Pulsation analysis device, pulsation analysis method, and program

Abstract

To provide a pulsation analysis device suitable for accurately analyzing pulsation intervals by measuring physical vibrations generated in a body due to contraction and / or relaxation of internal organs.SOLUTION: A pulsation analysis device 5 analyzes a measurement signal obtained by measuring periodic repetition of contraction and / or relaxation of internal organs of a subject 7 by a measuring device 3. A provisional detection peak processing unit 39 calculates indices for at least two extreme points in a waveform of measurement data obtained from the measurement signal on the basis of amplitudes of the respective extreme points and amplitudes and occurrence times of extreme points before and after the respective extreme points, and sets a part or all of the extreme points in the waveform of the calculated indices as peak points. An analysis processing unit 36 analyzes intervals of contraction and / or relaxation of the internal organs using intervals between the peak points.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a pulse analysis device, a pulse analysis method, and a program, and in particular to a pulse analysis device that analyzes measurement signals obtained by measuring the periodic contraction and / or relaxation of internal organs in a subject using a measuring device. [Background technology]

[0002] An electrocardiogram is a recording of electrical signals generated by the cardiac muscle as the heart beats, while a ballistocardiogram measures the physical vibrations of the body caused by the contraction and relaxation of the heart.

[0003] Referring to Fig. 22, as disclosed in Patent Document 1, the inventors have proposed a new system for obtaining a ballistocardiogram. In this system, measuring devices 103 and 105 are placed on a first cushion 101 placed on the seat of a chair. A second cushion 107 and a third cushion 109 are placed on measuring devices 103 and 105, respectively. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent application 2023-175892 Summary of the Invention [Problem to be solved by the invention]

[0005] Most conventional ballistocardiogram heart rate monitoring techniques focus on obtaining accurate heart rate data, making it difficult to measure the heartbeat interval. Similar challenges arise when measuring not only the heartbeat (heartbeat), but also physical vibrations of the body caused by the periodic contraction and / or relaxation of other internal organs.

[0006] Therefore, the present invention aims to provide a pulse analysis device suitable for accurately analyzing even the interval between pulses by measuring physical vibrations that occur in the body due to contraction and / or relaxation of internal organs. [Means for solving the problem]

[0007] A first aspect of the present invention is a pulse analysis device that analyzes measurement signals obtained by a measurement device measuring the periodic repetition of contraction and / or relaxation of internal organs in a subject, the pulse analysis device comprising a pre-processing unit and an analysis processing unit, the pre-processing unit comprising a provisional detection peak processing unit, the provisional detection peak processing unit calculating an index for at least two extreme points in the waveform of measurement data obtained from the measurement signal from the amplitude of each of the extreme points and the amplitudes and occurrence times of the extreme points before and after each of the extreme points, and setting some or all of the extreme points in the waveform of the calculated index as peak points, and the analysis processing unit analyzing the interval of contraction and / or relaxation of the internal organs using the interval between the peak points.

[0008] A second aspect of the present invention is the pulse analysis device of the first aspect, wherein the pre-processing unit includes an over-detection processing unit and an undetection processing unit, and the over-detection processing unit removes some of the peak points so that the amount of change before and after the waveform rearranged in order based on the interval between peak points is below a threshold, and the undetection processing unit adds peak points so that the amount of change before and after the waveform rearranged in order based on the interval between peak points is below a threshold.

[0009] A third aspect of the present invention is a pulse analysis device according to the first or second aspect, wherein there are a plurality of measuring devices, and the pulse analysis device is equipped with a parallel processing unit, and each of the measuring devices measures the periodic contraction and / or relaxation of the internal organs of the subject at different positions on the subject to obtain measurement signals, the pre-processing unit processes the measurement signals obtained by each of the measuring devices individually to obtain peak points corresponding to each of the measuring devices, and the parallel processing unit synchronizes the peak points corresponding to each of the measuring devices and maintains, adds, changes, and deletes the peak points corresponding to each of the measuring devices by referring to peak points corresponding to other measuring devices.

[0010] A fourth aspect of the present invention is a pulse analysis method in a pulse analysis device that analyzes measurement signals obtained by a measurement device measuring the periodic repetition of contraction and / or relaxation of internal organs in a subject, the pulse analysis device comprising a pre-processing unit and an analysis processing unit, the pre-processing unit comprising a provisional detection peak processing unit, the provisional detection peak processing unit calculating an index for at least two extreme points in the waveform of measurement data obtained from the measurement signal from the amplitude of each of the extreme points and the amplitudes and occurrence times of extreme points before and after each of the extreme points, and setting some or all of the extreme points in the waveform of the calculated index as peak points, and the analysis processing unit analyzing the intervals of contraction and / or relaxation of the internal organs using the intervals between peak points.

[0011] A fifth aspect of the present invention is the beat analysis method of the fourth aspect, wherein the pre-processing unit includes an over-detection processing unit and an undetection processing unit, and includes a step in which the over-detection processing unit removes some of the peak points so that the amount of change before and after the waveform rearranged in order based on the interval between peak points is below a threshold, and the undetection processing unit adds peak points so that the amount of change before and after the waveform rearranged in order based on the interval between peak points is below a threshold.

[0012] A sixth aspect of the present invention is a pulse analysis method according to the fourth or fifth aspect, wherein there are a plurality of measuring devices, each of which is equipped with a parallel processing unit, and each of the measuring devices measures the periodic contraction and / or relaxation of the internal organs of the subject at a different position on the subject to obtain measurement signals, the pre-processing unit individually processes the measurement signals obtained by each of the measuring devices to obtain peak points corresponding to each of the measuring devices, and the parallel processing unit synchronizes the peak points corresponding to each of the measuring devices, and maintains, adds, changes, and deletes the peak points corresponding to each of the measuring devices by referring to peak points corresponding to other measuring devices.

[0013] A seventh aspect of the present invention is a program for causing a computer to function as the pulse analyzer of any one of the first to third aspects. The present invention may also be understood as a computer-readable recording medium on which the program of the seventh aspect is recorded. [Effects of the Invention]

[0014] According to each aspect of the present invention, the tentatively detected peak processing unit calculates an index for at least two peaks in the waveform of the measurement data from the amplitude of each peak and the amplitudes and occurrence times of the peaks before and after each peak. Therefore, for example, an index indicating the sharpness of the waveform shape can be introduced by evaluating the height (amplitude) of each peak while taking into account the amplitude and width (width of occurrence time) of the adjacent peak. Then, some or all of the peaks in the waveform of this index can be used as peak points, and the interval between them can be used to analyze the interval between contractions and / or relaxations of the internal organs. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a block diagram showing an example of the configuration of a pulse measurement system 1 according to an embodiment of the present invention. [Figure 2] 3 is a flowchart showing an example of the operations of the preprocessing unit 33, the parallel processing unit 35, and the analysis processing unit 36 ​​in FIG. [Figure 3] 2 is a graph showing measurement data obtained by smoothing processing unit 37 of FIG. 1 performing smoothing processing on measurement raw data of one channel. [Figure 4] FIG. 10A is a diagram for explaining a waveform sharpness index, and FIG. 10B is a diagram showing an example of the waveform of the waveform sharpness index. [Figure 5] FIG. 10 is a first diagram showing the results of peak point detection by the tentatively detected peak processing unit 39 in FIG. 1 when the quality of the measured raw data is good. [Figure 6] FIG. 2 is a second diagram showing the results of peak point detection by the tentatively detected peak processing unit 39 in FIG. 1 when the quality of the measured raw data is good. [Figure 7]FIG. 10 is a first diagram showing the results of peak point detection by the tentatively detected peak processing unit 39 in FIG. 1 when the quality of the measured raw data is poor. [Figure 8] FIG. 2 is a second diagram showing the results of peak point detection by the tentatively detected peak processing unit 39 in FIG. 1 when the quality of the measured raw data is poor. [Figure 9] 2 is a diagram for explaining processing by the overdetection processing unit 41 of FIG. 1. FIG. [Figure 10] FIG. 10 shows the plateau region in the measurement of FIG. 9. [Figure 11] FIG. 2 is a first diagram for explaining processing by the undetection processing unit 43 of FIG. [Figure 12] FIG. 2 is a second diagram for explaining the processing by the undetection processing unit 43 of FIG. [Figure 13] FIG. 2 is a diagram for explaining an example of specific processing by the parallel processing unit 35 of FIG. 1, and is a diagram for explaining processing for synchronization. [Figure 14] FIG. 10 is a diagram showing an example in which correction is required due to a change amount between the front and rear directions, and is a diagram showing a state before correction. [Figure 15] 15 is a diagram in which the eighth peak point of channel CH1 in FIG. 14 has been corrected. [Figure 16] FIG. 10 is a first diagram for explaining an example in which correction is necessary when the amount of change between the front and rear directions is determined to be appropriate. [Figure 17] FIG. 2 is a second diagram for explaining an example in which correction is necessary when the amount of change before and after is determined to be appropriate. [Figure 18] This is a scatter plot of JJ intervals (vertical axis) against RR intervals (horizontal axis) for all four trials of four subjects. [Figure 19] FIG. 1 shows a scatter plot of the heart rate variability analysis index calculated from the JJ interval (vertical axis) against the heart rate variability analysis index calculated from the RR interval (horizontal axis) for all four trials of 16 subjects. [Figure 20] FIG. 2 shows a scatter plot of the heart rate variability analysis index calculated from the JJ interval (vertical axis) against the heart rate variability analysis index calculated from the RR interval (horizontal axis) for all four trials of 16 subjects. [Figure 21]FIG. 10 is a block diagram showing an example of the configuration of a pulse measurement system 51 which is another example of an embodiment of the present invention. [Figure 22] FIG. 1 is a diagram showing an overview of a system proposed by the inventors for obtaining a ballistocardiogram. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, embodiments of the present invention will be described with reference to the drawings, but the present invention is not limited to these embodiments.

[0017] FIG. 1 is a block diagram showing an example of the configuration of a pulse measurement system 1 according to an embodiment of the present invention.

[0018] The pulse measurement system 1 includes M (M is a natural number of 2 or more) measurement devices 31, . . . , 3 M and a pulse analyzer 5.

[0019] Measuring device 31,...,3 M The measuring devices 31, ..., 3 simultaneously measure the periodic contraction and / or relaxation of the internal organs of the subject 7 at different positions on the subject 7 to obtain measurement signals. M transmits the measurement signals to the pulse analyzer 5 via different channels.

[0020] The contraction and relaxation of the internal organs are, for example, the beating of the heart (heartbeat). M can be realized by, for example, the measurement devices 103 and 105 in FIG.

[0021] The pulse analysis device 5 can be realized by, for example, a computer that operates under the control of a program. The pulse analysis device 5 includes a storage device 11 and a processing device 13.

[0022] The storage device 11 can be realized by, for example, a storage device such as a memory, etc. The storage device 11 includes a measurement result storage unit 21, a post-smoothing storage unit 23, and a peak point storage unit 25.

[0023] The processing device 13 includes a measurement processing unit 31, a preprocessing unit 33, a parallel processing unit 35, and an analysis processing unit 36. The preprocessing unit 33 includes a smoothing processing unit 37, a tentatively detected peak processing unit 39, an overdetection processing unit 41, and an underdetection processing unit 43. The processing device and each processing unit can be realized, for example, by a processor operating under the control of a program.

[0024] The measurement processing unit 31 is connected to each of the measurement devices 31, . . . , 3 M The measurement result storage unit 21 stores time-series measurement raw data representing the measurement signals received from the respective channels.

[0025] FIG. 2 is a flow diagram showing an example of the operations of the preprocessing unit 33, the parallel processing unit 35, and the analysis processing unit 36 ​​in FIG.

[0026] FIG. 2(a) is a flow diagram showing the overall operation of the preprocessing unit 33, parallel processing unit 35, and analysis processing unit 36 ​​in FIG.

[0027] The pre-processing unit 33 performs pre-processing on each of the measuring devices 31, . . . , 3 M The measurement raw data is individually processed (step STA1). The peak point storage unit 25 stores peak point data that specifies the peak point group (a group of one or more peak points) obtained by the preprocessing.

[0028] The parallel processing unit 35 compares the measurement data of all channels, synchronizes the peaks of all channels, and acquires additional end point peaks (step STA2). The peak point storage unit 25 stores data related to synchronization and peak point data that identifies a peak point group including the additionally acquired peak points.

[0029] The parallel processing unit 35 determines whether the plateau ratio of any channel is 1 (step STA3). If the plateau ratio of any channel is 1, the process proceeds to step STA5. If the plateau ratios of all channels are not 1, the process proceeds to step STA4.

[0030] In step STA4, the parallel processing unit 35 performs parallel correction on all channels. The peak point storage unit 25 stores peak point data that identifies each peak point after parallel correction. The process proceeds to step STA5.

[0031] The peak point data stored in the peak point storage unit 25 at least identifies the time at which each peak point occurs. Therefore, the analysis processing unit 36 ​​can identify the intervals at which peak points occur over time (peak point intervals) using the peak point data. In step STA5, the analysis processing unit 36 ​​performs analysis using the peak points of the channel with a high plateau ratio. The analysis processing unit 36 ​​analyzes the intervals of contraction and / or relaxation of the internal organs of the subject 7, for example, using the intervals between peak points.

[0032] In this way, in the preprocessing (step STA1), the preprocessing unit 33 performs correction processing separately for the measurement results of each measurement device. The parallel processing unit 35 (steps STA2 to STA4) compares the measurement results of all the measurement devices and performs correction processing.

[0033] FIG. 2(b) is a flow diagram showing an example of pre-processing of step STA1.

[0034] The smoothing processor 37 performs smoothing processing on the raw measurement data of each channel to obtain time-series measurement data of each channel (step STB1). The post-smoothing memory 23 stores the measurement data of each channel.

[0035] The tentatively detected peak processing unit 39 acquires peak points for the measurement data of each channel (step STB2). The peak point storage unit 25 stores peak point data that specifies the peak point group of each channel.

[0036] The overdetection processing unit 41 determines whether the plateau ratio of each channel is 1 or not (step STB3). If the plateau ratios of all channels are 1, the preprocessing ends. If there are channels whose plateau ratios are not 1, the processing from step STB4 onwards is performed for the channels whose plateau ratios are not 1, and the processing from step STB4 onwards is not performed for channels whose plateau ratios are 1.

[0037] In step STB4, the overdetection processing unit 41 performs correction processing on the peak points that are overdetected in the channels whose plateau ratio is not 1. The peak point storage unit 25 stores peak point data that specifies the corrected peak point group for each channel.

[0038] The undetected signal processing unit 43 determines whether the plateau ratio of each channel is 1 or not (step STB5). If the plateau ratios of all channels are 1, the preprocessing is terminated. If there are channels whose plateau ratios are not 1, the processing of step STB6 is performed for the channels whose plateau ratios are not 1, and the processing of step STB6 is not performed for channels whose plateau ratios are 1.

[0039] In step STB6, the undetected processing unit 43 performs correction processing on undetected peak points of channels whose plateau ratios are not 1. The peak point storage unit 25 stores peak point data that specifies the corrected peak point group of each channel.

[0040] An example of specific processing by the pre-processing unit 33 in Fig. 1 will be described with reference to Fig. 3 to Fig. 12. In the following, the number of channels M is set to 2, and an example will be described in which waveform quality and accurate heartbeat interval measurement are attempted for a total of 64 data points measured from 16 subjects (four trials per subject) using the measuring devices 103 and 105 in Fig. 22.

[0041] Most conventional ballistocardiogram heart rate monitoring techniques focus on obtaining accurate heart rate data, but do not attempt to obtain accurate heartbeat intervals. In this example, the heartbeat interval is measured by continuously and accurately obtaining the wave corresponding to ventricular contraction (hereinafter referred to as the "J wave").

[0042] In raw measurement data, there are countless extreme points (maximum and / or minimum points). The smoothing process is a process that removes extreme points that do not affect the measurement of visceral contraction and / or relaxation, and reflects the influence of visceral contraction and / or relaxation in the remaining extreme points. In this example, the influence of ventricular contraction is reflected in the remaining extreme points.

[0043] The smoothing processor 37 applied a Gaussian filter as a smoothing process to remove small high-frequency vibrations and obtain appropriate maximum and minimum points. The number of points was determined by the plateau ratio. Note that other smoothing processes may be used as long as they reflect the influence of contraction and / or relaxation of the internal organs in the remaining peak points.

[0044] 3 is a graph showing measurement data obtained by smoothing the raw measurement data of one channel by the smoothing processor 37 of FIG. 1 (see step STB1). FIG. 3 is a ballistocardiogram. In FIG. 3, the horizontal axis represents time and the vertical axis represents amplitude [V]. Referring to FIG. 3, the measurement data includes local maximum points LX1, ..., LX 16 and the minimum points LN1,…,LN 15 exist alternately.

[0045] FIG. 4(a) is a diagram for explaining the waveform sharpness index, and FIG. 4(b) is a diagram showing an example of the waveform of the waveform sharpness index.

[0046] In ballistocardiograms, the J wave typically appears as the wave with the largest amplitude. Therefore, it is generally expected that the J wave can be analyzed by focusing on its amplitude. However, before and after the J wave, H waves, L waves, and N waves, which have the same polarity as the J wave, occur. The magnitude relationship between these amplitudes changes due to various factors during measurement. Therefore, focusing only on the amplitude will result in erroneous peak detection.

[0047] The waveform shape of the J wave is characterized by being sharper than the waveform shapes of other waves. Therefore, we introduce a waveform sharpness index as an index to indicate the amplitude characteristics of the waveform.

[0048] Specifically, referring to Figure 4(a), all the maximum and minimum points in the waveform of the measurement data are found. For a given maximum point (x[i+1], y[i+1]), there exists a minimum point (x[i], y[i]) just before it and a minimum point (x[i+2], y[i+2]) just after it.

[0049] The height is defined as the difference between the amplitude of the maximum (y[i+1]) and the average amplitude of the minimum points adjacent to the maximum ((y[i]+y[i+2]) / 2), i.e., height = y[i+1]-(y[i]+y[i+2]) / 2.

[0050] The width is the time span between the minimum points before and after the maximum point, i.e., width = x[i+2] - x[i].

[0051] The sharpness of a waveform is defined as the ratio of its height to its width, i.e., sharpness = height / width.

[0052] FIG. 4(b) shows the waveform of the sharpness index calculated at each maximum point in the measurement data in FIG. 3. The waveform of the sharpness index is obtained by arranging the sharpness indices calculated at each maximum point in time series and connecting them with a line. The sharpness index of the waveform can obtain a high value, which is characteristic of the J wave. In other words, the maximum value of the sharpness index waveform in FIG. 4(b) corresponds to the time of occurrence of the J wave. The tentatively detected peak processing unit 39 in FIG. 1 obtains the peak point by determining the maximum point of the sharpness index waveform. The peak point data can identify a group of peak points in the sharpness index waveform by the time at which the maximum point exists and the value of the sharpness index at the maximum point.

[0053] 5 and 6 show the results of peak point detection by the tentatively detected peak processing unit 39 in FIG. 1 when the quality of the measured raw data is good.

[0054] FIG. 5(a) is a graph showing a ballistocardiogram after smoothing. The horizontal axis is time, and the vertical axis is amplitude [mV]. FIG. 5(b) is a sharpness index obtained from the ballistocardiogram of FIG. 5(a), with circles marking the maximum points in the sharpness index waveform. The horizontal axis is time, and the vertical axis is the sharpness index. FIG. 5(c) is an electrocardiogram obtained simultaneously with the measured raw data. The horizontal axis is time, and the vertical axis is amplitude [mV].

[0055] Comparing Figure 5(b) and Figure 5(c), the maximum point in the sharpness index waveform in Figure 5(b) is perfectly synchronized with the peak of the ECG in Figure 5(c). Therefore, it can be evaluated that the time of occurrence of the J wave can be obtained from the maximum point in the sharpness index waveform in Figure 5(b).

[0056] Figure 6(a) shows the results of sorting the intervals between maximum points (peak intervals) in the waveform of the sharpness index. The waveform obtained by sorting in ascending order by peak interval is called a sorted waveform. The horizontal axis shows the order of the sorted peak interval data, and the vertical axis shows the peak interval [ms]. It can be seen that there is no significant change in the peak interval, and a constant interval is obtained.

[0057] Figure 6(b) shows the amount of change in the interval between each peak after sorting in Figure 6(a). The horizontal axis shows the order of each data point in the sorting, and the vertical axis shows the amount of change [ms] between the peaks before and after each sorted maximum point. It can be seen that the difference in the interval between the peaks before and after each sorted maximum point is small throughout the data.

[0058] 7 and 8 show the results of peak point detection by the tentatively detected peak processing unit 39 in FIG. 1 when the quality of the measured raw data is poor.

[0059] FIG. 7(a) is a graph showing a ballistocardiogram after smoothing. The horizontal axis is time, and the vertical axis is amplitude [mV]. FIG. 7(b) is a sharpness index obtained from the ballistocardiogram of FIG. 7(a), with circles marking the maximum points in the waveform of the sharpness index. The horizontal axis is time, and the vertical axis is the sharpness index. FIG. 7(c) is an electrocardiogram obtained simultaneously with the measured raw data. The horizontal axis is time, and the vertical axis is amplitude [mV].

[0060] In Figure 7(b), the peak is not immediately apparent. Therefore, even if the sharpness index is used, there will be some over-detection and under-detection. In fact, comparing Figure 7(b) and Figure 7(c), the maximum point in the sharpness index waveform in Figure 7(b) is not synchronized with the peak of the ECG in Figure 7(c).

[0061] Figure 8(a) shows the results of sorting the sharpness index waveform in ascending order by the distance between local maximum points (peak distance). The horizontal axis shows the order of each data point in the sort, and the vertical axis shows the peak distance [ms]. Sorting by peak distance reveals that there are areas where the distance changes significantly.

[0062] Figure 8(b) shows the amount of change in the interval between each peak after sorting in Figure 8(a). The horizontal axis shows the order of each data point in the sorting, and the vertical axis shows the amount of change between the interval between the peaks before and after [ms]. For each sorted maximum point, there are some that have a large change in the interval between the peaks before and after.

[0063] The overdetection processing unit 41 and the undetection processing unit 43 in Figure 1 use an index called the plateau ratio to perform processing to correct overdetection (mistakenly detecting the peak point of a wave that is not a J wave) and undetection (inability to detect the peak point of a J wave) peak points in cases such as those in Figures 7 and 8.

[0064] The plateau ratio is explained below. The intervals (JJ intervals) between the maximum points (peak points of the J wave) in the waveform of the sharpness index are sorted in ascending order and plotted (sort plot). In the sort plot, when the quality is good, a consistent interval is obtained over a wide range, whereas when the quality is poor, the range where a consistent interval is obtained is narrow due to false detection. The plateau ratio was introduced with an eye on this characteristic, and is a waveform evaluation index that indicates the consistency of the intervals between the maximum points in the waveform of the sharpness index.

[0065] Specifically, a threshold is determined for the amount of change before and after the sorted waveform using Equation 1. Here, Q3 is the third quartile (the data in the smallest 3 / 4 of the data sorted in ascending order), median is the median value, and mean is the average value. The plateau region (the region determined to be a plateau (flat)) is the range below this threshold. The number of intervals in the plateau region, N p can be determined by the number of intervals below this threshold.

[0066]

number

[0067] The plateau ratio is the number of intervals in the plateau region, N p This shows the ratio of the number of intervals N to the number of intervals N. In other words, the plateau ratio = N p / N.

[0068] However, since it is unlikely that the heartbeat intervals during this experiment in a sitting position would include intervals of less than 500 ms, if the plateau region includes intervals of less than 500 ms, the plateau ratio is left at 0. In addition, the median value of the intervals within the plateau region is extracted as the representative heartbeat interval.

[0069] Referring to Figure 6, the value PV1 in Figure 6(b) indicates the threshold calculated for the sorted waveform. The entire range is below the threshold, and the plateau region is the entire range. The range PA1 in Figure 6(a) indicates the plateau region. The plateau ratio is 1.

[0070] Referring to Figure 8, the value PV2 in Figure 8(b) indicates the threshold calculated for the sorted waveform. There are parts below the threshold and parts above it. The range PA2 in Figure 8(a) indicates the plateau region. The plateau ratio was 0.41.

[0071] In this way, the plateau ratio is evaluated as reflecting quality. That is, the maximum value of the plateau ratio is 1, and the larger the plateau ratio, the higher the quality can be evaluated.

[0072] Furthermore, by performing all of the processing steps up to this point for both polarities, the polarity with the highest plateau ratio and the number of points for the Gaussian filter are determined, and subsequent processing is performed using the data under the determined conditions.

[0073] If the plateau ratio of any channel was below 0.6, the data was deemed to be of insufficient quality for analysis and was therefore excluded from the study.

[0074] FIG. 9 is a diagram for explaining the processing by the overdetection processing unit 41 of FIG. 1. FIG. 9(a) is a graph showing a ballistocardiogram after smoothing processing. The horizontal axis is time, and the vertical axis is amplitude [mV]. FIG. 9(b) is a sharpness index obtained from the ballistocardiogram of FIG. 9(a), with circles marking the maximum points in the waveform of the sharpness index. The horizontal axis is time, and the vertical axis is the sharpness index. FIG. 9(c) is an electrocardiogram obtained simultaneously with the measured raw data. The horizontal axis is time, and the vertical axis is amplitude [mV].

[0075] Figure 10 shows the plateau region in the measurement of Figure 9. The lower limit of the peak interval in the plateau region is 985 ms.

[0076] The overdetection processor 41 removes peak points from overdetected locations that occur consecutively. The criterion for determining a short interval is an interval lower than the lower limit of the plateau region of the sorted results, and if such an interval occurs consecutively, the peak point that causes the short interval is removed. Region OS1 in Figure 9(b) indicates the range where overdetection occurs. In region OS1, due to overdetection of the target peak TP1, the intervals before and after it are 775 ms and 390 ms, respectively, so there are consecutive intervals lower than the lower limit of 985 ms for the peak interval. Therefore, region OS1 is the target of this process. The overdetection processor 41 removes the target peak TP1. As a result, the interval after removal is 1165 ms, which falls within the plateau range.

[0077] The overdetection processing unit 41 removes overdetected peak points from the peak point group extracted by the tentative detection peak processing unit 39, and sets the removed peak point data as new peak point data. The peak point storage unit 25 stores the new peak point data after the overdetected peak points have been removed.

[0078] If three or more short intervals occurred consecutively, the overdetected data that occurred earliest in chronological order was removed, and the remaining data were passed on to the next step without any processing. After this processing, the plateau ratio was calculated again, and data for which the plateau ratio was 1 in either channel were not subjected to further preprocessing (see step STB5 in Figure 2).

[0079] 11 and 12 are diagrams for explaining the processing by the undetection processing unit 43 in Fig. 1. The undetection processing unit 43 detects a plausible peak point for an undetected interval based on the sharpness index and information on the intervals before and after the undetected interval.

[0080] The undetection processing unit 43 defines an undetected interval as an interval that exceeds 1.27 times the representative heartbeat interval obtained from the sorted waveform. Here, "1.27 times" is determined based on the heartbeat interval (RR interval) obtained from the electrocardiogram, which is the reference signal.

[0081] The undetected interval processor 43 first determines the validity of the intervals before and after the undetected interval. The validity is also determined based on a magnification obtained from an electrocardiogram, and is conditioned on the range of 0.80 to 1.27 times the representative heartbeat interval.

[0082] 11 and 12 are diagrams for explaining the process of adding a new peak point when non-detection actually occurs and both the preceding and following intervals are within the range.

[0083] For each maximum point of the sharpness index (four points in FIG. 11) within the undetected range, the undetected processing unit 43 evaluates the two intervals that are divided when the point is set as a tentative peak point based on the valid intervals before and after it, and selects the point with the highest evaluation value as the new peak point.

[0084] Specifically, the valid interval before the undetected interval is defined as pre_interval, and the interval after is defined as post_interval. When the j-th maximum point is defined as a tentative peak point, the interval before the tentative peak point is defined as interval_1, and the interval after the tentative peak point is defined as interval_2. The undetection processing unit 43 evaluates interval_1 using equation 2. The undetection processing unit 43 evaluates interval_2 using equation 3. The undetection processing unit 43 calculates the evaluation value of the tentative peak point using equation 4, which is the summation formula. In FIG. 11, the evaluation value of the third peak point from the front is the highest, so this point is acquired as the new peak point, as shown in FIG. 12.

[0085]

number

[0086]

number

[0087]

number

[0088] If only one of the preceding and following points is a valid interval, the undetected point processor 43 determines a new peak point by calculating the peak value using only the information on the valid interval, rather than adding them up. If neither point is a valid interval, the undetected point processor 43 does not acquire a new peak point and proceeds to the next process.

[0089] The undetected point processor 43 repeats the process of adding new peak points until the plateau ratio and the representative heartbeat interval no longer change.

[0090] An example of specific processing by the parallel processing unit 35 in Fig. 1 will be described with reference to Fig. 13 to Fig. 17. Parallel processing unit 35 utilizes the fact that multiple channels are simultaneously measuring a ballistocardiogram, and performs corrections in parallel while mutually checking the intervals between the multiple channels. In the following, the number of channels is assumed to be two.

[0091] The parallel processing unit 35 first performs synchronization to determine which peak points of the two channels correspond to each other. FIG. 13 is a diagram for explaining the synchronization process. FIG. 13(a) shows the peak points obtained in channel CH1. FIG. 13(b) shows the peak points obtained in channel CH2. Note that FIG. 13 shows the peak points after processing for overdetected and undetected peak points has been performed.

[0092] The parallel processing unit 35 acquires the occurrence times of the first and second peak points of the channel that appears later than the first peak point. Next, the parallel processing unit 35 calculates the time difference between the first peak point and a predetermined number of peak points (10 in this example) of the other channel, and associates peak points with small differences. In FIG. 13, since the peak point of channel CH1 occurs later in time, the parallel processing unit 35 acquires times TA1 and TA2 of the first two peak points of channel CH1. Next, the parallel processing unit 35 calculates the difference between the first 10 peak points of channel CH2. As a result, the parallel processing unit 35 associates times TA1 and TA2 of channel CH1 with the times of the third and fourth peak points of channel CH2, respectively.

[0093] Next, the parallel processing unit 35 acquires additional endpoint peaks. In the example of FIG. 13, the first and second peak points of channel CH1 correspond to the third and fourth peak points of channel CH2, respectively. It can be seen that two peak points are present frequently in channel CH2. Therefore, if the two peak points present frequently in channel CH2 are not erroneous detections, it is highly likely that a similar peak point will be obtained in channel CH1, and therefore the parallel processing unit 35 acquires additional endpoint peaks in channel CH1.

[0094] The parallel processing unit 35 first calculates the rate of change around the interval for the channel with the larger number of peak points, and determines whether the interval is appropriate based on the change amount before and after the rate of change. The appropriateness is determined based on whether the rate of change exceeds 0.4. The value of "0.4" used as the criterion for judgment is set because it is unlikely that two consecutive interval changes that correspond to the rate of change standard of 0.2 determined based on an electrocardiogram would occur in a normal heartbeat pattern. If the peak points are not determined to be appropriate when the change amount is calculated, no additional acquisition is performed.

[0095] 13(b), the rate of change between the interval between the first and second peak points and the interval between the second and third peak points of channel CH2 is -0.065, the rate of change between the interval between the second and third peak points and the interval between the third and fourth peak points is 0.128, and the rate of change between the interval between the third and fourth peak points and the interval between the fourth and fifth peak points is 0.147. Therefore, the amount of change for the first peak point is 0.193, and the amount of change for the second peak point is 0.019. Since both are below 0.4, the parallel processing unit 35 determines that both the first and second peak points are valid.

[0096] If it is determined to be a valid peak point, the interval between the second and third peak points of channel CH2 is first set as pre_interval, and equation 2 is used to calculate the pre-peak value for the sharpness point before the first peak point of channel CH1, and the peak point with the highest value is determined as the tentative peak point. Similarly, additional data is acquired for the interval between the first and second peak points of channel CH2, and the rate and amount of change are calculated, including the tentative peak point. Only if the amount of change is below 0.4 is it determined to be a new peak point. If it exceeds 0.4, no additional data is acquired, and the first and second peak points of channel CH2 are removed to achieve synchronization.

[0097] The parallel processing unit 35 performs the parallel processing of step STA2 for the purpose of additional acquisition even when the plateau ratio is already 1. The peak point storage unit 25 stores data indicating the peak points of each channel after the processing for additional acquisition.

[0098] The parallel processing unit 35 does not perform the process of step STA4 because no correction is necessary when the plateau ratio of at least one of CH1 and CH2 is 1. In other words, there are also cases where the measurement interval in the ballistocardiogram is determined to be the interval obtained in step STA2.

[0099] In step STA4, the parallel processing unit 35 performs parallel correction on data for which the plateau ratios of both channels are not 1. There are two conditions for correction: one is that the amount of change in the rate of change before and after exceeds 0.4, as before, and the other is that the interval difference between both channels exceeds 80 ms.

[0100] 14 and 15 show examples where correction is required due to the amount of change between the front and rear.

[0101] Figure 14 shows the state before correction. Figures 14(a), (b), and (c) show the sharpness, rate of change, and amount of front-to-back change for channel CH1, respectively. Figures 14(d), (e), and (f) show the sharpness, rate of change, and amount of front-to-back change for channel CH2, respectively.

[0102] As shown in Figure 14(f), channel CH2 has a normal interval, and the change in the rate of change before and after is also less than 0.4. As shown in Figure 14(c), the eighth peak point of channel CH1 was erroneously detected, which also affected the change in the rate of change. In this case, the parallel processing unit 35 sets the synchronized interval of channel CH1 as pre_interval and attempts to obtain a new peak point using the value calculated using Equation 2. The rate of change and change in the rate of change are recalculated using the interval obtained here.

[0103] If the amount of change falls below 0.4, the parallel processing unit 35 performs a final check to confirm that the interval is within a range of 0.8 to 1.27 times the representative heartbeat interval and that the interval difference is within 80 ms. If these conditions are met, the parallel processing unit 35 confirms that the new peak point is detected. If these conditions are not met, there is a possibility that a sharpness point with a valid interval does not exist, or that the next peak point is a false positive. Therefore, the parallel processing unit 35 calculates the rate of change and amount of change for CH1 using the synchronization interval for channel CH2. If this is below 0.4, the parallel processing unit 35 determines that the former is true, assumes that the interval was achieved using channel CH2, and moves on to processing the next interval. If it is above 0.4, the parallel processing unit 35 determines that the latter is true, confirms the new peak point detected earlier, and then moves on to processing the next interval.

[0104] When the rate and amount of change are recalculated for a new peak point and the result exceeds 0.4, the same process is performed based on the two possibilities described above.

[0105] Figure 15 shows the eighth peak point of channel CH1 after correction. Figures 15(a), (b), and (c) show the sharpness, rate of change, and amount of change before and after correction for channel CH1, respectively. Figures 15(d), (e), and (f) show the sharpness, rate of change, and amount of change before and after correction for channel CH2, respectively. As shown in Figure 15(c), after correction, the value falls below 0.4, satisfying all conditions, and the peak point is confirmed.

[0106] Even when the front-to-back change amount is below 0.4 and is deemed appropriate, there are cases where correction is required, as shown in Figure 16. Figures 16(a), (b), and (c) respectively show the sharpness, rate of change, and front-to-back change amount for channel CH1 after correction. Figures 16(d), (e), and (f) respectively show the sharpness, rate of change, and front-to-back change amount for channel CH2. In this case, compared to CH2, which is correctly detected, erroneous detections occur continuously at almost equal intervals in CH1, so the erroneous detections cannot be recognized based on the change amount, and the normal peak points beyond them are judged to be erroneous detections.

[0107] To deal with cases like this, the difference in intervals between both channels has been added to the criteria for correction. Based on the idea that the minimum interval between H waves and J waves is 80 ms, if the interval difference exceeds 80 ms, it is possible that one of the channels is mistakenly detecting H waves or other waves. Therefore, for the interval on the channel with the largest difference from the representative heartbeat interval, the interval on the other channel is set as pre_interval, as before, and a new peak point is obtained using equation 2. Furthermore, after the new peak point is obtained, the rate and amount of change are recalculated and the conditions are confirmed, as before.

[0108] As shown in Figure 17, this process allows us to recognize that the second interval is a false positive. Once this point is corrected, we can automatically recognize that the next peak point is also a false positive.

[0109] In addition, the example shown here is a case where one of the channels has a valid interval, but if the change amount exceeds 0.4 at the same time in both channels, mutual correction is not possible, so correction is performed using the previous interval in each channel using Equation 2.

[0110] Through this process, the parallel processing unit 35 finally selects the one with the highest plateau ratio.

[0111] Figure 18 shows a scatter plot of the JJ interval (vertical axis) against the RR interval (horizontal axis) for all four trials of four subjects. Figure 18 shows a high correlation between the RR interval and the JJ interval, with the error rate remaining at most 2.68%.

[0112] 19 and 20 show scatter plots of the heart rate variability analysis index calculated from the JJ interval (vertical axis) against the heart rate variability analysis index calculated from the R-R interval (horizontal axis) for all four trials of 16 subjects. 19 and 20 show that a high correlation was obtained in the heart rate variability analysis, and the mean absolute error (MAE) and root mean square error (RMSE) were also small, suggesting that the JJ interval measured by this method collects information equivalent to that of the simultaneously measured R-R interval.

[0113] Therefore, for example, for normal measurement purposes, one can accumulate one's own heart rate information through daily continuous measurement using a commercially available blood pressure monitor, and understand the normal range. For abnormality detection purposes, one can compare the accumulated heart rate information and, if a deviation from the normal range is found, consult with one's family doctor or the like.

[0114] FIG. 21 is a block diagram showing an example of the configuration of a pulse measurement system 51 which is another example of an embodiment of the present invention.

[0115] The pulse measurement system 51 comprises a measurement device 53 and an interval analysis device 55 .

[0116] The pulse measurement system 51 has one measurement device 53 and the number of channels is 1. Therefore, the pulse measurement system 51 can basically be realized by omitting the processing of the parallel processing unit 35 in FIG.

[0117] The interval analyzer 55 comprises a storage device 61 and a processing device 63 .

[0118] Storage device 61 can be realized in the same manner as storage device 11 in Fig. 1. Storage device 61 includes a measurement result storage unit 71, a post-smoothing storage unit 73, and a peak point storage unit 75. Measurement result storage unit 71, post-smoothing storage unit 73, and peak point storage unit 75 are similar to measurement result storage unit 21, post-smoothing storage unit 23, and peak point storage unit 25 in Fig. 1, respectively.

[0119] The processing device 63 includes a measurement processing section 81, a preprocessing section 83, and an analysis processing section 86. The preprocessing section 83 includes a smoothing processing section 87, a tentatively detected peak processing section 89, an overdetection processing section 91, and an underdetection processing section 93.

[0120] Measurement processing section 81, preprocessing section 83, and analysis processing section 86 are similar to measurement processing section 31, preprocessing section 33, and analysis processing section 36 in FIG. 1, respectively.

[0121] The smoothing processing unit 87, the tentatively detected peak processing unit 89, the over-detection processing unit 91, and the undetection processing unit 93 are similar to the smoothing processing unit 37, the tentatively detected peak processing unit 39, the over-detection processing unit 41, and the undetection processing unit 43 in Figure 1, respectively.

[0122] Note that, for example, if a measurement device can perform measurements with high accuracy, there is no need to perform correction processing on the peak points obtained by the tentative detected peak processing unit. Therefore, the analysis processing unit can perform analysis processing on the peak points obtained by the tentative detected peak processing unit. The present invention can be understood as, for example, the tentative detected peak processing unit obtaining peak points from the maximum points of the waveform of the sharpness index, and the analysis processing unit analyzing the intervals of contraction and / or relaxation of the subject's internal organs based on the intervals between the obtained peak points.

[0123] Furthermore, whether to use the maximum point or the minimum point is a relative matter, and can be interpreted as the opposite. That is, the present invention can be interpreted as, for example, the tentatively detected peak processing unit obtaining peak points from minimum points in the waveform of the sharpness index, and the analysis processing unit analyzing the intervals of contraction and / or relaxation of the subject's internal organs based on the intervals between the obtained peak points.

[0124] Furthermore, two minimum points are adjacent to each maximum point, and two minimum points are adjacent to each maximum point. Therefore, the present invention can be understood by considering the maximum point and the minimum point together as "maximum points." That is, the present invention can be understood as a system in which the tentatively detected peak processing unit calculates an index for at least two maximum points in the waveform of measurement data obtained from the measurement signal from the amplitude of each maximum point and the amplitudes and occurrence times of the maximum points before and after each maximum point, and sets some or all of the maximum points in the waveform of the calculated index as peak points, and the analysis processing unit analyzes the intervals between the peak points to determine the contraction and / or relaxation intervals of the subject's internal organs.

[0125] Furthermore, the present invention can be understood as an overdetection processing unit removing some of the peak points so that the amount of change before and after in the sorted waveform (a waveform rearranged in order based on the spacing between peak points) is below a threshold, and an undetection processing unit adding peak points so that the amount of change before and after in the sorted waveform is below a threshold.

[0126] Furthermore, the present invention can be understood as a system in which the preprocessing unit individually processes the measurement signals obtained by each measuring device to obtain peak points corresponding to each measuring device, and the parallel processing unit synchronizes the peak points corresponding to each measuring device and maintains, adds, changes, and deletes the peak points corresponding to each measuring device by referring to peak points corresponding to other measuring devices. [Explanation of symbols]

[0127] 1. Pulse measurement system 3. Measuring equipment 5. Pulse analyzer 7 Person to be measured 11 Storage device 13 Processing equipment 21 Measurement result storage section 23 Smoothed memory section 25 Peak point memory section 31 Measurement processing section 33 Pretreatment section 35 Parallel Processing Unit 36 Analysis Processing Unit 37 Smoothing processing section 39 Tentatively detected peak processing section 41 Overdetection processing unit 43 Undetected processing unit 51 Pulse Measurement System 53 Measuring Equipment 55 Interval analyzer 57 Person to be measured 61 Storage device 63 Processing equipment 71 Measurement result storage section 73 Smoothed memory section 75 Peak point memory section 81 Measurement processing section 83 Pretreatment section 86 Analysis Processing Unit 87 Smoothing processing section 89 Tentatively detected peak processing section 91 Overdetection processing unit 93 Undetected processing unit 101 First Cushion 103 Measuring Equipment 105 Measuring Equipment 107 Second Cushion 109 Third Cushion

Claims

1. A pulse analysis device for analyzing a measurement signal obtained by measuring the periodic repetition of contraction and / or relaxation of an internal organ in a subject, The pulse analysis device includes a preprocessing unit and an analysis processing unit, the pre-processing unit includes a tentatively detected peak processing unit; the tentatively detected peak processing unit calculates an index from the amplitude of each of at least two extreme points in the waveform of the measurement data obtained from the measurement signal, and the amplitudes and occurrence times of extreme points before and after each of the extreme points, and determines some or all of the extreme points in the waveform of the calculated index as peak points; The analysis processing unit analyzes the intervals of contraction and / or relaxation of the internal organs using the intervals between peak points.

2. the pre-processing unit includes an over-detection processing unit and an under-detection processing unit, the overdetection processing unit excludes some of the peak points so that an amount of change before and after the waveform rearranged in order according to the interval between the peak points is equal to or less than a threshold value; The pulse analyzer according to claim 1 , wherein the undetected processing unit adds peak points so that the amount of change before and after the waveform, which has been rearranged in order based on the interval between peak points, is equal to or less than a threshold value.

3. There are a plurality of the measuring devices, The pulse analyzer includes a parallel processing unit, each measuring device measures the periodic repetition of contraction and / or relaxation of the internal organs of the subject at a different position on the subject to obtain a measurement signal; the pre-processing unit processes the measurement signals obtained by each of the measurement devices individually to obtain peak points corresponding to each of the measurement devices; The pulse analysis device of claim 1, wherein the parallel processing unit synchronizes peak points corresponding to each of the measurement devices and maintains, adds, changes, and deletes peak points corresponding to each of the measurement devices by referring to peak points corresponding to other measurement devices.

4. A pulse analysis method in a pulse analysis device, in which a measurement device measures the periodic repetition of contraction and / or relaxation of internal organs in a subject and analyzes a measurement signal obtained by the measurement device, comprising: The pulse analyzer includes a preprocessing unit and an analysis processing unit, the pre-processing unit includes a tentatively detected peak processing unit; the tentatively detected peak processing unit calculates an index from the amplitude of each of at least two extreme points in the waveform of the measurement data obtained from the measurement signal, and the amplitudes and occurrence times of extreme points before and after each of the extreme points, and sets some or all of the extreme points in the waveform of the calculated index as peak points; A pulsation analysis method including a step in which the analysis processing unit analyzes the intervals of contraction and / or relaxation of the internal organs using the intervals between peak points.

5. A program for causing a computer to function as the pulse analyzer of claim 1.

Images