JP7520670B2 - Biological information processing device, biological information processing method, and biological information processing program - Google Patents

Description

The present invention relates to a biometric information processing device, a biometric information processing method, and a biometric information processing program.

In the medical field, doctors use electrocardiograms to measure the condition of a patient's heart. In addition to regular electrocardiograms that record electrocardiograms for a few minutes, there are also Holter electrocardiograms that can record electrocardiograms over long periods of time (e.g., 1 to 14 days). Holter electrocardiogram tests using a Holter electrocardiogram are performed to record electrocardiogram waveforms over long periods of time throughout daily life, such as during sleep and exercise.

Some bioinformation processing devices that process and display electrocardiogram waveforms have the ability to detect heartbeats from collected electrocardiograms and classify each heartbeat as an arrhythmia. However, because there is a possibility that noise picked up by the electrodes attached to the patient may be included in the electrocardiogram waveform or that the bioinformation processing device may not properly classify the arrhythmia, currently, a medical technician visually checks each heartbeat waveform before the doctor makes a diagnosis. Typically, an electrocardiogram contains a huge number of heartbeat waveforms, so checking the heartbeat waveforms is a very burdensome task for medical technicians.

In relation to this, for example, Patent Document 1 listed below discloses a technology in which a user such as a doctor or technician can specify an area on a heart rate trend graph and collectively change the arrhythmia classification for multiple heart rates included in that area to a different arrhythmia classification.

JP 2007-357 A

However, while the technology in Patent Document 1 allows the user to change the classification of arrhythmia, the user still needs to check graphs such as trendgrams and Lorenz plots of a large number of classified heart rate waveforms. Therefore, there is still the problem that this places a heavy burden on the user.

The present invention has been made to solve the above-mentioned problems. Therefore, the main object of the present invention is to provide a biometric information processing device, a biometric information processing method, and a biometric information processing program that can reduce the user's work of checking classified graphs.

The above object of the present invention is achieved by the following:

The biometric information processing device has an acquisition unit, a detection unit, a calculation unit, a generation unit, and an output unit. The acquisition unit acquires biometric information of a living organism. The detection unit detects beats from the biometric information acquired by the acquisition unit. The calculation unit calculates the interval between successive beats detected by the detection unit. The generation unit generates information based on a variation in the interval for each measurement period obtained by dividing an entire measurement period of beats into a plurality of periods. The output unit outputs at least one of the information based on the variation and information related to the measurement period corresponding to the information for each measurement period.
In one aspect, the generation unit calculates a correlation coefficient related to the variation for each measurement period as the information based on the variation.
In another aspect, the information based on the variation includes at least one of the shape of the points when the variation is plotted on a graph and the area of the points when the variation is plotted on a graph.
In another aspect, the output section changes an output form of at least one of the pieces of information depending on a magnitude of the information based on the variation.
In yet another aspect, the generating unit generates information on the shape, area, or density of a point cloud when a plurality of the beats in the measurement period are plotted as the information based on the variation, and the output unit classifies the plurality of the measurement periods according to the information on the shape, area, or density, and outputs at least one of the information based on the classification result.

In addition, the biometric information processing method includes the steps of: (a) acquiring biometric information of a living body; (b) detecting beats from the biometric information acquired in the step (a); (c) calculating an interval between successive beats detected in the step (b); (d) generating information based on a variation in the intervals for each measurement period obtained by dividing a total measurement period of the beats into a plurality of periods; and (e) outputting, for each measurement period, at least one of the information based on the variation and information related to the measurement period corresponding to the information.
In one aspect, in the step (d), a correlation coefficient relating to the variation is calculated for each of the measurement periods as the information based on the variation.
In another aspect, in the step (e), an output form of at least one of the pieces of information is changed depending on the magnitude of the information based on the variation.

The biometric information processing program also causes a computer to execute the processes included in the biometric information processing method.

According to the present invention, at least one of information based on the variation in beat intervals and information related to the measurement period is output for each measurement period, so the user can easily check the magnitude of the variation in beats for each measurement period. This reduces the user's workload in checking classified graphs.

1 is a block diagram showing a schematic hardware configuration of a biological information processing system according to a first embodiment. 2 is a functional block diagram illustrating main functions of the biometric information processing apparatus shown in FIG. 1; 4 is a flowchart for explaining a processing procedure of a biological information processing method according to the first embodiment. FIG. 2 is a schematic diagram illustrating an example of an outline of an electrocardiogram waveform. FIG. 13 is a diagram illustrating an example of calculation results of heartbeat intervals for each measurement period. 1 is a graph illustrating a Lorentz plot of sinus rhythm. 1 is a graph illustrating a Lorentz plot of atrial fibrillation. FIG. 13 is a schematic diagram illustrating, as a comparative example, a list display of Lorenz plots for each measurement period. FIG. 13 is a schematic diagram illustrating a list display of Lorenz plots sorted according to correlation coefficients. FIG. 13 is a schematic diagram illustrating a list display in which the symbols of the measurement periods are added to the rearranged Lorenz plots. 13 is a schematic diagram illustrating an example of a list display of start times for each measurement period as information relating to the measurement period. FIG. 11 is a schematic diagram illustrating an example of a list display of measurement period codes as information relating to the measurement period; FIG. FIG. 13 is a schematic diagram illustrating a list display to which correlation coefficients are added as information regarding variation. FIG. 13 is a schematic diagram illustrating a case where information based on variation or information related to a measurement period is identified by changing the background color for display. FIG. 13 is a schematic diagram illustrating a case where identification information is added to a list display. 1 is a graph showing time series changes in the RR interval of heartbeats in terms of beat rate (density). 1 is a graph showing time series changes in the RR interval of heartbeats as points. 10 is a flowchart illustrating a processing procedure of a biological information processing method according to a second embodiment. 1 is a graph illustrating a Lorentz plot of tachycardia. 1 is a graph illustrating a Lorentz plot of atrial flutter. 1 is a graph illustrating a Lorenz plot including VPC compensated sinus arrhythmia. 1 is a graph illustrating a Lorentz plot including VT and the like. 1 is a graph illustrating an example of Lorenz plots during normal sleep and during sleep when sleep apnea syndrome occurs. FIG. 13 is a schematic diagram illustrating a list display of Lorenz plots classified according to the shapes of point clouds.

Below, an embodiment of the present invention will be described with reference to the attached drawings. Note that the same reference numerals are used for the same components in the drawings. Also, the dimensional ratios in the drawings have been exaggerated for the convenience of explanation and may differ from the actual ratios.

First Embodiment
Fig. 1 is a block diagram showing a schematic hardware configuration of a biological information processing system 100 according to the first embodiment. As shown in Fig. 1, the biological information processing system 100 includes a biological information processing device 110, an output device 120, and an input device 130.

The bio-information processing device 110 has a function of detecting beats (heartbeats or pulses) from bio-information (e.g., electrocardiogram or pulse waves) collected from a patient (living body), and outputting and/or classifying the magnitude of beat variation for each measurement period so that it is easy to check. For example, it is known that patients with certain cardiac diseases have greater variation in beat intervals for a certain number of beats in a chronologically uniform sequence compared to healthy individuals. Therefore, by checking the output results or classification results from the bio-information processing device 110, the user can easily identify measurement periods with large variations in beat intervals, reducing the effort required for the user to search for measurement periods suspected of cardiac disease. The user is, for example, a medical professional such as a doctor or a medical technician.

Note that the following describes an example in which an electrocardiogram is acquired as biometric information and the heart rate is detected from the electrocardiogram, but this embodiment is not limited to this example and can also be applied to a case in which a pulse wave signal is acquired as biometric information from a pulse oximeter or the like and the pulse rate is detected from the pulse wave signal.

The bioinformation processing system 100 of this embodiment may be, for example, a dedicated medical device (such as a bioinformation monitor) that can display heart rate waveforms, trend graphs and lists of bioinformation, etc. Furthermore, the bioinformation processing device 110 may be, for example, a personal computer, a smartphone, a tablet terminal, etc., on which a bioinformation processing program for executing the bioinformation processing method described below is installed. Furthermore, the bioinformation processing device 110 may be a wearable device that is attached to the user's body (for example, the arm, head, etc.).

The biometric information processing device 110 has a processor 101, a memory 102, a storage device 103, and a communication control device 104. The memory 102 has a ROM (Read Only Memory) and a RAM (Random Access Memory). Various programs, parameters, etc. are stored in the ROM. The RAM also has a work area in which various programs executed by the processor 101 are stored. The processor 101 is configured to load a specified program from the various programs stored in the ROM and storage device 103 onto the RAM, and to execute various processes by working together with the RAM.

The storage device 103 has storage such as a hard disk drive (HDD) or a solid state drive (SSD). The storage device 103 is configured to store a biometric information processing program and various data.

The storage device 103 can also store electrocardiogram data as bioinformation. An electrocardiogram includes multiple heartbeat waveforms that occur continuously on a time axis, information on the measurement time (time), and the like. A heartbeat waveform is a waveform that indicates a heartbeat, i.e., the beating of the heart. The electrocardiogram data is acquired by an electrocardiograph (e.g., a Holter monitor) and provided to the bioinformation processing device 110. The electrocardiogram data can include, for example, 14 days' worth of electrocardiograms of a patient.

The electrocardiograph also includes a posture sensor that measures the patient's posture, and provides information about the patient's posture to the bio-information processing device 110 along with the electrocardiogram data.

The communication control device 104 is configured to connect the biometric information processing device 110 to a communication network 140. Specifically, the communication control device 120 includes processing circuits for various interfaces for communicating with external devices such as a server (computer) via the communication network 140, and is configured to conform to a communication standard for communicating via the communication network 140. Here, the communication network 140 is a LAN (Local Area Network), a WAN (Wide Area Network), the Internet, or the like.

The biometric information processing device 110 can be configured to acquire the patient's electrocardiogram data and information regarding the body position from, for example, a server located on the communication network 140 via the communication control device 120.

The output device 120 has a display. The display displays the display data received from the bio-information processing device 110 on a screen. The display may be a liquid crystal display, an organic EL display, or the like. The display may also be a display device such as a transmissive or non-transmissive head-mounted display that is worn on the user's head. Note that while FIG. 1 illustrates an example in which the bio-information processing device 110 and the output device 120 are separate devices, this is not limited to the case, and the display of the output device 120 and the bio-information processing device 110 may be integrated.

The output device 120 also has a printer, which prints the printing data received from the biometric information processing device 110 onto paper.

The input device 130 is configured to receive an input operation from a user who operates the biometric information processing device 110, and to generate an instruction signal corresponding to the input operation. The input device 130 includes, for example, a touch panel superimposed on the display of the output device 120, operation buttons attached to the housing, a mouse, a keyboard, and the like. The instruction signal generated by the input device 130 is transmitted to the biometric information processing device 110. The biometric information processing device 110 executes a predetermined process in response to the instruction signal.

The input device 130 also includes a device for reading a storage medium such as a USB memory. The electrocardiogram data and information related to body position acquired by the electrocardiograph are stored in the storage medium and provided to the user. The input device 130 reads the electrocardiogram data and information related to body position stored in the storage medium and stores them in the storage device 103.

In addition, if the electrocardiograph and the bio-information processing device 110 can be connected by a wired or wireless connection means, the electrocardiograph transmits electrocardiogram data and information regarding body position to the bio-information processing device 110, and the bio-information processing device 110 receives the electrocardiogram data and information regarding body position and stores them in the storage device 103.

The processor 101 of the biometric information processing device 110 expands the biometric information processing program on the RAM and executes the biometric information processing program to control each part of the biometric information processing device 110 and realize various functions.

<Biometric information processing method>
Fig. 2 is a functional block diagram illustrating main functions of the biological information processing device 110 shown in Fig. 1, and Fig. 3 is a flowchart for explaining the processing procedure of the biological information processing method according to the first embodiment. Fig. 4 is a schematic diagram illustrating an outline of an electrocardiogram waveform, and Fig. 5 is a diagram illustrating the calculation results of the heartbeat interval for each measurement period.

As shown in FIG. 2, the biometric information processing device 110 functions as an acquisition unit 111, a detection unit 112, a calculation unit 113, a generation unit 114, and an output unit 115. Details of these functional units and the details of the processing procedure of the biometric information processing method using these functional units are described below. Each functional unit and its control process are realized by the processor 101 of the biometric information processing device 110 executing a biometric information processing program.

As shown in FIG. 3, first, the patient's electrocardiogram is acquired (step S101). The acquisition unit 111 acquires information about the patient's electrocardiogram and posture from the electrocardiograph, for example, via a USB memory. The acquired electrocardiogram data and information about posture are stored in the storage device 103.

Next, the heartbeat is detected from the electrocardiogram acquired in step S101 (step S102). The detection unit 112 detects the heartbeat from the electrocardiogram acquired by the acquisition unit 111. As shown in FIG. 4, the electrocardiogram includes a plurality of heartbeat waveforms each having an R wave as a peak. The detection unit 112 includes, for example, a peak detection unit that detects the peak value of the electrocardiogram waveform, and detects a wave in the electrocardiogram whose height exceeds a predetermined threshold as an R wave of the heartbeat.

Next, the interval between successive heartbeats detected in step S102 is calculated (step S103). The calculation unit 113 calculates, for the entire measurement period of the heartbeat (the period during which the detection unit 112 detects heartbeats in one measurement), the interval RRn between the reference beat and the forward beat, which is the heartbeat immediately preceding the reference beat, and the interval RRn+1 between the reference beat and the backward beat, which is the heartbeat immediately following the reference beat. An example of the calculation result of the interval RRn and the interval RRn+1 is shown in FIG. 5. In FIG. 5, the entire measurement period of the heartbeat is divided into a number of measurement periods T1 to Tn. The measurement periods T1 to Tn are periods that include a chronologically-series number of heartbeats (for example, about 30 minutes in terms of time, about 2000 beats in terms of beat rate).

Usually, after detecting the patient's heartbeat, the heartbeat data for the entire measurement period is automatically classified according to the type of arrhythmia. As a result, the heartbeat data can be labeled with the appropriate arrhythmia, such as normal, atrial fibrillation, tachycardia, atrial flutter, etc. However, automatic classification is not perfect, and there are cases where classification is incorrect or classification is not possible due to noise picked up by the electrodes during electrocardiogram measurement (for example, the data surrounded by a thick line in the figure). Detected noise can be automatically removed, but incorrectly classified or unclassified heartbeats must be reclassified by the user.

Next, information based on the variation in the heartbeat intervals is generated (step S104). The generation unit 114 generates information based on the variation in the heartbeat intervals (hereinafter referred to as "variation information") for each measurement period T1 to Tn. The variation information can be, for example, an index (hereinafter simply referred to as "index") for evaluating the magnitude of the variation in the heartbeat intervals.

The generation unit 114 calculates an index using, for example, the interval RRn and the interval RRn+1 for each measurement period T1 to Tn. This index can be, for example, correlation coefficients S1 to Sn for the interval RRn and the interval RRn+1, the area of the point cloud when the interval RRn and the interval RRn+1 are plotted on a graph, the density of the point cloud, the shape of the point cloud, etc. The graph can be, for example, a Lorentz plot.

For example, if a patient's heart is in atrial fibrillation, the variation in the intervals between heartbeats will be greater, and the correlation coefficient will be smaller than if the patient's heart is in sinus rhythm, which means that the heart is beating at a normal rhythm.

Figure 6 is a graph illustrating a Lorentz plot of sinus rhythm, and Figure 7 is a graph illustrating a Lorentz plot of atrial fibrillation. As shown in Figure 6, assuming a Lorentz plot with intervals RRn and RRn+1 on the vertical and horizontal axes, respectively, in sinus rhythm, each point is distributed roughly along the 45-degree line L1 in the figure. In the example of Figure 6, the measurement targets are the patient's heart rate when awake and asleep, so the heart rate intervals can be short (e.g., during exercise) and long (e.g., during sleep). Therefore, each point is distributed along L1 from a portion where the heart rate interval is short and close to the origin to a portion where the heart rate interval is long and far from the origin. In contrast, as shown in Figure 7, in atrial fibrillation, the shape has a spread in the direction intersecting with the 45-degree line L2 in the figure.

When calculating the area of the point cloud as an index, the generation unit 114 assumes a Lorentz plot similar to that shown in Figures 6 and 7, for example, and calculates the area of the point cloud in the Lorentz plot. More specifically, the generation unit 114 calculates the area of the point cloud as an index by counting the number of points that do not overlap with each other, i.e., are not plotted at the same position, with each point at coordinates (RRn, RRn+1) being an area of area 1. Therefore, since this process is equivalent to the process of adding up the number of points with different coordinates (RRn, RRn+1), there is no need to display a graph on the display, but if a Lorentz plot were displayed, the index would correspond to the area of the part covered by the point cloud (see, for example, Figures 6 and 7, etc.).

Note that in Figures 6 and 7, heat maps are displayed according to the density of the plotted points (the same applies to the Lorentz plots in Figures 8 to 10, 13 to 15, and 19 to 24). In these figures, the density is highest in the dark parts in the center of the point cloud, medium density in the light-colored parts around it, and lowest density in the dark parts on the periphery of the point cloud. However, the shade of the heat map represents the relative density within one heat map, but does not represent the relative relationship of density between other heat maps. In other words, the same shade in two heat maps does not necessarily represent the same level of density.

Note that Figures 6 and 7 are heat maps that express the density of points using grayscale shading, but heat maps that express the density of points are not limited to this, and for example, the density of points can also be expressed according to changes in color. In other words, it is sufficient to display a heat map that uses any expression that can express the density of a point cloud (changes in color, changes in shading, changes in brightness, changes in saturation, etc.).

As for the density of the point cloud, for example, in FIG. 6, high-density point cloud areas are concentrated near the 45-degree line L1, whereas in FIG. 7, medium-density point cloud areas are distributed widely near the 45-degree line L2 and its surrounding areas. The density of the point cloud is highest on the 45-degree line L2 and decreases the further away from the 45-degree line L2. Therefore, the density near the 45-degree line on the Lorenz plot can be used as an index.

In addition, the generation unit 114 can obtain information related to the patient's body position and measurement time during heartbeat measurement, and generate variation information based on the information by excluding heartbeats that do not satisfy at least any of the conditions regarding the patient's body position, the conditions regarding the time period of measurement, and the conditions regarding the shape and interval of the heartbeat.

For example, when generating variability information during wakefulness, it is determined whether the following conditions are met: the patient is in a standing or sitting position; the patient is not usually asleep; and the shape and interval of the heartbeats meet certain conditions. In this case, the output unit 115 uses information about the patient's position and measurement time to exclude, for example, heartbeats measured in the supine, prone, or lateral position, and heartbeats measured during times when the patient is asleep (nighttime). It also excludes heartbeats whose shape and interval do not meet the specified conditions. This allows variability information to be generated using heartbeat data that meets the conditions desired by the user.

Next, at least one of the variance information and information related to the measurement period corresponding to the variance information (hereinafter simply referred to as "measurement period information") (hereinafter referred to as "output information") is output for each measurement period (step S105).

As a comparative example, FIG. 8 is a schematic diagram illustrating a list display of Lorenz plots for each measurement period, FIG. 9 is a schematic diagram illustrating a list display of Lorenz plots sorted according to correlation coefficients, and FIG. 10 is a schematic diagram illustrating a list display in which the sign of the measurement period is added to the sorted Lorenz plots.

In Figure 8, using conventional technology, the patient's heart rate is detected for the entire measurement period from 03:12 on the 15th to 05:42 on the 16th, and based on the detection results, a Lorentz plot of the heart rate for each measurement period is created and displayed in a list. The entire measurement period includes measurement periods T1 to T54, but due to space limitations, Figure 8 only shows periods T1 to T36 of the entire measurement period. Each measurement period is 30 minutes long, and the measurement start time for each measurement period is displayed at the top of each Lorentz plot. Also, in Figure 8, for ease of explanation, the measurement periods are labeled T1 to T36 to distinguish them from one another.

In this embodiment, the output unit 115 changes the output format of the output information depending on the magnitude of the correlation coefficient. More specifically, the output unit 115 generates display data in which the order of the variance information and/or the measurement period information for the measurement periods T1 to T54 is rearranged in ascending or descending order of the correlation coefficient, for example, and transmits the data to the output device 120 to display the data on the display. The output unit 115 can also generate print data in which the order of the variance information and/or the measurement period information for the measurement periods T1 to T54 is rearranged in ascending or descending order of the correlation coefficient, and transmits the data to the output device 120 to have the printer print the data.

For example, as shown in FIG. 9, the output unit 115 displays on the display of the output device 120 display data in which the Lorentz plots for measurement periods T1 to T54 are sorted in ascending order of correlation coefficient. Note that the measurement start time for each measurement period is displayed above each Lorentz plot as measurement period information. Therefore, after sorting by correlation coefficient, the order may differ from the chronological order, but the user can understand the correspondence between the displayed Lorentz plots for each measurement period and the time series by the measurement start time.

As described above, typically, after the patient's heartbeat is detected, the heartbeat data is automatically classified according to the type of arrhythmia. In this embodiment, the output unit 115 displays the data on the display in ascending or descending order of correlation coefficients. Therefore, if the heartbeat corresponding to atrial fibrillation is classified, the measurement periods in the two left columns after sorting in ascending order of correlation coefficients as shown in FIG. 9 are presumed to be arrhythmias of atrial fibrillation. This allows the user to immediately determine that the arrhythmias in the measurement periods in the above two columns correspond to atrial fibrillation without checking the results of the moving classification, making it possible to respond quickly to the arrhythmias in the measurement periods in these two columns.

For example, when a user creates a report on an electrocardiogram test, the user can select the above two columns (e.g., by touching a touch panel) to copy the Lorenz plots for these two columns together and include them in the "Findings" section of the report. This reduces the effort and time required for the user to create a report. The report can also include a graph (e.g., see Figure 16) that plots the change over time in the RR intervals that correspond to the Lorenz plots.

In addition, as shown in FIG. 10, the output unit 115 can display the Lorentz plots for each measurement period with the symbols T1 to T54. This makes it easier to understand the correspondence between the displayed Lorentz plots for each measurement period and the time series.

Fig. 11 is a schematic diagram illustrating a list display of the start times of each measurement period as measurement period information, and Fig. 12 is a schematic diagram illustrating a list display of the codes of the measurement periods as measurement period information. The measurement period information includes the time or time period when multiple heart rate measurements corresponding to the variability information were performed, or codes for distinguishing multiple measurement periods from one another. Fig. 13 is a schematic diagram illustrating a list display to which a correlation coefficient is added as information regarding variability, Fig. 14 is a schematic diagram illustrating a case where the background color is displayed in a color different from the color in which the variability information or measurement period information is displayed, and Fig. 15 is a schematic diagram illustrating a case where identification information is added to the list display.

If the user does not need to check the Lorentz plot, as shown in Figure 11, the Lorentz plot may not be displayed, and the measurement start times may be displayed as measurement period information for measurement periods T1 to T54 in ascending or descending order of correlation coefficients. Figure 11 shows an example of a case where the start times for each measurement period are displayed in ascending order of correlation coefficients (due to space limitations, only the measurement start times for some of the measurement periods are shown).

In addition, if the user does not need to check the Lorentz plot, as shown in FIG. 12, the Lorentz plot may not be displayed, and the measurement period information may be configured to display the symbols of the measurement periods T1 to T54 in ascending or descending order of correlation coefficient. FIG. 12 shows an example of a case where the symbols of the measurement periods T1 to T54 are displayed in ascending order of correlation coefficient (due to space limitations, only the symbols of some of the measurement periods are shown).

As shown in FIG. 13, each Lorenz plot shown in FIG. 9 can be configured to have a correlation coefficient value added as variation information. The user can check the correlation coefficient value, making it easier to understand the magnitude of variation in the heartbeat intervals during each measurement period.

The output unit 115 can also display the background color of the information in a color different from the color in which the variance information or the measurement period information is displayed, without changing the order of the Lorenz plots for each measurement period from the chronological order shown in FIG. 8. For example, as shown in FIG. 14, the output unit 115 displays the background color of the measurement start time in a color different from the color in which the measurement start time is displayed, depending on the magnitude of the correlation coefficient. More specifically, the output unit 115 displays the background color of the measurement start time in red when the magnitude of the correlation coefficient is smaller than a predetermined first threshold, in yellow when the magnitude of the correlation coefficient is smaller than a predetermined second threshold and equal to or greater than the first threshold, and in green when the magnitude of the correlation coefficient is equal to or greater than the second threshold. As a result, the order of the Lorenz plots for each measurement period displayed on the display does not change, but the user can grasp at a glance the magnitude of the correlation coefficient, i.e., the magnitude of the variance in the heartbeat interval.

The output unit 115 can also assign identification information for identifying the variance information or the measurement period information to the information without changing the order of the Lorenz plots for each measurement period from the chronological order shown in FIG. 8. For example, as shown in FIG. 15, the output unit 115 displays the Lorenz plots for the measurement periods in which the correlation coefficient is smaller than a predetermined third threshold value by enclosing them in a square frame mark. This allows the user to easily identify the magnitude of the variance in the heart rate by relying on the mark. Note that the mark is not limited to a square frame, and may have any other shape that is easily identifiable. The magnitude of the correlation coefficient can also be divided into multiple stages, and marks of different forms (shapes and colors) can be displayed according to the stage.

The above describes an example of displaying a point cloud on a Lorentz plot, but the graph is not limited to a Lorentz plot, and any graph that can display the variation in heartbeat intervals can be applied.

As shown in FIG. 16, when the RR interval and time of the heartbeat are plotted on the vertical and horizontal axes, respectively, and the number of beats (density) corresponding to the RR interval is displayed, in section (A) where the density of points is low, the RR intervals vary widely, and the correlation coefficient for the intervals RRn and RRn+1 is small. On the other hand, in sections where the density of points is higher than section (A), the RR intervals vary less, and the correlation coefficient is high.

As shown in FIG. 17, when the RR intervals of the heartbeats and the time are plotted on the vertical and horizontal axes, respectively, and the points corresponding to the RR intervals are displayed one by one, the RR intervals in section (B) have a large variation, and the correlation coefficients for the intervals RRn and RRn+1 are small. On the other hand, in sections where the points are concentrated, the RR intervals have a small variation and the correlation coefficients are high.

The output unit 115 can, for example, display the background color of the variation information or the measurement period information in a color different from the color in which the information is displayed. Alternatively, it can provide identification information for identifying the variation information or the measurement period information.

Figure 16 shows an example in which an "*" is added to the measurement start time included in the low density section (A). Also, Figure 17 shows an example in which the background color of the measurement start time included in the section (B) where the dots are scattered is displayed in a color different from the color in which the measurement start time is displayed. Therefore, the user can infer that some kind of arrhythmia is occurring in the sections (A) and (B) where the dots are scattered widely, and that the sections with small scattering, i.e., high density, correspond to sinus rhythm.

According to the bio-information processing device 110 of this embodiment described above, at least one of information on the variation in heartbeat intervals and information on the measurement period is output for each measurement period. More specifically, the output form of the output information is varied according to an index for evaluating the magnitude of the variation in heartbeat intervals, so that the user can easily check the magnitude of the variation in heartbeats for each measurement period. This reduces the user's workload of checking classified graphs.

Second Embodiment
In the second embodiment, a case will be described in which a plurality of measurement periods is classified using the above-mentioned index or information on the shape of a group of points when a plurality of heartbeats during a measurement period are plotted on a graph as variation information, and output information is displayed based on the classification results. Note that in the following description, detailed description of the same configuration as that of the first embodiment will be omitted.

Fig. 18 is a flowchart for explaining the processing procedure of the biological information processing method according to the second embodiment. Figs. 19 and 20 are graphs illustrating Lorentz plots of tachycardia and atrial flutter, respectively, and Figs. 21 and 22 are graphs illustrating Lorentz plots including VPC (Ventricular Premature Contraction) compensated sinus arrhythmia and VT (Ventricular Tachycardia), etc., respectively. Fig. 23 is a graph illustrating Lorentz plots during normal sleep and during sleep with sleep apnea syndrome. Fig. 24 is a schematic diagram illustrating a list display of Lorentz plots classified according to the shape of the point cloud.

In FIG. 18, the processing in steps S201 to S203 is the same as the processing in steps S101 to S103 in the first embodiment, so detailed explanation is omitted.

In step S204, information on the variation in the heartbeat intervals is generated. In this embodiment, the variation information can be an index, or information on the shape of a point cloud when multiple heartbeats during the measurement period are plotted on a graph. Below, a typical shape of a point cloud is described assuming a Lorentz plot in which the vertical and horizontal axes represent the intervals RRn and RRn+1, respectively, as the graph.

As described with reference to FIG. 6 in the first embodiment, in sinus rhythm, the points are distributed generally along the 45-degree line L1 in the figure. Therefore, the point cloud has a band-like (string-like) shape along L1. In contrast, as described with reference to FIG. 7, in atrial fibrillation, the point cloud has a shape that spreads in the direction intersecting with the 45-degree line L2 in the figure. More specifically, the point cloud has a triangular shape that spreads less near the origin on L2 and spreads more as it moves away from the origin.

Also, as shown in FIG. 19, in tachycardia, the points are distributed roughly along the 45-degree line L3 in the figure, as in the case of sinus rhythm (FIG. 6), but compared to the case of sinus rhythm, they are generally distributed in an area closer to the origin. This is because in the case of tachycardia, the interval between heartbeats tends to be shorter than in the case of sinus rhythm. Therefore, the point cloud has a band-like (string-like) shape that follows L3. The length of the band (string) is shorter than that of the point cloud in sinus rhythm.

Also, as shown in FIG. 20, in atrial flutter, the point cloud has a shape that spreads in the direction intersecting with the 45-degree line L4 in the figure. More specifically, the point cloud has a shape in which the spread is small in the parts close to and far from the origin on L4, and the spread becomes larger in between these parts. Also, in atrial flutter, there are parts where the points are densely distributed and parts where they are sparsely distributed, so that the point cloud has a roughly tuft-like shape centered on L4.

Also, as shown in FIG. 21, when VPC-compensated sinus arrhythmia is included, the point cloud has a shape that spreads in a direction intersecting with the 45-degree line L5 in the figure. More specifically, the point cloud has a portion D1 that is distributed roughly along the 45-degree line L5 in the figure, and a portion D2 that extends from a portion close to the origin on L5 in a direction different from L5. Therefore, the point cloud has a roughly palm-shaped shape centered on L5.

Also, as shown in FIG. 22, when VT and the like are included, the point cloud has a shape that spreads in a direction intersecting with the 45-degree line L6 in the figure. More specifically, the point cloud has a portion D3 that is distributed roughly along the 45-degree line L6 in the figure, and a portion D4 that extends from D3 in a direction different from L6. Therefore, the point cloud has a tree-like shape centered roughly on L6.

As shown in the upper graph of Figure 23, in the Lorenz plot of normal sleep in which no arrhythmia occurs and breathing is stable, the points tend to be distributed roughly along the 45-degree line L7 in the figure, in an area where RRn and RR+1 are both greater than 1.0. This is because sympathetic nerve activity decreases, causing the heart rate to drop. Therefore, the point cloud has a very short band-like (string-like) shape centered roughly on L7.

On the other hand, as shown in the lower graph, when a person is sleeping and has sleep apnea syndrome (SAS), the sympathetic nervous system continues to be active in the same way as when awake in order to compensate for the lack of oxygen, causing the heart rate to rise and periods of time between heartbeats to become shorter. This causes the point cloud to have a band-like (string-like) shape centered on L8. The length of the band (string) is longer than during normal sleep.

In this way, the shape of the points on the Lorenz plot changes depending on the presence or absence of arrhythmia during the measurement period and the type of arrhythmia. Therefore, it is possible to identify the presence or absence of arrhythmia and the type of arrhythmia from the shape of the points. In addition to the arrhythmias described above, it is often possible to identify various arrhythmias such as supraventricular/ventricular arrhythmias, bradycardia, and block from the shape of the points.

As described in the first embodiment, the area of the point cloud can be found by adding up the number of different coordinates (RRn, RRn+1). Therefore, from Figures 6, 7, and 19 to 23, the area of the point cloud is considered to be roughly tachycardia < sinus rhythm < atrial fibrillation, atrial flutter, etc. This makes it possible to determine at least the presence or absence of arrhythmia from the area of the point cloud.

Next, in step S205, the multiple measurement periods are classified using the variation information. For example, the output unit 115 classifies the measurement periods T1 to T54 according to the presence or absence of arrhythmia and the type of arrhythmia using information about the shape of the point cloud of the Lorenz plot.

More specifically, the output unit 115 extracts the shape characteristics of the point cloud for the Lorenz plot selected from the measurement periods T1 to T54, and links Lorenz plots having the same characteristics among the measurement periods T1 to T54 other than the selected Lorenz plot. This allows Lorenz plots having the same characteristics to be classified into the same group. The processor 101 of the bio-information processing device 110 generates folders on the OS or application software corresponding to the presence or absence of arrhythmia and the type of arrhythmia. The Lorenz plots for each measurement period T1 to T54 classified into groups are sorted into the corresponding folders.

The user can check the classified Lorenz plot, for example, on the display of the output device 120 or on the display of a computer terminal (e.g., a personal computer, a smartphone, etc.) connected to the bio-information processing device 110 via the communication network 140.

The selection of the Lorenz plot can be performed, for example, by the user through the input device 130. The user uses the touch panel, mouse, or the like of the input device 130 to select the Lorenz plot to be classified, for example, from the list of Lorenz plots displayed.

For example, the output unit 115 calculates the similarity between the features of a Lorentz plot selected from among the measurement periods T1 to T54 and the features of other Lorentz plots from the measurement periods T1 to T54. If the similarity is equal to or greater than a predetermined threshold, it is determined that the shapes of the point clouds have the same features, and if the similarity is less than the predetermined threshold, it is determined that the shapes of the point clouds do not have the same features. The similarity can be calculated by considering the similarity of the shapes of the point clouds from multiple perspectives (for example, the shape of the contour of the point cloud, the angle between the contour of the point cloud and the 45-degree line, etc.).

In addition, instead of extracting the shape characteristics of the point cloud from the Lorenz plot for the measurement periods T1 to T54, the shape characteristics of the point cloud for normal, atrial fibrillation, atrial flutter, tachycardia, etc. can be stored in advance in the storage device 103 of the bio-information processing device 110. The output unit 115 can also determine whether the shape of the point cloud of the Lorenz plot for the measurement periods T1 to T54 matches these characteristics stored in the storage device 103, thereby determining whether the shape of the point cloud has the same characteristics.

For example, the points of the Lorenz plots of sinus rhythm (FIG. 6) and tachycardia (FIG. 19) are generally band-shaped (string-shaped) along the 45-degree line L1 (L3), while the points of the Lorenz plots of other arrhythmias have a spread in the direction intersecting with the 45-degree line. Therefore, the output unit 115 can determine the presence or absence of arrhythmia based on the difference in whether the points have a spread or not. Also, the points of the Lorenz plots of atrial fibrillation (FIG. 7) and atrial flutter (FIG. 20), and the Lorenz plot including VPC-compensated sinus arrhythmia (FIG. 21) have a triangular shape centered on the 45-degree line L2, a bunch shape centered on the 45-degree line L4, and a palm shape centered on the 45-degree line L5, respectively. In addition, the point cloud of the Lorenz plot ( FIG. 22 ) including VT has a tree-like shape centered on L6, and the point cloud shape when SAS occurs has a band-like (string-like) shape centered on L8. Therefore, the output unit 115 can distinguish the type of arrhythmia based on the difference in the shape of the point cloud.

In addition, when classifying SAS, the output unit 115 can use information about the patient's posture to, for example, exclude measurement results from standing, turning over, and waking up, and classify only measurement results from a specific posture (e.g., supine). This makes it possible to avoid fluctuations caused by the patient turning over or waking up being mistaken for heart rate variations caused by SAS.

Next, in step S206, the output information is output for each measurement period. As shown in FIG. 24, the output unit 115 displays, for example, a list of Lorenz plots classified according to the type of arrhythmia for each measurement period.

In this way, in the process of the flowchart shown in FIG. 18, biometric information of a living body is acquired, a heartbeat is detected from the acquired biometric information, and the interval between successive detected heartbeats is calculated. Next, the entire measurement period of the beat is divided into multiple measurement periods, and interval variation information for each measurement period is generated. The multiple measurement periods are classified using the variation information, and output information is output based on the classification results.

In the above, we have mainly described the case where Lorenz plots are classified using the shape of the points of the Lorenz plot as the variation information, but Lorenz plots can also be classified using the correlation coefficient, area of the points, or density as the variation information. For example, when using the correlation coefficient as the variation information, Lorenz plots can be classified according to the value of the correlation coefficient (-1 to 1). In addition to Lorenz plots, graphs (Figures 16 and 17) that display the change over time in the RR interval can also be configured to be classified by measurement period.

In addition, standard values for the area or density of the point clouds in the Lorenz plot can be preset in the storage device 103 of the bioinformation processing device 110 for each type of arrhythmia, and the output unit 115 can classify measurement periods corresponding to point clouds of an area or density equivalent to the area or density of the preset point cloud into the same group. Alternatively, measurement periods of Lorenz plots having point clouds of equivalent areas or densities can be classified into the same group.

In the above description, the target Lorenz plot was selected and classification was performed based on the similarity with the selected Lorenz plot, but this is not necessarily limited to this. The output unit 115 performs arbitrary machine learning (unsupervised learning/supervised learning) using the shapes of the point clouds for the Lorenz plots for the measurement periods T1 to T54. As a result, the output unit 115 may classify each Lorenz plot into multiple groups. Here, Lorenz plots that belong to the same group have similar tendencies. At this time, the number of groups to be classified may be specified in advance. Here, an example of arbitrary machine learning is clustering classification (K-Means method, etc.). Note that the Lorenz plots may be imaged and then clustered.

The bioinformation processing device 110 of this embodiment described above classifies multiple measurement periods using information on the shape of the point cloud when multiple heartbeats during the measurement periods are plotted on a graph as variation information, and outputs output information based on the classification results. Therefore, the user can easily check graphs with different levels of variation or graphs with different point cloud shapes. This reduces the user's workload in checking classified graphs.

As described above, the biometric information processing device 110, the biometric information processing method, and the biometric information processing program of the present invention have been described in the embodiments. However, it goes without saying that a person skilled in the art can appropriately add to, modify, and omit aspects of the present invention within the scope of the technical concept thereof.

For example, in the first and second embodiments described above, the bio-information processing device 110 is configured separately from the electrocardiograph, but the present invention is not limited to such a case, and the bio-information processing device 110 and the electrocardiograph can also be configured as an integrated unit.

In addition, in the embodiment, some or all of the processing executed by the biometric information processing program may be replaced with hardware such as a circuit.

In addition, in the first and second embodiments described above, the living body is assumed to be a human (patient), but the living body is not limited to being a human, and may be an animal such as a dog or cat.

100 Biometric information processing system,
101 processor,
102 memory,
103 storage device,
104 communication control device,
110 Biometric information processing device,
111 Acquisition unit,
112 detection unit,
113 calculation unit,
114 Generation unit,
115 output unit,
120 output device,
130 input device,
140 communications network.

Claims (16)

An acquisition unit for acquiring biometric information of a living body;
a detection unit that detects a beat from the biological information acquired by the acquisition unit;
a calculation unit that calculates an interval between successive beats detected by the detection unit;
a generation unit that generates information based on the variation in the interval for each measurement period obtained by dividing an entire measurement period of the beat;
an output unit that outputs at least one of information based on the variation and information related to the measurement period corresponding to the information, for each measurement period ;
The generation unit calculates a correlation coefficient related to the variation for each measurement period as information based on the variation . An acquisition unit for acquiring biometric information of a living body;
a detection unit that detects a beat from the biological information acquired by the acquisition unit;
a calculation unit that calculates an interval between successive beats detected by the detection unit;
a generating unit that generates information based on the variation in the interval for each measurement period obtained by dividing an entire measurement period of the beat;
an output unit that outputs at least one of information based on the variation and information related to the measurement period corresponding to the information, for each measurement period ;
A bioinformation processing device, wherein the information based on the variation includes at least one of a shape of a point cloud when the variation is plotted on a graph and an area of the point cloud when the variation is plotted on a graph . The bioinformation processing device according to claim 1 or 2, wherein the output unit classifies the measurement periods using information based on the variation, and outputs at least one of the pieces of information based on the classification result. An acquisition unit for acquiring biometric information of a living body;
a detection unit that detects a beat from the biological information acquired by the acquisition unit;
a calculation unit that calculates an interval between successive beats detected by the detection unit;
a generation unit that generates information based on the variation in the interval for each measurement period obtained by dividing an entire measurement period of the beat;
an output unit that outputs at least one of information based on the variation and information related to the measurement period corresponding to the information, for each measurement period ;
The output unit changes an output form of at least any of the information depending on a magnitude of the information based on the variation . The biometric information processing apparatus according to claim 4 , wherein the output unit rearranges the order of at least any of the information based on the variation in ascending or descending order and outputs the information. An acquisition unit for acquiring biometric information of a living body;
a detection unit that detects a beat from the biological information acquired by the acquisition unit;
a calculation unit that calculates an interval between successive beats detected by the detection unit;
a generating unit that generates information based on the variation in the interval for each measurement period obtained by dividing an entire measurement period of the beat;
an output unit that outputs at least one of information based on the variation and information related to the measurement period corresponding to the information, for each measurement period ;
the generating unit generates, as the information based on the variation, information regarding a shape, an area, or a density of a point cloud when a plurality of the beats in the measurement period are plotted;
The output unit classifies the plurality of measurement periods according to information relating to the shape, area, or density, and outputs at least any one of the pieces of information based on a result of the classification . The biometric information processing device according to any one of claims 1 to 6, wherein the generation unit acquires information relating to the body position of the living body when measuring the beats and the time of the measurement, and determines based on the information whether at least any of conditions regarding the body position of the living body, conditions regarding the time period of the measurement, and conditions regarding the shape and interval of the beats are satisfied, and if at least any of the conditions are not satisfied, excludes the beat from the multiple beats and generates information based on the variation. The biometric information processing device according to claim 6 , wherein the output unit classifies measurement periods corresponding to a point cloud having a shape equivalent to a shape of a preset point cloud, or measurement periods corresponding to point clouds having shapes similar to each other, into the same group. The biometric information processing device according to claim 6 , wherein the output unit classifies measurement periods corresponding to point clouds having an area or density equivalent to an area or density of a preset point cloud, or measurement periods of point clouds having areas or densities equivalent to each other, into the same group. The biometric information processing apparatus according to claim 4 , wherein the output unit outputs a background color of the at least one piece of information in a color different from a color in which the at least one piece of information is output. The biometric information processing apparatus according to claim 4 , wherein the output unit assigns identification information for identifying at least one of the pieces of information to the piece of information. The bioinformation processing device according to any one of claims 1 to 11 , wherein the information relating to the measurement period includes a time or a time period during which the measurements of the multiple beats corresponding to the information based on the variation were performed, or a code for distinguishing the multiple measurement periods from each other. (a) acquiring biometric information of a living body;
a step (b) of detecting a beat from the biological information acquired in the step (a);
a step (c) of calculating an interval between successive beats detected in the step (b);
a step (d) of generating information based on the variation in the interval for each measurement period obtained by dividing the entire measurement period of the beat;
and (e) outputting , for each measurement period, at least one of information based on the variation and information related to the measurement period corresponding to the information based on the variation,
In the step (d), a correlation coefficient relating to the variation is calculated for each measurement period as information based on the variation . The biological information processing method according to claim 13 , wherein in said step (e), the plurality of measurement periods are classified using information based on the variation, and at least any of the information is output based on a result of the classification. (a) acquiring biometric information of a living body;
a step (b) of detecting a beat from the biological information acquired in the step (a);
a step (c) of calculating an interval between successive beats detected in the step (b);
a step (d) of generating information based on the variation in the interval for each measurement period obtained by dividing the entire measurement period of the beat;
and (e) outputting , for each measurement period, at least one of information based on the variation and information related to the measurement period corresponding to the information based on the variation,
In the step (e), an output form of at least any of the information is changed depending on a magnitude of the information based on the variation . A biometric information processing program for causing a computer to execute the processes included in the biometric information processing method according to any one of claims 13 to 15 .