Ultrahigh-frequency electrocardio mapping analysis method and system

Abstract

According to the ultrahigh-frequency electrocardiogram mapping analysis method and system, through the Hilbert change algorithm, the time-frequency characteristics of high-frequency electrocardiogram signals can be reserved, and the influence of complex clutters on the high-frequency electrocardiogram signals can be reduced to the maximum extent; the time-space distribution of electrical activity is described in a graphical manner by drawing an electrical activation diagram, so that the problem of atrial / ventricular desynchrony or myocardial ischemia can be revealed more accurately; the model is helped to better understand and process new data through transfer learning, and the problem of poor model training effect caused by less high-frequency electrocardiogram data is effectively solved; by constructing an intelligent diagnosis system, the automation level of diagnosis is improved, and auxiliary diagnosis of inputting an electrical activation diagram and outputting a classification result is realized.

Description

Technical Field

[0001] The present invention belongs to the field of biomedical electronic technology, and in particular relates to an ultra-high frequency electrocardiogram mapping and analysis method and system. Background Art

[0002] Electrocardiogram technology is a basic first-line tool for clinical diagnosis of cardiovascular diseases, and the technology has been developed for more than a hundred years. The frequency response range of ECG acquisition devices is usually 0-150Hz. Since the 1980s, Goldberg et al. have collected high-frequency ECG with a frequency response ECG device of 150-250Hz, and proposed the corresponding evaluation index Root Mean Square (RMS) to judge myocardial ischemia, which improved the accuracy of myocardial ischemia diagnosis. In the following twenty years, Abboud S et al. enriched the detection methods of high-frequency QRS, studied the high-frequency ECG information with a frequency upper limit of 250Hz or 300Hz, proposed the Reduced Amplitude Zone (RAZ) indicator, and explored its application in the field of myocardial ischemia detection. With the development of digital signal processing and electronic technology, higher frequency Ultra-High-Frequency Electrocardiogram (Ultra-High-Frequency Electrocardiogram) mapping technology (frequency response exceeding 300 Hz) has begun to attract attention, and is expected to provide a potential non-invasive and quantitative new mapping method for clinical detection of cardiac electrophysiology, myocardial ischemia, and atrial / ventricular asynchrony. Summary of the invention

[0003] In view of the current lack of ultra-high frequency ECG mapping and analysis devices, especially the lack of ultra-high frequency ECG mapping and analysis devices for non-invasive and quantitative analysis of cardiac asynchrony and myocardial ischemia, the present invention proposes an ultra-high frequency ECG mapping and analysis method and system, which is non-invasive and can provide quantitative and positioning analysis to assist doctors in evaluating the suitability of myocardial ischemia and cardiac resynchronization therapy; at the same time, it combines machine learning and deep learning technologies to improve diagnostic accuracy and automation level.

[0004] The specific technical solutions are as follows:

[0005] A method for ultra-high frequency electrocardiogram mapping and analysis,

[0006] S1: Acquire high-frequency ECG data: Use a high-frequency ECG acquisition device to acquire high-frequency ECG signal data; the high-frequency ECG signal data refers to ECG waveform data acquired by a high-frequency ECG signal acquisition device with a frequency response range of 0 to 1000 Hz, an ADC accuracy greater than or equal to 24 bits, and 6 to 18 leads; the leads include chest leads, limb leads, and pressurized limb leads;

[0007] S2: Data preprocessing: Use Butterworth IIR filter and notch filter to preprocess the high-frequency ECG signal data collected by S1;

[0008] S3: Heart beat division based on R wave: Use Pan-Tompkins algorithm to find R wave in the high-frequency ECG signal data collected by limb leads or pressurized limb leads after S2 preprocessing; and record the time nodes at the waveform of the specified time length before and after the peak of the R wave; according to the recorded time nodes, the high-frequency ECG signal collected by the chest lead is cut to obtain a high-frequency ECG signal of a heart beat, i.e., the heart beat;

[0009] S4: Feature extraction: using the Hilbert transform algorithm to calculate the amplitude envelope of the QRS wave in the high-frequency electrocardiogram signal of each frequency band in the continuous frequency band for the heartbeat obtained in S3; the continuous frequency band refers to a series of continuous frequency ranges, one frequency range is one frequency band, and the difference between the upper limit and the lower limit of each frequency range is the same;

[0010] S5: Signal averaging: The amplitude envelope of the QRS wave obtained in S4 is averaged by the signal averaging technique; the averaged amplitude envelope is low-pass filtered to form a feature vector corresponding to the amplitude envelope of each lead;

[0011] S6: Draw the electrical activation map: Use the characteristic vectors of all chest leads in S1 to form a characteristic matrix, namely the electrical activity conduction matrix; calculate the time difference between lead excitations, namely the asynchronous time difference, through the electrical activity conduction matrix; use the normalization algorithm to normalize the electrical activity conduction matrix so that its amplitude is between 0 and 1; use the colormap algorithm to draw the normalized electrical activity conduction matrix into a heat map to obtain the electrical activation map;

[0012] S7: Pattern recognition: The S6 electrical activation map is input into the VGG19 model based on transfer learning, and the preliminary auxiliary diagnosis results of atrial / ventricular dyssynchrony or myocardial ischemia or non-detection are output;

[0013] S8: Result visualization: The high-frequency ECG signals collected by 6-18 leads obtained by S1, the electrical activation map and asynchronous time difference obtained by S6, and the preliminary auxiliary diagnosis results obtained by S7 are displayed on the user interface.

[0014] Preferably, the 6-18 leads include: 6, 9, 12, 14 and 18 leads; wherein:

[0015] The 6 leads are six chest leads V1-V6;

[0016] The 9 leads are the 6 leads plus three limb leads I, II, and III;

[0017] The 12 leads are based on the 9 leads and have three additional limb leads, i.e., the standard 12 leads.

[0018] The 14 leads are based on the 12 leads and have V7 and V8 chest leads added;

[0019] The 18 leads are based on the 14 leads and add four more leads, namely V9, V3R, V4R and V5R. The V3R lead is located on the right chest symmetrical to V3; the V4R lead is located on the right chest symmetrical to V4; the V5R lead is located on the right chest symmetrical to V5; and V9 is the left posterior chest wall lead.

[0020] Preferably, the data preprocessing method described in S2 is: using a Butterworth IIR filter to filter out low-frequency baseline drift of high-frequency ECG signal data; using a notch filter to filter out power frequency interference of a specified frequency in the high-frequency ECG signal data.

[0021] Preferably, the Hilbert transform algorithm described in S4 is to transform the divided high-frequency ECG signal into an amplitude envelope signal, and the specific steps are as follows:

[0022] S41: Calculate Fourier transform: Set the Fourier transform of the high-frequency ECG signal x(t) after the heartbeat division to X(jw), and according to the Fourier transform formula, we get:

[0023]

[0024] In formula (1), x * (t) is the conjugate of x(t), j is the imaginary unit; if the phase is changed by 90°, that is, ω is -ω, then x can be calculated * The Fourier transform of (t) is X * (-jω);

[0025] S42: Calculate Hilbert transform: The Fourier transform of a real signal contains negative frequencies. If you want to remove the negative frequencies but keep the total power unchanged, multiply X(jw) by 2u(t), where u(t) is a step function, to obtain a high-frequency ECG signal after Hilbert transform:

[0026]

[0027] The frequency response characteristics are:

[0028]

[0029]

[0030] It can be seen from equation (4) that the Hilbert transform shifts the phase of all positive frequency components by 90 degrees and shifts the phase of all negative frequency components by 90 degrees.

[0031] S43: Construct analytical signal: Convert the real signal into a complex signal, take the high-frequency ECG signal x(t) after the original heartbeat division as the real part, and the high-frequency ECG signal after Hilbert transformation As the imaginary part, we get:

[0032]

[0033] The high-frequency ECG signal x(t) after heart beat division has both amplitude information and phase information. Let x(t) be:

[0034] x(t)=A(t)cos(ωt+θ(t)) (6)

[0035] Among them, A(t) represents the amplitude information of the signal, and θ(t) represents the phase information;

[0036] Substituting formula (6) into formula (5), we get:

[0037]

[0038] S44: Calculate and obtain the amplitude envelope signal: Take the absolute value to get the amplitude envelope signal:

[0039]

[0040] Preferably, the specific steps of signal averaging described in S5 are:

[0041] S51: Calculate the average value of the amplitude envelope of the QRS wave corresponding to all the heart beats of S4 in the same lead and the same frequency band to obtain the simplified amplitude envelope of the QRS wave;

[0042] S52: Calculate the median of each time point in the simplified QRS wave amplitude envelope in all frequency bands, i.e., the feature vector of each lead; the median is a one-dimensional feature vector of 1×divided time length.

[0043] Preferably, the asynchronous time difference DYS described in S6 is obtained by calculating the time difference between the lead that starts to be excited earliest and the time difference between all the leads in the electrical activity conduction matrix; the group with the largest time difference between adjacent leads is taken as the main delay lead group, i.e. MDL, and this time difference is the main delay time difference, i.e. δDYS; the value of the asynchronous time difference represents the degree of atrial / ventricular asynchrony, and the larger the value, the stronger the degree of atrial / ventricular asynchrony.

[0044] Preferably, the specific method for obtaining the VGG19 model based on transfer learning in step S7 is:

[0045] S61: pre-training the VGG19 model using the EGC data set to obtain the pre-training model weight; the intermediate variable of the VGG19 model is a replaceable feature variable, which is a key carrier for transmitting and processing information between layers of the VGG19 model;

[0046] S62: Establish a pattern recognition model, namely, a VGG19 model; use a transfer learning algorithm to transfer the weights of the S61 pre-trained model to the VGG19 model, and use the electrical activation map dataset to fine-tune the VGG19 model to obtain a VGG19 model for detecting atrial / ventricular dyssynchrony or myocardial ischemia; the fine-tuning is to use the electrical activation map as the input of the model, and the output is a preliminary auxiliary diagnosis result of whether atrial / ventricular dyssynchrony or myocardial ischemia is detected or not;

[0047] S63: Verify the fine-tuned model. The indicators for verifying model performance include accuracy, F1 score, precision and recall.

[0048] An ultra-high frequency electrocardiogram mapping and analysis system, comprising:

[0049] Data acquisition module: using high-frequency ECG acquisition equipment to collect high-frequency ECG signal data;

[0050] Data preprocessing module: The Butterworth IIR filter unit is used to filter out the low-frequency baseline drift in the collected high-frequency ECG signal data; and the notch filter is used to filter out the power frequency interference of the specified frequency in the collected ECG signal data;

[0051] Feature extraction module: using the Hilbert transform unit to calculate the amplitude envelope of the ECG signal data processed by the data preprocessing module within the specified frequency range; using the signal averaging technology unit to average the amplitude envelope obtained by the Hilbert transform unit; using the low-pass filter unit to filter the amplitude envelope averaged by the signal averaging technology unit to form a waveform corresponding to the amplitude envelope of each lead;

[0052] Visualization module: construct a feature matrix, i.e., an electrical activity conduction matrix, by using a matrix construction unit to construct the feature vector of the specified lead; input the electrical activity conduction matrix into a calculation unit for calculating the asynchronous time difference to calculate the time difference between the lead excitations; and use a colormap algorithm unit to draw the electrical activity conduction matrix into a heat map to obtain an electrical activation map;

[0053] Pattern recognition module: input the electrical activation map of the visualization module into the VGG19 model unit to obtain the preliminary auxiliary diagnosis result of whether atrial / ventricular asynchrony or myocardial ischemia is detected or not;

[0054] User interface: displays the high-frequency ECG signals of the data acquisition module, the electrical activation diagram and asynchronous time difference of the visualization module, and the preliminary auxiliary diagnosis results and operation options of the pattern recognition module.

[0055] Beneficial effects:

[0056] The present invention proposes an ultra-high frequency electrocardiogram mapping analysis method and system, which provides quantitative and localized analysis to assist doctors in evaluating the suitability of myocardial ischemia and cardiac resynchronization therapy; at the same time, it combines machine learning and deep learning technologies to improve diagnostic accuracy and automation level. Specifically,

[0057] By applying the Hilbert transform algorithm, the signal-to-noise ratio of the UHF signal can be significantly improved, and the noise interference in the UHF band can be effectively suppressed, thereby protecting the effective signal from being affected. This algorithm can not only preserve the time-frequency characteristics of the signal, but also minimize the interference of complex clutter on the signal, ensuring the accuracy and reliability of signal analysis.

[0058] Drawing an electrical activation map can accurately reveal delays or blockages in cardiac electrical conduction and indicate which leads these problems occur near. This map not only provides qualitative information about the degree of myocardial ischemia and cardiac asynchrony, but also provides quantitative and localized analysis to assist doctors in assessing whether patients are suitable for cardiac resynchronization therapy. By calculating the delay time of each lead in the electrical activation map and the difference in delays between each other, doctors can determine the degree of electrical asynchrony in a specific ventricular area.

[0059] By building an intelligent diagnosis system, pre-training the model can learn rich feature representations, and then transfer learning can help the model better understand and process new data, which can effectively solve the problem of poor model training results caused by insufficient high-frequency ECG data, and significantly improve the model recognition accuracy; machine learning and deep learning technologies are used to improve the accuracy and automation level of diagnosis. According to the presented electrical activation map, deep learning based on VGG19 is performed to achieve automated and intelligent diagnosis of atrial / ventricular asynchrony or myocardial ischemia. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A flow chart of an ultra-high frequency electrocardiogram mapping and analysis method.

[0061] Figure 2 A schematic diagram of the results of each step in the process of ultra-high frequency electrocardiogram mapping and analysis method.

[0062] Figure 3 A schematic diagram of the results of each step of a specific example of an ultra-high frequency electrocardiogram mapping and analysis method.

[0063] Figure 4 Schematic diagram of electrical activation diagram of atrial / ventricular dyssynchrony group.

[0064] Figure 5 Schematic diagram of electrical activation map of normal group.

[0065] Figure 6 An architecture diagram of an ultra-high frequency ECG mapping and analysis system. DETAILED DESCRIPTION

[0066] like Figure 1 As shown, an ultra-high frequency electrocardiogram mapping and analysis method comprises:

[0067] S1: Acquire high-frequency ECG data: Use a high-frequency ECG acquisition device to acquire high-frequency ECG signal data; the high-frequency ECG signal data refers to ECG waveform data acquired by a high-frequency ECG signal acquisition device with a frequency response range of 0 to 1000 Hz, an ADC accuracy greater than or equal to 24 bits, and 6 to 18 leads; the leads include chest leads, limb leads, and pressurized limb leads;

[0068] S2: Data preprocessing: Use Butterworth IIR filter and notch filter to preprocess the high-frequency ECG signal data collected by S1;

[0069] S3: Heart beat division based on R wave: Use Pan-Tompkins algorithm to find R wave in the high-frequency ECG signal data collected by limb leads or pressurized limb leads after S2 preprocessing; and record the time nodes at the waveform of the specified time length before and after the peak of the R wave; according to the recorded time nodes, the high-frequency ECG signal collected by the chest lead is cut to obtain a high-frequency ECG signal of a heart beat, i.e., the heart beat;

[0070] S4: Feature extraction: using the Hilbert transform algorithm to calculate the amplitude envelope of the QRS wave in the high-frequency electrocardiogram signal of each frequency band in the artificially set continuous frequency band for the heartbeat obtained in S3; the continuous frequency band refers to a series of continuous frequency ranges, one frequency range is one frequency band, and the difference between the upper limit and the lower limit of each frequency range is the same;

[0071] S5: Signal averaging: The amplitude envelope of the QRS wave obtained in S4 is averaged by the signal averaging technique; the averaged amplitude envelope is low-pass filtered to form a feature vector corresponding to the amplitude envelope of each lead;

[0072] S6: Draw the electrical activation map: Use the characteristic vectors of all chest leads in S1 to form a characteristic matrix, namely the electrical activity conduction matrix; calculate the time difference between lead excitations, namely the asynchronous time difference, through the electrical activity conduction matrix; use the normalization algorithm to normalize the electrical activity conduction matrix so that its amplitude is between 0 and 1; use the colormap algorithm to draw the normalized electrical activity conduction matrix into a heat map to obtain the electrical activation map;

[0073] S7: Pattern recognition: The S6 electrical activation map is input into the VGG19 model based on transfer learning, and the preliminary auxiliary diagnosis results of atrial / ventricular dyssynchrony or myocardial ischemia or non-detection are output;

[0074] S8: Result visualization: The high-frequency ECG signals collected by 6-18 leads obtained by S1, the electrical activation map and asynchronous time difference obtained by S6, and the preliminary auxiliary diagnosis results obtained by S7 are displayed on the user interface.

[0075] An ultra-high frequency electrocardiogram mapping and analysis system, comprising:

[0076] Data acquisition module: using high-frequency ECG acquisition equipment to collect high-frequency ECG signal data;

[0077] Data preprocessing module: The Butterworth IIR filter unit is used to filter out the low-frequency baseline drift in the collected high-frequency ECG signal data; and the notch filter is used to filter out the power frequency interference of the specified frequency in the collected ECG signal data;

[0078] Feature extraction module: using the Hilbert transform unit to calculate the amplitude envelope of the ECG signal data processed by the data preprocessing module within the specified frequency range; using the signal averaging technology unit to average the amplitude envelope obtained by the Hilbert transform unit; using the low-pass filter unit to filter the amplitude envelope averaged by the signal averaging technology unit to form a waveform corresponding to the amplitude envelope of each lead;

[0079] Visualization module: construct a feature matrix, i.e., an electrical activity conduction matrix, by using a matrix construction unit to construct the feature vector of the specified lead; input the electrical activity conduction matrix into a calculation unit for calculating the asynchronous time difference to calculate the time difference between the lead excitations; and use a colormap algorithm unit to draw the electrical activity conduction matrix into a heat map to obtain an electrical activation map;

[0080] Pattern recognition module: input the electrical activation map of the visualization module into the VGG19 model unit to obtain the preliminary auxiliary diagnosis result of whether atrial / ventricular asynchrony or myocardial ischemia is detected or not;

[0081] User interface: displays the high-frequency ECG signals of the data acquisition module, the electrical activation diagram and asynchronous time difference of the visualization module, and the preliminary auxiliary diagnosis results and operation options of the pattern recognition module.

[0082] Specific methods:

[0083] like Figure 2 and Figure 3 As shown, a specific process of an ultra-high frequency electrocardiogram mapping and analysis method, wherein Figure 2 The electrical asynchrony map and electrical activation map in the MRI are different names for the same map:

[0084] 1. Data acquisition:

[0085] Use a 6-18 lead ECG acquisition device with a sampling rate of no less than 4000Hz, a frequency response range of 0-1000Hz, and an ADC accuracy of no less than 24 bits to complete at least 1 minute of acquisition to obtain sufficient data. Taking 14 leads as an example, the leads include the standard 12 leads (I lead, II lead, III lead, aVL lead, aVR lead, aVF lead, and six chest leads of V1-V6) and 2 additional chest leads (V7, V8).

[0086] 2. Data preprocessing:

[0087] The Butterworth IIR filter was used for preprocessing to filter out low-frequency baseline drift, and the 50Hz power frequency interference was filtered out using a notch filter. The Pan-Tompkins algorithm was used to divide the heart beats based on the R wave, and the waveform 120ms before and after the R wave peak was cut.

[0088] 3. Feature matrix solution is feature extraction:

[0089] The amplitude envelope was calculated in the frequency range of 100-150Hz, 150-250Hz, 200-300Hz, 250-350Hz, 300-400Hz, 350-450Hz, 400-500Hz, …, 850-950Hz, and 900-1000Hz by Hilbert transform. Signal averaging technology was used to improve the signal-to-noise ratio. The waveform of 120ms before and after each QRS wave divided in the time period was averaged, and the signal was low-pass filtered (passband 0–40Hz) to form the waveform corresponding to each lead. The waveforms were arranged from top to bottom according to leads V1-V8. The formed 2D matrix represents the distribution of electrical activity in time (relative to the time of the R wave) and space (V lead).

[0090] Specifically, a specific example of applying signal averaging technology is to extract 2 leads of ECG for a total of 10 seconds, divide it into 10 heart beats, and then divide it into 9 frequency bands. First, the 10 heart beats in each frequency band are averaged, so that each frequency band is simplified to 1 heart beat, and then the median of the simplified heart beats is calculated, that is, for each point in the heart beat, the median is taken in the 9 frequency bands.

[0091] After constructing the electrical activity conduction matrix, the time difference between the earliest lead to be excited in the matrix and the time of all leads being excited is taken as the asynchronous time difference DYS, and the group with the largest time difference (asynchronous time difference) between adjacent leads is taken as the major delayed leads group (Major Delayed Leads, MDL), and this time difference is the major delayed time difference (Major Delayed Duration, denoted as δDYS).

[0092] Specifically, in the process of index calculation, since the signal-to-noise ratio of the UHF signal is too low, the amplitude of the noise in the UHF band is much larger than the amplitude of the effective signal. Therefore, an algorithm is needed that can both retain the time-frequency characteristics and minimize the impact of complex clutter on the signal - the Hilbert transform algorithm.

[0093] Assume that the Fourier transform of x(t) is X(jw), then according to the Fourier transform formula:

[0094]

[0095] In formula (1), *(t) is the conjugate of x(t), j is the imaginary unit, and if the phase is changed by 90°, that is, ω is -ω, then x can be calculated as * The Fourier transform of (t) is X * (-jω); The Fourier transform of a real signal contains negative frequencies. If you want to remove the negative frequencies but keep the total power unchanged, multiply X(jw) by 2u(t), where u(t) is a step function, and get the Hilbert transform:

[0096]

[0097] The frequency response characteristics are:

[0098]

[0099] It can be seen from equation (4) that the Hilbert transform shifts all positive frequency components by 90 degrees and shifts all negative frequency components by 90 degrees. Constructing an analytical signal, the real signal is converted into a complex signal, and the original signal x(t) is taken as the real part. The signal after the Hilbert transform is As the imaginary part, we get:

[0100]

[0101] The signal x(t) has both amplitude information and phase information. Let

[0102]

[0103] Among them, A(t) represents the amplitude information of the signal, and θ(t) represents the phase information;

[0104] Substituting into formula (5) we get:

[0105]

[0106] right Taking the absolute value gives us the envelope signal:

[0107]

[0108] It should be noted that in practical applications, in order to perform envelope calculations for the ultra-high frequency band separately, bandpass filtering is required first. However, the enveloped signal still has different random errors due to noise, so the divided heartbeat signals within 5 minutes are superimposed and divided by the sum to obtain a section of averaged signal, and filtered to below 40Hz to ensure that the waveform is not affected by noise. The matrix obtained after processing the 8 leads is converted into a heat map (linear interpolation between rows) and is the electrical activation map (EAD). Linear interpolation is an interpolation method for one-dimensional data. It estimates the value based on the two data points adjacent to the point to be interpolated in the one-dimensional data sequence. The average value is usually used for interpolation between leads.

[0109] EAD technology can accurately reveal delays or blockages in cardiac electrical conduction and indicate which leads these problems occur near. This map not only provides qualitative information on the degree of myocardial ischemia and cardiac asynchrony, but also performs quantitative and location analysis to assist doctors in assessing whether patients are suitable for cardiac resynchronization therapy. By calculating the delay time of each lead in the EAD and the delay difference between each other, doctors can determine the degree of electrical asynchrony in a specific ventricular area. These calculation results constitute the Dyssynchronization Parameter System (DPS), which provides a set of standardized indicators for evaluating and monitoring the effectiveness of treatment. If certain parameters in the DPS are alleviated, it indicates that the electrical asynchrony in a specific ventricular area (such as between the ventricular septum and the left ventricle, between the left and right ventricles, or inside the left ventricle) has improved, and vice versa.

[0110] 4. Pattern Recognition:

[0111] A machine learning algorithm is used to classify and locate atrial / ventricular dyssynchrony or myocardial ischemia, where atrial / ventricular dyssynchrony includes: atrial dyssynchrony, ventricular dyssynchrony, and atrial and ventricular dyssynchrony. The electrical activation map can be used as an image classification vector to input the VGG19 model of transfer learning for deep learning to achieve image classification, and output a preliminary auxiliary diagnosis result of whether atrial / ventricular dyssynchrony or myocardial ischemia is detected or not. This is because the intermediate variable of the VGG19 network is a replaceable feature vector, which is convenient for transfer learning and can maintain a high accuracy rate. Construct an intelligent diagnostic system, use machine learning and deep learning technology to improve the accuracy and automation level of diagnosis, and realize auxiliary diagnosis of input electrical activation map output classification results.

[0112] 5. Result visualization:

[0113] The intuitively designed user interface displays high-frequency ECG signals, interventricular conduction time differences, electrical activation sequence diagrams (electrical activation diagrams), dyssynchrony time differences (DYS) and operation options.

[0114] like Figure 4 , Figure 5 As shown in the figure, the electrical activation diagram of the patient with atrial / ventricular asynchrony is significantly different in leads V4 and V6 from the remaining leads. The V6 lead of the asynchrony patient has a more obvious delay, so that the red strip of the activation point between it and the V5 lead is broken. According to the algorithm, the main delayed leads are leads V5 and V6, and the maximum lead delay time difference is 24ms.

[0115] like Figure 6 As shown, the composition of an ultra-high frequency ECG mapping and analysis system is as follows:

[0116] Preprocessing module:

[0117] The Butterworth IIR filter was used for preprocessing to filter out low-frequency baseline drift, and the 50Hz power frequency interference was filtered out using a notch filter. The Pan-Tompkins algorithm was used to divide the heart beats based on the R wave, and the waveform 120ms before and after the R wave peak was cut.

[0118] Feature extraction module:

[0119] Quantitative indicators for achieving atrial / ventricular asynchrony include QRS complex width, interventricular conduction time difference, electrical activation sequence diagram (electrical activation diagram), asynchronous time difference (DYS), etc.

[0120] First, the amplitude envelope was calculated in the frequency range of 100-150Hz, 150-250Hz, 200-300Hz, 250-350Hz, 300-400Hz, 350-450Hz, 400-500Hz, …, 850-950Hz, and 900-1000Hz by Hilbert transform. Signal averaging technology was used to improve the signal-to-noise ratio. The waveform of 120ms before and after each QRS wave divided in the time period was averaged, and the signal was low-pass filtered (passband 0–40Hz) to form the waveform corresponding to each lead. The waveforms were arranged from top to bottom according to leads V1-V8. The formed 2D matrix represents the distribution of electrical activity in time (relative to the time of the R wave) and space (V lead).

[0121] After constructing the electrical activity conduction matrix, the time difference between the earliest excited lead in the matrix and the excited time of all leads was taken as the asynchronous time difference DYS, and the group with the largest time difference between adjacent leads was taken as the major delayed lead group (Major Delayed Leads, MDL), and this time difference was the major delayed time difference (Major Delayed Duration, recorded as δDYS). In subsequent research, the diagnostic and evaluation effect of this set of indicators on atrial / ventricular asynchrony or atrial / ventricular asynchrony was evaluated, and compared with indicators such as asynchrony parameters of echocardiography.

[0122] Pattern Recognition Module:

[0123] Use machine learning algorithms to classify and locate atrial / ventricular dyssynchrony or myocardial ischemia. The electrical activation map can be used as an image classification vector and deep learning can be performed using the VGG19 model. Build an intelligent diagnosis system and use machine learning and deep learning techniques to improve the accuracy and automation of diagnosis. Perform deep learning based on VGG19 based on the presented electrical activation map to achieve automated and intelligent diagnosis of atrial / ventricular dyssynchrony or myocardial ischemia.

[0124] Visualization module:

[0125] After obtaining the two-dimensional matrix, due to the inconsistency between the amplitude and positive and negative of the V1-V8 leads, the normalization algorithm is used to normalize each V lead so that its amplitude is between 0 and 1. Based on the normalized matrix, an electrical activation map (thermal map, the maximum amplitude is displayed in red and the minimum amplitude is displayed in blue) is created, and the envelope signals are arranged in order into rows from V1 to V8. Since there are only 8 leads, that is, 8 rows of data, linear interpolation is performed between the rows to graphically describe the time-space distribution of electrical activity (depolarization process). The time region presented is ±120ms around the R wave. In the electrical activation map, the color represents the activation time and order. The V3 lead appears red first, that is, it is activated first, followed by V1, V2, and so on.

[0126] user interface:

[0127] After obtaining the electrical activation map, an intuitive user interface is designed to display high-frequency ECG signals, interventricular conduction time differences, electrical activation sequence diagrams (electrical activation maps), asynchronous time differences (DYS) and operation options.

[0128] Specifically, by exploring the changes of various DPS under different conditions of atrial / ventricular asynchrony or myocardial ischemia through modeling simulation and clinical experiments, the cardiac electrical activity in patients with different types of asynchrony or myocardial ischemia can be determined. For a specific patient, a set of DPS suitable for diagnosis and mapping can always be determined, thereby realizing the verification of the indicator system.

[0129] Finally, it should be noted that the above is only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred arrangement scheme, a person skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for ultra-high frequency electrocardiogram mapping and analysis, characterized in that: S1: Acquire high-frequency ECG data: Use a high-frequency ECG acquisition device to acquire high-frequency ECG signal data; the high-frequency ECG signal data refers to ECG waveform data acquired by a high-frequency ECG signal acquisition device with a frequency response range of 0 to 1000 Hz, an ADC accuracy greater than or equal to 24 bits, and 6 to 18 leads; the leads include chest leads, limb leads, and pressurized limb leads; S2: Data preprocessing: Use Butterworth IIR filter and notch filter to preprocess the high-frequency ECG signal data collected by S1; S3: Heart beat division based on R wave: Use Pan-Tompkins algorithm to find R wave in the high-frequency ECG signal data collected by limb leads or pressurized limb leads after S2 preprocessing; and record the time nodes at the waveform of the specified time length before and after the peak of the R wave; according to the recorded time nodes, the high-frequency ECG signal collected by the chest lead is cut to obtain a high-frequency ECG signal of a heart beat, i.e., the heart beat; S4: Feature extraction: using the Hilbert transform algorithm to calculate the amplitude envelope of the QRS wave in the high-frequency electrocardiogram signal of each frequency band in the continuous frequency band for the heartbeat obtained in S3; the continuous frequency band refers to a series of continuous frequency ranges, one frequency range is one frequency band, and the difference between the upper limit and the lower limit of each frequency range is the same; S5: Signal averaging: The amplitude envelope of the QRS wave obtained in S4 is averaged by the signal averaging technique; the averaged amplitude envelope is low-pass filtered to form a feature vector corresponding to the amplitude envelope of each lead; S6: Draw the electrical activation map: Use the characteristic vectors of all chest leads in S1 to form a characteristic matrix, namely the electrical activity conduction matrix; calculate the time difference between lead excitations, namely the asynchronous time difference, through the electrical activity conduction matrix; use the normalization algorithm to normalize the electrical activity conduction matrix so that its amplitude is between 0 and 1; use the colormap algorithm to draw the normalized electrical activity conduction matrix into a heat map to obtain the electrical activation map; S7: Pattern recognition: The S6 electrical activation map is input into the VGG19 model based on transfer learning, and the preliminary auxiliary diagnosis results of atrial / ventricular dyssynchrony or myocardial ischemia or non-detection are output; S8: Result visualization: The high-frequency ECG signals collected by 6-18 leads obtained by S1, the electrical activation map and asynchronous time difference obtained by S6, and the preliminary auxiliary diagnosis results obtained by S7 are displayed on the user interface.

2. An ultra-high frequency electrocardiogram mapping and analysis method according to claim 1, characterized in that: The 6-18 leads include: 6, 9, 12, 14 and 18 leads; wherein: The 6 leads are six chest leads V1-V6; The 9 leads are the 6 leads plus three limb leads I, II, and III; The 12 leads are based on the 9 leads and have three additional limb leads, i.e., the standard 12 leads. The 14 leads are based on the 12 leads and have two chest leads, V7 and V8, added. The 18 leads are based on the 14 leads and add four more leads, namely V3R, V4R, V5R and V9. The V3R lead is located on the right chest symmetrical to V3; the V4R lead is located on the right chest symmetrical to V4; the V5R lead is located on the right chest symmetrical to V5; and V9 is the left posterior chest wall lead.

3. An ultra-high frequency electrocardiogram mapping and analysis method according to claim 1, characterized in that: The data preprocessing method described in S2 is: using a Butterworth IIR filter to filter out the low-frequency baseline drift of the high-frequency ECG signal data; using a notch filter to filter out the power frequency interference of a specified frequency in the high-frequency ECG signal data.

4. An ultra-high frequency electrocardiogram mapping and analysis method according to claim 1, characterized in that: The Hilbert transform algorithm described in S4 is to convert the divided high-frequency ECG signal into an amplitude envelope signal. The specific steps are as follows: S41: Calculate Fourier transform: Set the Fourier transform of the high-frequency ECG signal x(t) after the heartbeat division to X(jw), and according to the Fourier transform formula, we get: In formula (1), x * (t) is the conjugate of x(t), j is the imaginary unit; if the phase is changed by 90°, that is, ω is -ω, then x can be calculated * The Fourier transform of (t) is X * (-jω); S42: Calculate Hilbert transform: The Fourier transform of a real signal contains negative frequencies. If you want to remove the negative frequencies but keep the total power unchanged, multiply X(jw) by 2u(t), where u(t) is a step function, to obtain a high-frequency ECG signal after Hilbert transform: The frequency response characteristics are: It can be seen from equation (4) that the Hilbert transform shifts the phase of all positive frequency components by 90 degrees and shifts the phase of all negative frequency components by 90 degrees. S43: Construct analytical signal: Convert the real signal into a complex signal, take the high-frequency ECG signal x(t) after the original heartbeat division as the real part, and the high-frequency ECG signal after Hilbert transformation As the imaginary part, we get: The high-frequency ECG signal x(t) after heart beat division has both amplitude information and phase information. Let x(t) be: x(t)=A(t)cos(ωt+θ(t)) (6) Among them, A(t) represents the amplitude information of the signal, and θ(t) represents the phase information; Substituting formula (6) into formula (5), we get: S44: Calculate and obtain the amplitude envelope signal: Take the absolute value to get the amplitude envelope signal:

5. An ultra-high frequency electrocardiogram mapping and analysis method according to claim 1, characterized in that: The specific steps of signal averaging described in S5 are: S51: Calculate the average value of the amplitude envelope of the QRS wave corresponding to all the heart beats of S4 in the same lead and the same frequency band to obtain the simplified amplitude envelope of the QRS wave; S52: Calculate the median of each time point in the simplified QRS wave amplitude envelope in all frequency bands, i.e., the feature vector of each lead; the median is a one-dimensional feature vector of 1×divided time length.

6. An ultra-high frequency electrocardiogram mapping and analysis method according to claim 1, characterized in that: The asynchronous time difference DYS described in S6 is obtained by calculating the time difference between the lead that starts to be excited earliest and the time difference between all the leads in the electrical activity conduction matrix; the group with the largest time difference between adjacent leads is taken as the main delay lead group, i.e. MDL, and this time difference is the main delay time difference, i.e. δDYS; the value of the asynchronous time difference represents the degree of atrial / ventricular asynchrony, and the larger the value, the stronger the degree of atrial / ventricular asynchrony.

7. An ultra-high frequency electrocardiogram mapping and analysis method according to claim 1, characterized in that: The specific method for obtaining the VGG19 model based on transfer learning described in step S7 is: S61: Pre-training the VGG19 model to obtain the pre-training model weight; the intermediate variable of the VGG19 model is a replaceable feature variable, which is a key carrier for transmitting and processing information between layers of the VGG19 model; S62: Establish a pattern recognition model, namely, a VGG19 model; use a transfer learning algorithm to transfer the weights of the S61 pre-trained model to the VGG19 model, and use the electrical activation map dataset to fine-tune the VGG19 model to obtain a VGG19 model for detecting atrial / ventricular dyssynchrony or myocardial ischemia; the fine-tuning is to use the electrical activation map as the input of the model, and the output is a preliminary auxiliary diagnosis result of whether atrial / ventricular dyssynchrony or myocardial ischemia is detected or not; S63: Verify the fine-tuned model. The indicators for verifying model performance include accuracy, F1 score, precision and recall.

8. An ultra-high frequency electrocardiogram mapping and analysis system formed based on the method described in any one of claims 1 to 7, characterized in that: include: Data acquisition module: using high-frequency ECG acquisition equipment to collect high-frequency ECG signal data; Data preprocessing module: The Butterworth IIR filter unit is used to filter out the low-frequency baseline drift in the collected high-frequency ECG signal data; and the notch filter is used to filter out the power frequency interference of the specified frequency in the collected ECG signal data; Feature extraction module: uses the Hilbert transform unit to calculate the amplitude envelope of the ECG signal data processed by the data preprocessing module within the specified frequency range; The amplitude envelope obtained by the Hilbert transform unit is averaged by the signal averaging technology unit; the amplitude envelope averaged by the signal averaging technology unit is filtered by the low-pass filtering unit to form a waveform corresponding to the amplitude envelope corresponding to each lead; Visualization module: construct a feature matrix, i.e., an electrical activity conduction matrix, by using a matrix construction unit to construct the feature vector of the specified lead; input the electrical activity conduction matrix into a calculation unit for calculating the asynchronous time difference to calculate the time difference between the lead excitations; and use a colormap algorithm unit to draw the electrical activity conduction matrix into a heat map to obtain an electrical activation map; Pattern recognition module: input the electrical activation map of the visualization module into the VGG19 model unit to obtain the preliminary auxiliary diagnosis result of whether atrial / ventricular asynchrony or myocardial ischemia is detected or not; User interface: displays the high-frequency ECG signals of the data acquisition module, the electrical activation diagram and asynchronous time difference of the visualization module, and the preliminary auxiliary diagnosis results and operation options of the pattern recognition module.

Images