JP6403311B2 - Heart rate analysis device - Google Patents

Description

  The present invention relates to a heartbeat state analyzing apparatus that analyzes a heartbeat state based on a heart sound signal of a human body such as a fetus.

  In the diagnosis of fetal health, the heart rate of the fetus is one of the examination items, and the importance of the fetal heartbeat variability has been pointed out. Regarding the heartbeat state of the fetus, the heartbeat signal of the fetus is measured using an ultrasonic Doppler method in a delivery monitoring device or the like. In the ultrasonic Doppler method, a contact is attached to the abdominal wall of a pregnant woman, a weak ultrasonic pulse is emitted from the contact with a low sound pressure toward the heart of the fetus in the body, and a reflected wave reflected by the heart is generated. The heartbeat signal of the fetus can be measured based on the received signal received by the contact.

  Regarding the processing of the signal waveform obtained by the ultrasonic Doppler method, for example, in Patent Document 1, an autocorrelation method is set by setting a short-time window corresponding to approximately one beat of the heartbeat in order to calculate the instantaneous heartbeat period. The points to be calculated are described. Further, in Patent Document 2, an autocorrelation function of the obtained fetal heartbeat signal is obtained, an effective correlation value peak is detected from a plurality of correlation values, and the fetal heartbeat is calculated based on the time difference corresponding to the detected correlation value peak. The point to count the number is described. Moreover, in patent document 3, an ultrasonic signal is irradiated to the fetal heart, an echo signal is detected to detect a heart valve signal, and a meniscal valve opening signal and an atrioventricular valve closing signal are extracted from the detected signals. A device for displaying the stored half-moon valve opening signal and the atrioventricular valve closing signal on the same time axis on a display device and measuring the time axis position of both signals and the accommodating systole is described. Has been. Moreover, in patent document 4, based on the audio | voice signal obtained from the body sound sensor, the heart rate and heart sound level of a mother and a fetus are calculated, and the state of a fetus is referred with reference to the classification memorize | stored beforehand about the calculated data. A fetal observation support device that acquires data to be shown is described.

JP-A-10-28686 JP-A 63-277034 JP 2000-197633 A JP 2001-276079 A Japanese Patent No. 5278952

  When measuring heart rate variability of the fetus, it is important to measure the heart rate state in real time. On the other hand, in the above-described patent document, a heart rate state such as a fetal heart rate is measured based on a fetal heart sound signal, but real-time measurement is technically difficult. For example, in Patent Documents 1 and 2, a heartbeat waveform is analyzed using an autocorrelation function. However, when an autocorrelation function is used, it is necessary to store and process a signal waveform once. It has become difficult to do. Also in Patent Documents 3 and 4, it is necessary to store and process signal waveforms in the same manner, and it is technically difficult to process in real time.

  Methods such as electrocardiography and magnetocardiography are used as a method for measuring the heart rate of an adult in real time. In such a method, it is necessary to attach a sensor directly to the body in order to perform accurate measurement. In the case of a fetus, since it is difficult to attach a sensor directly to the body, techniques such as electrocardiography are not suitable for measuring the fetal heart rate. In particular, since the fetus moves around in the womb, accurate measurement must be difficult.

  As described in Patent Document 5, the present inventor has proposed an infant emotion diagnosis apparatus. This emotion diagnosis apparatus uses a system that measures an instantaneous pitch period in real time for an infant's voice.

  Therefore, an object of the present invention is to provide a heartbeat state analysis apparatus capable of analyzing a heartbeat state in real time based on a fetal heart sound signal using a technique for processing such a sound signal waveform in real time.

The heartbeat state analyzing apparatus according to the present invention detects and reflects an ultrasonic signal transmitted toward the body in a contact detection sensor attached to a mother who is a pregnant woman and a contact detection sensor attached to an abdominal wall near the fetus. A detection processing unit for acquiring a detection signal relating to a heartbeat state of the mother and fetus obtained, and a heart sound processing for obtaining a period of the detection signal of the mother and fetus in real time based on amplitude data of the detection signal of the mother and fetus obtained And a heart sound analysis unit for analyzing the correlation between the analysis results of the heart rate of the mother and the fetus while processing the obtained period of the detection signal of the mother and the fetus by fast Fourier transform and performing frequency analysis in real time. ing. Further, the heart sound processing unit obtains a period related to the I sound waveform generated when the atrioventricular valve is closed at the beginning of the systole of the heart as the period of the detection signal. Furthermore, the heart sound processing unit extracts the starting position of the I sound waveform in said classified set signal block based on the average value of the amplitude data by 4 block segments. Further, the heart sound analysis unit performs a fluctuation analysis process of the peak value obtained by frequency analysis.

  According to the present invention, since the heartbeat state can be analyzed in real time based on the fetal heart sound signal, it is possible to accurately diagnose the fetal heartbeat variability.

1 is a schematic block configuration diagram relating to an embodiment of a heart rate analysis apparatus according to the present invention. It is explanatory drawing which shows typically the waveform of a heart sound signal and a heart noise. It is a wave form diagram regarding the detection signal of the mother and fetus actually detected. It is explanatory drawing regarding the principle of an average amplitude difference function method. It is a graph which shows the average value with respect to a detection waveform, and the reset signal block. It is a graph which shows transition of RR time calculated based on the heart sound waveform and electrocardiogram waveform which were obtained from the mother. It is the graph which showed the time series data of RR time calculated | required from the mother's and fetus's heart sound waveform converted into the heart rate for 1 minute. 6 is a graph showing an analysis result in an analysis section A. 6 is a graph showing an analysis result in an analysis section B. 6 is a graph showing an analysis result in an analysis section C. It is a graph which shows the fluctuation | variation of the peak value of the spectrum waveform of a fetus. It is a graph which shows the fluctuation | variation of the peak value of a mother's spectrum waveform. It is a graph which shows transition of the heart rate at the time of letting music to a mother and a fetus. It is an example of a screen which shows an example regarding a display screen.

  Hereinafter, the present invention will be specifically described. FIG. 1 is a schematic block configuration diagram relating to an embodiment of a heart rate analysis device according to the present invention. The heart rate analysis apparatus includes a processing unit 10, a storage unit 11, and a display unit 12. The processing unit 10 is attached to a contact detection sensor 13 attached to a mother who is a pregnant woman and an abdominal wall in the vicinity of the fetus. A contact detection sensor 14 is connected.

  The processing unit 10 reflects the ultrasonic signals transmitted toward the inside of the body by the contact detection sensors 13 and 14, acquires the detection signal detected, and the detection signal obtained by the detection processing unit 100. A heart sound processing unit 101 that extracts a heart sound signal by processing, a heart sound analysis unit 102 that analyzes a heart sound signal extracted by the heart sound processing unit 101, a processing result of the heart sound processing unit 101, and an analysis result of the heart sound analysis unit 102 are displayed. The unit 12 includes a display processing unit 103 that performs display processing.

  The storage unit 11 includes a detection DB 110 that stores the obtained detection signal and a processing DB 111 that stores data obtained by processing the detection signal.

  In addition, about the process part 10, the memory | storage part 11, and the display part 12, by installing the program and data required in order to implement | achieve a function using a well-known personal computer, the function of a heartbeat state analyzer is added to a personal computer. Can be realized.

  The detection processing unit 100 ultrasonically vibrates the contact detection sensors 13 and 14 attached to the mother to transmit an ultrasonic signal toward the inside of the body, and the reflected signal received by reflecting the transmitted ultrasonic signal inside the body. The beat sound generated by the Doppler effect is acquired as a detection signal. Then, the acquired detection signal is stored in the detection DB 110 of the storage unit 11.

  The detection signal includes heart noise (noise) generated along with the heartbeat in addition to the heart sound signal corresponding to the heartbeat. FIG. 2 is an explanatory diagram schematically showing waveforms of a heart sound signal and a heart noise. In the heart sound signal, the I sound waveform generated when the atrioventricular valve is closed at the beginning of the systole, the II sound waveform generated when the arterial valve is closed at the end of the systole, and the III sound wave generated when the heart is diastole. It is known that four types of patterns occur: shape and IV sound waveform. Further, in the electrocardiogram, it is known that an R wave is generated at the beginning of the systole of the heart and a T wave is generated at the end of the systole. When the waveform of the heart sound signal is compared with the waveform of the electrocardiogram, the I sound waveform corresponds to the R wave and the II sound waveform corresponds to the T wave. Therefore, the time between the R wave and the next R wave (hereinafter referred to as “RR time”), which is the evaluation standard of instantaneous heartbeat variability in the ECG waveform analysis, is the I sound waveform and the next I sound wave. It can be measured as the time between shapes.

  FIG. 3 is a waveform diagram regarding detection signals of the mother and fetus that are actually detected. The upper waveform B shows the detection waveform of the fetus, and the lower waveform M shows the detection waveform of the mother. As described in FIG. 2, the detection signal including both heart sounds and heart noise appears in the waveform.

  The heart sound processing unit 101 samples the detection signal acquired by the detection processing unit 100 every predetermined time, and uses an average amplitude difference function (AMDF) method based on the amplitude data of the detection signal obtained by sampling. The period of the detection signal is obtained by signal correlation. In this example, the position on the time axis of the I sound waveform is determined based on the signal correlation, and the time between the I sound waveform and the next I sound waveform is calculated to thereby calculate the RR time that is the basic period of the detection signal. Ask for.

FIG. 4 is an explanatory diagram regarding the principle of the average amplitude difference function method. In FIG. 4, the detected waveform of the sound is schematically drawn with the vertical axis representing amplitude and the horizontal axis representing time. A waveform A in a section from a position i on an arbitrary time axis sampled in the detected waveform to N sampled positions, and a section in waveform B in which the same N sections as the waveform A are shifted by a time delay m. And the basic period can be extracted by looking at the correlation between the waveforms based on the average amplitude difference between the waveforms in the two sections. The correlation value D (m) of the two waveforms is calculated by the following mathematical formula, and the time delay m at which the correlation value D (m) becomes the minimum value can be estimated as the RR time.

  As described with reference to FIG. 2, the detected waveform includes noise, but the amplitude is maximum in the I sound waveform. Therefore, when calculating the correlation value D (m) of the waveform, the amplitude data is appropriately set. If the threshold is set, the accuracy of the correlation value of the I sound waveform can be increased. Then, the time correlation data of the RR time is obtained by performing the signal correlation process while shifting the section in accordance with the acquisition timing of the sampling data of the detected waveform to obtain the RR time. Therefore, it becomes possible to inspect the fluctuation of the RR time in real time, and the heartbeat state such as the heartbeat variability can be accurately grasped. The obtained time series data is stored in the processing DB 111 of the storage unit 11.

  Since the average amplitude difference function method described above can perform calculation processing at high speed, it is possible to perform RR time calculation processing in real time. In addition, if the position on the time axis for calculating the amplitude difference is set at every sampled position, stable high-speed processing is possible, and real-time heart sound processing is performed to stably display the test results. It can be carried out.

  As described in FIG. 2, the RR time extraction method can be obtained by paying attention to the fact that the heart sound waveform is composed of four patterns. Specifically, first, a detected waveform is sampled, an average value of amplitude data is calculated for each predetermined number of samplings, and the average value is set as a threshold value. What level of sampling is set may be set as appropriate according to the heartbeat state.

  Next, signal block setting processing is performed using the average value as a threshold value. A waveform section of amplitude data that is equal to or greater than the threshold is defined as a section with a signal, a rectangular wave signal block having an average amplitude value, a waveform section of amplitude data that is smaller than the threshold is defined as a section without a signal, and the amplitude is 0.

  Next, signal block resetting processing is performed. When a rising position and a falling position on the time axis of a signal block are detected, and the interval between the falling position of the signal block and the rising position of the next signal block is smaller than a predetermined interval, the signal block It is determined that the signal block has continued, and the signal block is reset. When it is determined that the length between the rising position and the falling position of the reset signal block is smaller than a predetermined length, the reset signal block is referred to as “no signal section” as noise. Set to. FIG. 5 is a graph showing an average value and a reset signal block with respect to the detected waveform.

  Next, the RR section is set and the start position of the RR time is determined. The finally set signal block is divided into four blocks each in the RR section, the signal block having the longest length between the rising position and the falling position in the RR section is extracted, and the rising position is determined as the RR time. Save as start position on time axis.

  Then, an RR time calculation process is performed. The length between the stored start position of the RR time and the next start position is obtained as the RR time, and time series data of the RR time can be obtained.

  In the RR time extraction method described above, processing is performed for each of the four blocks based on the characteristics of the heart sound waveform, so that highly accurate data can be obtained by high-speed processing, and the heart rate state including heartbeat variability Can be accurately captured in real time.

  FIG. 6 is a graph showing the transition of the RR time calculated based on the electrocardiogram and the electrocardiogram waveform obtained from the mother. In FIG. 6, the vertical axis represents the RR time (milliseconds), and the horizontal axis represents the waveform section. Then, the RR time is calculated by processing the ECG waveform and ECG waveform obtained from the mother at the same time by the above-mentioned average amplitude difference function method, the calculated value by the ECG is shown as a line graph with ◆, and the calculated value by the ECG waveform Is shown by a line graph with a mark.

  The two line graphs are almost the same as a whole although there are sections in which the calculated values by the electrocardiogram greatly deviate due to the influence of noise, and the RR time obtained from the heart sound waveform is replaced with the RR time obtained from the electrocardiogram waveform. It can be used for.

  Accordingly, although it is technically difficult to obtain an accurate electrocardiogram waveform from the fetus, the RR time can be accurately calculated by using the heart sound waveform obtained from the fetus, It becomes possible to diagnose a heartbeat condition such as heartbeat variability in real time.

  The heart sound analysis unit 102 performs frequency analysis processing by fast Fourier transform (FFT) for each predetermined analysis section on the basis of the RR time obtained by the heart sound processing unit 101, so that the state of fluctuation of the RR time is represented by a frequency characteristic. Analyze in real time from the viewpoint. In the following description, the heart sound analysis process will be described in detail based on a specific example.

  FIG. 7 is a graph showing time-series data of RR times obtained from mother and fetal heart waveforms converted to a one-minute heart rate. The ordinate represents the heart rate (bpm) for 1 minute, and the abscissa represents the time (minutes). In the graph, the heart rate of the mother is stable at around 70, but the heart rate of the fetus varies between 100 and 140, and it can be seen that instantaneous heartbeat variability clearly appears. For the graph shown in FIG. 7, an analysis interval having a predetermined time width is set, and an analysis interval having the same width shifted by 10 seconds from the analysis interval is set. In FIG. 7, an analysis interval A having a predetermined time width is set, analysis intervals B and C shifted by 10 seconds are set, and thereafter, the analysis intervals are sequentially set shifted by 10 seconds. Then, frequency analysis processing is performed on each set analysis section by FFT. 8 to 10 are graphs showing the analysis results of the analysis sections A to C, respectively. In each graph, the vertical axis represents the spectral intensity, and the horizontal axis represents the frequency.

  By analyzing the fluctuation of the peak value of the spectrum waveform of each analysis section obtained in this way, it can be used for diagnosis of the health condition of the mother and the fetus. Moreover, by observing the correlation between the analysis results of both, it is possible to make a comprehensive diagnosis by associating the health states of the two. FIG. 11 is a graph showing the fluctuation of the peak value of the fetal spectrum waveform. In FIG. 11, the vertical axis represents the spectral intensity, and the horizontal axis represents the time (minutes). VLF (0-0.04 Hz; very low frequency component), LF (0.04-0.15 Hz; low frequency) Component) and HF (0.15-0.4 Hz; high frequency component), the transition of the peak value in each frequency range is shown.

  The peak values of these frequency ranges are known to be closely related to the activity of the autonomic nervous system, and when the peak values of the VLF and LF frequency ranges are large, the sympathetic nervous system When the activity is in an active stress state and the peak value in the HF frequency range is large, the parasympathetic nervous system activity is in an active relaxed state. Therefore, by analyzing the graph regarding the spectral intensity at the peak as shown in FIG. 11, it can be used as one of important data for diagnosing the current health condition of the fetus.

  FIG. 12 is a graph showing fluctuations in the peak value of the mother's spectrum waveform. FIG. 12 shows the transition of the peak values for the three frequency ranges as in FIG. 11, and can be used as important data when diagnosing the current health condition of the mother, as in the case of the fetus. .

  Then, by analyzing the correlation between the analysis results of the heartbeat status of the mother and the fetus, it is possible to obtain useful data when making a comprehensive diagnosis by associating the health status of both. As a specific example, we listened to mother's favorite music and prenatal music alternately for a predetermined time, and conducted an experiment to find a correlation between fluctuations in the heart rate of both. FIG. 13 is a graph showing changes in heart rate when music is given to a mother and a fetus. In FIG. 13, the mother's favorite music and fetal music are heard in the time frame surrounded by a thick frame. Then, when the frequency analysis and the peak value fluctuation analysis by the above-mentioned FFT were performed, the fluctuation of the peak value in which the activity of the autonomic nervous system of the fetus and the mother became active was observed, and the state of such fluctuation varies depending on the type of music. I understood it. In this way, it is possible to quantitatively analyze the degree of the influence of various external factors on the mother and the fetus in real time.

  The display processing unit 103 displays the above-described detection signals and processing results on the display unit 12 as a list and performs display processing so that the heartbeat states of the mother and the fetus can be accurately confirmed. FIG. 14 is a screen example showing an example of the display screen. The operation display unit 120 displays a process start or end button, a selection button such as file selection, and a setting button such as a threshold value. The detection display unit 121 displays the detection signal in real time on a scale set as a signal waveform. The processing display unit 122 displays the time series data of the RR time obtained by the heart sound processing unit 101 on a set scale. In this example, a graph of average RR time for each predetermined section and a continuous graph of RR time are displayed in real time. The analysis display unit 123 dynamically displays the spectrum waveform shown in FIG. 8 in real time.

  Since the heart rate analysis apparatus described above acquires a heart sound signal by the ultrasonic Doppler method, a known labor monitoring device using the ultrasonic Doppler method can have the function of the heart rate analysis apparatus. . Therefore, it is possible to monitor the health status of both the mother and the fetus while confirming the heartbeat status of the mother and the fetus in real time on the display screen.

DESCRIPTION OF SYMBOLS 10 ... Processing part, 11 ... Memory | storage part, 12 ... Display part, 100 ... Detection processing part, 101 ... Heart sound processing part, 102 ... Heart sound analysis part, 103 ... Display Processing part

Claims (5)

Detection of the heart rate of the mother and fetus detected by reflecting ultrasonic signals transmitted toward the body in the contact detection sensor attached to the mother who is a pregnant woman and the contact detection sensor attached to the abdominal wall near the fetus a detection processing unit that acquires a signal, a heart sound processor for determining the period of the mother and the detection signal of a fetus in real time based on the amplitude data of the obtained mother and the detection signal of the fetus was determined maternal and fetal A heartbeat state analysis apparatus comprising: a heart sound analysis unit that processes a cycle of the detection signal by fast Fourier transform to perform frequency analysis in real time and analyzes a correlation between analysis results of heartbeat states of a mother and a fetus .   2. The heartbeat state analyzing apparatus according to claim 1, wherein the heart sound processing unit obtains a period related to an I sound waveform generated when an atrioventricular valve is closed at the beginning of a systole of a heart as a period of the detection signal. The heartbeat state analysis apparatus according to claim 2, wherein the heart sound processing unit extracts a start position of the I sound waveform in a section obtained by dividing a signal block set based on an average value of the amplitude data into four blocks.   The heartbeat analysis device according to any one of claims 1 to 3, wherein the heart sound analysis unit performs a fluctuation analysis process of a peak value obtained by frequency analysis.   A delivery monitoring device comprising the heartbeat state analysis device according to any one of claims 1 to 4.