Anxiety index and life quality influence assessment method based on HRV index

Abstract

The invention belongs to the technical field of medical research, and particularly discloses an anxiety index and life quality influence assessment method based on an HRV index, and the method comprises the steps: collecting the HRV data of an individual; the collected HRV data are preprocessed; the preprocessed HRV data are input into a Transform neural network model, and training and testing are carried out; and evaluating the anxiety index and the life quality influence index of the individual by using the trained Transform neural network model, and outputting an evaluation result including the anxiety index and the life quality influence index, thereby realizing more comprehensive and accurate evaluation of the anxiety state and the life quality of the individual. The problem that an existing method for evaluating the psychological state, the physiological health and the life quality of an individual excessively depends on simple analysis of single or multiple indexes and lacks comprehensiveness and accuracy is solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of medical research, and specifically relates to a design of an anxiety index and life quality influence evaluation method based on HRV indicators. BACKGROUND

[0002] Heart rate variability (HRV) refers to the small changes in consecutive heart beat intervals (RR intervals), which reflects the regulation of the autonomic nervous system on heart rhythm. HRV indicators are widely used to assess individual psychological state, physiological health and life quality. However, existing evaluation methods often rely on simple analysis of single or multiple indicators, lacking comprehensiveness and accuracy. Transformer neural network is an advanced neural network structure with strong learning and generalization ability, capable of handling complex nonlinear relationships. SUMMARY

[0003] The purpose of the present application is to solve the problem that existing methods for evaluating individual psychological state, physiological health and life quality rely too much on simple analysis of single or multiple indicators, lacking comprehensiveness and accuracy. A method for evaluating anxiety index and life quality influence based on HRV indicators is proposed.

[0004] The technical solution of the present application is a method for evaluating anxiety index and life quality influence based on HRV indicators, which includes the following steps:

[0005] S1. Collecting HRV data of individuals;

[0006] S2. Preprocessing the collected HRV data;

[0007] S3. Inputting the preprocessed HRV data into the Transformer neural network model for training and testing;

[0008] S4. Using the trained Transformer neural network model to evaluate the anxiety index and life quality influence index of individuals, and outputting the evaluation results including anxiety index and life quality influence index.

[0009] The beneficial effects of the present application are:

[0010] By analyzing multiple indicators of HRV and combining the computing power of Transformer neural network model, the anxiety state and life quality of individuals can be more comprehensively and accurately evaluated. Meanwhile, the present application has good generalization ability and is suitable for individuals of different ages, genders and health conditions.

[0011] As a preferred embodiment, the HRV data in step S1 is obtained by a physiological signal monitoring device.

[0012] The above preferred scheme has the beneficial effect that:

[0013] The HRV data of the individual obtained by the physiological signal monitoring device can more scientifically and accurately evaluate the anxiety state and life quality of the individual.

[0014] Preferably, the HRV data in step S1 includes time domain indicators, frequency domain indicators and nonlinear domain indicators.

[0015] Preferably, the time domain indicators include Mean_NNI, SDNN, SDSD, NN50, pNN50, NN20, pNN20, RMSSD, Median_NN, Range_NN, CVSD, CV_NNI, Mean_HR, Max_HR, Min_HR and STD_HR.

[0016] Preferably, the frequency domain indicators include LF, HF, VLF, LFn, HFn and Total_Power.

[0017] Preferably, the nonlinear domain indicators include CSI, CVI, Modified_CSI, SD1, SD2, SD1 / SD2 ratio and SampEn.

[0018] The above preferred scheme has the beneficial effect that:

[0019] Collecting the HRV data of the individual including time domain indicators, frequency domain indicators and nonlinear domain indicators can more comprehensively and accurately evaluate the anxiety state and life quality of the individual.

[0020] Preferably, step S2 specifically includes the following steps:

[0021] S21. Filtering the collected HRV data and removing outliers;

[0022] S22. Standardizing the HRV data after filtering and removing outliers;

[0023] S23. Mapping the standardized HRV data into a vector of standard length.

[0024] Preferably, the standardization formula in step S22 is as follows:

[0025]

[0026] Wherein, y represents the indicator value after standardization; x represents the indicator value before standardization; mean represents the mean of all samples under the indicator; std represents the standard deviation of all samples under the indicator.

[0027] The above preferred scheme has the beneficial effect that:

[0028] The preprocessing can clean the noise and outliers possibly contained in the collected HRV data, thereby improving the quality of the data and the accuracy of the model.

[0029] As preferred, in the training phase in step S3, the Transformer neural network minimizes the prediction error by adjusting the weights and biases between the input layer nodes, the hidden layer nodes, and the output layer nodes.

[0030] The above preferred scheme has the beneficial effect that:

[0031] The prediction error is minimized by adjusting the weights and biases between the input layer nodes, the hidden layer nodes, and the output layer nodes, thereby accurately capturing the complex relationship between the HRV indicators and the anxiety state and the quality of life. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 A flowchart of a method for evaluating the anxiety index and the quality of life impact based on HRV indicators is shown. DETAILED DESCRIPTION

[0033] Exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be understood that the embodiments shown and described in the drawings are merely exemplary and are intended to illustrate the principles and spirit of the present application, and are not intended to limit the scope of the present application.

[0034] As shown in Figure 1 A method for evaluating the anxiety index and the quality of life impact based on HRV indicators, the steps of which include:

[0035] S1. Collecting HRV data of an individual;

[0036] S2. Preprocessing the collected HRV data;

[0037] S3. Inputting the preprocessed HRV data into a Transformer neural network model for training and testing;

[0038] S4. Using the trained Transformer neural network model to evaluate the anxiety index and the quality of life impact index of the individual, and outputting the evaluation results including the anxiety index and the quality of life impact index.

[0039] In this embodiment, the HRV data in step S1 is obtained by a physiological signal monitoring device.

[0040] In this embodiment, the HRV data in step S1 includes time domain indicators, frequency domain indicators, and nonlinear domain indicators.

[0041] In the present embodiment, the time domain indicators include: mean of all RR intervals (Mean_NNI), standard deviation of all RR intervals (SDNN), standard deviation of successive RR intervals differences (SDSD), number of adjacent RR intervals differences greater than 50 milliseconds (NN50), percentage of NN50 over all RR intervals (pNN50), number of adjacent RR intervals differences greater than 20 milliseconds (NN20), percentage of NN20 over all RR intervals (pNN20), root mean square of successive differences (RMSSD), median of all RR intervals (Median_NN), range of all RR intervals (Range_NN), ratio of SDSD to Mean_NNI (CVSD), ratio of SDNN to Mean_NNI (CV_NNI), mean heart rate (Mean_HR), maximum heart rate (Max_HR), minimum heart rate (Min_HR), and standard deviation of heart rate (STD_HR).

[0042] In the present embodiment, the frequency domain indicators include: low frequency power (LF), high frequency power (HF), very low frequency power (VLF), ratio of low frequency power to high frequency power (LH / HF ratio), percentage of low frequency power over total power (Lfnu), percentage of high frequency power over total power (Hfnu), and total power (Total_Power).

[0043] In the present embodiment, the non-linear domain indicators include: complexity entropy (CSI), consecutive variation index (CVI), modified complexity entropy (Modified_CSI), standard deviation 1 of Poincaré plot (SD1), standard deviation 2 of Poincaré plot (SD2), ratio of SD1 to SD2 (SD1 / SD2 ratio), and sample entropy (SampEn).

[0044] In the present embodiment, the step S2 specifically includes the following steps:

[0045] S21. filtering the collected HRV data and removing outliers;

[0046] S22. performing standardization processing on the filtered and outlier-removed HRV data;

[0047] S23. mapping the standardized HRV data into a vector of standard length, multiplying the values of all indicators by a matrix of (N*64), where N is the number of indicators, so that all indicators become a 64-dimensional vector.

[0048] In the present embodiment, the standardization processing formula in step S22 is as follows:

[0049]

[0050] where y represents the standardized index value; x represents the index value before standardization; mean represents the mean of all samples under the index; and std represents the standard deviation of all samples under the index.

[0051] In the training phase in step S3, the Transformer neural network model includes an input layer, multiple hidden layers, and an output layer. The neural network minimizes the prediction error by adjusting the weights and biases between the input layer nodes, the hidden layer nodes, and the output layer nodes.

[0052] The specific working principle and process of the application are as follows:

[0053] The HRV of 12580 anonymous participants was continuously collected by electrocardiogram (ECG) and pulse wave physiological signal detection equipment for one year at irregular intervals, with a total of 13278795 data. The collected data include but are not limited to time domain indicators (Mean_NNI, SDNN, SDSD, NN50, pNN50, NN20, pNN20, RMSSD, Median_NN, Range_NN, CVSD, CV_NNI, Mean_HR, Max_HR, Min_HR, and STD_HR), frequency domain indicators (LF, HF, VLF, LH / HF ratio, Lfnu, Hfnu, and Total_Power), and nonlinear domain indicators (CSI, CVI, Modified_CSI, SD1, SD2, SD1 / SD2 ratio, and SampEn). Then, the collected HRV data were cleaned through filtering, removing outliers, and standardization, to remove possible high-frequency noise, low-frequency drift, and outliers. The preprocessed HRV data were divided into a training set and a test set, the training set was used to train the Transformer neural network model, the optimization algorithm was AdamW, the learning rate was 0.00001, the betas were (0.9, 0.95), the weight decay was 0.01, the loss function was cross-entropy, and the training was performed for 2000 rounds. The test set was used to evaluate the performance of the trained Transformer neural network model, and the prediction accuracy of the anxiety index and the life quality impact index was calculated. Finally, the evaluation results of the anxiety index and the life quality impact index of each participant in the test set were output.

[0054] Through the above embodiments, the effectiveness and accuracy of the method of the application in evaluating the anxiety index and the life quality impact of individuals can be verified. In addition, the method of the application can be adjusted and optimized according to different application scenarios and requirements to adapt to different data sets and evaluation targets.

[0055] Those skilled in the art will appreciate that the embodiments described herein are presented for purposes of illustration and understanding of the principles of the application and should not be construed as limiting the scope of the application to such specifically outlined embodiments and examples. Various other specific embodiments and examples not described herein will be apparent to those skilled in the art in view of these teachings. The scope of the application should be determined from the claims.

Claims

1. A method for assessing the impact of anxiety index and quality of life based on the HRV indicator, characterized in that, Includes the following steps: S1. Collect individual HRV data; S2. Preprocess the collected HRV data; S3. Input the preprocessed HRV data into the Transformer neural network model for training and testing; S4. Use the trained Transformer neural network model to assess the individual's anxiety index and quality of life impact index, and output the assessment results including the anxiety index and quality of life impact index.

2. The method for assessing the impact of anxiety index and quality of life based on HRV index according to claim 1, characterized in that: The HRV data in step S1 is obtained through a physiological signal monitoring device.

3. The method for assessing the impact of anxiety index and quality of life based on HRV index according to claim 2, characterized in that: The HRV data in step S1 includes time-domain indicators, frequency-domain indicators, and nonlinear domain indicators.

4. The method for assessing the impact of anxiety index and quality of life based on HRV index according to claim 3, characterized in that, The time-domain metrics include: Mean_NNI, SDNN, SDSD, NN50, pNN50, NN20, pNN20, RMSSD, Median_NN, Range_NN, CVSD, CV_NNI, Mean_HR, Max_HR, Min_HR, and STD_HR.

5. The method for assessing the impact of HRV index on anxiety index and quality of life according to claim 3, characterized in that, The frequency domain metrics include: LF, HF, VLF, LH / HF ratio, Lfnu, Hfnu, and Total_Power.

6. The method for assessing the impact of anxiety index and quality of life based on HRV index according to claim 3, characterized in that, The nonlinear domain indices include: CSI, CVI, Modified_CSI, SD1, SD2, SD1 / SD2 ratio, and SampEn.

7. The method for assessing the impact of anxiety index and quality of life based on HRV indicators according to claim 1, characterized in that, Step S2 specifically includes the following steps: S21. Filter the collected HRV data and remove outliers; S22. Standardize the HRV data after filtering and outlier removal; S23. Map the standardized HRV data into a standard-length vector.

8. The method for assessing the impact of anxiety index and quality of life based on HRV index according to claim 7, characterized in that, The standardization formula described in step S22 is as follows: Where y represents the index value after standardization; x represents the index value before standardization; mean represents the mean of all samples under this index; and std represents the standard deviation of all samples under this index.

9. The method for assessing the impact of anxiety index and quality of life based on HRV index according to claim 1, characterized in that: In the training phase of step S3, the neural network minimizes the prediction error by adjusting the weights and biases between each input layer node, each hidden layer node, and each output layer node.

Images