cumulative distribution function as the myocardial activity parameter, regardless of whether the fifth unimodal distribution and the sixth unimodal distribution are Gaussian distributions.

[0183] [Acquisition of Myocardial Activity Parameters by Calculation of Parameters Obtained by Fitting] The characteristics of the activity of the myocardium of the object heart are not only manifested in the above-mentioned parameters obtained by fitting, but also clearly manifested in the values ​​obtained by the operation of the parameters obtained by fitting. Therefore, the values ​​obtained by the operation of the parameters obtained by fitting can also be obtained as myocardial activity parameters by the signal analysis device 1. The parameters obtained by fitting refer to at least one of the parameters for determining the unimodal distribution or the parameters for determining the cumulative distribution function, the weights, and the level values. The parameters for determining the unimodal distribution or the parameters for determining the cumulative distribution function are at least one of the mean value and the standard deviation (or variance) when the unimodal distribution is a Gaussian distribution. Regardless of whether the unimodal distribution is a Gaussian distribution, the moment corresponding to the maximum value of the unimodal distribution is an example of determining the parameters of the unimodal distribution, and the moment corresponding to the maximum slope of the cumulative distribution function of the unimodal distribution is an example of determining the parameters of the cumulative distribution function.

[0184] For example, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 can obtain parameters showing the activity of the myocardium of the object heart (parameters different from the above-mentioned parameters) by calculating the myocardial activity parameters obtained by the above-mentioned fitting of the waveform of the time interval of the R wave included in the waveform showing one cycle of the cardiac cycle of the object heart (i.e., parameters showing the activity of the myocardium of the object heart in the time interval of the R wave) and the myocardial activity parameters obtained by the above-mentioned fitting of the waveform of the time interval of the T wave included in the waveform showing one cycle (i.e., parameters showing the activity of the myocardium of the object heart in the time interval of the T wave).

[0185] In addition, for example, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 may obtain parameters showing the activity of the myocardium of the object heart (parameters different from the above-mentioned parameters) by calculating the parameters showing the activity of the inner layer of the myocardium of the object heart obtained by the above-mentioned fitting of the waveform of the time interval of the R wave included in the waveform showing one cycle of the cardiac cycle of the object heart and the parameters showing the activity of the outer layer of the myocardium of the object heart obtained by the fitting.

[0186] In addition, for example, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 may obtain parameters showing the activity of the myocardium of the object heart (parameters different from the above-mentioned parameters) by calculating the parameters showing the activity of the inner layer of the myocardium of the object heart obtained by the above-mentioned fitting of the waveform of the time interval of the T wave included in the waveform showing one cycle of the cardiac cycle of the object heart and the parameters showing the activity of the outer layer of the myocardium of the object heart obtained by the fitting.

[0187] Hereinafter, an example will be described in which, when all the first to fourth unimodal distributions are Gaussian distributions, a value obtained by calculating the average values ​​obtained by fitting is used as a myocardial activity parameter. When the first to fourth unimodal distributions are not Gaussian distributions, the "average value" in the following example may be replaced with "the time when the value is the largest in the unimodal distribution", i.e., "the time corresponding to the maximum value of the unimodal distribution", "the time when the slope is the largest in the cumulative distribution function", i.e., "the time corresponding to the maximum slope of the cumulative distribution function", etc.

[0188] (1) Parameters indicating the time from depolarization to repolarization It is known that when the time from depolarization of the inner or outer layer of the myocardium to repolarization of the inner or outer layer of the myocardium is extremely short or extremely long, sudden death may occur due to arrhythmia. That is, shortening or prolonging the time from depolarization of the myocardium to repolarization sometimes indicates the possibility of a certain pathological condition of the myocardium. Therefore, the signal analysis device 1 can obtain the time from depolarization of the inner or outer layer of the myocardium to repolarization of the inner or outer layer of the myocardium as a myocardial activity parameter. Specifically, at least any one of the following four parameters (1A) to (1D) can be obtained as a myocardial activity parameter.

[0189] (1A) Parameter representing the time from depolarization of the inner myocardial layer to repolarization of the inner myocardial layer The myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 may also acquire as the myocardial activity parameter the difference between the average value of the first unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the R wave included in the waveform representing one cycle of the cardiac cycle of the target heart, i.e., the first target time waveform, and the average value of the third unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the T wave included in the waveform representing one cycle, i.e., the second target time waveform. For example, if the average value of the first unimodal distribution is μ , as in the above-mentioned example, a , the mean of the third unimodal distribution is μ c , the myocardial activity information parameter acquisition unit 132 may also set |μ a -μ c | is obtained as myocardial activity parameter. In addition, if μ c In terms of time, μ a If the myocardial activity information parameter acquisition unit 132 is late, c -μ a The myocardial activity parameter is obtained. The myocardial activity parameter is a parameter indicating the time from depolarization of the inner myocardial layer to repolarization of the inner myocardial layer.

[0190] (1B) Parameter representing the time from depolarization of the outer myocardial layer to repolarization of the outer myocardial layer The myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 may also acquire as the myocardial activity parameter the difference between the average value of the second unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the R wave included in the waveform representing one cycle of the cardiac cycle of the target heart, i.e., the first target time waveform, and the average value of the fourth unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the T wave included in the waveform representing one cycle, i.e., the second target time waveform. For example, if the average value of the second unimodal distribution is μ , as in the above example, b , the mean of the fourth unimodal distribution is μ d , the myocardial activity information parameter acquisition unit 132 may also set |μ b -μ d | is obtained as myocardial activity parameter. In addition, if μ d In terms of time, μ b If the myocardial activity information parameter acquisition unit 132 is late, d -μ b The myocardial activity parameter is obtained. The myocardial activity parameter is a parameter indicating the time from depolarization of the myocardial outer layer to repolarization of the myocardial outer layer.

[0191] (1C) Parameter representing the time from depolarization of the outer myocardial layer to repolarization of the inner myocardial layer The myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 may also acquire as the myocardial activity parameter the difference between the average value of the second unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the R wave included in the waveform representing one cycle of the cardiac cycle of the target heart, i.e., the first target time waveform, and the average value of the third unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the T wave included in the waveform representing one cycle, i.e., the second target time waveform. For example, if the average value of the second unimodal distribution is μ , as in the above example, b , the mean of the third unimodal distribution is μ c , the myocardial activity information parameter acquisition unit 132 may also set |μ b -μ c | is obtained as myocardial activity parameter. In addition, if μ c In terms of time, μ b If the myocardial activity information parameter acquisition unit 132 is late, c -μ b The myocardial activity parameter is obtained. The myocardial activity parameter is a parameter indicating the time required to switch from depolarization of the myocardial outer layer to repolarization of the myocardial inner layer.

[0192] (1D) Parameter representing the time from the depolarization of the inner myocardium layer to the repolarization of the outer myocardium layer The myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 may also acquire as the myocardial activity parameter the difference between the average value of the first unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the R wave included in the waveform representing one cycle of the cardiac cycle of the target heart, i.e., the first target time waveform, and the average value of the fourth unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the T wave included in the waveform representing one cycle, i.e., the second target time waveform. For example, if the average value of the first unimodal distribution is μ , as in the above-mentioned example, a , the mean of the fourth unimodal distribution is μ d , the myocardial activity information parameter acquisition unit 132 may also set |μ a -μ d | is obtained as myocardial activity parameter. In addition, if μ d In terms of time, μ a If the myocardial activity information parameter acquisition unit 132 is late, d -μ a The myocardial activity parameter is obtained. The myocardial activity parameter is a parameter indicating the time required to switch from depolarization of the inner myocardial layer to repolarization of the outer myocardial layer.

[0193] (2) Parameters representing the time difference and order of inner layer activity and outer layer activity in depolarization When the time difference between the inner layer activity and the outer layer activity in depolarization is longer than normal, the conduction of myocardial excitation may be delayed or blocked, or the place where the excitation starts or the order of the excitation transmission may be different from the normal mode, especially indicating the obstruction of the myocardial stimulation conduction system or the ischemic state of the myocardium, and the existence of extrasystoles. Therefore, the signal analysis device 1 can obtain the parameters representing the time difference and order of the inner layer activity and the outer layer activity in depolarization as myocardial activity parameters. Specifically, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 can also obtain the difference between the average value of the first unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the R wave included in the waveform showing one cycle of the cardiac cycle of the object heart, that is, the first object time waveform, and the average value of the second unimodal distribution as the myocardial activity parameter. For example, if, as in the above example, the average value of the first unimodal distribution is μ a , the mean of the second unimodal distribution is μ b , then the myocardial activity information parameter acquisition unit 132 may also set μ a -μ b or μ b -μ a Obtained as myocardial activity parameters.

[0194] (3) Parameters representing the time difference and order of inner layer activity and outer layer activity in repolarization When the time difference between the inner layer activity and the outer layer activity in repolarization is prolonged or shortened, certain abnormalities may occur in the myocardium. Therefore, the signal analysis device 1 can obtain the parameters representing the time difference and order of the inner layer activity and the outer layer activity in repolarization as myocardial activity parameters. Specifically, the myocardial activity information parameter acquisition unit 132 of the analysis unit 130 of the signal analysis device 1 can also obtain the difference between the average value of the third unimodal distribution obtained by the above-mentioned fitting of the waveform of the time interval of the T wave included in the waveform showing one cycle of the cardiac cycle of the object heart, that is, the second object time waveform, and the average value of the fourth unimodal distribution as the myocardial activity parameter. For example, if, as in the above example, the average value of the third unimodal distribution is μ c , the mean of the fourth unimodal distribution is μ d , then the myocardial activity information parameter acquisition unit 132 may also set μ c -μ d or μ d -μ c Obtained as myocardial activity parameters.

[0195] (Correction of myocardial activity parameters) In addition, among the myocardial activity parameters related to the width in the time direction among the above-mentioned myocardial activity parameters acquired by the myocardial activity information parameter acquisition unit 132, there are characteristics that the influence of the variation caused by the heart rate and the influence of individual differences such as age and gender are seen, similarly to the parameters related to the width in the time direction among the previous parameters representing cardiac activity (e.g., QT interval). It is known that the previous parameters related to the width in the time direction are corrected to reduce the influence of the variation caused by the heart rate or the influence of individual differences, and the corrected values ​​are used as parameters for evaluating cardiac activity. Therefore, for the myocardial activity parameters acquired by the myocardial activity information parameter acquisition unit 132, corrections can also be performed to reduce the influence of the variation caused by the heart rate or the influence of individual differences, and the corrected values ​​can be used as parameters for evaluating myocardial activity. For example, for the variation caused by the heart rate, the myocardial activity parameters acquired by the above-mentioned fitting can also be corrected by linear or nonlinear calculation based on the time interval between adjacent R spines and R spines (hereinafter referred to as "RR interval"), and the corrected values ​​can be used as myocardial activity parameters. In addition, in order to account for individual differences, the myocardial activity parameters obtained by the fitting can be corrected by finding the relationship between the RR interval and the myocardial activity parameters based on the electrocardiogram data of each person, and the corrected values ​​can be used as the myocardial activity parameters. This correction can be performed by other devices after the signal analysis device 1 outputs the myocardial activity parameters obtained by the fitting, or it can be performed by the signal analysis device 1. In the case where the signal analysis device 1 corrects the myocardial activity parameters, for example, the myocardial activity information parameter acquisition unit 132 corrects the myocardial activity parameters obtained by the fitting, and outputs the corrected values ​​as the myocardial activity parameters. In addition, in the case where the correction is performed in the signal analysis device 1, the myocardial activity parameters obtained by the fitting become intermediate parameters obtained in the device in the end, but they are parameters representing myocardial activity in the same way as when they are output outside the device, and this point does not change.

[0196] (Variation Example) In addition, the electrocardiogram is preferably an induction electrocardiogram close to the electromotive force vector of the heart. The induction electrocardiogram close to the electromotive force vector is preferably an electrocardiogram that can obtain three-dimensional information of cardiac potentials such as II induction, V4 induction, and V5 induction. The more channels the electrocardiogram has, the more information it contains, so the more channels the electrocardiogram has, the better. That is, the waveforms of each channel of the multi-channel electrocardiogram can also be taken as objects and analyzed in the signal analysis device 1 to obtain myocardial activity parameters as the analysis results for each channel.

[0197] The signal analysis device 1 may be implemented using a plurality of information processing devices that are communicably connected via a network. In this case, each functional unit included in the signal analysis device 1 may be distributed and implemented in the plurality of information processing devices.

[0198] The electrocardiogram acquisition unit 110 is an example of a biological information acquisition unit.

[0199] <Second Embodiment> The first embodiment is to obtain myocardial activity parameters from biological information in a time series related to cardiac pulsation to obtain information useful for understanding the state of the heart. However, it is also possible to obtain information useful for understanding whether the heart is normal or what kind of disease is caused by an abnormal heart from biological information in a time series related to cardiac pulsation. This method is described as the second embodiment.

[0200] As described in the background technology section, even if the waveforms of electrocardiograms are similar, the pathogenesis of heart-related diseases may be different. That is, even if an estimation model is learned that uses biological information of a time series related to heart beats as input to obtain estimation results such as whether the heart is normal or what kind of disease is caused when the heart is abnormal, the estimation accuracy may not be improved if the estimation result is obtained by estimation using the learned estimation model.

[0201] On the other hand, according to the signal analysis device 1 of the first embodiment, information that clearly shows the activity characteristics of the myocardium can be obtained as a myocardial activity parameter from the biological information of the time series of one channel related to the heart beat. In addition, as described in the first embodiment, the myocardial activity parameter obtained by the signal analysis device 1 of the first embodiment is strongly related to whether the heart is normal or what kind of disease it corresponds to when the heart is abnormal. Therefore, if the myocardial activity parameter is obtained from the biological information of the time series related to the heart beat by the same processing as that performed by the signal analysis device 1 of the first embodiment, an estimation model is learned that uses the obtained myocardial activity parameter as an input to obtain estimation results such as whether the heart is normal or what kind of disease it corresponds to when the heart is abnormal, and an estimation result is obtained by estimation using the learned estimation model, it is expected that a high-precision estimation result can be obtained. The information providing system of the second embodiment performs this learning and estimation. When the doctor uses the estimation result of the information providing system of the second embodiment as a reference when making a diagnosis, it is expected that a doctor can more accurately diagnose whether the subject's heart is normal or what kind of disease it corresponds to when the subject's heart is abnormal.

[0202] like Fig.60 As shown in FIG. 1 , the information providing system 200 of the second embodiment includes a learning device 300 and an information providing device 400. The hardware configurations of the information providing system 200, the learning device 300, and the information providing device 400 are similar to those of FIG. Figure 1 The hardware structure of the signal analysis devices 1 shown is the same.

[0203] First, the learning device 300 will be described. [Learning device 300] like Fig.60 As shown, the learning device 300 includes a signal analysis unit 310 and a learning unit 320. The learning device 300 performs Fig.61 The processing of steps S310 and S320 is shown.

[0204] [Learning source data group, learning source data] A learning source data group is input to the learning device 300. The learning source data group includes a plurality of learning source data. Each learning source data includes at least biological information of a time series related to the pulsation of the heart of the object of acquisition of the learning source data (hereinafter referred to as "the acquired object heart"), and information indicating the state of the acquired object heart, i.e., heart state information. Each learning source data may also include information for identifying the learning source data, such as an identification number uniquely assigned to the learning source data.

[0205] An example of biological information of a time series related to the pulsation of the heart of the acquired object is an electrocardiogram waveform of the heart of the acquired object. The heart state information of the acquired object heart is information that uniquely specifies the diagnosis result of the heart state by a doctor who has seen the biological information of a time series related to the pulsation of the heart of the acquired object (e.g., electrocardiogram waveform). The diagnosis result of the heart state by the doctor specifically refers to whether the heart is in a normal state, and what disease the heart is in when it is not in a normal state. That is, for example, the heart state information is "0" when the doctor who has seen the electrocardiogram waveform determines that it is in a normal state, "1" when the doctor who has seen the electrocardiogram waveform determines that it is Burgardt syndrome, "2" when the doctor who has seen the electrocardiogram waveform determines that it is ischemic heart disease, "3" when the doctor who has seen the electrocardiogram waveform determines that it is early repolarization syndrome, etc. In this way, it is sufficient to pre-assign a certain numerical value to whether the heart is in a normal state or which disease it is.

[0206] Alternatively, for example, the heart status information may be "0" when the doctor who sees the ECG waveform determines that it is normal sinus rhythm, "1" when the doctor who sees the ECG waveform determines that it is supraventricular extrasystole, "2" when the doctor who sees the ECG waveform determines that it is atrial fibrillation, "3" when the doctor who sees the ECG waveform determines that it is ventricular extrasystole, and "4" when the doctor who sees the ECG waveform determines that it is ventricular fibrillation... In this way, it is also possible to pre-assign specific numerical values ​​to indicate whether the heart rhythm is normal or what kind of arrhythmia it is.

[0207] Alternatively, for example, the heart status information may be "0" when the doctor who sees the electrocardiogram waveform determines that the coronary blood flow state is normal, "1" when the doctor who sees the electrocardiogram waveform determines that the inner layer of the myocardium is in an ischemic state, "2" when the doctor who sees the electrocardiogram waveform determines that the inner and outer layers of the myocardium are in an ischemic state (transmural ischemic state), "3" when the doctor who sees the electrocardiogram waveform determines that it is anterior wall myocardial infarction, and "4" when the doctor who sees the electrocardiogram waveform determines that it is inferior wall myocardial infarction… In this way, it is also possible to pre-assign specific numerical values ​​to determine whether the blood flow state of the heart is normal, whether it is ischemic, and where the ischemic state is located.

[0208] If the number of learning source data included in the learning source data group is set to Y, and each integer greater than 1 and less than Y is set to y, then the yth learning source data includes at least biological information of the time series related to the pulsation of the heart of the acquisition object of the yth learning source data, and the heart state information of the heart of the acquisition object of the yth learning source data. The Y learning source data included in the learning source data group may be data respectively acquired from the acquisition object hearts of Y individuals, or may be data acquired from the acquisition object hearts of less than Y individuals. In the case where the learning source data acquired from the acquisition object hearts of less than Y individuals are included in the learning source data group, the learning source data group includes a plurality of learning source data acquired from the acquisition object heart of the same person, but in this case, the plurality of learning source data acquired from the acquisition object heart of the same person may be learning source data acquired at different times, and may also be learning source data with different heart state information.

[0209] The biological information of the time series related to the heart beat included in each learning source data can be a waveform of one cycle or a waveform of multiple cycles. For example, when a doctor observes a waveform of any one cycle in the electrocardiogram waveform to diagnose the state of the heart, the waveform of one cycle is included in the learning source data as the biological information of the time series related to the heart beat, and the information of a diagnosis result obtained based on the waveform of one cycle is included in the learning source data as the heart state information. In addition, for example, when a doctor observes a waveform of multiple cycles in the electrocardiogram waveform to diagnose the state of the heart, the waveform of the multiple cycles is included in the learning source data as the biological information of the time series related to the heart beat, and the information of a diagnosis result obtained based on the waveform of the multiple cycles is included in the learning source data as the heart state information.

[0210] [Signal analysis unit 310] The signal analysis unit 310 is input with the biological information of the time series related to the cardiac pulsation included in each learning source data included in the learning source data group input to the learning device 300. The signal analysis unit 310 performs the same processing as the signal analysis device 1 of the first embodiment, and obtains a group of myocardial activity parameters from the biological information of the time series related to the cardiac pulsation (step S310). The myocardial activity parameters obtained by the signal analysis unit 310 are any one or more of the myocardial activity parameters described in the first embodiment, and are myocardial activity parameters of a type selected in advance as myocardial activity parameters used by the learning unit 320 described later. Hereinafter, the group of myocardial activity parameters obtained by the signal analysis unit 310 will be described as a "myocardial activity parameter group". Sometimes, the myocardial activity parameters obtained by the signal analysis unit 310 are only one type of myocardial activity parameters for one cycle, but it can be understood that "a group of myocardial activity parameters" or "a myocardial activity parameter group" or "each myocardial activity parameter included in the myocardial activity parameter group" can be appropriately replaced with "myocardial activity parameters". The myocardial activity parameter group acquired by the signal analysis unit 310 is output from the signal analysis unit 310 and input to the learning unit 320 .

[0211] When the biological information of the time series related to the cardiac pulsation included in each learning source data is a waveform of one cycle, the signal analysis unit 310 obtains the group formed by the above-mentioned one or more myocardial activity parameters from the waveform of one cycle as the myocardial activity parameter group. When the biological information of the time series related to the cardiac pulsation included in each learning source data is a waveform of multiple cycles, the signal analysis unit 310 obtains the group formed by the above-mentioned one or more myocardial activity parameters from the waveform of each cycle, and obtains the group formed by the myocardial activity parameters of multiple cycles as the myocardial activity parameter group.

[0212] [Learning Department 320] The learning unit 320 is input with a myocardial activity parameter group obtained by the signal analysis unit 310 from biological information of a time series related to the pulsation of the heart of the acquired object included in each learning source data included in the learning source data group input to the learning device 300, and cardiac state information included in each learning source data included in the learning source data group input to the learning device 300. The learning unit 320 uses a group formed by a myocardial activity parameter group and cardiac state information obtained based on the same learning source data as learning data. That is, in each learning data, the myocardial activity parameter group and the cardiac state information are biological information based on the same time series related to the pulsation of the heart of the acquired object. In other words, each myocardial activity parameter group is a myocardial activity parameter group corresponding to each cardiac state information, and each cardiac state information is cardiac state information corresponding to each myocardial activity parameter group.

[0213] If a group consisting of all the learning data used by the learning unit 320 is called a learning data group, then the learning data group used by the learning unit 320 is composed of a plurality of (Y) learning data. The y-th learning data includes at least a myocardial activity parameter group obtained by the signal analysis unit 310 from the biological information of the time series related to the pulsation of the heart of the acquisition object included in the y-th learning source data, and the heart state information indicating the state of the heart of the acquisition object of the y-th learning source data.

[0214] The learning unit 320 uses the learning data group, more specifically, the group formed by the myocardial activity parameter group included in each learning data included in the learning data group and the heart state information corresponding to the myocardial activity parameter group, to learn an estimation model that uses the myocardial activity parameter group as input to obtain the heart state information corresponding to the myocardial activity parameter group, and obtains the learned estimation model (step S320). The learned estimation model obtained by the learning unit 320 is output from the learning unit 320 and is output from the learning device 300. In learning the estimation model, a known learning technique can be used, and the number of learning data can be sufficient as long as it is a number sufficient for learning the estimation model. In other words, the number of learning source data can be the same as the number of learning data.

[0215] [First Example of Learning Device 300] The outline of each data and operation of the learning device 300 will be described by taking as a first example a case where the time-series biological information related to the heart beat included in each learning source data is a waveform corresponding to one cycle.

[0216] If the biological information of the time series related to the heart beat included in the y-th learning source data is set to Wave(y) and the heart state information is set to Info(y), then the y-th learning source data is [Wave(y), Info(y)]. That is, the learning source data group is [[Wave(1), Info(1)], [Wave(2), Info(2)], ..., [Wave(Y), Info(Y)]].

[0217] The signal analysis unit 310 obtains a myocardial activity parameter group Para(y) from the biological information Wave(y) of the time series related to the cardiac pulsation for each y of 1, ..., Y. Each myocardial activity parameter group Para(y) is a group formed by one or more myocardial activity parameters obtained from the biological information Wave(y) of the time series related to the cardiac pulsation by the same processing as that performed by the signal analysis device 1 of the first embodiment, and therefore has one or more elements.

[0218] Since the learning unit 320 uses the group formed by the myocardial activity parameter group and the heart state information obtained based on the same learning source data as learning data, the y-th learning data is [Para(y), Info(y)]. That is, the learning data group is [[Para(1), Info(1)], [Para(2), Info(2)], ..., [Para(Y), Info(Y)]]. The learning unit 320 uses [Para(1), Info(1)], [Para(2), Info(2)], ..., [Para(Y), Info(Y)] to learn an estimation model that uses the myocardial activity parameter group as an input to obtain the heart state information corresponding to the myocardial activity parameter group, and obtains a learned estimation model.

[0219] [Second Example of Learning Device 300] As a second example, the case where the biological information of the time series related to the heart beat included in each learning source data is a waveform of the amount of multiple cycles will be described to outline each data and operation of the learning device 300. In the second example, it is assumed that the biological information of the time series related to the heart beat included in each learning source data is a waveform of the amount of Z cycles, and each of the indexes 1, ..., Z of the cycle is set to z. That is, Z is an integer greater than 2, and z is each integer greater than 1 and less than Z.

[0220] If the waveform of each cycle included in the biological information of the time series related to the heart beat included in the yth learning source data is set to Wave(y, z), and the heart state information included in the yth learning source data is set to Info(y), then the yth learning source data is [Wave(y, 1), Wave(y, 2),…, Wave(y, Z), Info(y)].

[0221] The signal analysis unit 310 obtains one or more myocardial activity parameters Para(y, z) from each Wave(y, z) for each y of 1, ..., Y and each z of 1, ..., Z, thereby obtaining a myocardial activity parameter group [Para(y, 1), Para(y, 2), ..., Para(y, Z)] for each y of 1, ..., Y. Each myocardial activity parameter Para(y, z) is one or more myocardial activity parameters obtained from the waveform Wave(y, z) of each cycle included in the biological information of the time series related to the cardiac pulsation by the same processing as that performed by the signal analysis device 1 of the first embodiment, and therefore has one or more elements.

[0222] Since the learning unit 320 uses the group formed by the myocardial activity parameter group and the heart state information obtained based on the same learning source data as learning data, an example of the y-th learning data is [Para(y, 1), Para(y, 2), ..., Para(y, Z), Info(y)]. That is, the learning data group is [[Para(1, 1), Para(1, 2), ..., Para(1, Z), Info(1)], [Para(2, 1), Para(2, 2), ..., Para(2, Z), Info(2)], ..., [Para(Y, 1), Para(Y, 2), ..., Para(Y, Z), Info(Y)]. The learning unit 320 uses [Para(1,1), Para(1,2), ..., Para(1,Z), Info(1)], [Para(2,1), Para(2,2), ..., Para(2,Z), Info(2)], ..., [Para(Y,1), Para(Y,2), ..., Para(Y,Z), Info(Y)] to learn an estimation model that uses a myocardial activity parameter group as input to obtain cardiac state information corresponding to the myocardial activity parameter group, obtains the learned estimation model and outputs it.

[0223] Next, the information providing device 400 will be described. [Information providing device 400] The information providing device 400 is input with biological information of a time series related to the pulsation of the heart to which the information is provided (hereinafter referred to as the "information providing target heart"). The information providing device 400 obtains myocardial activity parameters from the biological information of a time series related to the pulsation of the information providing target heart, uses the learned estimation model, takes the obtained myocardial activity parameters as input, obtains information indicating the state of the heart corresponding to the myocardial activity parameters, i.e., heart state information, and outputs the information. Fig.60 As shown in FIG. 4 , the information providing device 400 includes a signal analyzing unit 410 and a status information generating unit 420. Fig.62 The processing of steps S410 and S420 is shown.

[0224] [Signal analysis unit 410] The time series biological information related to the pulsation of the heart of the information providing object input to the information providing device 400 is input to the signal analyzing unit 410. The signal analyzing unit 410 obtains a myocardial activity parameter group from the time series biological information related to the pulsation of the heart of the information providing object by performing the same processing as the signal analyzing unit 310 of the learning device 300, that is, by performing the same processing as the signal analyzing device 1 of the first embodiment (step S410). The myocardial activity parameter group obtained by the signal analyzing unit 410 is output from the signal analyzing unit 410 and input to the state information generating unit 420. The myocardial activity parameter group obtained by the signal analyzing unit 410 is a group formed by any one or more myocardial activity parameters described in the first embodiment. The one or more myocardial activity parameters are myocardial activity parameters of a type selected in advance as myocardial activity parameters used by the state information generating unit 420 described later, and are myocardial activity parameters of a type used by the learning unit 320 of the second embodiment in learning the estimation model.

[0225] When the input biological information of the time series related to the pulsation of the heart of the information providing object is a waveform of one cycle, the signal analysis unit 410 obtains the group formed by the above-mentioned one or more myocardial activity parameters from the waveform of one cycle as the myocardial activity parameter group. When the input biological information of the time series related to the pulsation of the heart of the information providing object is a waveform of multiple cycles, the signal analysis unit 410 obtains the group formed by the above-mentioned one or more myocardial activity parameters from the waveform of each cycle, and obtains the group formed by the one or more myocardial activity parameters of multiple cycles as the myocardial activity parameter group.

[0226] [Status Information Generating Unit 420] like Fig.60 As shown, the state information generating unit 420 includes a model storage unit 425. The learned estimation model output by the learning device 300 of the second embodiment is pre-stored in the model storage unit 425. The learned estimation model pre-stored in the model storage unit 425 is an estimation model that receives a myocardial activity parameter group as an input and obtains information representing the state of the heart corresponding to the myocardial activity parameter group, that is, heart state information.

[0227] The myocardial activity parameter group of the heart of the information providing target outputted by the signal analysis unit 410 is inputted to the state information generating unit 420. The state information generating unit 420 uses the learned estimation model pre-stored in the model storage unit 425 and takes the myocardial activity parameter group of the heart of the information providing target as input, and obtains the heart state information, which is information indicating the state of the heart of the information providing target (step S420). The heart state information obtained by the state information generating unit 420 is outputted from the state information generating unit 420 as the output of the information providing device 400.

[0228] [First Example of Information Providing Device 400] Taking as a first example a case where the time-series biological information related to the pulsation of the heart of the information provision target input into the information provision apparatus 400 is a waveform corresponding to one cycle, an overview of each data and operation of the information provision apparatus 400 will be described.

[0229] If the biological information of the time series related to the pulsation of the heart of the information providing object input to the information providing device 400 is WAVE, the signal analyzing unit 410 obtains the myocardial activity parameter group PARA from the biological information WAVE of the time series related to the pulsation of the heart of the information providing object. The myocardial activity parameter group PARA is a group formed by one or more myocardial activity parameters obtained from the biological information WAVE of the time series related to the pulsation of the heart of the information providing object by the same processing as that performed by the signal analyzing device 1 of the first embodiment, and therefore has one or more elements.

[0230] The state information generating unit 420 obtains one piece of heart state information INFO corresponding to the myocardial activity parameter group PARA from the myocardial activity parameter group PARA using the learned estimation model stored in advance in the model storage unit 425 .

[0231] [Second Example of Information Providing Device 400] The case where the biological information of the time series related to the pulsation of the heart of the information provision object input to the information provision device 400 is a waveform of a plurality of cycles is described as a second example, and the outline of each data and operation of the information provision device 400 is described. In the second example, it is assumed that the biological information of the time series related to the pulsation of the heart of the information provision object is a waveform of Z cycles, and each of the indexes 1, ..., Z of the cycle is set to z. That is, Z is an integer greater than 2, and z is each integer greater than 1 and less than Z.

[0232] If the waveform of each cycle included in the biological information of the time series related to the pulsation of the heart of the information provision object input to the information provision device 400 is set to WAVE(z), the signal analysis unit 410 obtains one or more myocardial activity parameters PARA(z) from each WAVE(z) for each z of 1, ..., Z, thereby obtaining a myocardial activity parameter group [Para(1), Para(2), ..., Para(Z)]. The myocardial activity parameter PARA(z) is one or more myocardial activity parameters obtained from the waveform WAVE(z) of each cycle included in the biological information of the time series related to the pulsation of the heart of the information provision object by the same processing as that performed by the signal analysis device 1 of the first embodiment, and therefore has one or more elements.

[0233] The state information generating unit 420 uses the learned estimation model pre-stored in the model storage unit 425 to obtain one heart state information INFO corresponding to the myocardial activity parameter group [Para(1), Para(2), ..., Para(Z)] from the myocardial activity parameter group [Para(1), Para(2), ..., Para(Z)].

[0234] The heart state information obtained by the state information generating unit 420 is the result of estimation of the state of the heart of the information providing target by the information providing apparatus 400. That is, for example, the heart state information obtained by the state information generating unit 420 indicates the estimation result of whether the heart of the information providing target is in a normal state or what disease the heart is suffering from by using predetermined numerical values ​​assigned in advance, such as "0" corresponding to the normal state of the heart of the information providing target, "1" corresponding to Buddah syndrome, "2" corresponding to ischemic heart disease, "3" corresponding to early repolarization syndrome, and so on.

[0235] Alternatively, for example, the heart status information obtained by the status information generating unit 420 uses a pre-assigned specific numerical value to represent an estimated result of whether the heart rhythm of the heart of the information providing object is normal and what kind of arrhythmia it is in, such as "0" corresponding to normal sinus rhythm of the heart of the information providing object, "1" corresponding to supraventricular extrasystole, "2" corresponding to atrial fibrillation, "3" corresponding to ventricular extrasystole, "4" corresponding to ventricular fibrillation... and so on.

[0236] Alternatively, for example, the heart state information obtained by the state information generating unit 420 uses a pre-assigned fixed numerical value to represent an estimated result of whether the blood flow state of the heart of the information providing object is a normal state or an ischemic state, and where the ischemic state is located, such as "0" corresponding to a normal coronary blood flow state of the heart of the information providing object, "1" corresponding to an ischemic state of the inner layer of the myocardium, "2" corresponding to an ischemic state of the inner and outer layers of the myocardium (a state of transmural ischemia), "3" corresponding to anterior wall myocardial infarction, "4" corresponding to inferior wall myocardial infarction, and so on.

[0237] <Third Embodiment> The second embodiment is to obtain a myocardial activity parameter group from biological information of a time series related to cardiac pulsation and then perform learning of an estimation model or estimation of cardiac state information, but it is also possible to use a myocardial activity parameter group obtained in advance to perform learning of an estimation model or estimation of cardiac state information. This method is described as the third embodiment, focusing on the differences from the second embodiment.

[0238] like Fig.63As shown in FIG. 1 , the information providing system 201 of the third embodiment includes a learning device 301 and an information providing device 401. The hardware structures of the information providing system 201, the learning device 301 and the information providing device 401 are as follows. Figure 1 The hardware structure of the signal analysis devices 1 shown is the same.

[0239] First, the learning device 301 will be described. [Learning device 301] like Fig.63 As shown, the learning device 301 includes a learning unit 320. The learning device 301 performs Fig.64 The process of step S320 is shown.

[0240] [Learning data set, learning data] A learning data group is input to the learning device 300. The learning data group includes a plurality of learning data. Each learning data includes at least a heart that is the object of acquisition of the learning data, i.e., a myocardial activity parameter group of the object heart, and information indicating the state of the object heart, i.e., heart state information. Each learning data may also include information for identifying the learning data, such as an identification number uniquely assigned to the learning data.

[0241] If the number of learning data included in the learning data group is set to Y, and each integer greater than 1 and less than Y is set to y, then the Yth learning data includes at least a myocardial activity parameter group of the heart of the object for acquiring the yth learning data, and information representing the state of the heart of the object for acquiring the yth learning data, namely, heart state information.

[0242] The myocardial activity parameter group of the acquired target heart is obtained from biological information of a time series related to the pulsation of the acquired target heart by the same processing as that performed by the signal analysis unit 310 of the learning device 300 of the second embodiment, that is, by the same processing as that performed by the signal analysis device 1 of the first embodiment. The cardiac state information of the acquired target heart is the same as that of the acquired target heart in the second embodiment.

[0243] [Learning Department 320] The learning data group input to the learning device 301 is input to the learning unit 320. That is, the myocardial activity parameter group included in each learning data included in the learning data group input to the learning device 301, and the heart state information included in each learning data included in the learning data group input to the learning device 301 are input to the learning unit 320. The learning unit 320 uses the group formed by the myocardial activity parameter group and the heart state information included in the same learning data as learning data. That is, in each learning data, the myocardial activity parameter group and the heart state information are biological information based on the same time series related to the heart beat. In other words, the myocardial activity parameter group is a myocardial activity parameter group corresponding to the heart state information, and the heart state information is the heart state information corresponding to the myocardial activity parameter group.

[0244] The learning unit 320, like the learning unit 320 of the second embodiment, uses the learning data group, more specifically, uses the group formed by the myocardial activity parameter group included in each learning data included in the learning data group and the heart state information corresponding to the myocardial activity parameter group, to learn an estimation model that uses the myocardial activity parameter group as an input to obtain the heart state information corresponding to the myocardial activity parameter group, and obtains the learned estimation model (step S320). The learned estimation model obtained by the learning unit 320 is output from the learning unit 320 as the output of the learning device 300. In learning the estimation model, a known learning technique can be used, and the number of learning data can be any number sufficient for learning the estimation model. The specific example of the learning unit 320 is as described in the second embodiment.

[0245] Next, the information providing device 401 will be described. [Information providing device 401] The myocardial activity parameter group of the heart to be provided as the information is input to the information providing device 401. The myocardial activity parameter group of the heart to be provided as the information is obtained from biological information of a time series related to the pulsation of the heart to be provided as the information is provided by the same processing as that performed by the signal analysis unit 410 of the information providing device 400 of the second embodiment, that is, by the same processing as that performed by the signal analysis device 1 of the first embodiment. The information providing device 401 uses the learned estimation model to obtain information indicating the state of the heart corresponding to the myocardial activity parameter group from the input myocardial activity parameter group, that is, heart state information, and outputs it. Fig.63 As shown, the information providing device 401 includes a status information generating unit 420. The information providing device 401 performs Fig.65 The processing of step S420 is shown.

[0246] [Status Information Generating Unit 420] like Fig.63As shown, the state information generating unit 420 includes a model storage unit 425. The learned estimation model output by the learning device 301 of the third embodiment is pre-stored in the model storage unit 425. The learned estimation model pre-stored in the model storage unit 425 is an estimation model that takes a myocardial activity parameter group as an input to obtain information representing the state of the heart corresponding to the myocardial activity parameter group, that is, heart state information.

[0247] The state information generating unit 420 inputs the myocardial activity parameter group of the heart of the information providing object input to the information providing device 401. The state information generating unit 420 uses the learned estimation model pre-stored in the model storage unit 425 and takes the myocardial activity parameter group of the heart of the information providing object as input to obtain information representing the state of the heart of the information providing object, i.e., heart state information (step S420). The heart state information obtained by the state information generating unit 420 is output from the state information generating unit 420 as an output of the information providing device 401. A specific example of the state information generating unit 420 is as described in the second embodiment. The heart state information obtained by the state information generating unit 420 is the same as the heart state information obtained by the state information generating unit 420 of the second embodiment, and is an estimation result of the state representing the state of the heart of the information providing object by the information providing device 401, and a specific example is as described in the second embodiment.

[0248] <Modifications of the Second and Third Embodiments> As can be seen from the description of the third embodiment, the estimated model obtained by the learning device 300 of the second embodiment and the estimated model obtained by the learning device 301 of the third embodiment are equivalent. Therefore, the learned estimated model output by the learning device 301 of the third embodiment can be pre-stored in the model storage unit 425 of the information providing device 400 of the second embodiment, and the learned estimated model output by the learning device 300 of the learning device of the second embodiment can also be pre-stored in the model storage unit 425 of the information providing device 401 of the third embodiment. That is, the information providing system can be as follows Fig.66 The information providing system 202 including the learning device 301 of the third embodiment and the information providing device 400 of the second embodiment as shown in FIG. Fig.67 As shown in FIG. 1 , the information providing system 203 is configured to include the learning device 300 of the second embodiment and the information providing device 401 of the third embodiment. The hardware configurations of the information providing system 202, the information providing system 203, the learning device 300, the learning device 301, the information providing device 400, and the information providing device 401 are similar to the hardware configurations of the information providing system 202, the information providing system 203, the learning device 300, the learning device 301, the information providing device 400, and the information providing device 401. Figure 1 The hardware structure of the signal analysis device 1 shown is the same.

[0249] <Summary of the Second and Third Embodiments and Their Modifications> [Learning device] The learning device 300 of the second embodiment and the learning device 301 of the third embodiment both assume that the learning data group is composed of Y (Y is a plurality of) learning data, each integer greater than 1 and less than Y is set to y, each of the Y learning data includes a myocardial activity parameter group of the heart as the object of the y-th learning data, and heart state information indicating the state of the heart as the object of the y-th learning data, and both include a learning unit 320, which uses the learning data group to learn an estimation model, and the estimation model uses the myocardial activity parameter group as input to obtain information indicating the state of the heart corresponding to the myocardial activity parameter group, that is, heart state information. In addition, the learning device 300 of the second embodiment further includes a signal analysis unit 310, which uses biological information of a time series related to the pulsation of the heart as the object of each y-th learning data as input to obtain a myocardial activity parameter group. However, in the myocardial activity parameter group of the heart that is the object of the y-th learning data in the learning device 300 of the second embodiment and the learning device 301 of the third embodiment, there is included one or more predetermined types of myocardial activity parameters among a plurality of myocardial activity parameters included in the first group to be described later, a plurality of myocardial activity parameters included in the second group to be described later, a plurality of myocardial activity parameters included in the third group to be described later, and one or more myocardial activity parameters included in the fourth group to be described later, for each cycle of the time series of biological information related to the beating of the heart that is the object of the y-th learning data.

[0250] [Information providing device] The information providing device 400 of the second embodiment and the information providing device 401 of the third embodiment both include a state information generating unit 420, which pre-stores an estimation model that uses a myocardial activity parameter group as input to obtain information indicating the state of the heart corresponding to the myocardial activity parameter group, namely, heart state information, and uses the estimation model to use the myocardial activity parameter group of the heart of the information providing object as input to obtain information indicating the state of the heart of the information providing object, namely, heart state information. In addition, the information providing device 400 of the second embodiment also includes a signal analyzing unit 410, which uses a time series of biological information related to the pulsation of the heart of the information providing object as input to obtain a myocardial activity parameter group. However, the myocardial activity parameter group of the heart of the information providing object in the information providing device 400 of the second embodiment and the information providing device 401 of the third embodiment includes one or more predetermined types of myocardial activity parameters of each cycle of the biological information of the time series related to the beating of the heart of the information providing object, including a plurality of myocardial activity parameters included in the first group described later, a plurality of myocardial activity parameters included in the second group described later, a plurality of myocardial activity parameters included in the third group described later, and one or more myocardial activity parameters included in the fourth group described later.

[0251] [Various myocardial activity parameters included in the first group] The multiple myocardial activity parameters included in the first group refer to: when the waveform of the time interval of the R wave included in the waveform of one cycle of the biological information of the time series related to the cardiac pulsation is taken as the first object time waveform, a time waveform obtained by the difference or weighted difference between the first cumulative distribution function as the cumulative distribution function of the first unimodal distribution and the second cumulative distribution function as the cumulative distribution function of the second unimodal distribution, that is, a first approximate time waveform; or when the first object time waveform is approximated by the first approximate time waveform obtained by adding the difference or weighted difference between the first cumulative distribution function as the cumulative distribution function of the first unimodal distribution and the second cumulative distribution function as the cumulative distribution function of the second unimodal distribution, that is, a first approximate time waveform, determining the parameters of the first unimodal distribution, determining the parameters of the first cumulative distribution function, determining the parameters of the second unimodal distribution, determining the parameters of the second cumulative distribution function, the weight of the first cumulative distribution function, the weight of the second cumulative distribution function, the ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, and the first level value.

[0252] [Various myocardial activity parameters included in the second group] The plurality of myocardial activity parameters included in the second group refer to, when the waveform of the time interval of the T wave included in the waveform of one cycle of the biological information of the time series related to the cardiac pulsation is taken as the second target time waveform and the waveform after the time axis of the second target time waveform is reversed is taken as the second target inverse time waveform, a waveform obtained by the difference or weighted difference between the third cumulative distribution function as the cumulative distribution function of the third unimodal distribution and the fourth cumulative distribution function as the cumulative distribution function of the fourth unimodal distribution, that is, a second approximate inverse time waveform, or a waveform obtained by adding the difference or weighted difference between the third cumulative distribution function as the cumulative distribution function of the third unimodal distribution and the fourth cumulative distribution function as the cumulative distribution function of the fourth unimodal distribution plus the second level value to approximate the second target inverse time waveform, determining parameters of the third unimodal distribution, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, weight of the third cumulative distribution function, weight of the fourth cumulative distribution function, ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value.

[0253] Alternatively, the plurality of myocardial activity parameters included in the second group refer to the case where a waveform of a time interval of a T wave included in a waveform of one cycle of biological information of a time series related to cardiac pulsation is used as the second object time waveform, a third unimodal distribution cumulative distribution function is used as the third cumulative distribution function, a fourth unimodal distribution cumulative distribution function is used as the fourth cumulative distribution function, a function obtained by subtracting the third cumulative distribution function from 1 is used as the third inverse cumulative distribution function, and a function obtained by subtracting the fourth cumulative distribution function from 1 is used as the fourth inverse cumulative distribution function, and the third inverse cumulative distribution function and the fourth inverse cumulative distribution function are used as the fourth inverse cumulative distribution function. When the second object time waveform is approximated by the waveform obtained by the difference or weighted difference of the integral distribution function, that is, the second approximate time waveform, or the time waveform obtained by adding the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function plus the second level value, that is, the second approximate time waveform, determine the parameters of the third unimodal distribution, determine the parameters of the third cumulative distribution function, determine the parameters of the fourth unimodal distribution, determine the parameters of the fourth cumulative distribution function, the weight of the third inverse cumulative distribution function, the weight of the fourth inverse cumulative distribution function, the ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value.

[0254] [Various myocardial activity parameters included in the third group] The multiple myocardial activity parameters included in the third group refer to, when any one of the time waveform of the difference between the first object time waveform and the first approximate time waveform, the time waveform of the difference between the second object time waveform and the second approximate time waveform, and the time waveform after the time axis of the waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform is inverted as the residual time waveform, and the residual time waveform is approximated by the fifth cumulative distribution function which is the cumulative distribution function of the fifth unimodal distribution or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, that is, an approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function is weighted.

[0255] Alternatively, the multiple myocardial activity parameters included in the third group refer to, when at least any one of the time waveform of the difference between the first object time waveform and the first approximate time waveform, the time waveform of the difference between the second object time waveform and the second approximate time waveform, and the time waveform after the time axis of the waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform is reversed as the residual time waveform, and the residual time waveform is approximated by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function as the cumulative distribution function of the fifth unimodal distribution and the sixth cumulative distribution function as the cumulative distribution function of the sixth unimodal distribution, determining the parameters of the fifth cumulative distribution function, determining the parameters of the sixth unimodal distribution, determining the parameters of the sixth cumulative distribution function, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function, using at least one of the time waveforms of the difference between the first object time waveform and the first approximate time waveform, the time waveform of the difference between the second object time waveform and the second approximate time waveform after the time axis is reversed as the residual time waveform, and approximating the residual time waveform by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function as the cumulative distribution function of the fifth unimodal distribution and the sixth cumulative distribution function as the cumulative distribution function of the sixth unimodal distribution.

[0256] [One or more myocardial activity parameters included in the fourth group] The one or more myocardial activity parameters included in the fourth group are myocardial activity parameters obtained by calculating multiple myocardial activity parameters included in the first group, multiple myocardial activity parameters included in the second group, and multiple myocardial activity parameters included in the third group.

[0257] In addition, all or part of the functions of the signal analysis device 1 and / or the information providing systems 200, 201, 202, 203 and / or the learning devices 300, 301 and / or the information providing devices 400, 401 may also be implemented using hardware such as ASIC (Application Specific Integrated Circuit), PLD (Programmable Logic Device), FPGA (Field Programmable Gate Array). The program may be recorded in a computer-readable recording medium. The computer-readable recording medium is a storage device such as a removable medium such as a floppy disk, a magneto-optical disk, a ROM, a CD-ROM, or a hard disk built into a computer system. The program may also be sent via a telecommunications line.

[0258] As mentioned above, although the embodiment of the present invention is described in detail with reference to the drawings, the specific structure is not limited to this embodiment, and also includes the design etc. within the scope which does not deviate from the summary of the present invention.

[0259] Description of Reference Numerals 1…signal analysis device, 11…control unit, 12…input unit, 13…communication unit, 14…storage unit, 15…output unit, 91…processor, 92…memory, 110…electrocardiogram acquisition unit, 120…fitting information acquisition unit, 130…analysis unit, 131…fitting unit, 132…myocardial activity information parameter acquisition unit, 140…recording unit, 200, 201, 202, 203…information providing system, 300, 301…learning device, 310…signal analysis unit, 320…learning unit, 400, 401…information providing device, 410…signal analysis unit, 420…state information generation unit, 425…model storage unit.

Claims

1. A learning device, wherein: It includes a learning department, which performs the following operations: The waveform of the time interval of the R wave included in the waveform representing one cycle of the cardiac cycle of the heart is taken as the first target time waveform. The waveform of the time interval of the T wave included in the waveform is used as the second target time waveform, The second object time waveform is obtained by inverting the time axis as a second object inverse time waveform. The cumulative distribution function of the first unimodal distribution is used as the first cumulative distribution function, and the cumulative distribution function of the second unimodal distribution is used as the second cumulative distribution function, The following contents are included as multiple myocardial activity parameters in the first group: when the first object time waveform is approximated by a time waveform obtained by a difference or weighted difference between the first cumulative distribution function and the second cumulative distribution function, that is, a first approximate time waveform, or a time waveform obtained by adding a difference or weighted difference between the first cumulative distribution function and the second cumulative distribution function and a first level value, that is, a first approximate time waveform, determining parameters of the first unimodal distribution, determining parameters of the first cumulative distribution function, determining parameters of the second unimodal distribution, determining parameters of the second cumulative distribution function, a weight of the first cumulative distribution function, a weight of the second cumulative distribution function, a ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, and the first level value, using the cumulative distribution function of the third unimodal distribution as the third cumulative distribution function, using the cumulative distribution function of the fourth unimodal distribution as the fourth cumulative distribution function, subtracting the third cumulative distribution function from 1 as the third inverse cumulative distribution function, and subtracting the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function, The following contents are used as multiple myocardial activity parameters included in the second group: when the second object inverse time waveform is approximated by a waveform obtained by the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function, that is, a second approximate inverse time waveform, or a waveform obtained by adding the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function plus a second level value, that is, a second approximate inverse time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, the weight of the third cumulative distribution function, the weight of the fourth cumulative distribution function, the ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value; or, a time waveform obtained by the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, or a time waveform obtained by adding the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function plus a second level value. when the second object time waveform is approximated by a time waveform obtained by adding a second level value to the difference or weighted difference of the functions, that is, a second approximate time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, a weight of the third inverse cumulative distribution function, a weight of the fourth inverse cumulative distribution function, a ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value, taking any one of the time waveforms of the difference between the first object time waveform and the first approximate time waveform, the time waveform of the difference between the second object time waveform and the second approximate time waveform, and the time waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform after the time axis is reversed as the residual time waveform, taking the cumulative distribution function of the fifth unimodal distribution as the fifth cumulative distribution function, and taking the cumulative distribution function of the sixth unimodal distribution as the sixth cumulative distribution function, The following contents are used as multiple myocardial activity parameters included in the third group: when the residual time waveform is approximated by the fifth cumulative distribution function or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, i.e., an approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function is determined; or, when the residual time waveform is approximated by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, the parameters of the sixth unimodal distribution are determined, the parameters of the sixth cumulative distribution function, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function, using parameters obtained by calculating the multiple myocardial activity parameters included in the first group, the multiple myocardial activity parameters included in the second group, and the multiple myocardial activity parameters included in the third group as one or more myocardial activity parameters included in the fourth group, a group formed by one or more predetermined myocardial activity parameters among the plurality of myocardial activity parameters included in the first group, the plurality of myocardial activity parameters included in the second group, the plurality of myocardial activity parameters included in the third group, and one or more myocardial activity parameters included in the fourth group, obtained from waveforms showing more than one cycle of the cardiac cycle of the heart, as a myocardial activity parameter group; The learning data set consists of Y (Y is a plurality of) learning data. Let each integer greater than 1 and less than Y be y, and each of the Y learning data includes a myocardial activity parameter group of a heart that is the object of the y-th learning data, and information indicating the state of the heart that is the object of the y-th learning data, i.e., heart state information, The learning data group is used to learn an estimation model that receives the myocardial activity parameter group as an input and obtains heart state information that is information indicating the state of the heart corresponding to the myocardial activity parameters.

2. The learning device according to claim 1, wherein: The system further includes a signal analyzing unit that obtains the myocardial activity parameter group based on each of waveforms representing one or more cardiac cycles of the heart that is the target of the Y learning data.

3. The learning device according to claim 1 or 2, wherein: The unimodal distributions are Gaussian distributions.

4. An information providing device, wherein: It includes a status information generating unit, which performs the following operations: The waveform of the time interval of the R wave included in the waveform representing one cycle of the cardiac cycle of the heart is taken as the first target time waveform. The waveform of the time interval of the T wave included in the waveform is used as the second target time waveform, The second object time waveform is obtained by inverting the time axis as a second object inverse time waveform. The cumulative distribution function of the first unimodal distribution is used as the first cumulative distribution function, and the cumulative distribution function of the second unimodal distribution is used as the second cumulative distribution function, The following contents are included as multiple myocardial activity parameters in the first group: when the first object time waveform is approximated by a time waveform obtained by a difference or weighted difference between the first cumulative distribution function and the second cumulative distribution function, that is, a first approximate time waveform, or a time waveform obtained by adding a difference or weighted difference between the first cumulative distribution function and the second cumulative distribution function and a first level value, that is, a first approximate time waveform, determining parameters of the first unimodal distribution, determining parameters of the first cumulative distribution function, determining parameters of the second unimodal distribution, determining parameters of the second cumulative distribution function, a weight of the first cumulative distribution function, a weight of the second cumulative distribution function, a ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, and the first level value, using the cumulative distribution function of the third unimodal distribution as the third cumulative distribution function, using the cumulative distribution function of the fourth unimodal distribution as the fourth cumulative distribution function, subtracting the third cumulative distribution function from 1 as the third inverse cumulative distribution function, and subtracting the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function, The following contents are used as multiple myocardial activity parameters included in the second group: when the second object inverse time waveform is approximated by a waveform obtained by the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function, that is, a second approximate inverse time waveform, or a waveform obtained by adding the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function plus a second level value, that is, a second approximate inverse time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, the weight of the third cumulative distribution function, the weight of the fourth cumulative distribution function, the ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value; or, a time waveform obtained by the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, or a time waveform obtained by adding the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function plus a second level value. when the second object time waveform is approximated by a time waveform obtained by adding a second level value to the difference or weighted difference of the functions, that is, a second approximate time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, a weight of the third inverse cumulative distribution function, a weight of the fourth inverse cumulative distribution function, a ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value, taking any one of the time waveforms of the difference between the first object time waveform and the first approximate time waveform, the time waveform of the difference between the second object time waveform and the second approximate time waveform, and the time waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform after the time axis is reversed as the residual time waveform, taking the cumulative distribution function of the fifth unimodal distribution as the fifth cumulative distribution function, and taking the cumulative distribution function of the sixth unimodal distribution as the sixth cumulative distribution function, The following contents are used as multiple myocardial activity parameters included in the third group: when the residual time waveform is approximated by the fifth cumulative distribution function or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, i.e., an approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function is determined; or, when the residual time waveform is approximated by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, the parameters of the sixth unimodal distribution are determined, the parameters of the sixth cumulative distribution function, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function, using parameters obtained by calculating the multiple myocardial activity parameters included in the first group, the multiple myocardial activity parameters included in the second group, and the multiple myocardial activity parameters included in the third group as one or more myocardial activity parameters included in the fourth group, a group formed by one or more predetermined myocardial activity parameters among the plurality of myocardial activity parameters included in the first group, the plurality of myocardial activity parameters included in the second group, the plurality of myocardial activity parameters included in the third group, and one or more myocardial activity parameters included in the fourth group, obtained from waveforms showing more than one cycle of the cardiac cycle of the heart, as a myocardial activity parameter group; An estimation model is prestored that receives a myocardial activity parameter group as an input and obtains heart state information corresponding to the myocardial activity parameter group, that is, information indicating the heart state. The estimation model is used to obtain cardiac state information of the target heart, by taking as input a group of myocardial activity parameters of the target heart, which is a heart to be provided with information.

5. The information providing device according to claim 4, wherein: The system further includes a signal analyzing unit that obtains the myocardial activity parameter group based on a waveform representing one or more cardiac cycles of the heart of the information providing target.

6. The information providing device according to claim 4 or 5, wherein: The unimodal distributions are Gaussian distributions.

7. A learning method performed by a learning device, wherein: The method comprises a learning step of performing the following operations: taking a waveform of a time interval of an R wave included in a waveform representing one cycle of a cardiac cycle of a heart as a first target time waveform, The waveform of the time interval of the T wave included in the waveform is used as the second target time waveform, The second object time waveform is obtained by inverting the time axis as a second object inverse time waveform. The cumulative distribution function of the first unimodal distribution is used as the first cumulative distribution function, and the cumulative distribution function of the second unimodal distribution is used as the second cumulative distribution function, The following contents are included as multiple myocardial activity parameters in the first group: when the first object time waveform is approximated by a time waveform obtained by a difference or weighted difference between the first cumulative distribution function and the second cumulative distribution function, that is, a first approximate time waveform, or a time waveform obtained by adding a difference or weighted difference between the first cumulative distribution function and the second cumulative distribution function and a first level value, that is, a first approximate time waveform, determining parameters of the first unimodal distribution, determining parameters of the first cumulative distribution function, determining parameters of the second unimodal distribution, determining parameters of the second cumulative distribution function, a weight of the first cumulative distribution function, a weight of the second cumulative distribution function, a ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, and the first level value, using the cumulative distribution function of the third unimodal distribution as the third cumulative distribution function, using the cumulative distribution function of the fourth unimodal distribution as the fourth cumulative distribution function, subtracting the third cumulative distribution function from 1 as the third inverse cumulative distribution function, and subtracting the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function, The following contents are used as multiple myocardial activity parameters included in the second group: when the second object inverse time waveform is approximated by a waveform obtained by the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function, that is, a second approximate inverse time waveform, or a waveform obtained by adding the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function plus a second level value, that is, a second approximate inverse time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, the weight of the third cumulative distribution function, the weight of the fourth cumulative distribution function, the ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value; or, a time waveform obtained by the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, or a time waveform obtained by adding the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function plus a second level value. when the second object time waveform is approximated by a time waveform obtained by adding a second level value to the difference or weighted difference of the functions, that is, a second approximate time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, a weight of the third inverse cumulative distribution function, a weight of the fourth inverse cumulative distribution function, a ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value, taking any one of the time waveforms of the difference between the first object time waveform and the first approximate time waveform, the time waveform of the difference between the second object time waveform and the second approximate time waveform, and the time waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform after the time axis is reversed as the residual time waveform, taking the cumulative distribution function of the fifth unimodal distribution as the fifth cumulative distribution function, and taking the cumulative distribution function of the sixth unimodal distribution as the sixth cumulative distribution function, The following contents are used as multiple myocardial activity parameters included in the third group: when the residual time waveform is approximated by the fifth cumulative distribution function or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, i.e., an approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function is determined; or, when the residual time waveform is approximated by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, the parameters of the sixth unimodal distribution are determined, the parameters of the sixth cumulative distribution function, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function, using parameters obtained by calculating the multiple myocardial activity parameters included in the first group, the multiple myocardial activity parameters included in the second group, and the multiple myocardial activity parameters included in the third group as one or more myocardial activity parameters included in the fourth group, a group formed by one or more predetermined myocardial activity parameters among the plurality of myocardial activity parameters included in the first group, the plurality of myocardial activity parameters included in the second group, the plurality of myocardial activity parameters included in the third group, and one or more myocardial activity parameters included in the fourth group, obtained from waveforms showing more than one cycle of the cardiac cycle of the heart, as a myocardial activity parameter group; The learning data set consists of Y (Y is a plurality of) learning data. Let each integer greater than 1 and less than Y be y, and each of the Y learning data includes a myocardial activity parameter group of a heart that is the object of the y-th learning data, and information indicating the state of the heart that is the object of the y-th learning data, i.e., heart state information, The learning data group is used to learn an estimation model that receives the myocardial activity parameter group as an input and obtains heart state information that is information indicating the state of the heart corresponding to the myocardial activity parameters.

8. An information providing method performed by an information providing device, wherein: The process includes a state information generation step, which performs the following operations: The waveform of the time interval of the R wave included in the waveform representing one cycle of the cardiac cycle of the heart is taken as the first target time waveform. The waveform of the time interval of the T wave included in the waveform is used as the second target time waveform, The second object time waveform is obtained by inverting the time axis as a second object inverse time waveform. The cumulative distribution function of the first unimodal distribution is used as the first cumulative distribution function, and the cumulative distribution function of the second unimodal distribution is used as the second cumulative distribution function, The following contents are included as multiple myocardial activity parameters in the first group: when the first object time waveform is approximated by a time waveform obtained by a difference or weighted difference between the first cumulative distribution function and the second cumulative distribution function, that is, a first approximate time waveform, or a time waveform obtained by adding a difference or weighted difference between the first cumulative distribution function and the second cumulative distribution function and a first level value, that is, a first approximate time waveform, determining parameters of the first unimodal distribution, determining parameters of the first cumulative distribution function, determining parameters of the second unimodal distribution, determining parameters of the second cumulative distribution function, a weight of the first cumulative distribution function, a weight of the second cumulative distribution function, a ratio of the weight of the first cumulative distribution function to the weight of the second cumulative distribution function, and the first level value, using the cumulative distribution function of the third unimodal distribution as the third cumulative distribution function, using the cumulative distribution function of the fourth unimodal distribution as the fourth cumulative distribution function, subtracting the third cumulative distribution function from 1 as the third inverse cumulative distribution function, and subtracting the fourth cumulative distribution function from 1 as the fourth inverse cumulative distribution function, The following contents are used as multiple myocardial activity parameters included in the second group: when the second object inverse time waveform is approximated by a waveform obtained by the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function, that is, a second approximate inverse time waveform, or a waveform obtained by adding the difference or weighted difference between the third cumulative distribution function and the fourth cumulative distribution function plus a second level value, that is, a second approximate inverse time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, the weight of the third cumulative distribution function, the weight of the fourth cumulative distribution function, the ratio of the weight of the third cumulative distribution function to the weight of the fourth cumulative distribution function, and the second level value; or, a time waveform obtained by the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function, that is, a second approximate time waveform, or a time waveform obtained by adding the difference or weighted difference between the third inverse cumulative distribution function and the fourth inverse cumulative distribution function plus a second level value. when the second object time waveform is approximated by a time waveform obtained by adding a second level value to the difference or weighted difference of the functions, that is, a second approximate time waveform, determining parameters of the third unimodal distribution, determining parameters of the third cumulative distribution function, determining parameters of the fourth unimodal distribution, determining parameters of the fourth cumulative distribution function, a weight of the third inverse cumulative distribution function, a weight of the fourth inverse cumulative distribution function, a ratio of the weight of the third inverse cumulative distribution function to the weight of the fourth inverse cumulative distribution function, and the second level value, taking any one of the time waveforms of the difference between the first object time waveform and the first approximate time waveform, the time waveform of the difference between the second object time waveform and the second approximate time waveform, and the time waveform of the difference between the second object inverse time waveform and the second approximate inverse time waveform after the time axis is reversed as the residual time waveform, taking the cumulative distribution function of the fifth unimodal distribution as the fifth cumulative distribution function, and taking the cumulative distribution function of the sixth unimodal distribution as the sixth cumulative distribution function, The following contents are used as multiple myocardial activity parameters included in the third group: when the residual time waveform is approximated by the fifth cumulative distribution function or the time waveform obtained by multiplying the fifth cumulative distribution function by a weight, i.e., an approximate residual time waveform, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, and the weight of the fifth cumulative distribution function is determined; or, when the residual time waveform is approximated by the time waveform obtained by the difference or weighted difference between the fifth cumulative distribution function and the sixth cumulative distribution function, the parameters of the fifth unimodal distribution are determined, the parameters of the fifth cumulative distribution function are determined, the parameters of the sixth unimodal distribution are determined, the parameters of the sixth cumulative distribution function, the weight of the fifth cumulative distribution function, the weight of the sixth cumulative distribution function, and the ratio of the weight of the fifth cumulative distribution function to the weight of the sixth cumulative distribution function, using parameters obtained by calculating the multiple myocardial activity parameters included in the first group, the multiple myocardial activity parameters included in the second group, and the multiple myocardial activity parameters included in the third group as one or more myocardial activity parameters included in the fourth group, a group formed by one or more predetermined myocardial activity parameters among the plurality of myocardial activity parameters included in the first group, the plurality of myocardial activity parameters included in the second group, the plurality of myocardial activity parameters included in the third group, and one or more myocardial activity parameters included in the fourth group, obtained from waveforms showing more than one cycle of the cardiac cycle of the heart, as a myocardial activity parameter group; An estimation model is prestored that receives a myocardial activity parameter group as an input and obtains heart state information corresponding to the myocardial activity parameter group, that is, information indicating the heart state. The estimation model is used to obtain cardiac state information of the target heart, by taking as input a group of myocardial activity parameters of the target heart, which is a heart to be provided with information.

9. A program for causing a computer to function as the learning device according to claim 1 or 2.

10. A program for causing a computer to function as the information providing device according to claim 4 or 5.

Images