HEART RATE VARIABILITY ANALYSIS METHOD, DEVICE AND USE THEREOF

Abstract

A method and an apparatus for analyzing heart rate Variability (HRV), and use thereof are provided. A low-cost, portable and wearable signal acquisition device is utilized to acquire electrocardiography (ECG) signals of epilepsy patients for 24 hours before treatment, and a time domain index, a frequency domain index and a nonlinear index of the ECG during a long term and during a short term are calculated with a programmed HRV analysis method, and the efficacy of vagus nerve stimulation (VNS) treatment for patients with medically intractable epilepsy is accurately and efficiently predicted based on characteristic parameters for characterizing an effect level of the vagus nerve regulating the heart rate, i.e., vagus nerve activity, thereby avoiding unnecessary costs and avoiding the delay of the optimal treatment timing. In addition, the characteristic parameters obtained by the HRV analysis on the ECG may be utilized to clearly select VNS treatment indication patients.

Claims

1-10. (canceled)

11. A method for analyzing heart rate variability, comprising: 1) acquiring electrocardiography data externally; 2) performing a digitization process and a denoising process on the electrocardiography data; 3) forming a sinus NN interval sequence with the processed electrocardiography data; 4) selecting sinus NN interval data of a subject during a long term of more than 20 hours and during short terms of 2 to 10 minutes in an awake state and in a sleep state; and 5) calculating at least one of a time domain index, a frequency domain index and a nonlinear index for the sinus NN interval data of the subject during the long term of more than 20 hours and during the short terms of 2 to 10 minutes in the awake state and in the sleep state.

12. The method for analyzing the heart rate variability according to claim 11, wherein 1) calculation formulas for time domain indexes are expressed as follows, and one or more of the following indexes are selected to be calculated during the analysis; TABLE-US-00003 Name of index Definition of index Unit mean value (MEAN) $\mathrm{MEAN}=\stackrel{\_}{\mathrm{RR}}=\underset{i=1}{\overset{N}{.Math.}}.Math.{\mathrm{RR}}_i/N$ ms overall standard deviation (SDNN) $\mathrm{SDNN}=\sqrt{\frac{1}{N}.Math.\underset{i=1}{\overset{N}{.Math.}}.Math.{\left({\mathrm{RR}}_i-\stackrel{\_}{\mathrm{RR}}\right)}^2}$ ms root-mean-square of difference of adjacent interval (RMSSD) $\mathrm{RMSSD}=\sqrt{\frac{1}{N-1}.Math.\underset{i=1}{\overset{N-1}{.Math.}}.Math.{\left({\mathrm{RR}}_{i+1}-{\mathrm{RR}}_i\right)}^2}$ ms percentage of adjacent pNN50 = NN50/TotalNN 100% % intervals with difference greater than 50 ms (pNN50) 2) calculation formulas for frequency domain indexes are expressed as follows, and one or more of the following indexes are selected to be calculated during the analysis; TABLE-US-00004 Name of index Definition of index Unit Frequency range TP Total signal power, sum ms.sup.2 0.4 Hz of VLF, LF and HF VLF Power in very low ms.sup.2 0.0033 to 0.04 Hz frequency band LF Power in low frequency ms.sup.2 0.04 to 0.15 Hz band HF Power in high frequency ms.sup.2 0.15 to 0.4 Hz band LF/HF Ratio of power in low frequency band to power in high frequency band and 3) the nonlinear index comprises SD1 and SD2 indexes in a Poincare Plot, and one or more of the indexes are selected to be calculated during the analysis.

13. The method for analyzing the heart rate variability according to claim 11, further comprising: calculating one of RMSSD, pNN50, SD1 and HF.

14. The method for analyzing the heart rate variability according to claim 13, further comprising: determining a threshold of one of RMSSD, pNN50, SD1 and HF.

15. The method for analyzing the heart rate variability according to claim 11, wherein a time period for the long term is set as 24 hours, and a time period for the short long term is set as 5 minutes.

16. The method for analyzing the heart rate variability according to claim 14, wherein the threshold of RMSSD is 37 ms, the threshold of pNN50 is 27%, the threshold of SD1 is 35 ms, and the threshold of HF is 798 ms.

17. The method for analyzing the heart rate variability according to claim 11, wherein the method are utilized to analyze vagus nerve stimulation (VNS) treatment for medically intractable epilepsy.

18. An apparatus for analyzing heart rate variability, comprising: a calculating module configured to perform calculating with the method according to claim 11.

19. The apparatus for analyzing the heart rate variability according to claim 18, further comprising one or more of a data acquiring module, a digitization processing module, a denoising module and a determining module.

20. The apparatus for analyzing the heart rate variability according to claim 18, wherein the apparatus are utilized to analyze vagus nerve stimulation (VNS) treatment for medically intractable epilepsy.

21. A wearable electrocardiography monitoring device, comprising: the apparatus for analyzing the heart rate variability according to claim 18.

22. The wearable electrocardiography monitoring device according to claim 21, wherein the device are utilized to analyze vagus nerve stimulation (VNS) treatment for medically intractable epilepsy.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0019] FIG. 1 is a flowchart of screening VNS indication patients in the conventional technology;

[0020] FIG. 2 is a schematic diagram showing 12-lead ECG acquisition;

[0021] FIG. 3 is a flowchart of processing an ECG signal;

[0022] FIG. 4 is a schematic diagram showing calculation for HRV nonlinear indexes SD1 and SD2;

[0023] FIG. 5 is a flowchart showing a determining process;

[0024] FIG. 6 is diagram showing parameters for a responder group Responder50 and a non-responder group Non-responder50; and

[0025] FIG. 7 is a curve showing parameter characteristics for a subject.

DETAILED DESCRIPTION OF EMBODIMENTS

First Embodiment

[0026] As shown in FIG. 2, standard 12-lead ECG acquisition is performed for 24 hours before treatment. In the acquisition process, an electrocardiography acquisition device is required to have a sampling frequency greater than or equal to 500 Hz. During the ECG recording, subjects should not make activities such as strenuous exercise and dose that may affect the cardiac function. The recording time period is set as 24 hours. Recording environments or conditions in which the subjects are located should be substantially similar to each other. In addition, it is required that data for HRV analysis is normal sinus NN interval data. In the HRV analysis, a time domain index, a frequency domain index and a nonlinear index not only during a time period of 24 hours, but also during short terms of 5 minutes in an awake state and in a sleep state are extracted. A specific flow of processing an ECG signal is as shown in FIG. 3, which includes:

[0027] 1) acquiring an ECG signal and digitizing the signal;

[0028] 2) performing a denoising process and a de-artifact process on the digitized signal;

[0029] 3) automatically detecting QRS waves in the signal;

[0030] 4) artificially viewing QRS waves in the detected signal;

[0031] 5) removing QRS waves caused by ectopic pacing in the signal;

[0032] 6) forming a sinus NN interval sequence;

[0033] 7) selecting sinus NN interval data of each subject during a long term of 24 hours and during short terms of 5 minutes in an awake state and in a sleep state;

[0034] 8) calculating, with the method as shown in Table 1, Table 2 and FIG. 4, a time domain index, a frequency domain index and a nonlinear index for the sinus NN interval data of the subject during the long term of 24 hours and during the short terms of 5 minutes in the awake state and in the sleep state; and

[0035] 9) selecting characteristic parameters RMSSD, pNN50, SD1 and HF for characterizing vagus nerve activity.

[0036] The electrocardiography acquisition is performed on patients with medically intractable epilepsy for 24 hours before treatment. The electrocardiography data acquired during 24 hours is processed with the above method to obtain normal sinus NN interval sequences during the long term of 24 hours and during the short terms of 5 minutes in the awake state and in the sleep state. HRV time domain analysis, frequency domain analysis, and nonlinear analysis are performed on the NN interval sequences during the long term and the short terms with the method as shown in Table 1, Table 2 and FIG. 4, and characteristic parameters RMSSD, pNN50, SD1 and HF for characterizing vagus nerve activity are selected. Thresholds corresponding to the characteristic parameters are determined to make a comprehensive determination and selection, so as to predict and determine an efficacy of VNS treatment for medically intractable epilepsy (as shown in FIG. 5). The characteristic parameters RMSSD, pNN50, SD1, and HF for characterizing the vagus nerve activity during a time period of 5 minutes at night, are individually utilized to predict the efficacy of VNS treatment for medically intractable epilepsy, and the selected corresponding thresholds and corresponding prediction accuracies are described as follows.

[0037] The threshold for RMSSD is selected as 37 ms. That is, in a case that the ECG analysis result on a patient before treatment is greater than the threshold, the epileptic seizure after VNS treatment is decreased by 50% or more. In addition, the prediction accuracy is 88.2%. The threshold may also be utilized for screening patients.

[0038] The threshold for pNN50 is selected as 27%. That is, in a case that the ECG analysis result on a patient before treatment is greater than the threshold, the epileptic seizure after VNS treatment is decreased by 50% or more. In addition, the prediction accuracy is 82.7%. The threshold may also be utilized for screening patients.

[0039] The threshold for SD1 is selected as 35 ms. That is, in a case that the ECG analysis result on a patient before treatment is greater than the threshold, the epileptic seizure after VNS treatment is decreased by 50% or more. In addition, the prediction accuracy is 70.6%. The threshold may also be utilized for screening patients.

[0040] The threshold for HF is selected as 798 ms.sup.2. That is, in a case that the ECG analysis result on a patient before treatment is greater than the threshold, the epileptic seizure after VNS treatment is decreased by 50% or more. In addition, the prediction accuracy is 76.5%. The threshold may also be utilized for screening patients.

Second Embodiment

[0041] In the HRV analysis method according to the first embodiment, other indexes that can characterize the vagus nerve activity may also be utilized to predict the efficacy of the VNS treatment for medically intractable epilepsy and screen patients.

[0042] According to the present disclosure, the ECG acquisition is performed on patients with medically intractable epilepsy for 24 hours before treatment, and the HRV time domain analysis, frequency domain analysis and nonlinear analysis are performed, so that the efficacy for the patients with medically intractable epilepsy can be predicted before VNS treatment, and the patients with medically intractable epilepsy can be correctly guided to determine whether to receive the VNS treatment, thereby avoiding unnecessary costs and avoiding the delay of the optimal treatment timing. In addition, the characteristic parameters for characterizing the vagus nerve activity obtained by the HRV analysis on the ECG are utilized to predict the efficacy of VNS treatment and clearly select VNS treatment indication patients, so that the overall efficacy of the VNS treatment can be improved.

Third Embodiment

[0043] Based on the above screening method, 32 patients with medically intractable epilepsy who have completed the VNS treatment at a Beijing Tiantan Hospital from Aug. 13, 2014 to Dec. 31, 2014 are selected for verification. A comprehensive evaluation (including analysis on demographic characteristics, clinical history, of antiepileptic drug history, video electroencephalography for 24 hours, MRI, and dynamic electrocardiography for 24 hours) is performed on the 32 patients with medically intractable epilepsy before the VNS treatment.

[0044] The HRV time domain analysis, frequency domain analysis, and nonlinear analysis are performed on the dynamic electrocardiography data for 24 hours before treatment with the ECG signal processing method described above, and the characteristic parameters RMSSD, pNN50, SD1, and HF that characterize the vagus nerve activity are extracted for each patient. It is found from the follow-up for one year after treatment that, among the 32 patients with medically intractable epilepsy who have received the VNS treatment, epileptic seizures in 17 patients (where epileptic seizures in 6 patients can be controlled completely) are decreased by 50% or more, which are referred to as a responder group Responder50, and epileptic seizures in other 15 patients (where 4 patients have no change in epileptic seizures after receiving the VNS treatment) are decreased by 50% or less, which are referred to as a non-responder group Non-responder50. The parameters RMSSD, pNN50, SD1 and HF for the responder group Responder50 and the non-responder group Non-responder50 are shown in FIG. 6. Since the parameters for characterizing the vagus nerve activity for the two groups of patients have significant statistically differences, it is indicated that, the efficacy of the VNS treatment is related to vagus nerve activity in patients before treatment, and the efficacy of the VNS treatment can be predicted with the HRV analysis method. Further, efficacy prediction abilities of RMSSD, pNN50, SD1 and HF during the long term of 24 hours and during the short terms of 5 minutes in the awake state and in the sleep state are evaluated. It is found from the ROC curve analysis that prediction of the RMSSD during the short term of 5 minutes in the sleep state is most accurate among the above four parameters (as shown in FIG. 7). In the case that the threshold for RMSSD is set as 37 ms, only 2 patients among the 17 patients in the responder group have RMSSD less than 37 ms, and the other patients all have RMSSD greater than 37 ms, and the prediction accuracy is over 85%. Therefore, it can be proved that the characteristics parameter for characterizing the vagus nerve activity obtained by the above HRV analysis can be utilized to predict the efficacy of VNS treatment accurately and effectively and screen the VNS indication patients.

[0045] The above description shows only preferred embodiments of the present disclosure. It should be noted that those skilled in the art may make improvements and modifications to the present disclosure without departing from the principle of the present disclosure. The improvements and modifications should also be included in the protection scope of the present disclosure. In addition, although some specific terms are utilized in this specification, the terms are described only for convenience of description and do not intended to limit the present disclosure.