Signal analysis apparatus, signal analysis method, and program

Abstract

This signal analysis apparatus comprises: a biological signal acquisition unit that acquires a target time waveform representing an activity of an analysis target; a distribution information acquisition unit that acquires a plurality of cumulative distribution functions including a first cumulative distribution function associated with a first unimodal distribution and a second cumulative distribution function associated with a second unimodal distribution; a fitting unit that approximates the target time waveform using a function composite wave that is a composite wave of the plurality of cumulative distribution functions; and a feature analysis unit that analyzes the features of the target time waveform on the basis of the function composite wave that approximates the target time waveform.

Description

Signal analysis apparatus, signal analysis method and program

[0001] The present invention relates to a signal analysis apparatus, a signal analysis method and a program.

[0002] A pulse wave representing myocardial activity is one of useful biological signals for grasping the state of the heart. A pulse wave is a composite wave of an ejection wave and a reflected wave. An ejection wave is a wave generated when the heart pumps out blood. A reflected wave is a wave generated when blood pumped from the heart is reflected by blood vessels. The pulse wave is measured using a photoplethysmograph, an impedance plethysmograph, a pressure gauge for a vascular catheter, or the like.

[0003] Based on the amplitude and time difference of inflection points of a pulse wave, the ratio of the ejection wave to the reflected wave may be indexed (see Non-Patent Document 1). Pulse wave velocity (PWV) may be calculated based on the distance and time difference between two points of an artery detection unit (see Non-Patent Document 2). Central blood pressure may be estimated based on the result of frequency analysis (Generalized Transfer Function) of a pulse wave (see Non-Patent Document 3). Further, based on the result of second-order differentiation of a pulse wave, the acceleration of the pulse wave may be classified (see Non-Patent Document 4).

[0004] Yoshio Matsui, 3 others, "Blood Pressure Pulse Wave Examination as a New Arteriosclerosis Index", Medical Instrumentation, Vol.80, No.4 (2010), pp.63-69Junichiro Hashimoto, "Arterial Stiffness and Pressure Pulse Wave Reflection", Monthly "Heart", Vol.42, No.8 (2010), pp.1021-1025TAKAZAWA. et al., "Relationship between Radial and Central Arterial Pulse Wave and Evaluation of Central Aortic Pressure Using the Radial Arterial Pulse Wave," Hypertension Research, Vol.30, No.3 (2007), pp.219-228Hiroto Seki, 2 others, "On Clinical Application of Accelerated Pulse Wave", Vol.3, No.1 (1994), pp.32-39

[0005] However, measuring the amplitude and time difference of inflection points in biological signals (e.g., pulse waves or electrocardiograms) is susceptible to fluctuations specific to those biological signals (e.g., vasoconstriction, fluctuations in myocardial ejection force, and waveform distortion due to respiratory fluctuations and body movement). Furthermore, frequency analysis and differential analysis of biological signals are difficult due to waveforms specific to the blood vessels of the heart (e.g., waveforms resulting from overlapping notches during aortic valve closure). Thus, there is a problem in that it is not possible to analyze biological signals while ensuring the versatility of the analysis.

[0006] In view of the above circumstances, the present invention aims to provide a signal analysis device, a signal analysis method, and a program that can analyze biological signals while ensuring versatility of analysis.

[0007] One aspect of the present invention is a signal analysis device comprising: a biosignal acquisition unit that acquires a target time waveform representing the activity to be analyzed; a distribution information acquisition unit that acquires a plurality of cumulative distribution functions including a first cumulative distribution function associated with a first unimodal distribution and a second cumulative distribution function associated with a second unimodal distribution; a fitting unit that approximates the target time waveform with a function composite wave which is a composite wave of the plurality of cumulative distribution functions; and a feature analysis unit that analyzes the characteristics of the target time waveform based on the function composite wave which approximates the target time waveform.

[0008] One aspect of the present invention is a signal analysis method performed by the above-described signal analysis device, comprising the steps of: acquiring a target time waveform representing the activity to be analyzed; acquiring a plurality of cumulative distribution functions including a first cumulative distribution function associated with a first unimodal distribution and a second cumulative distribution function associated with a second unimodal distribution; approximating the target time waveform with a function composite wave which is a composite wave of the plurality of cumulative distribution functions; and analyzing the characteristics of the target time waveform based on the function composite wave that approximates the target time waveform.

[0009] One aspect of the present invention is a program for causing a computer to perform the following steps: a step to acquire a time waveform representing the activity to be analyzed; a step to acquire a plurality of cumulative distribution functions including a first cumulative distribution function associated with a first unimodal distribution and a second cumulative distribution function associated with a second unimodal distribution; a step to approximate the time waveform with a function composite wave which is a composite wave of the plurality of cumulative distribution functions; and a step to analyze the characteristics of the time waveform based on the function composite wave which approximates the time waveform.

[0010] This invention makes it possible to analyze biological signals while ensuring versatility in analysis.

[0011] This figure shows an example of the configuration of the signal analysis device in the embodiment. This figure shows an example of pulse wave and electrocardiogram in the embodiment. This figure shows an example of cumulative distribution function, function composite wave, and probability density function in the embodiment. This figure shows a detailed example of cumulative distribution function, function composite wave, and probability density function in the embodiment. This figure shows an example of the relationship between volume pulse wave, pulse rate, mean value of the first cumulative distribution function, standard deviation of the third cumulative distribution function, and amplitude of the first cumulative distribution function in the embodiment. This flowchart shows an example of the operation of the signal analysis device in the embodiment. This figure shows an example of the hardware configuration of the signal analysis device in the embodiment.

[0012] Embodiments of the present invention will be described in detail with reference to the drawings. Figure 1 is a diagram showing an example of the configuration of the signal analysis device 1 in an embodiment. The signal analysis device 1 is a device that analyzes biological signals (target time waveforms). The biological signal is not limited to a specific signal as long as it is a time waveform signal obtained from a living organism. The biological signal is, for example, a pulse wave (a composite wave of ejection and reflected waves) that represents the activity of the myocardium in the heart that is the subject of analysis. The biological signal may also be, for example, an electrocardiogram (time waveform of electrocardiogram potentials) that represents the activity of the myocardium, an electroencephalogram (event-related potential) that represents the activity in the brain that is the subject of analysis, or a time waveform that represents the contraction and relaxation of the muscle that is the subject of analysis. Hereinafter, the biological signal is a pulse wave as an example. The pulse wave may be, for example, a time waveform of blood pressure values, a time waveform of measurements taken with a volume pulse wave meter, or a time waveform of a standardized value based on these.

[0013] The signal analysis device 1 comprises a control unit 11, an input unit 12, a communication unit 13, a storage device 14, and an output unit 15. The control unit 11 comprises a biosignal acquisition unit 110, a distribution information acquisition unit 120, and an analysis unit 130. The analysis unit 130 comprises a fitting unit 131 and a feature analysis unit 132.

[0014] The control unit 11 models the pulse wave using a plurality of cumulative distribution functions (CDFs). The cumulative distribution functions are associated with a unimodal distribution. Such a unimodal distribution can be represented, for example, using a probability density function (PDF). A desired distribution function such as a normal distribution, beta distribution, or gamma distribution may be used for the unimodal distribution. The desired distribution function may be, for example, the cumulative distribution function of a normal distribution.

[0015] A cumulative distribution function whose function value (amplitude) increases over time corresponds to a unimodal distribution with a local maximum. Such a cumulative distribution function corresponds to the contraction (tension) of the heart being analyzed and the increase in pulse pressure. Conversely, a cumulative distribution function whose function value (amplitude) decreases over time corresponds to a unimodal distribution with a local minimum. Such a cumulative distribution function corresponds to the expansion (relaxation) of the heart being analyzed and the decrease in pulse pressure.

[0016] The number of cumulative distribution functions used by the control unit 11 to model the pulse wave is not limited to a specific number. The control unit 11 may, for example, use two or more cumulative distribution functions to model the pulse wave.

[0017] If the control unit 11 uses three cumulative distribution functions to model the pulse wave, these three cumulative distribution functions may be, for example, one cumulative distribution function whose function value increases over time and two cumulative distribution functions whose function value decreases over time.

[0018] If the control unit 11 uses four cumulative distribution functions to model the pulse wave, these four cumulative distribution functions may be, for example, two cumulative distribution functions whose function values ​​increase over time and two cumulative distribution functions whose function values ​​decrease over time.

[0019] Constraints may be imposed on the amplitude (weights), standard deviation, and mean of the cumulative distribution function. That is, constraints may be imposed on the amplitude, standard deviation, and mean of the unimodal distribution associated with the cumulative distribution function. Furthermore, constraints may be imposed on the interval of the cumulative distribution function. The constraints imposed on the interval of the cumulative distribution function may be defined, for example, using the standard deviation σ, such as 2σ or 3σ. The mean of the cumulative distribution function is the interval between the time of the peak of the pulse wave and the time of the peak of the unimodal distribution, as illustrated in Figures 3 and 4 below. Alternatively, the mean of the cumulative distribution function may also be the interval between the times of the peaks of adjacent unimodal distributions, as illustrated in Figures 3 and 4 below. The mean of the cumulative distribution function may also be the median of the cumulative distribution function. The timing of the mean corresponds to the timing when the blood flow velocity is maximum. With respect to time, rules may be imposed on the order in which multiple cumulative distribution functions approximating the pulse wave are arranged.

[0020] Hereinafter, the composite wave of multiple cumulative distribution functions will be referred to as a "function composite wave." The control unit 11 generates a function composite wave such that the function composite wave (model equation) representing the temporal transition process of the ejection wave and reflected wave substantially matches (approximates) the pulse wave.

[0021] The control unit 11 adjusts each parameter (amplitude, standard deviation, and mean value) of the multiple cumulative distribution functions that constitute the function composite wave, for example using the least squares method, so that the error of the function composite wave relative to the pulse wave (measured value) is minimized. That is, the control unit 11 adjusts each parameter (amplitude, standard deviation, and mean value) of each unimodal distribution associated with the multiple cumulative distribution functions that constitute the function composite wave, for example using the least squares method.

[0022] The timing and characteristics of each cumulative distribution function constituting the generated function composite wave serve as indicators of the timing and characteristics (such as transition velocity) of the ejection wave and reflected wave that constitute the pulse wave. In other words, the function composite wave obtained by adjusting each parameter represents the respective features of the ejection wave and reflected wave. For example, the mean value of the cumulative distribution function (unimodal distribution) corresponding to the ejection wave approximately coincides with the maximum velocity of the ejection wave. Similarly, the mean value of the cumulative distribution function (unimodal distribution) corresponding to the reflected wave approximately coincides with the maximum velocity of the reflected wave. Furthermore, the characteristics of the cardiovascular system may be evaluated based on the mean value, amplitude, and standard deviation of the cumulative distribution function. In addition, the pulse wave propagation velocity and intravascular pressure may be estimated based on the mean value, amplitude, and standard deviation of the cumulative distribution function.

[0023] The control unit 11 analyzes the precise timing and amplitude (indicators) of the ejection wave and reflected wave independently for each acquired pulse wave (parameter analysis) based on the obtained function composite wave. The control unit 11 may also analyze the time interval between the ejection wave and the reflected wave for each acquired pulse wave based on the obtained function composite wave. The control unit 11 may also analyze the fluctuations in the time interval and amplitude of the acquired pulse wave (heartbeat) independently for each acquired pulse wave (ejection wave and reflected wave) based on the obtained function composite wave.

[0024] The input unit 12 is an input device, such as a mouse, keyboard, or touch panel. The input unit 12 receives input operations for the signal analysis device 1. For example, the input unit 12 may receive an operation to input a biological signal (e.g., a pulse wave) to the signal analysis device 1.

[0025] The communication unit 13 performs communication with an external device (not shown). For example, the communication unit 13 may receive biological signals (e.g., pulse waves) from an external device (not shown).

[0026] The storage device 14 stores a program executed by the control unit 11. The storage device 14 may also store biological signals. The biological signals stored in the storage device 14 may be biological signals received by the communication unit 13.

[0027] The output unit 15 is a display device having a display device such as a liquid crystal display or an organic electro-luminescence (OLED) display. The output unit 15 outputs various types of information. For example, the output unit 15 may display biological signals and function composite waves. For example, the output unit 15 may display the analysis results by the control unit 11 (e.g., the timing of ejection waves and reflected waves).

[0028] Next, the details of the control unit 11 will be described. Figure 2 is a diagram showing an example of a pulse wave and electrocardiogram in an embodiment. The pulse wave (time waveform of pulse pressure) is a composite wave of ejection wave and reflected wave. That is, the pulse wave represents the time-dependent transition process (time waveform) of the ejection wave and reflected wave. The biosignal acquisition unit 110 acquires such a pulse wave as a biosignal from the input unit 12 or the storage device 14.

[0029] A pulse wave may be associated with an electrocardiogram (a time-based waveform of the electrocardiogram). The electrocardiogram may include not only one or more R waves 21, but also, for example, T waves 22 and P waves 23 during the period from R wave 21-1 to R wave 21-2.

[0030] Figure 3 is a diagram, "Pulse Wave Peak - Cumulative Distribution Function," showing an example of the Cumulative Distribution Function (CDF), the function composite wave, and the probability density function in the embodiment. Figure 4 is a diagram, "Pulse Wave Peak - Cumulative Distribution Function," showing a detailed example of the Cumulative Distribution Function, the function composite wave, and the probability density function in the embodiment. The pulse wave peak 201 is the highest point of the function composite wave 41 (pulse wave). The first amplitude 231 is the amplitude of the first cumulative distribution function 31. The second amplitude 232 is the amplitude of the second cumulative distribution function 32. The third amplitude 233 is the amplitude of the third cumulative distribution function 33. The first standard deviation 251 is the standard deviation of the first cumulative distribution function 31. The second standard deviation 252 is the standard deviation of the second cumulative distribution function 32. The third standard deviation 253 is the standard deviation of the third cumulative distribution function 33. The first mean 351 is the mean of the first cumulative distribution function 31. The second mean 352 is the mean of the second cumulative distribution function 32. The third mean 353 is the mean of the third cumulative distribution function 33.

[0031] The distribution information acquisition unit 120 acquires multiple cumulative distribution functions from the input unit 12 or the storage device 14. In Figure 3, the distribution information acquisition unit 120 acquires the first cumulative distribution function 31, the second cumulative distribution function 32, and the third cumulative distribution function 33 from the input unit 12 or the storage device 14.

[0032] The analysis unit 130 adjusts the parameters of the multiple cumulative distribution functions that make up the function composite wave (model equation) so that the error of the function composite wave (model equation) with respect to the pulse wave (measured value), which is the waveform of the target time, is minimized. The analysis unit 130 also analyzes the characteristics of the waveform of the target time based on the function composite wave 41 which approximates the pulse wave, which is the waveform of the target time.

[0033] The fitting unit 131 adjusts each parameter (amplitude, standard deviation, and mean value) of the multiple cumulative distribution functions that constitute the function composite wave (model equation) using, for example, the least squares method, so that the error of the function composite wave (model equation) with respect to the pulse wave (measured value), which is the time waveform of the target, is minimized.

[0034] The fitting unit 131 may perform function fitting to the biological signal using a function composite wave (model equation) and method using tensor analysis for the biological signal. The function composite wave and method using tensor analysis for the biological signal may be, for example, the function composite wave and method disclosed in Reference 1 (International Publication No. 2022 / 149381) or Reference 2 (International Publication No. 2022 / 149382).

[0035] In Figure 3, the fitting unit 131 generates a function composite wave 41 from the first cumulative distribution function 31, the second cumulative distribution function 32, and the third cumulative distribution function 33. Here, the fitting unit 131 adjusts the parameters of the first cumulative distribution function 31 so as to approximate the ejection wave during systole. The fitting unit 131 adjusts the parameters of the second cumulative distribution function 32 so as to approximate the early component of the reflected wave during diastole. The fitting unit 131 adjusts the parameters of the third cumulative distribution function 33 so as to approximate the late component of the reflected wave during diastole.

[0036] The feature analysis unit 132 analyzes the features of the target time waveform based on a function composite wave 41 that approximates the pulse wave, which is the target time waveform. For example, the feature analysis unit 132 analyzes the features of the pulse wave based on at least one of the first cumulative distribution function 31, the second cumulative distribution function 32, and the third cumulative distribution function 33 that constitute the function composite wave 41.

[0037] For example, the feature analysis unit 132 analyzes the characteristics of the ejection wave (e.g., the timing and characteristics of the ejection wave (such as the transition speed between the ejection wave and the reflected wave)) based on at least one of the amplitude, standard deviation, and mean of the first cumulative distribution function 31 of the function composite wave 41. That is, the feature analysis unit 132 analyzes the characteristics of the ejection wave based on at least one of the amplitude, standard deviation, and mean of the first unimodal distribution 51 associated with the first cumulative distribution function 31.

[0038] For example, the feature analysis unit 132 may analyze the features of the early component of the reflected wave (e.g., the timing and characteristics of the reflected wave (such as the speed of transition between the ejection wave and the reflected wave)) based on at least one of the amplitude, standard deviation, and mean of the second cumulative distribution function 32 of the function composite wave 41. That is, the feature analysis unit 132 may analyze the features of the early component of the reflected wave based on at least one of the amplitude, standard deviation, and mean of the second unimodal distribution 52 associated with the second cumulative distribution function 32.

[0039] For example, the feature analysis unit 132 may analyze the characteristics of the later component of the reflected wave based on at least one of the amplitude, standard deviation, and mean of the third cumulative distribution function 33 of the function composite wave 41. That is, the feature analysis unit 132 may analyze the characteristics of the later component of the reflected wave based on at least one of the amplitude, standard deviation, and mean of the third unimodal distribution 53 associated with the third cumulative distribution function 33.

[0040] The mean value of the first cumulative distribution function 31 (first unimodal distribution 51) corresponds to the pulse wave propagation velocity. Therefore, the mean value of the first cumulative distribution function 31 (first unimodal distribution 51) is approximately equal to the blood pressure or blood flow rate. The feature analysis unit 132 may estimate the blood pressure or blood flow rate (for example, heart rate and pulse rate correlated with blood flow rate) based on the mean value of the first cumulative distribution function 31 (first unimodal distribution 51).

[0041] The amplitude of the first cumulative distribution function 31 (first unimodal distribution 51) corresponds to the amplitude of the ejection wave. Therefore, the amplitude of the first cumulative distribution function 31 (first unimodal distribution 51) is approximately equal to the ejection force of the myocardium. The feature analysis unit 132 may estimate the ejection force of the myocardium based on the amplitude of the first cumulative distribution function 31 (first unimodal distribution 51).

[0042] The amplitude of the third cumulative distribution function 33 (third unimodal distribution 53) corresponds to the amplitude of the reflected wave. Therefore, the amplitude of the third cumulative distribution function 33 (third unimodal distribution 53) is approximately equal to the stiffness of the blood vessel contraction. The feature analysis unit 132 may estimate the stiffness of the blood vessel contraction based on the amplitude of the third cumulative distribution function 33 (third unimodal distribution 53).

[0043] Figure 5 shows an example of the relationship between volume pulse wave, pulse rate, mean value of the first cumulative distribution function, mean value of the first cumulative distribution function, standard deviation of the third cumulative distribution function, and amplitude of the first cumulative distribution function in an embodiment. In Figure 5, a subject with the heart being analyzed is experiencing a virtual reality (VR) training simulator as an example. Pulse rate and other parameters were measured during the period of experiencing this training simulator. The times indicated by the black triangles are when a stressful event was presented to the subject. A stressful response is observed at the times indicated by the black triangles. The times indicated by the white arrows are when a frustration event was presented to the subject. A frustration response is observed at the times indicated by the white arrows.

[0044] When a subject is mentally stressed (not relaxed), not only does the heart rate increase, but at least one of the standard deviation of the third cumulative distribution function, the mean value of the first cumulative distribution function, and the amplitude of the first cumulative distribution function fluctuates. That is, at least one of the standard deviation of the third cumulative distribution function, the mean value of the first cumulative distribution function, and the amplitude of the first cumulative distribution function correlates with the subject's mental stress (relaxation). Accordingly, each parameter of these cumulative distribution functions may be used as an indicator of mental stress (relaxation) of a subject that has been set as an analysis target. Further, when a subject feels frustration, at least one of the standard deviation of the third cumulative distribution function, the mean value of the first cumulative distribution function, and the amplitude of the first cumulative distribution function fluctuates without being accompanied by an increase in heart rate. The feature analysis unit 132 may estimate an indicator of mental stress (relaxation) or frustration of the subject set as the analysis target based on at least one of the standard deviation of the third cumulative distribution function (e.g., the amount of change or the number of changes in the standard deviation), the mean value of the first cumulative distribution function (e.g., the amount of change or the number of changes in the mean value), and the amplitude of the first cumulative distribution function (e.g., the amount of change or the number of changes in the amplitude).

[0045] Next, an example of the operation of the signal analysis device 1 will be described. FIG. 6 is a flowchart showing an example of the operation of the signal analysis device 1 according to the embodiment. A biological signal acquisition unit 110 acquires a biological signal from the input unit 12 or the storage device 14 (step S101). A distribution information acquisition unit 120 acquires a plurality of cumulative distribution functions from the input unit 12 or the storage device 14 (step S102). A fitting unit 131 approximates the target time waveform with a composite function wave 41 that is a composite wave of the obtained plurality of cumulative distribution functions (step S103). A feature analysis unit 132 analyzes the features of the target time waveform based on the composite function wave 41 that approximates the target time waveform (step S104).

[0046] As described above, the biological signal acquisition unit 110 acquires a target time waveform (biological signal) representing the activity of an analysis target from the input unit 12 or the storage device 14. The target time waveform is, for example, a pulse wave (a composite wave of an ejection wave and a reflected wave) representing the activity of cardiac muscle in the heart that has been set as an analysis target.

[0047] A distribution information acquisition unit 120 acquires a plurality of cumulative distribution functions including a first cumulative distribution function 31 associated with a first unimodal distribution 51 and a second cumulative distribution function 32 associated with a second unimodal distribution 52. The distribution information acquisition unit 120 may acquire a plurality of cumulative distribution functions including more cumulative distribution functions. For example, the distribution information acquisition unit 120 may acquire a plurality of cumulative distribution functions including the first cumulative distribution function 31, the second cumulative distribution function 32, and a third cumulative distribution function 33 associated with a third unimodal distribution 53.

[0048] A fitting unit 131 approximates a target time waveform (e.g., a pulse wave sample point group) with a function composite wave 41 that is a composite wave of the plurality of acquired cumulative distribution functions.

[0049] The fitting unit 131 approximates at least a part of an ejection wave with the first cumulative distribution function 31 by adjusting at least one of an amplitude, a standard deviation, and an average value of the first cumulative distribution function 31. That is, the fitting unit 131 approximates at least a part of the ejection wave (e.g., the ejection wave in a systolic period) with the first cumulative distribution function 31 by adjusting at least one of an amplitude, a standard deviation, and an average value of the first unimodal distribution 51 (unimodal distribution having a maximum value).

[0050] The fitting unit 131 approximates at least a part of a reflected wave (e.g., an early-phase component of the reflected wave in a diastolic period) with the second cumulative distribution function 32 by adjusting at least one of an amplitude, a standard deviation, and an average value of the second cumulative distribution function 32. That is, the fitting unit 131 approximates at least a part of the reflected wave with the second cumulative distribution function 32 by adjusting at least one of an amplitude, a standard deviation, and an average value of the second unimodal distribution 52 (unimodal distribution having a minimum value).

[0051] The fitting unit 131 approximates at least a portion of the reflected wave (for example, the later component of the reflected wave during the expansion phase) with the third cumulative distribution function 33 by adjusting at least one of the amplitude, standard deviation, and mean of the third cumulative distribution function 33. That is, the fitting unit 131 approximates at least a portion of the reflected wave with the third cumulative distribution function 33 by adjusting at least one of the amplitude, standard deviation, and mean of the third unimodal distribution 53 (a unimodal distribution with a local minimum).

[0052] The feature analysis unit 132 analyzes the features of the time waveform based on a function composite wave 41 that approximates the time waveform. For example, the feature analysis unit 132 analyzes the features of the pulse wave, which is a composite wave of ejection and reflected waves, based on at least one of the first cumulative distribution function 31, the second cumulative distribution function 32, and the third cumulative distribution function 33. The feature analysis unit 132 may estimate the blood pressure or blood flow rate in the heart under analysis based on the mean value of the first cumulative distribution function 31. The feature analysis unit 132 may estimate the ejection force by the myocardium in the heart under analysis based on the amplitude of the first cumulative distribution function 31. The feature analysis unit 132 may estimate the vasoconstriction stiffness in the heart under analysis based on the amplitude of the third cumulative distribution function 33. The features of the time waveform are, for example, the blood pressure or blood flow rate in the heart under analysis, the ejection force by the myocardium in the heart, the vasoconstriction stiffness in the heart, or an indicator of the mental tension of a subject with a heart.

[0053] This makes it possible to analyze biological signals while ensuring the versatility of the analysis.

[0054] By using a combination of multiple cumulative distribution functions, it is possible to efficiently approximate biological signals such as pulse waves. It is possible to improve the accuracy of detecting the time (timing) of biological signals. It is possible to separate the pulse wave into ejection and reflected waves and analyze each of them. It is possible to improve the accuracy of detecting the time difference between the average value (time) of the cumulative distribution function representing the ejection wave and the average value (time) of the cumulative distribution function representing the reflected wave. Furthermore, it is possible to precisely analyze pulse wave fluctuations due to respiratory fluctuations, mental stress, exercise, or cardiovascular load.

[0055] (Hardware Configuration) Figure 7 shows an example of the hardware configuration of the signal analysis device 1 in an embodiment. The signal analysis device 1 is implemented as software by a processor 91 such as a CPU (Central Processing Unit) executing a program stored in a storage device 14 having a non-volatile recording medium (non-temporary recording medium) and a memory 92. The program may be recorded on a computer-readable recording medium. A computer-readable recording medium is a non-temporary recording medium such as a portable medium such as a flexible disk, magneto-optical disk, ROM (Read Only Memory), CD-ROM (Compact Disc Read Only Memory), or a storage device such as a hard disk or solid-state drive (SSD) built into a computer system. The communication unit 13 executes predetermined communication processing.

[0056] The signal analysis device 1 may be implemented via hardware including an electronic circuit (or circuitry) using, for example, an LSI (Large Scale Integrated Circuit), ASIC (Application Specific Integrated Circuit), PLD (Programmable Logic Device), or FPGA (Field Programmable Gate Array).

[0057] Although embodiments of this invention have been described in detail above with reference to the drawings, the specific configuration is not limited to these embodiments and includes designs and the like that do not depart from the spirit of this invention. Furthermore, each embodiment may be combined.

[0058] This invention is applicable to devices that analyze biological signals such as pulse waves, electrocardiograms, or electroencephalograms.

[0059] 1...Signal analysis device, 11...Control unit, 12...Input unit, 13...Communication unit, 14...Storage device, 15...Output unit, 21...R wave, 22...T wave, 23...P wave, 31...First cumulative distribution function, 32...Second cumulative distribution function, 33...Third cumulative distribution function, 41...Function composite wave, 51...First unimodal distribution, 52...Second unimodal distribution, 53...Third unimodal distribution, 91...Processor, 92...Memory 110... Biosignal acquisition unit, 120... Distribution information acquisition unit, 130... Analysis unit, 131... Fitting unit, 132... Feature analysis unit, 201... Pulse wave peak, 231... First amplitude, 232... Second amplitude, 233... Third amplitude, 251... First standard deviation, 252... Second standard deviation, 253... Third standard deviation, 351... First mean, 352... Second mean, 353... Third mean

Claims

1. A signal analysis device comprising: a biosignal acquisition unit that acquires a target time waveform representing the activity to be analyzed; a distribution information acquisition unit that acquires a plurality of cumulative distribution functions including a first cumulative distribution function associated with a first unimodal distribution and a second cumulative distribution function associated with a second unimodal distribution; a fitting unit that approximates the target time waveform with a function composite wave which is a composite wave of the plurality of cumulative distribution functions; and a feature analysis unit that analyzes the characteristics of the target time waveform based on the function composite wave that approximates the target time waveform.

2. The signal analyzer according to claim 1, wherein the target time waveform is a pulse wave which is a composite wave of ejection waves and reflected waves representing the activity of the myocardium in the heart that is the subject of analysis, an electrocardiogram which represents the activity of the myocardium, an electroencephalogram which represents the activity of the brain that is the subject of analysis, or a time waveform which represents the contraction and relaxation of the muscle that is the subject of analysis.

3. The signal analysis device according to claim 1 or 2, wherein the target time waveform is a pulse wave which is a composite wave of ejection waves and reflected waves representing the activity of the myocardium in the heart that is the subject of analysis, the fitting unit approximates the ejection wave with the first cumulative distribution function by adjusting at least one of the amplitude, standard deviation, and mean of the first cumulative distribution function, and approximates the reflected wave with the second cumulative distribution function by adjusting at least one of the amplitude, standard deviation, and mean of the second cumulative distribution function, and the feature analysis unit analyzes the features of the target time waveform based on at least one of the first cumulative distribution function and the second cumulative distribution function.

4. The signal analysis device according to claim 3, wherein the distribution information acquisition unit acquires a plurality of cumulative distribution functions including the first cumulative distribution function, the second cumulative distribution function, and the third cumulative distribution function associated with the third unimodal distribution; the fitting unit approximates the ejection wave with the first cumulative distribution function by adjusting at least one of the amplitude, standard deviation, and mean of the first cumulative distribution function; approximates the early component of the reflected wave with the second cumulative distribution function by adjusting at least one of the amplitude, standard deviation, and mean of the second cumulative distribution function; approximates the late component of the reflected wave with the third cumulative distribution function by adjusting at least one of the amplitude, standard deviation, and mean of the third cumulative distribution function; and the feature analysis unit analyzes the features of the target time waveform based on at least one of the first cumulative distribution function, the second cumulative distribution function, and the third cumulative distribution function.

5. The signal analysis device according to claim 1, wherein the characteristics of the target time waveform are blood pressure or blood flow in the heart being analyzed, the ejection force by the myocardium in the heart, the stiffness of the vasoconstriction in the heart, or an indicator of the mental tension of a subject having the heart.

6. A signal analysis method performed by a signal analysis device, comprising: a step of acquiring a target time waveform representing the activity to be analyzed; a step of acquiring a plurality of cumulative distribution functions including a first cumulative distribution function associated with a first unimodal distribution and a second cumulative distribution function associated with a second unimodal distribution; a step of approximating the target time waveform with a function composite wave which is a composite wave of the plurality of cumulative distribution functions; and a step of analyzing the characteristics of the target time waveform based on the function composite wave that approximates the target time waveform.

7. A program for causing a computer to execute the following steps: a procedure for acquiring a time waveform representing the activity to be analyzed; a procedure for acquiring a plurality of cumulative distribution functions, including a first cumulative distribution function associated with a first unimodal distribution and a second cumulative distribution function associated with a second unimodal distribution; a procedure for approximating the time waveform with a function composite wave, which is a composite wave of the plurality of cumulative distribution functions; and a procedure for analyzing the characteristics of the time waveform based on the function composite wave that approximates the time waveform.

Images