WO2019153579A1

WO2019153579A1 - Electrocardiographic signal-based airbag-free blood pressure measurement method and system

Info

Publication number: WO2019153579A1
Application number: PCT/CN2018/088430
Authority: WO
Inventors: 王新安; 李冉; 刘彦伶; 赵天夏; 李秋平; 陈红英; 何春舅; 马浩; 孙贺
Original assignee: 北京大学深圳研究生院
Priority date: 2018-02-06
Filing date: 2018-05-25
Publication date: 2019-08-15
Also published as: CN108652604A; CN108652604B

Abstract

Disclosed in the present invention are an electrocardiographic signal-based airbag-free blood pressure measurement method and system. Said method comprises acquiring an electrocardiographic signal, and acquiring a pRRx sequence corresponding to the electrocardiographic signal, performing linear analysis and/or non-linear analysis on the pRRx sequence to obtain a corresponding characteristic index, performing machine learning using the characteristic index and the corresponding blood pressure value, which are obtained by means of calculation, as an input and a label, and performing training to obtain a model function of the correlation between the characteristic index of the electrocardiographic signal and the blood pressure value; and when the blood pressure value at a certain time point is to be detected, acquiring an electrocardiographic signal prior to the time point, performing calculation and obtaining, according to the characteristic index of the electrocardiographic signal, the blood pressure value at the time point by means of the model function. Compared with the prior art, the present invention acquires an electrocardiographic signal as a source signal in a non-invasive and airbag-free manner, having a low cost, being safe and effective, facilitating continuous measurement operations, and providing a better user experience; in addition, this method involves a small amount of calculation in the blood pressure measurement process, enabling low algorithm complexity, and high efficiency.

Description

Method and system for detecting airbag blood pressure based on ECG signal

Technical field

The invention relates to the technical field of air bagless blood pressure detection, and particularly relates to a method and system for detecting air balloon blood pressure based on an electrocardiogram signal.

Background technique

Blood Pressure (BP), as an important physiological parameter of the human body, is an important basis for judging cardiovascular diseases in humans. Blood pressure measurement methods are divided into direct measurement method and indirect measurement method.

Direct measurement is a measurement of the direct contact of the measurement system with blood, which is also referred to as damage measurement because it damages the skin and blood vessels.

Indirect measurement, also known as non-invasive measurement, can be subdivided into intermittent measurement and continuous measurement. The intermittent airbag measurement method has experienced more than 100 years of history. The measured blood pressure is close to the intra-aortic pressure (arterial tension method), which is the most common and most common examination method in the clinical diagnosis process. At the same time, due to the balloon to the muscles and blood vessels. With a large squeezing force, this method cannot be extended to continuous measurement.

In the non-invasive continuous blood pressure measurement method, mainly includes: arterial tension method and volume compensation method. Among them, it is difficult to keep the measurement position of the sensor relatively fixed when the arterial tension method measures blood pressure for a long time. The long-term measurement by the volume compensation method may cause venous congestion, which brings uncomfortable feeling or even tenderness to the subject, and the measuring device is complicated.

Summary of the invention

The technical problem mainly solved by the present invention is a commonly used method for indirect blood pressure measurement of a balloon. Because the balloon has a large pressing force on muscles and blood vessels, it cannot be extended to continuous measurement, and the existing indirect non-invasive continuous blood pressure detecting method also has operational difficulties. Problems such as poor user experience.

In order to solve the above technical problem, the present invention provides a new indirect blood pressure measurement method, that is, a method for detecting an air bagn blood pressure based on an electrocardiogram signal, comprising: acquiring an electrocardiogram signal; and calculating a corresponding blood pressure according to the ECG signal value.

In another aspect, the present invention also provides a balloonless blood pressure detecting system based on an electrocardiographic signal, comprising: an electrocardiographic signal collecting device for collecting an electrocardiogram signal of a subject to be detected; and a processor for performing the above method.

In another aspect, the present invention also provides a non-balloon blood pressure detecting product based on an electrocardiographic signal, comprising: a memory for storing a program; a processor for implementing the method as described above by executing the program stored by the memory .

In another aspect, the invention also provides a computer readable storage medium comprising a program executable by a processor to implement the method as described above.

Compared with the prior art, the ECG-free airbag blood pressure detecting method adopted by the present invention uses a non-invasive and non-balloon to collect an ECG signal as a source signal, which is low in cost, safe and effective, easy to measure continuously, and has a good user experience, and The blood pressure detecting process of the method has a small amount of calculation, a low complexity of the algorithm, and high efficiency.

DRAWINGS

1 is a flow chart of a method for detecting blood pressure without airbag based on an electrocardiogram signal;

2 is a flow chart showing a method for establishing a model function corresponding to a characteristic index of an electrocardiographic signal and a blood pressure value;

3 is a schematic diagram of a balloonless blood pressure detecting system based on an electrocardiographic signal;

Figure 4 is a schematic diagram of a non-balloon blood pressure detecting product based on an electrocardiographic signal.

Detailed ways

The present invention will be further described in detail below with reference to the accompanying drawings. Similar elements in different embodiments employ associated similar component numbers. In the following embodiments, many of the details are described in order to provide a better understanding of the application. However, those skilled in the art can easily realize that some of the features may be omitted in different situations, or may be replaced by other components, materials, and methods. In some cases, some operations related to the present application have not been shown or described in the specification, in order to avoid that the core portion of the present application is overwhelmed by excessive description, and those skilled in the art will describe these in detail. Related operations are not necessary, they can fully understand the relevant operations according to the description in the manual and the general technical knowledge in the field.

In addition, the features, operations, or characteristics described in the specification may be combined in any suitable manner to form various embodiments. At the same time, the various steps or actions in the method description can also be sequentially reversed or adjusted in a manner apparent to those skilled in the art. Therefore, the various sequences in the specification and the drawings are only for the purpose of describing a particular embodiment, and are not intended to

The serial numbers themselves for the components herein, such as "first", "second", etc., are only used to distinguish the described objects, and do not have any order or technical meaning. As used herein, "connected" or "coupled", unless otherwise specified, includes both direct and indirect connections (joining).

The ECG-based balloonless blood pressure detecting method proposed by the present invention is mainly based on an RR interval sequence of an electrocardiographic signal, wherein the RR interval refers to a time interval between adjacent R peaks and R peaks in the waveform of the electrocardiogram signal, and the RR interval The sequence includes all RR intervals in a segment of the ECG signal.

Embodiment 1 of the present invention: Please refer to FIG. 1 , a method for detecting blood pressure without airbag based on an electrocardiogram signal, which includes steps A000 to A100, which are specifically described below:

A000: Acquire the ECG signal of the person to be tested.

A100: Calculate a corresponding blood pressure value according to the ECG signal.

In an embodiment, the step A100 includes: calculating one or more characteristic indicators of the electrocardiographic signal according to the electrocardiographic signal, and calculating a corresponding blood pressure value according to the characteristic index of the electrocardiographic signal.

In an embodiment, the characteristic index of the electrocardiographic signal comprises: performing linear analysis on the pRRx sequence of the electrocardiographic signal to obtain one or more linear characteristic indicators, and/or performing nonlinear analysis to obtain one or more Nonlinear feature indicators. The pRRx sequence of any one of the ECG signals is calculated by calculating a ratio of the number of adjacent RR intervals in the segment of the ECG signal that is greater than the threshold x milliseconds to the number of all RR intervals, and the setting values are different. The threshold x is obtained as a ratio corresponding to each threshold x, and these ratios constitute the pRRx sequence. In this embodiment, the ratio is expressed as a percentage, as shown in equation (1):

One or more characteristic indicators can be obtained by performing linear analysis and/or nonlinear analysis based on the pRRx sequence of the electrocardiographic signal.

For example, the characteristic indicators obtained by the linear analysis may include: the mean AVRR of the pRRx sequence, the standard deviation SDRR of the pRRx sequence, the root mean square rMSSD of the adjacent pRRx difference in the pRRx sequence, and the standard deviation SDSD of the adjacent pRRx difference in the pRRx sequence. .

The nonlinear analysis of the pRRx sequence of each segment of the ECG signal is performed by entropy analysis, that is, according to the prior art, for the random variable set A of the probability distribution function p(x), the definition of entropy is as shown in equation (2). Show:

H(A)=-∑p _A (x)logp _A (x) (2)

The available metrics include:

(1) The pRRx sequence histogram distribution information entropy S _dh is the numerical distribution information entropy of the pRRx sequence;

(2) pRRx sequence power spectrum histogram distribution information entropy S _ph is a discrete Fourier transform of the pRRx sequence to obtain the power spectrum, and then calculate its information entropy according to the numerical distribution of the power spectrum sequence;

(3) pRRx sequence power spectrum full-band distribution information entropy S _pf is a discrete Fourier transform of the pRRx sequence to obtain the power spectrum in the full frequency band [f _s /N, f _s /2] (the sampling frequency of the signal is f _s The number of sampling points is N). The i-1 sub-points f ₁ , f ₂ , . . . , f _m-1 are inserted, and the full frequency band is divided into i sub-bands. Taking the sum of the power densities in each band as the power density of the band, m power densities are obtained. The i power densities are normalized to obtain the probability p _i appearing in each frequency band, then ∑ _i p _i =1, and the corresponding power spectrum full-band entropy is as shown in equation (3):

For the nonlinear analysis of the pRRx sequence of each segment of the ECG signal, the following four fractal dimension calculation methods can also be used to obtain the following characteristic indicators:

(1) The fractal dimension D _sf calculated by the structural function method, wherein the structural function method means that for a given sequence z(x), the incremental variance is defined as a structural function, and the relationship is:

For a number of scales τ, calculate the corresponding S(τ) for the discrete values of the sequence z(x), then plot the logS(τ)-logτ function curve and perform a linear fit in the scale-free region to obtain the slope. α, then the transformation relationship between the fractal dimension D _sf and the slope α is as shown in equation (5):

(2) The fractal dimension D _cf calculated by the correlation function method, wherein the correlation function method means that for a given sequence z(x), the correlation function C(τ) is defined as equation (6):

C(τ)=AVE(z(x+τ)*z(x)),τ=1,2,3,K,N-1 (6)

Among them, AVE (·) represents the average, and τ represents the two-point distance. At this time, the correlation function is a power type. Since there is no feature length, the distribution is fractal, and there is C(τ)ατ ^-α . At this time, the function curve of logC(τ)-logτ is drawn, and linear fitting is performed in the scale-free region to obtain the slope α. The conversion relationship between the fractal dimension D _cf and the slope α is as shown in equation (7):

D _cf =2-α (7)

(3) The fractal dimension D _vm calculated by the variation method, wherein the variation method covers the fractal curve with the width of τ and the rectangular frame end-to-end, so that the maximum and minimum values of the curve in the i-th frame are obtained. The difference is H(i), which is the height of the rectangle. Multiply the height and width of all rectangles to obtain the total area S(τ). Change the size of τ to get a series of S(τ). As shown in equation (8):

The function curve of logN(τ)-logτ is drawn, and the slope α is obtained by linear fitting in the scale-free region. The conversion relationship between the corresponding fractal dimension D _vm and the slope α is as shown in equation (7).

(4) The fractal dimension D _rms calculated by the root mean square method, wherein the root mean square method covers the fractal curve with the rectangular frame of width τ end to end, so that the maximum value of the curve in the i-th frame is The difference between the minimum values is H(i), which is the height of the rectangle. Calculate the root mean square value S(τ) of these rectangular heights. Change the size of τ to get a series of S(τ). The function curve of logS(τ)-logτ is drawn, and the slope α is obtained by linear fitting in the scale-free region. The conversion relationship between the corresponding fractal dimension D _rms and the slope α is as shown in equation (7).

The ECG signal characteristic index used for calculating the blood pressure value is one or more of the above-mentioned linear and/or non-linear analysis characteristic indexes, or a set of several of them, or may be listed in the present embodiment. Corresponding feature indicators obtained from existing analytical methods.

In an embodiment, when calculating the corresponding blood pressure value according to the characteristic index of the electrocardiographic signal, the A100 step may pre-establish a model function corresponding to the characteristic index of the ECG signal and the blood pressure value, and input the characteristic index of the ECG signal into the model. Function, get the corresponding blood pressure value. For example, the A100 step can establish a model function corresponding to the characteristic index of the ECG signal and the blood pressure value through machine learning and training, as shown in FIG. 2 .

As shown in FIG. 2, the A100 step establishes the above model function, which may include steps A110 to A112, which are specifically described below.

A110: Acquire a plurality of blood pressure values in advance, and an ECG signal before the time point of each blood pressure value. Wherein, the obtaining blood pressure values, including exercise and meditation, different emotional states, before and after taking antihypertensive drugs, morning and afternoon, different sleep states, and the like, may also increase the blood pressure value as needed. Time point; the method for obtaining the blood pressure value in this step may adopt a method with high precision commonly used in the prior art, such as an invasive or air bag blood pressure meter, and at the same time, an electrocardiogram is required for each blood pressure value. Signals, due to differences in individual metabolic conditions, the time length of ECG signals required by each sampler is not the same. The actual modeling effect is correct. In this embodiment, ECG signals of different lengths of 1 to 30 minutes are selected.

A111: Obtain the characteristic indicators of these ECG signals.

A112: Taking the characteristic indexes of the electrocardiographic signals as input, the blood pressure values corresponding to the electrocardiographic signals are used as labels, and machine learning is performed to obtain a model function corresponding to the characteristic index of the electrocardiographic signal and the blood pressure value. Wherein, the pre-acquired blood pressure values and the electrocardiographic signals are all taken from the same subject, and the obtained blood pressure detection model is also used for the non-balloon blood pressure detection of the same subject. In addition, a separate model function is required for each person's diastolic and systolic pressures. Further, in the present embodiment, when performing machine learning, the output blood pressure value is divided into sections according to a step size of 5 mmHg, the diastolic pressure section is 40 to 150 mmHg, and the systolic pressure section is 50 to 300 mmHg, thereby dividing the systolic pressure and the diastolic blood pressure. It has become a number of intervals.

After obtaining the model function corresponding to the characteristic index of the ECG signal and the blood pressure value according to the above steps, the ECG signal of the person to be detected obtained in the step A000 is input into the model function, and the blood pressure value is obtained, and the blood balloon blood pressure detection is completed. By continuously acquiring several segments of ECG signals and inputting the model function, the blood pressure values corresponding to the continuous detection can be obtained. In this embodiment, the specific output values of systolic pressure and diastolic blood pressure are taken as the median value of the interval and rounded up, for example, the systolic blood pressure measured by the model [70, 75], the median value of the interval is 72.5, and the upward rounding is 73, the output systolic pressure is a value of 73 mmHg.

Embodiment 2: A non-balloon blood pressure detecting system based on an electrocardiographic signal, as shown in FIG. 3, includes an ECG signal collecting device B00 and a processor B10, which are specifically described below:

The ECG signal collecting device B00 is configured to collect an ECG signal of the person to be detected;

The processor B10 is configured to perform the ECG-based airbag-free blood pressure detecting method according to any of the above embodiments. For example, the processor B10 may calculate one or more characteristic indicators of the electrocardiographic signal according to the electrocardiographic signal, and calculate a corresponding blood pressure value according to the characteristic index of the electrocardiographic signal. On the other hand, the processor B10 can establish a model function corresponding to the characteristic index of the electrocardiographic signal and the blood pressure value in advance, and input the characteristic index of the electrocardiographic signal into the model function to obtain a corresponding blood pressure value. The processor B10 obtains a plurality of blood pressure values and an ECG signal before the time point of each blood pressure value in advance; acquires characteristic indexes of the ECG signals; and takes the characteristic indexes of the ECG signals as inputs, and the ECG signals correspond to The blood pressure value is used as a label, and machine learning is performed to obtain a model function corresponding to the characteristic index of the electrocardiographic signal and the blood pressure value.

Embodiment 3: A non-balloon blood pressure detecting product C00 based on an electrocardiographic signal, as shown in FIG. 4, includes a memory C01 and a processor C02, which are specifically described below:

a memory C01 for storing a program;

The processor C02 is configured to implement the ECG-based airbag-free blood pressure detecting method according to any one of the embodiments described above by executing the program stored in the memory. For example, the processor C02 executes the program stored in the memory C01, and can calculate one or more characteristic indicators of the electrocardiographic signal according to the electrocardiographic signal, and calculate the corresponding blood pressure value according to the characteristic index of the electrocardiographic signal. On the other hand, the program stored in the memory C01 can also be used to pre-establish a model function corresponding to the characteristic index of the electrocardiographic signal and the blood pressure value, and input the characteristic index of the electrocardiographic signal into the model function to obtain the corresponding blood pressure value. On the other hand, the processor C02 executes the program stored in the memory C01, by acquiring a plurality of blood pressure values in advance, and an electrocardiographic signal before the time point of each blood pressure value; acquiring characteristic indexes of the electrocardiographic signals; The characteristic index of the signal is used as an input, and the blood pressure value corresponding to these electrocardiographic signals is used as a label, and machine learning is performed to obtain a model function corresponding to the characteristic index of the electrocardiographic signal and the blood pressure value.

By combining the method of the first embodiment, using the device in the system of the second embodiment, the blood pressure value can be obtained based on non-invasive and non-balloon detection of the electrocardiographic signal. Such a device is low in cost, safe and effective, easy to measure continuously, and has a good user experience, and the blood pressure detecting process has a small amount of calculation, a low complexity of the algorithm, and high efficiency.

Those skilled in the art can understand that all or part of the functions of the various methods in the above embodiments may be implemented by hardware or by a computer program. When all or part of the functions in the above embodiments are implemented by a computer program, the program may be stored in a computer readable storage medium, and the storage medium may include: a read only memory, a random access memory, a magnetic disk, an optical disk, a hard disk, etc. The computer executes the program to implement the above functions. For example, the program is stored in the memory of the device, and when the program in the memory is executed by the processor, all or part of the above functions can be realized. In addition, when all or part of the functions in the above embodiment are implemented by a computer program, the program may also be stored in a storage medium such as a server, another computer, a magnetic disk, an optical disk, a flash disk or a mobile hard disk, and may be saved by downloading or copying. The system is updated in the memory of the local device, or the system of the local device is updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments may be implemented.

The invention has been described above with reference to specific examples, which are merely intended to aid the understanding of the invention and are not intended to limit the invention. For the person skilled in the art to which the invention pertains, several simple derivations, variations or substitutions can be made in accordance with the inventive concept.

Claims

A method for detecting blood pressure without airbag based on an electrocardiogram signal, comprising:

Acquire an ECG signal;

Based on the ECG signal, a corresponding blood pressure value is calculated.
The method according to claim 1, wherein the calculating the corresponding blood pressure value according to the ECG signal comprises: calculating one or more characteristic indicators of the ECG signal according to the ECG signal, according to the characteristic index of the ECG signal, Calculate the corresponding blood pressure value.
The method according to claim 2, comprising: pre-establishing a model function corresponding to the characteristic index of the electrocardiographic signal and the blood pressure value, and inputting the characteristic index of the electrocardiographic signal into the model function to obtain a corresponding blood pressure value.
The method according to claim 2 or 3, wherein the characteristic index of the electrocardiographic signal comprises: linearly analyzing the pRRx sequence of the electrocardiographic signal to obtain one or more linear characteristic indicators, and/or performing nonlinearity. Analysis to obtain one or more nonlinear characteristic indicators; wherein the pRRx sequence of any one of the ECG signals is calculated by calculating the number of adjacent RR intervals in the segment of the ECG signal that is greater than the threshold x milliseconds and The ratio of the number of all RR intervals is obtained by setting a threshold x having a different value to obtain a ratio corresponding to each threshold x, and these ratios constitute the pRRx sequence.
The method of claim 4, wherein the characteristic indicator of the electrocardiographic signal further comprises:

The characteristic indexes obtained by the linear analysis are: the mean AVRR of the pRRx sequence, the standard deviation SDRR of the pRRx sequence, the root mean square rMSSD of the difference of the adjacent pRRx in the pRRx sequence, and the standard deviation SDSD of the difference of the adjacent pRRx in the pRRx sequence. At least one; and/or,

The nonlinear characteristic index includes a feature index obtained by entropy analysis of the pRRx sequence, including: pRRx sequence histogram distribution information entropy S dh , pRRx sequence power spectrum histogram distribution information entropy S ph , pRRx sequence power spectrum At least one of the full-band distribution information entropy S pf ; and/or the non-linear feature index includes a feature index obtained by performing the fractal dimension calculation and analysis of the pRRx sequence, including: a fractal obtained by the structural function method The dimension D sf , the fractal dimension D cf calculated by the correlation function method, the fractal dimension D vm calculated by the variogram method, and at least one of the fractal dimension D rms calculated by the root mean square method.
The method according to claim 3, wherein said model function for pre-establishing a correspondence between a characteristic index of the electrocardiographic signal and a blood pressure value comprises:

Obtaining a plurality of blood pressure values in advance, and an ECG signal before the time point of each blood pressure value;

Obtaining the characteristic indicators of these ECG signals;

The characteristic indexes of these electrocardiographic signals are input, and the blood pressure values corresponding to the electrocardiographic signals are used as labels, and machine learning is performed to obtain a model function corresponding to the characteristic index of the electrocardiographic signal and the blood pressure value.
The method according to claim 6, wherein when the model function is established by machine learning, the output blood pressure value is divided into sections according to a step size of 5 mmHg, and the specific output blood pressure value is a median value of the interval and rounded up.
A balloonless blood pressure detecting system based on an electrocardiographic signal, comprising:

An ECG signal collecting device for collecting an ECG signal of a person to be detected;

A processor for performing the method of any of claims 1-7.
A non-balloon blood pressure detecting product based on an electrocardiographic signal, characterized in that it comprises:

Memory for storing programs;

A processor for performing the method of the memory storage to implement the method of any of claims 1-7.
A computer readable storage medium, comprising a program executable by a processor to implement the method of any of claims 1-7.