JP2023032476A - Cardiac beat data analysis device and program - Google Patents

Abstract

To provide a cardiac beat data analysis device capable of coping with the case that there is data omission when a relatively short time function is an analysis object, so as to reduce impact of the data omission, by performing appropriate correction.SOLUTION: There is provided a cardiac beat data analysis device 10 comprising: an intermediate function generation part 3 for reading an R-R interval row which is a row of pulse intervals extracted from a cardiac beat data section which is cut out by a window from time series cardiac beat data, converting the R-R interval row to generate an intermediate function; and a frequency analysis part 4 for acquiring a result of frequency analysis of the intermediate function. The cardiac beat data analysis device further comprises an omission determination part 2 for determining whether or not there is any pulse omission position in the cardiac beat data section. Upon determination by the omission determination part 2 that there is pulse omission, the intermediate function generation part 3 converts the R-R interval row to generate the intermediate function, then, corrects and bring the position corresponding to the pulse omission in the intermediate function into a state with no pulse omission, so as to generate the intermediate function to be subjected to the frequency analysis in the frequency analysis part.SELECTED DRAWING: Figure 4

Description

The present invention relates to a heartbeat data analysis device and program.

As a method for analyzing heartbeat data, there is an existing method that conforms to the following framework, and analysis that conforms to this existing method is performed in Patent Document 1 and the like. FIG. 1 is a schematic diagram for explaining the framework of this existing method.

This existing method is based on the following findings. It is known that the R-wave generation interval (R-R interval) that indicates the start of ventricular contraction of heartbeat varies with periodicity, and the frequency component of this variation represents the working condition of the autonomic nerves. This relationship can be used to evaluate the functioning of the autonomic nerves. One method (Fig. 1) is to sequentially measure the R-R intervals in the heartbeat data, frequency-analyze the R-R intervals (spectrum analysis), and evaluate the fluctuations. A frequency component is obtained, and the working condition of the autonomic nerves is evaluated from the frequency component.

Data D1 in FIG. 1 shows an example of a graph of heartbeat data acquired by a heartbeat sensor such as an electrocardiogram sensor, and the heartbeat data is configured as a row of R waves in time series with the horizontal axis as the direction of time progression. . Here, the R wave represents each pulsation by representing the start of ventricular contraction, and appears with R-R intervals RR1, RR2, RR3, ... in the time axis direction (x axis direction), and the intervals fluctuate periodically. do.

In general, it is difficult to perform frequency analysis directly on data along the time axis. That is, it is difficult to directly obtain the data example D3 as the frequency analysis result from the data example D1 in FIG. 1 as indicated by the arrow A. Therefore, when frequency analysis is performed on the R-R interval variation as shown in data example D1, time-series data is generated as intermediate data by applying an intermediate function to data D1 as shown in data example D2, and this intermediate By frequency-analyzing the time series data D2 as data, a frequency analysis result as shown in the data example D3 is obtained.

As an intermediate function for obtaining the intermediate data D2 for enabling frequency analysis in this way, many types of intermediate functions have been proposed according to the purpose of heartbeat data analysis and the like. In FIG. 1, as an example of the intermediate function, for each of the R-R intervals RR1, RR2, RR3, . An example of an intermediate function that associates RR2, P3=1/RR3, . In Non-Patent Document 1, as described above, between adjacent R-wave occurrence times, an intermediate function is defined as the reciprocal of the R-R interval (this reciprocal means the instantaneous heart rate), and spectral analysis is performed on it. It is described that heart rate fluctuations are captured in real time.

In FIG. 1, the result data D3 obtained by frequency analysis of the above intermediate data D2 represents the power spectral density, the horizontal axis is the frequency (Hz), and the vertical axis is the power spectral density (msec ² /Hz). show. As is known in the field of signal processing, the power spectral density represents the power of time-series data by decomposing it into each frequency component. Here, as schematically indicated by using the same reference numerals in data D3, the sum of 0.04 to 0.15 Hz components is LF as the predetermined low-frequency band, and the predetermined high-frequency band is 0.15 to 0.40. Assuming that the sum of Hz components is HF, the quotient LF/HF indicates the degree of sympathetic nerve activity, and HF indicates the degree of parasympathetic nerve activity, so these values are used as autonomic nerve indices. Thus, analysis results of the heartbeat data can be obtained.

JP 2015-080624 A

BME Vol. 8, No. 10, 1994 p. 13-16 "1/f Spectrum Estimation of Heart Rate Fluctuation"

Regarding heartbeat data analysis of the above-mentioned conventional methods, in general, in the medical field, it is necessary to "accurately" evaluate the "(stationary) state" of the autonomic nerve index and to obtain high accuracy and reproducibility. purpose. Therefore, in order to obtain an autonomic nerve index, frequency analysis is performed on time functions (heartbeat data) at relatively long time intervals (for example, about 90 to 180 seconds).

On the other hand, depending on the application of the analysis, by subjecting the time function (heartbeat data) of relatively short time intervals (for example, about 10 to 80 seconds) to the frequency analysis, the instantaneous state or state change of the autonomic nerve index can be obtained. In some cases, it may be desirable to evaluate. In this case, the frequency analysis is performed continuously along the progress of time by subjecting the time function of, for example, 20 seconds as the shortest possible interval to the frequency analysis.

When the time function of such a short time interval is subjected to frequency analysis, there is an occurrence of failure to detect pulsation as a matter that can become a significant problem. If one pulsation is not detected and the frequency analysis is performed without taking measures against it, the result of the frequency analysis will be greatly disturbed.

Fig. 2 shows an example of the effect of detection omission (data loss) when the time function of short intervals is subject to frequency analysis. FIG. 10 is a diagram showing analysis results of data in which frequency components exist only in the latter (low-frequency side band LF) of the side band LF; (That is, the data in FIG. 2 is arbitrarily used for experimental demonstration as data in which frequency components exist only in the low frequency side band LF in order to make it possible to easily see the effect of detection omission. generated in the time domain.)

In Fig. 2, the graph GA is the frequency analysis result of the time interval with no data loss, and the graph examples of five power spectral densities are 12, 18, 24, 48, 96 pulses (pulsation number ), with window size time functions at intervals of about 10, 14, 19, 38 and 77 seconds, respectively. (That is, Graph GA lists the results of frequency analysis of a common time function (heartbeat data) to be analyzed with five different window sizes.) Graph GB is similar to Graph GA. An example of the results of applying a similar analysis (with similar 5-interval window sizes) to a time function with missing data (heart rate data). Although almost the same heart rate data is analyzed in Graph GA and Graph GB, the only difference is that there is one accidental missing data in the heart rate data of Graph GB. It can be seen that the frequency analysis results of .

Even when dealing with a time function (heartbeat data) with such data omissions, as shown by graph GC, a method of correcting the frequency analysis results so that they are almost the same as graph GA without data omissions is desired, but the prior art has not been able to adequately correct for this. Note that in FIG. 2, regarding the graphs GA, GB, and GC, the correspondence relationship between graphs having a common window size is shown by connecting the graphs GA and GB and the graphs GA and GC with dotted lines.

In addition, as a cause of missing data (loose detection of pulsation), there are omissions of detection due to noise during measurement by the heart rate data measuring device (when pulsation is present in the heart rate data but cannot be measured). , the presence of arrhythmias in heart rate data. Arrhythmia is known to occur accidentally due to temporary factors such as stress even in healthy individuals. Regardless of the cause of such data loss, it is desirable to appropriately correct the effects of data loss.

Note that one of the reasons why the frequency analysis result varies greatly as in the example of FIG. 2 is that when data is missing and converted to an intermediate function, peculiar data is generated. FIG. 3 is a diagram showing an example of an intermediate function when there is data loss, in comparison with a case where there is no data loss. Data D4 is an example of normal heartbeat data (time function without omission), and an intermediate function obtained therefrom is shown as data D5. These data D4 and D5 are examples similar to the data D1 and D2 in FIG. 1, and are examples in which the reciprocal of the RR interval is used as an intermediate function.

On the other hand, in FIG. 3, as a comparison of data D4 and D5, data D6 is an example of heartbeat data with arrhythmia (time function with missing points), and an intermediate function obtained therefrom is shown as data D7. In the time function of data D6, since data missing occurs at location R6, the R-R interval suddenly widens to approximately double. At R7, the R-R interval abruptly widens about twice, and at the same time, the amplitude value at that time decreases about half.

In this way, when comparing the data D5 of the intermediate function in the normal case with the data D7 of the intermediate function in the case of missing data, the data D7 has a peculiar value at the point R7 affected by the missing data. As a result, as shown in FIG. 2, frequency analysis of the intermediate function having such a peculiar value will greatly disturb the result.

When considering how to deal with such data missing, for example, in Patent Document 1, an interpolation pattern of a data string that does not disturb the phase of the data string before and after is inserted for a noise location on a time function (noise location is an interpolation pattern ), it is said that short-term interpolation will improve the restoration of biological fluctuations and improve the restoration of spectral structure, but in order for such interpolation to be performed appropriately, a long time function is used. needed and could not be adequately compensated for at short intervals. (That is, in the case of a short interval, in order to sufficiently interpolate with data strings before and after, it is necessary to interpolate outside the range of the short interval, so the instantaneous state or state change of the autonomic nerve index is If you need it, it will greatly affect the amount of change in that frequency component.Also, even if you try to interpolate only within a short interval, there is not enough data for interpolation in the first place, and you can not interpolate properly. Therefore, the frequency analysis result cannot be corrected appropriately.) Also, the interpolation processing is intended for a time function (generally a function showing complex changing behavior) as exemplified by the data D1 in FIG. There is also a problem that interpolation processing cannot always be performed simply.

In view of the above-described problems of the prior art, the present invention is capable of reducing the effects of data loss by performing appropriate corrections even when data loss exists when a relatively short time function is to be analyzed. It is an object of the present invention to provide a heartbeat data analysis device and a program that can

In order to achieve the above object, the present invention reads an RR interval string, which is a string of pulsation intervals extracted from a windowed heartbeat data interval from time-series heartbeat data, and converts the RR interval string. A heartbeat data analysis device comprising: an intermediate function generation unit for generating an intermediate function by performing a frequency analysis on the intermediate function; and a frequency analysis unit for obtaining a frequency analysis result of the intermediate function. a lack determination unit for determining, and when the lack determination unit determines that there is a lack of pulsation, the intermediate function generation unit converts the RR interval sequence to generate an intermediate function, and further, The intermediate function to be subjected to frequency analysis by the frequency analysis unit is generated by correcting a portion of the intermediate function corresponding to the lack of pulsation so that there is no lack of pulsation. Further, the program is characterized by causing a computer to function as the heartbeat data analysis device.

According to the present invention, by correcting the location affected by the dropout on the intermediate function so that there is no pulsation dropout, even if the data dropout exists, appropriate correction is performed to reduce the effect of the data dropout. can do.

FIG. 2 is a schematic diagram for explaining the framework of an existing technique for frequency analysis of heartbeat data; FIG. 10 is a diagram showing an example of the influence (and the effect of correction) of omission of detection (missing data) when a time function of short intervals is subjected to frequency analysis; FIG. 10 is a diagram showing an example of an intermediate function when there is data loss in comparison with an intermediate function when there is no data loss; 1 is a functional block diagram of a heartbeat data analysis device according to one embodiment; FIG. 4 is a flowchart of the operation of the heartbeat data analysis device according to one embodiment; FIG. 10 is a diagram showing an example of setting timing and window width when processing in real time; FIG. 10 is a diagram showing an example of generating an intermediate function as the reciprocal of the R-R interval; FIG. 10 is a diagram illustrating an example for explaining intermediate function correction processing by a correction unit; It is a figure which shows the example of a graph of the power spectral density in real time obtained by the analysis part at each time. FIG. 10 is a diagram showing an example for explaining correction when a beat function is used as an intermediate function; It is a figure which shows the hardware constitutions in a common computer apparatus.

FIG. 4 is a functional block diagram of the heartbeat data analysis device 10 according to one embodiment. The heartbeat data analysis device 10 includes an intermediate function generation section 3 including an R-R interval extraction section 1, a missing determination section 2, a generation section 31 and a correction section 32, and an analysis section 4. FIG. As an overall operation, the heartbeat data analysis device 10 reads the heartbeat data as an input from the measuring device 20 for measuring the heartbeat data, converts the heartbeat data (time function) into an intermediate function, and then converts the frequency analysis result. Output. 4, the measurement device 20 is an external component of the heartbeat data analysis device 10, but the measurement device 20 may be included in the heartbeat data analysis device 10. FIG.

FIG. 5 is a flow chart of the operation of the heartbeat data analysis device 10 according to one embodiment. Hereinafter, while explaining each step of FIG. 5, the processing of each functional unit of the heartbeat data analysis apparatus 10 of FIG. 4 will be explained.

When the flow of FIG. 5 is started, in step S1, the latest heartbeat data (time function) from the measuring device 20 up to the current time is read in the R-R interval extractor 1, and then the process proceeds to step S2. The most recent heartbeat data to be read covers the update interval of step S6. That is, assuming that the number of updates is i (i=1, 2, 3, . The latest heartbeat data up to i) should be read (additionally, assuming that it has not been read yet). (For the first time when i=1, wait from the past time T(0) to the current time T(1) before reading.) Set this latest current time T(i) to be read to coincide with the edge of the window. You can leave it.

That is, in the flow of FIG. 5, heartbeat data is measured in real time for a user whose heartbeat data is to be acquired and analyzed by the measuring device 20, and the most recent time on the side of the current time is measured from this heartbeat data (time function). This embodiment relates to an embodiment that cuts out a window size and repeats frequency analysis of the time function within this window size in real time at each time (each processing timing) in real time. (Note that step S6 in FIG. 5, which will be described later, is a step of performing update management at each time in this real-time repetitive processing.) It is also possible to sequentially analyze already measured heartbeat data using a sliding window in the heartbeat data analysis apparatus 10 in the same manner as in the flow of FIG.

Here, as shown in the schematic example of FIG. 6, various predetermined settings may be used for the timing and window width when processing the flow in FIG. 5 in real time. That is, as shown in example EX1 of FIG. 6, windows W1, W2, W3, . . . (the last time of one window coincides with the first time of the next window). Also, as shown in example EX2, windows W4, W5, W6, . . . There may be an interval between the last time of a window and the first time of the next window). Also, as shown in example EX3, windows W7, W8, W9, . . . (the last part of one window overlaps the first part of the next window). In example EX3, the overlapping ratio between adjacent windows is about 20% of the window width, but this overlapping ratio can also be set arbitrarily. (For example, 90% of the window width overlaps, and each window may also overlap a window further than the two neighboring windows on both sides.)

The measuring device 20 is worn in the form of a wearable device or the like by a user whose heartbeat data is to be acquired, and can acquire heartbeat data in any existing format by any existing method. For example, as an ECG (electrocardiogram) sensor, heartbeat data may be obtained in the form of an electrocardiogram through electrodes attached on the skin, or as a PPG (optical heartbeat) sensor, the blood flow volume in blood vessels is optically detected. heart rate data may be obtained in the form of pulse waves. Other types of heart rate sensors may be used.

In step S2, the following processes (1) to (3) are performed on the time function up to the current time including the latest time function (heartbeat data) read in step S1, and then the process proceeds to step S3.

(1) The R-R interval extraction unit 1 extracts the R-R interval from the time function up to the current time, and the R-R interval is generated by the generation unit 31 (and depending on the embodiment, the lack determination unit as indicated by the line L1 in FIG. 4) 2). As described with respect to data example D1 in FIG. It can be extracted as a time series formed by intervals of occurrence times of R waves.

(2) The generator 31 converts the R-R interval calculated by the R-R interval extractor 1 to generate an intermediate function as intermediate data, and outputs this intermediate function to the analyzer 4 or the corrector 32 . Here, the intermediate function generated by the generation unit 31 is output to the correction unit 32 when a positive determination is made in step S3 described later and the process proceeds to step S4, and when a negative determination is made in step S3 described later and step S4 is skipped. , is output to the analysis unit 4. The intermediate function generated by the generation unit 31 is also output to the missing determination unit 2 as indicated by line L2 in FIG. 4, depending on the embodiment.

As described above, there are various types of intermediate functions for converting to intermediate data. In the following, as an example of explanation, the generating unit 31 generates an intermediate function as the reciprocal of the RR interval, as illustrated in FIG. The case will be explained. As shown in FIG. 7 (similar to the examples of FIGS. 1 and 3 above), the extraction result of the RR interval extraction unit 1 for the time function (heartbeat data) within the read window is the peak time t _k (k=0, 1 _{, 2, 3, 4 ,} . 2,3,4...) are extracted. Converting this RR interval into an intermediate function whose amplitude is its reciprocal yields an intermediate function of the form f(t)=1/I _k+1 (t _k ≦t<t _k+1 ).

(3) The omission judging section 2 judges whether there is omission in the time function (heartbeat data) read in step S1, and corrects the judgment result (in the case of the omission judgment result, the information of the omission occurrence point is also included). output to unit 32; As shown by line L1 in FIG. If there is an R-R interval that is determined to be longer than a threshold, it may be determined that the portion of the R-R interval that is determined to be long is missing. (In addition, regarding a single "R-R interval" and a "R-R interval sequence" in which a plurality of these are arranged in chronological order, both are called "R-R intervals" when the distinction is clear from the context. )

Here, it is generally said that the human heartbeat interval changes only about 30%. Therefore, as an example of threshold setting, when the average R-R interval exceeds 30% (when the R-R interval exceeds 1.3 times the average R-R interval), or when the R-R interval before and after (at least one of the two R-R intervals before and after , or both, or the average of both), it may be determined that there is data missing. This average R-R interval may be determined in advance as a fixed value, or the R-R interval extraction unit 1 records the history of the R-R interval in the most recent fixed period (for example, the period from the present to the past three minutes ago). Then, the average R-R interval may be dynamically calculated at each time based on this recording, and the lack presence/absence determination unit 2 may perform threshold determination according to this.

Alternatively, the missing determination unit 2 may perform the same determination by directly analyzing the time function (heartbeat data) up to the current time obtained by reading and updating in step S1 as indicated by the line L0. Alternatively, it may be performed by analyzing the intermediate function obtained by the generator 31 as indicated by the line L2.

In step S3, cases are classified according to whether or not the determination result of the missing determination unit 2 in the processing of (3) in step S2 above is "missing". proceeds to step S4, and if negative (if it is determined that there is no omission), skips step S4 and proceeds to step S5 as shown in FIG.

In step S4, the correction unit 32 corrects the intermediate function obtained from the generation unit 31, outputs the corrected intermediate function to the analysis unit 4, and then proceeds to step S5. According to the determination result of step S3 above, this intermediate function has a missing part in the initial time function (heartbeat data), and is inappropriate for frequency analysis as it is, as explained in Figures 2 and 3 above. Therefore, the correcting section 32 corrects this to obtain a corrected intermediate function suitable for frequency analysis.

FIG. 8 is a diagram showing an example for explaining the intermediate function correction processing by the correction unit 32. As shown in FIG. Data D7 is shown as an example of an intermediate function to be corrected. This data example D7 is the same as that shown in FIG. R7 is a peculiar value due to the presence of missing data. For the sake of explanation, as shown in data D7 in _FIG . The amplitude at the point of occurrence R7 is f(t)=1/I _n (this time t is within the range belonging to the point R7 with the time width I _n ). It should be noted that the information of the dropout occurrence location R7 (time range within the window) as shown in the data example D7 in FIG. This information is output to the unit 32 .

As further shown in FIG. 8, the correction unit 32 obtains a correction result by processing the dropout occurrence location R7 in the data D7 into an intermediate function in the form of the data D7C. Regarding the data D7C, FIG. 8 shows the overall appearance on the upper side, and an enlarged view of the dropout occurrence location R7 to be corrected as a part of this and its vicinity on the lower side. In this correction process, as indicated by the data D7C, only the dropout occurrence location R7 is subject to correction, and within the window size, locations other than the dropout occurrence location R7 are not corrected. Further, the information of the intermediate function referred to in the correction is the R-R interval in the closest vicinity of the dropout location R7 and the amplitude value as its reciprocal.

That is, the dropout occurrence point R7 (consisting of the time width I _n and the constant amplitude 1/I _n ) should not have occurred originally, as described with reference to the data examples D6 and D7 in FIG. Since one pulsation that should be present is missing, the time width I _n and constant amplitude 1/I reflecting an unnaturally long RR interval (and unnaturally small amplitude) as a missing state _n , corresponding to the two natural short RR intervals (and amplitudes of natural magnitude) that should have occurred, "time width I _m1 and constant amplitude 1/I _m1 ” and “time width I _m2 and constant amplitude 1/I _m2 ”.

Specifically, the following equations (1) and (2) are satisfied in order to obtain a natural correction result by referring only to the information of the intermediate function closest to the defect occurrence location R7 as the original data for correction. , two time widths I _m1 and I _m2 corresponding to two RR intervals (and their amplitudes as reciprocals) are obtained to obtain a corrected intermediate function.

Equation ( ₁ ) is based on two RR intervals I _m1 and I _m2 (I _m1 is the first half of , I _m2 is meant to replace the latter half). When performing this replacement, the amplitude of the RR interval I _m1 in the first half is the reciprocal 1/I _m1 , and the amplitude of the RR interval I _m2 in the second half is the reciprocal 1/I _m2 as an intermediate function having the reciprocal amplitude. However, the constraint for making these amplitudes natural values as a result of correction is Equation (2). By imposing the two constraints of equations (1) and (2) on the two RR intervals I _m1 and I _m2 as two variables, the values of these two variables are determined and the corrected intermediate function (data example D7C) can be obtained.

Geometrically, the constraint of expression (2) is as illustrated in data example D7C (enlarged view on the lower side) of FIG. That is, the middle point m _n- 1 of the RR interval R7B (consisting of the time width I _n-1 and the constant amplitude 1/I _n-1 ) located immediately before the missing point R7 to be corrected, and the correction target A straight line L connecting the middle point m _n+1 of the RR interval R7A (consisting of time width I _n+1 and constant amplitude 1/I _n+1 ) located immediately after the missing point R7 is When drawn on the graph of the function, the midpoint m m1 _of the corrected first half RR interval I _m1 and the midpoint m _m2 of the corrected second half RR interval I _m2 are placed on this straight line L. It is a constraint that

As is clear from FIG. 8, the above midpoints m _n−1 , m _m1 , m _m2 , and m _n+1 are defined as intermediate functions having reciprocal amplitudes, and their graph shapes are “rectangular The upper side of each "strip" (rectangle) in the intermediate function "function" (a function of the shape of rectangles (strips) whose height is the reciprocal of the width that varies according to the RR interval and the reciprocal of this width is lined up on the time axis) It means the midpoint of the side formed by the amplitude values). (That is, the strip function varied the step position x=0 and the step height 1 of the step function f(x)=0(x<0), f(x)=1(x≧1)( The step height can be negative) and can be expressed as the sum of multiple step functions.)

As described above, the correction method of the correction unit 32 explained with reference to FIG. 8 is used as the first embodiment, and second and third embodiments, which are modifications thereof, will be explained.

In the second embodiment, instead of using the constraints of the equations (1) and (2) of the first embodiment, the first half side corrected by using the equation (1) and the following equation (3) An RR interval I _m1 and a corrected RR interval I _m2 on the rear half side can be determined.
I _m1 =I _m2 …(3)

That is, in the second embodiment, it is estimated that the pulsation that should have occurred was present at the midpoint on the time axis of the time width I _n (RR interval I _n ) of the dropout occurrence location R7 shown in FIG. 8 and the like. to obtain the correction result. From equations (3) and (1), the corrected two equal RR intervals and the corresponding amplitudes are as follows, and it is possible to correct the dropout occurrence point R7 of data example D7 to a natural value using an intermediate function. Become.
I _m1 =I _m2 =I _n /2
→ RR intervals I _m1 and I _m2 are half of the initial interval, which is twice.
1/I _m1 =1/I _m2 =2*(1/I _n )
→The amplitude values 1/I _m1 and 1/I _m2 are double the initial unnatural value 1/I _n and become natural values.

In the second embodiment, unlike the first embodiment (and the third embodiment), when correcting the intermediate function, the data of the intermediate function other than the dropout occurrence point R7 is referred to as the correction source data. becomes unnecessary.

In the third embodiment, the correction result is obtained by estimating the original pulsation as the midpoint on the time axis (that is, the midpoint in the horizontal axis direction of the intermediate function) in the above-described second embodiment. A correction result of the intermediate function corresponding to the original pulsation is obtained as the midpoint of the intermediate function in the vertical axis direction.

That is, as shown in FIG. 8, at the RR interval R7B (time width I _n-1 and constant amplitude 1/I _n-1 (= y _n-1 ) located immediately before the missing point R7, ), an RR interval R7A (consisting of a time width I _n+1 and a constant amplitude 1/I _n+1 (= y _n+1 ) located immediately after the dropout occurrence point R7), is used as reference data for correction, and the corrected amplitude y _n of the missing point R7 is calculated as the average value of the amplitudes y _n-1 and y _n+1 before and after these as follows. Get results.
y _n =(y _n-1 +y _n+1 )/2

In the third embodiment, the corrected amplitude may be the average value y _n and the length in the time axis direction may _be the same length as In before correction. (That is, for the intermediate function dropout occurrence location R7 (length I _n on the time axis), the amplitude 1/I _n before correction is changed to amplitude y _n . At this point, since "amplitude x time axis length = y _n x I _n ≠ 1", assuming that one RR interval corresponds to this point after correction, the reciprocal of the time interval is Although the amplitude value is contrary to the definition of an intermediate function having as the amplitude, the amplitude value itself is corrected to a natural value, so it is possible to obtain a natural frequency analysis result (or corrected In the later part, within the range of the time axis length _In , there are multiple RR intervals of the time axis length such that each is "amplitude y _n × time axis length = 1" (it does not have to be an integer ) is present.) That is, in the third embodiment, the point at which the original pulsation exists in the dropout occurrence point R7 whose length in the direction of the time axis is I _n is determined. Although it is not possible to obtain a clear estimation result as in the first and second embodiments, it is possible to obtain a natural correction result in the form of an intermediate function, which is the target of frequency analysis.)

In step S5, the analysis unit 4 performs frequency analysis on the intermediate function generated by the intermediate function generation unit 3, outputs the result, and proceeds to step S6. The intermediate function to be analyzed by the analysis unit 4 is generated by the generation unit 31 when the process proceeds from step S4 to step S5 (when there is data loss), and further corrected by the correction unit 32. It is an intermediate function corrected to be reduced, and if step S4 is skipped and step S3 is advanced to step S5 (when there is no data missing), it is the intermediate function generated by the generation unit 31. In addition, the range on the time axis of the intermediate function that becomes the frequency analysis in step S5 is the predetermined window range as described with reference to FIG. The range (the window range that was not processed in the past step S5) should be made a new analysis target.

When there is data loss, the correction unit 32 corrects the intermediate function so that it corresponds to a state in which data loss does not exist, as described above. It is possible to obtain an appropriate result with reduced influence. In other words, even if there is data missing as explained in Fig. 2 above, it is not an unnatural frequency analysis result like the graph GB, but a natural frequency analysis like the graph GC reflecting the correction. It is possible to obtain a result (a frequency result close to the original graph GA).

The analysis unit 4 can obtain the frequency analysis result by the same processing as that for obtaining the data D3 from the data D2 in the example of FIG. That is, the intermediate function is frequency-analyzed to obtain the power spectral density F(f), and the sum of 0.04 to 0.15 Hz components is defined as LF as a predetermined low-frequency band as shown in the following equation, and a predetermined high-frequency band For example, HF is the sum of 0.15 to 0.40 Hz components (the following formula).

It is known that the quotient LF/HF indicates the working condition of the sympathetic nerves, and HF indicates the working condition of the parasympathetic nerves. It is possible to obtain analysis results of heartbeat data based on frequency analysis.

FIG. 9 shows an example of a graph of the power spectrum density at each time in real time obtained by the analysis unit 4. In FIG. Since the power spectral density is available in real time, the autonomic index values (LF/HF and HF) can likewise be obtained in real time.

In step S6, the time is updated to the next time (next real-time processing timing), and then the process returns to step S1, so that the heartbeat data analysis device 10 repeatedly executes the above-described processing for each real-time time. .

As described above, according to the embodiments of the present invention, it is possible to reduce the influence of data loss (missed detection of pulsation) due to arrhythmia, measurement error, or the like when obtaining an autonomic nerve index. At this time, appropriate and simple correction is possible by correcting not the heartbeat interval (R-R interval) but the intermediate function data so as not to affect the frequency component. For this correction, by not referring to the data of other intervals (windows), it is possible to make corrections in line with real-time analysis targeting relatively short windows and corrections for requests that lead to instantaneous states or state changes of autonomic nerve indices. be.

As will be described later in (4) below, when data loss occurs at the edge of the window, the intermediate functions outside the window are exceptionally referred to for correction. Only one R-R interval is required as the minimum range when expanding the range, and it is unnecessary to refer to a wide range of intermediate functions outside the window, so it is virtually the same as correcting within a single window. is.

Various additions, supplements, alternatives, etc. are described below.

(1) According to the embodiment of the present invention, since it is possible to grasp the autonomic nerve index with high accuracy in real time, it can be used, for example, to prevent diseases caused by mental stress. , it will be possible to contribute to Goal 3 of the Sustainable Development Goals (SDGs) led by the United Nations, "Ensure healthy lives and promote well-being for all at all ages." (In addition, if the arrhythmia that contributes to the lack of data is not accidental, separate medical treatment is desired.)

(2) In the above description, the generation unit 31 describes the correction in the case of using the intermediate function whose amplitude value is the reciprocal of the R-R interval, but similar correction processing is possible with other intermediate functions. For example, even if the intermediate function determines the amplitude value by something other than the reciprocal (the -1 power of the R-R interval), if the graph shape is a rectangular function as in the case of the reciprocal, individual constraints for the intermediate function Based on this, it is possible to make corrections according to the first embodiment described with reference to FIG. 8 and modifications thereof according to the second and third embodiments.

For example, in the case of the first embodiment, in the same manner as described with reference to FIG. Only the portion R7B and the immediately following intermediate function portion R7A corresponding to one R-R interval should be referred to as correction source data. As a modification, an intermediate function portion corresponding to two or more R-R intervals may be referred to as correction source data in each of immediately before and immediately after. (However, it is desirable to use only the range as close as possible to the data missing point R7 as the source data for correction.)

(3) Correction in the corrector 32 when the beat function is used as the intermediate function in the generator 31 will be described with reference to FIG. The beat function is the _intermediate The function is obtained as it is as a pulse train I _k (k=1, 2, 3, . (In other words, unlike the intermediate function of the reciprocal, where the time axis was the same as the initial heart rate data, the intermediate function of the beat function converts the time axis into pulse numbers, and the pulse height reflects the time width information. The Rukoto.)

In FIG. 10, this discrete time pulse train is shown as a data example D10, and since the third pulse _I3 is significantly higher than the other pulses, it is assumed that the dropout determination unit 2 determines that the dropout occurs. . In this case, as shown in data example D10C, by replacing the missing pulse _I3 with two corrected pulses _I3-1 and _I3-2 , we obtain the following corrected intermediate function: be able to.
Intermediate function before correction (data D10) Four pulse trains arranged in order of _I1 , _I2 , _I3 , _I4 Intermediate function after correction (data D10C) _I1 , _I2 , _I3-1 , _I3- 5 pulse trains in order of ₂ ,I ₄

For the heights of the two corrected pulses _I3-1 and _I3-2 , the heights of the pulse _I2 immediately before the third pulse _I3 and the pulse _I4 immediately after the third pulse I3, which was the missing point, were calculated. By interpolation, it is possible to calculate as follows. Calculation by the interpolation is schematically shown as a line L10 in the post-correction data example D10C in FIG.
_I3-1 =(2* _I2 + _I4 )/3
_I3-2 = ( _I2 +2* _I4 )/3

It should be noted that the example of correction when the above beat function is used as an intermediate function applies the following replacement, and corrects in the same policy as the correction when using the reciprocal as an intermediate function described with reference to FIG. 8 .
3rd pulse I ₃ = data missing point R7
Second pulse I ₂ = RR interval point R7B immediately before data dropout occurrence point R7
4th pulse I ₄ = RR interval point R7A immediately after data missing point R7

Although the case where the beat function is used as the intermediate function has been described above, the case where spline interpolation and linear interpolation are used may be roughly the same. In spline interpolation, the RR interval is I _k (k=1, 2, ...), and the range of time t in this interval I _k is t _k ≤ t ≤ t _{k + 1} , (procedure 1) time t Find a pulse train function whose value of _k (k=1, 2, …) is the interval I _k (the value of time t other than this is 0), (Step 2) perform cubic spline interpolation on this pulse train function, (Procedure 3) A pulse train function (re)sampled at an average RR interval from this cubic spline interpolation is used as an intermediate function.

Regarding the intermediate function of the above spline interpolation, the pulse train function in (Procedure 1) can be corrected. And it is sufficient. In other words, one erroneous pulse can be corrected to two natural pulses as a missing point, and the position of the two pulses on the time axis is the position obtained by dividing the preceding and following pulse positions into three equal parts. , the heights of the two pulses are interpolated with the heights of the preceding and succeeding pulses.

Correction can be performed in the same way when linear interpolation is used as an intermediate function. That is, the intermediate function of the linear interpolation is obtained by replacing the spline interpolation with the linear interpolation in the steps 2 and 3 among the steps 1 to 3 in the spline interpolation. It is possible to correct the case of the intermediate function of linear interpolation in the same manner as the method of dividing and interpolating.

(4) In the example of FIG. 8, when correcting the data missing point R7, the intermediate functions of the immediately preceding R-R interval point R7B and the immediately following R-R interval point R7A are referred to as correction source data. It is unlikely that this will occur frequently, but as an exceptional case, if the data loss occurrence location R7 is located at the edge of the window used for analysis in step S5, the immediately preceding R-R interval location R7B is the one before the window. A situation may occur where the R-R interval point R7A immediately after is within the window or within the window one window after the window in question. In the correction, only the nearest neighbor of the data missing point R7 is referred to, so even if it is outside the window, the frequency analysis result is not significantly affected.

(5) The data newly read in step S1 may be a constant time interval, that is, a predetermined time range on the time function (heartbeat data), or a predetermined number of R-R intervals (a predetermined number of R-R intervals). . If it is determined that there is data loss, the number of R-R intervals may be determined in the R-R intervals increased from 1 to 2 by correcting the data loss.

(6) FIG. 11 is a diagram showing an example of hardware configuration in a general computer device 70. As shown in FIG. Each of the heartbeat data analysis device 10 and the measurement device 20 (they may be integrally configured as described above) can be implemented as one or more computer devices 70 having such a configuration. When each of the heartbeat data analysis device 10 and the measurement device 20 is implemented by two or more computer devices 70, information necessary for processing may be transmitted and received via a network. The computer device 70 includes a CPU (central processing unit) 71 that executes predetermined instructions, and a GPU (graphics processing unit) as a dedicated processor that executes part or all of the execution instructions of the CPU 71 instead of the CPU 71 or in cooperation with the CPU 71. ) 72, RAM 73 as a main storage device that provides a work area to CPU 71 (and GPU 72), ROM 74 as an auxiliary storage device, communication interface 75, display 76, input interface 77 that accepts user input by mouse, keyboard, touch panel, etc. A heartbeat sensor 78 and a bus BS for exchanging data therebetween.

Each functional unit of the heartbeat data analysis device 10 and the measurement device 20 can be realized by the CPU 71 and/or the GPU 72 that reads and executes a predetermined program corresponding to the function of each unit from the ROM 74 . Both the CPU 71 and the GPU 72 are a kind of arithmetic unit (processor). Here, when display-related processing is performed, the display 76 further operates in conjunction, and when communication-related processing relating to data transmission/reception is performed, the communication interface 75 further operates in conjunction. A heartbeat sensor 78 configured as an ECG sensor, a PPG sensor, or the like can be used as a sensor that provides a heartbeat measurement function in the measuring device 20 . Regarding the measuring device 20 having the heart rate sensor 78, instead of the CPU 71 and / or GPU 72, a dedicated integrated circuit (ASIC (application specific IC) or FPGA (Field Programmable Gate Array), etc.).

10...heartbeat data analysis device, 20...measurement device, 1...R-R interval extraction unit, 2...missing determination unit, 3...intermediate function generation unit, 4...analysis unit, 31...generation unit, 32...correction unit

Claims (7)

An intermediate function generation unit that reads an RR interval string, which is a string of pulsation intervals extracted from a windowed heartbeat data interval from time-series heartbeat data, converts the RR interval string, and generates an intermediate function. and,
A heartbeat data analysis device comprising a frequency analysis unit that obtains a result of frequency analysis of the intermediate function,
a missing determination unit that determines whether or not there is a missing portion of pulsation in the heartbeat data interval;
When the lack determination unit determines that there is a lack of pulsation, the intermediate function generation unit converts the RR interval sequence to generate an intermediate function, and further corresponds to the lack of pulsation in the intermediate function. A heartbeat data analysis apparatus, wherein the frequency analysis unit generates an intermediate function to be subjected to frequency analysis by correcting a portion where the pulsation is absent so that there is no missing pulsation. The lack determination unit determines that the heartbeat data section is long when an RR interval determined to be long by threshold determination exists in a plurality of RR intervals included in the RR interval sequence. 2. The heartbeat data analysis apparatus according to claim 1, wherein the determination is that there is a lack of pulsation at the RR interval. The intermediate function generation unit uses the portion (R7) corresponding to the missing pulsation in the intermediate function as correction source data used for correcting the portion corresponding to the missing pulsation (R7) to a state in which there is no missing pulsation. Heart rate data analysis device according to claim 1 or 2, characterized in that it refers only to the part (R7B, R7A) of the intermediate function corresponding to the RR interval located at . The intermediate function generation unit sets the location (R7) corresponding to the missing pulsation in the intermediate function to a position immediately before the location corresponding to the missing pulsation as correction source data used for correcting the missing pulsation so that there is no missing pulsation. and an intermediate function portion (R7B) corresponding to one RR interval, and an intermediate function portion (R7A) corresponding to one RR interval located immediately after the location corresponding to the missing pulsation. 3. The heartbeat data analysis device according to claim 1, wherein only . 5. The heartbeat data analysis apparatus according to claim 1, wherein said intermediate function generator generates said intermediate function as a rectangular function. 5. The heartbeat data analysis apparatus according to claim 1, wherein said intermediate function generator generates said intermediate function as a pulse train function. A program that causes a computer to function as the heartbeat data analysis device according to any one of claims 1 to 6.

Images