Autonomic nervous function evaluation method based on photoelectric plethysmogram to suppress local interference

Abstract

The application discloses a kind of based on photoelectric volume pulse wave inhibits local interference autonomic nerve function evaluation method and system, belong to biomedical signal processing field.The method includes: obtaining PPG signal S (t) ;From it respectively extracts the first parameter of the oscillation component N (t) of characterizing blood vessel tension, and the second parameter of characterizing local time-varying gain factor G (t) ;The first parameter is associated with correction using the second parameter, to inhibit the local interference introduced by G (t), to correct and generate stabilized autonomic nerve function evaluation index, which corresponds to the pure evaluation of nerve component N (t).The present application overcomes the defects in the prior art that the evaluation index is easily disturbed by local physiological physical factors, and only single PPG signal can realize high robustness, high stability autonomic nerve function quantitative evaluation.

Description

Technical Field

[0001] This invention relates to the fields of biomedical signal processing, autonomic nervous function assessment, and non-invasive health monitoring. Specifically, it relates to a method and system for assessing autonomic nervous function based on photoplethysmography (PPG) signals, by suppressing local tissue perfusion interference and the influence of photoelectric physical gain, in order to stabilize physiological oscillation indicators such as Mayer waves and Traube-Hering waves (respiratory waves). Background Technology

[0002] The rhythmic regulation of the cardiovascular system by the autonomic nervous system is mainly manifested in physiological signals as two characteristic oscillatory components: respiratory waves (i.e., Traube-Hering waves) and Mayer waves.

[0003] Photoplethysmography (PPG) offers advantages such as being non-invasive and portable, making it a signal source for detecting respiratory and Mayer wave components, replacing invasive blood pressure monitoring. However, existing assessment methods typically use the absolute power of the Mayer wave band in the PPG signal directly as an indicator, which has significant drawbacks: this signal is a light intensity signal, and the amplitude of its AC component reflects not only the intensity of vasomotor activity driven by the central nervous system but is also modulated by local physiological and physical interferences. These interference factors can be divided into two categories: first, physiological interferences such as changes in local tissue perfusion, including vasodilation, venous filling, and hydrostatic pressure changes caused by heat application; and second, physical interferences such as fluctuations in ambient light, sensor contact pressure, skin color differences, and differences in device photoelectric gain. Both constitute a time-varying signal transmission gain, and its fluctuations modulate the target neural signal through a multiplicative effect, leading to ineffective fluctuations in the assessment indicators. Experimental data shows that simply altering local perfusion (such as heat application) significantly affects existing assessment indicators, even when central autonomic nervous activity remains stable. This experimental data quantifies the impact of local interference on existing assessment indicators.

[0004] Table 1 shows the degree of interference of local disturbances on existing technical indicators. When the central autonomic nervous activity of the subjects remained stable, applying heat to the left hand to change local perfusion resulted in significant deviations in Mayer wave power, mean amplitude, and low-frequency and high-frequency power based on heart rate intervals compared to the measurements of the right hand at room temperature. The measured ratios of each indicator differed significantly from the ideal ratios (which should be close to 1.0), confirming that local physiological changes introduce significant measurement biases unrelated to neural activity. The indicators in Table 1 are explained as follows: 1) Mayer wave frequency band power: Mayer wave frequency band power value directly extracted from the PPG signal using existing technical methods; 2) Mean amplitude of the PPG signal pulse wave: This indicator is used to characterize the level of local tissue perfusion; 3) Heart rate variability low-frequency power HRV_LF: refers to the low-frequency power calculated based on the heart rate interval sequence IBI, reflecting autonomic nervous regulation activity. 4) Heart rate variability high-frequency power HRV_HF: refers to the high-frequency power calculated based on the heart rate interval sequence IBI, which mainly reflects respiratory-related parasympathetic nerve activity.

[0005] Table 1

[0006]

[0007] According to the data in Table 1, the existing PPG power spectrum method is extremely sensitive to local physiological and physical conditions, and its measurement error is significant, making it unreliable as a clinical quantitative indicator. Furthermore, the low-frequency to high-frequency power ratio in traditional heart rate variability analysis is limited by the inherent defects of the inter-heartbeat interpolation algorithm, exhibiting significant instability in short-time analysis. Therefore, there is an urgent need in this field for a signal processing method capable of quantitatively removing local interference and accurately correcting the intensity of central nervous system modulation.

[0008] Existing technologies employ various technical approaches to achieve physiological signal analysis and evaluation, but their core concepts differ significantly from those of this invention.

[0009] Several prior art technologies (such as CN114173647B) have adopted a multimodal / dual-source analysis approach. Their core lies in simultaneously acquiring two heterogeneous signals: electrocardiogram (ECG) signals and photoplethysmography (PPG). By analyzing the synchronicity, statistical correlation, and other external dynamic characteristics between these two signals, they reflect the coordinated working state of the cardiac and vascular systems. Their primary application is the diagnosis of organic cardiovascular diseases.

[0010] The core technology of patent WO2017089921A1 also relies on synchronous electrocardiogram signals (such as R waves) as a time reference. Its purpose is to preprocess and control the quality of photoplethysmography (PPG) signals to ensure the quality of the input signal, without involving the extraction of specific physiological parameters from the signal for evaluation.

[0011] Patent CN118102970A employs a feature engineering and machine learning classification approach. It extracts a large number of features from photoplethysmography (PPG) waves and inputs them into a classifier, aiming to classify different hemodynamic states rather than to provide a continuous quantitative assessment of autonomic nervous function.

[0012] Patent US6805673B2 discloses a method for extracting Mayer wave (low-frequency) components from photoplethysmography (PPG) signals. However, this technical solution does not address the practical problem that the amplitude or power of the extracted Mayer wave components can fluctuate significantly due to interference such as local perfusion level and sensor contact, and therefore does not propose any corresponding solutions.

[0013] Unlike the technologies mentioned above, the unique approach of this invention lies in relying entirely on the internal information of a single photoplethysmography (PPG) signal. By extracting two specific parameters with a clear physiological correspondence and establishing a deterministic mathematical correlation model between them, the invention aims to correct for local interference and ultimately achieve stable and quantitative assessment of autonomic nervous function.

[0014] Existing autonomic nervous system function assessment technologies based on photoplethysmography (PPG) signals face a long-standing technical bias: because this signal is inherently susceptible to significant interference from multiple physical and physiological factors such as local tissue perfusion levels, sensor contact conditions, and photoelectric gain, it is generally considered difficult to use alone for stable and reliable quantitative assessment of neural regulatory activities. Traditional solutions typically rely on introducing additional physiological signals for multimodal fusion, which increases the complexity of the system and the barrier to entry for its use.

[0015] Therefore, there is an urgent need in this field for an innovative signal processing method that can quantitatively isolate local physiological and physical gain interference and accurately correct the intensity of central nervous system regulation. Summary of the Invention

[0016] To address the shortcomings of the existing technologies, the present invention aims to provide a method and system for assessing autonomic nervous function based on photoplethysmography (PPG) to suppress local interference. By introducing a second parameter that reflects the local tissue perfusion level or optical coupling strength, the first parameter that characterizes the physiological oscillation intensity related to autonomic nervous regulation is corrected to eliminate the interference of local time-varying gain, thereby obtaining a stable autonomic nervous assessment index.

[0017] To achieve the above-mentioned objectives, the present invention proposes the following technical solution:

[0018] In a first aspect, the present invention proposes a method for assessing autonomic nervous function based on suppressing local interference using photoplethysmography (PPG), comprising the following steps:

[0019] Signal acquisition steps: Acquire the photoplethysmography (PPG) signal S(t) of the target object;

[0020] Feature extraction steps: Extract a first parameter and a second parameter from the photoplethysmography (PPG) signal S(t), wherein the first parameter characterizes the intensity of the vascular tension oscillation component N(t) driven by the central nervous system in the signal S(t), and the second parameter characterizes the intensity of the gain factor G(t) affected by local physiological and physical factors, including at least one of local tissue perfusion level, ambient light interference, or device photoelectric gain level;

[0021] Association generation step: The first parameter is correlated using the second parameter to suppress the local interference represented by the gain factor G(t), thereby correcting and generating a stable autonomic nerve function assessment index. The assessment index is a quantitative representation of the vascular tension oscillation component N(t), and its value has excluded the interference of the local gain factor G(t).

[0022] In some implementations, the second parameter is any one of the DC component of the signal, the injection index, or a transformed form thereof.

[0023] In some implementations, the feature extraction step includes extracting the first parameter as follows:

[0024] Power spectral density analysis was performed on the PPG signal;

[0025] In the power spectral density spectrum, the power value within the preset autonomic nervous system modulation-related frequency band is calculated by integration and used as the first parameter;

[0026] The autonomic nervous system modulation-related frequency bands include at least one of the Mayer wave frequency band and the respiratory wave frequency band.

[0027] In some implementations, the Mayer wave frequency band is 0.04 Hz to 0.15 Hz, and the respiratory wave frequency band is 0.15 Hz to 0.4 Hz.

[0028] In some implementations, the feature extraction step includes extracting the second parameter in at least one of the following ways:

[0029] Method A: Based on time-domain amplitude statistics, calculate the statistics of the pulse wave amplitude sequence of the PPG signal in the time domain, and use it as the second parameter;

[0030] Method B: Based on the frequency domain heart rate band power, the power value in the frequency band where the heart rate main frequency is located is calculated by integration in the power spectral density spectrum of the PPG signal, and used as the second parameter;

[0031] Method C: Based on gain estimation of envelope energy per beat, the amplitude envelope signal is constructed by extracting the envelope of the photoplethysmography pulse wave signal; the square of the amplitude of the envelope signal is statistically calculated within a preset time window, and the obtained statistical value is defined as the envelope energy parameter, and this parameter is used as the second parameter.

[0032] In some implementations, method A specifically includes:

[0033] Identify the systolic peak and diastolic trough of the PPG signal;

[0034] Calculate the pulse wave amplitude for each cardiac cycle to form an amplitude sequence;

[0035] The mean square value or functional form of the amplitude sequence is calculated as the second parameter; wherein the functional form includes the perfusion index.

[0036] In some implementations, the association generation step includes at least one of the following methods:

[0037] Ratio calculation: Calculate the ratio of the first parameter to the second parameter to generate a dimensionless index;

[0038] Functional relationship modeling: Construct a functional relationship model with the second parameter as the independent variable and the first parameter as the dependent variable, and use the model output or residual as the evaluation index;

[0039] Amplitude standardization: Before extracting the first parameter, the PPG signal is subjected to amplitude standardization using the second parameter or its functional form.

[0040] In some implementations, the association generation step further includes numerical mapping, nonlinear transformation, or scaling of the evaluation index to improve its statistical distribution characteristics.

[0041] In some implementations, a dynamic evaluation step is also included:

[0042] The PPG signal is divided into multiple continuous sub-segments;

[0043] For each sub-segment, the signal acquisition step, feature extraction step, and association generation step are performed to generate an evaluation index sequence for the sub-segment;

[0044] Based on the evaluation index sequence, statistical or dynamic characteristics are calculated to quantify the short-term dynamic characteristics of autonomic nervous system regulation.

[0045] The statistical or dynamic characteristics include at least one of standard deviation, coefficient of variation, adjacent fluctuation intensity, trend of change, or dynamic pattern.

[0046] Secondly, an autonomic nervous system function assessment system based on photoplethysmography (PPG) to suppress local interference includes:

[0047] The signal acquisition module is configured to acquire the photoplethysmography (PPG) signal S(t) of the target object.

[0048] A processing module, connected to the signal acquisition module, is configured to process the photoplethysmography (PPG) signal S(t), wherein the processing module includes:

[0049] The feature extraction unit is configured to extract a first parameter and a second parameter from the signal S(t), wherein the first parameter corresponds to the intensity characterization of the vascular tension oscillation component N(t) driven by the central nervous system in the signal S(t), and the second parameter corresponds to the estimated value of the time-varying gain factor G(t) affected by local physiological and physical factors.

[0050] A signal correction unit is configured to receive the first parameter and the second parameter, and to perform correlation operations between the first parameter and the second parameter to suppress the local interference represented by the gain factor G(t), thereby correcting and generating a stable autonomic nerve function assessment index. The assessment index is a quantitative representation of the vascular tension oscillation component N(t), and its value has excluded the interference of the local gain factor G(t).

[0051] An output module, connected to the processing module, is configured to output the stabilized autonomic nervous function assessment indicators.

[0052] In some embodiments, the signal acquisition module includes a photoelectric sensor and a signal conditioning circuit, and the processing module includes a processor and a memory. The memory stores program instructions, and when the program instructions are executed by the processor, the method steps are implemented.

[0053] Thirdly, a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any of the autonomic nervous function assessment methods based on photoplethysmography to suppress local interference.

[0054] Fourthly, an electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements any of the autonomic nervous function assessment methods based on photoplethysmography (PPG) suppression of local interference.

[0055] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0056] To address the aforementioned technical biases, this invention proposes an innovative solution based on the intrinsic characteristics of a single photoplethysmography (PPG) signal. Its core mechanism lies in revealing and utilizing for the first time the inherent correlation between two physiologically significant parameters within the signal: a "first parameter" reflecting the intensity of central nervous system modulation oscillations, and a "second parameter" characterizing the level of local time-varying interference. By constructing a deterministic mathematical model between the two, local interference components can be effectively removed at the algorithmic level, thereby stably extracting pure indicators of neuromodulation function. The following examples and experimental data will demonstrate that this correlation correction method can significantly overcome the shortcomings of traditional methods and achieve highly robust evaluation results.

[0057] Compared with the prior art, the present invention can achieve the following beneficial technical effects:

[0058] 1. Significantly Enhanced Resistance to Local Interference: By introducing a characterization parameter of local tissue perfusion level (the second parameter) to correlate and correct the intensity of autonomic nervous system-regulated oscillations (the first parameter), non-neurogenic interferences such as changes in local tissue perfusion, environmental temperature fluctuations, and probe coupling status are effectively suppressed. Experiments show that this method significantly improves the stability of physiological oscillation indicators (such as Mayer waves), avoiding the problem of large fluctuations in indicators caused by local interference in existing technologies.

[0059] 2. Optimize the statistical distribution characteristics of the indicators: Through correlation correction and optional statistical correction, the dispersion and outlier ratio of physiological oscillation indicators in the population sample are significantly reduced, improving their statistical stability and interpretability, thus providing them with the basic conditions for constructing clinical reference ranges.

[0060] 3. Enhanced cross-individual comparability: By suppressing non-neurogenic interferences such as local perfusion and optical coupling, the measurement bias introduced by individual anatomical differences (such as skin color, subcutaneous fat thickness, finger thickness), sensor wearing pressure, or device hardware differences is significantly reduced, providing a technical basis for establishing universal diagnostic criteria.

[0061] The above-mentioned beneficial effects are all directly generated based on the core technical features of the present invention (the correlation correction mechanism between the first parameter and the second parameter). The actual test verification of 455 PPG signal data (each signal duration ≥100 seconds) fully demonstrates the innovation and practicality of the present invention in the field of autonomic nervous function assessment. Attached Figure Description

[0062] Figure 1 is a flowchart of the overall process of the autonomic nervous function assessment method based on photoplethysmography to suppress local interference according to the present invention.

[0063] Figure 2 shows the frequency domain comparison of the existing technical indicators (Mayer wave power) and the local interference reference indicators for the left and right hands under local heat application interference. The figure demonstrates that, under heat application to the left hand, both the first parameter (Mayer wave power, left hand heat application 3161 V² / Hz vs right hand normal temperature 1147 V² / Hz) and the second parameter (heart rate power, left hand heat application 886 V² / Hz vs right hand normal temperature 338 V² / Hz) exhibit significant energy differences, confirming the overall boosting effect of local perfusion on frequency domain energy.

[0064] Figure 3 is a comparison chart of the second parameter (amplitude) of the present invention for the left and right hands. The left side shows the detection results of the left hand under hot compress conditions, and the right side shows the detection results of the right hand under normal temperature conditions.

[0065] Figure 4 is a schematic diagram of the autonomic nervous function assessment system module based on photoplethysmography to suppress local interference according to the present invention.

[0066] Figure 5 This is a schematic diagram of the technical route of the autonomic nervous function assessment method based on photoplethysmography to suppress local interference according to the present invention.

[0067] Figure 6 This is a schematic diagram of a platform implementation embodiment of the autonomic nervous function assessment system based on photoplethysmography pulse wave suppression of local interference according to the present invention.

[0068] Figure 7 This is a schematic diagram of the process for envelope extraction and perfusion gain estimation based on photoplethysmography (PPG) signals. Detailed Implementation

[0069] To enable those skilled in the art to clearly understand and implement the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the scope of protection of the present invention.

[0070] Example 1: As Figure 1 As shown, the specific process of the autonomic nervous function assessment method based on photoplethysmography (PPG) to suppress local interference according to the present invention includes the following steps:

[0071] Step S1, Signal Acquisition Step: Acquire the photoplethysmography (PPG) signal of the target object;

[0072] Step S2, Feature Extraction Step: Extract a first parameter and a second parameter from the photoplethysmography (PPG) signal S(t), wherein the first parameter characterizes the intensity of the vascular tension oscillation component N(t) driven by the central nervous system in the signal S(t), and the second parameter characterizes the intensity of the gain factor G(t) affected by local physiological and physical factors, including at least one of local tissue perfusion level, ambient light interference, or device photoelectric gain level;

[0073] This step includes extracting the first parameter reflecting neural modulation. and the second parameter reflecting local gain :

[0074] Extract the first parameter The power spectral density calculated above In the spectrum, the power value is calculated by integrating the power within the frequency band related to autonomic nervous system regulation. .

[0075] For example, the power calculation formula for the Mayer band is as follows:

[0076] ;

[0077] The lower integration limit of 0.04 Hz and the upper integration limit of 0.15 Hz correspond to the standard frequency band boundaries of the Mayer wave, respectively. For the signal at frequency The power spectral density value at that location.

[0078] Extract the second parameter The second parameter is used to quantize the local amplitude gain level of the PPG signal (i.e., the signal reference gain).

[0079] The second parameter is any one of the DC component of the signal, the injection index, or a transformed form thereof.

[0080] The following implementation method is adopted:

[0081] Extract the second parameter from the preprocessed PPG signal. This is used to quantize the local amplitude gain level of the signal (i.e., the signal reference gain). The following description uses a preferred implementation method (Method A and Method B) with low computational complexity and high real-time performance as an example.

[0082] Method A: Based on time-domain amplitude statistics:

[0083] 1. Time-domain waveform feature identification: In the preprocessed PPG time-domain waveform, each cardiac cycle is identified one by one. The peak value of the contraction phase is denoted as The trough of diastole is denoted as .

[0084] 2. Constructing a pulse wave amplitude sequence: Based on the identified peak and trough values, calculate the pulse wave amplitude for each cardiac cycle using the following formula to form an amplitude sequence. :

[0085] ;

[0086] 3. Calculate the amplitude baseline parameter: Within the set time window, perform statistical calculations on the above amplitude sequence, and the resulting statistic is the second parameter. It represents the local amplitude baseline level of the signal within that time period. The statistic can be the mean square value, the arithmetic mean, or a function thereof.

[0087] Further, the mean square value or functional form of the amplitude sequence is calculated as the second parameter; wherein the functional form includes the perfusion index, which is the ratio of the mean of the amplitude to the DC component (DC) of the PPG signal.

[0088] Method B: Based on frequency domain heart rate band power, as an alternative to Example 1:

[0089] As another preferred embodiment, the second parameter Alternatively, it can be extracted directly from the frequency domain representation of the signal. The specific steps are as follows:

[0090] 1. Identify the dominant heart rate frequency: In the power spectrum of the PPG signal, identify the frequency component with the highest energy, i.e., the dominant heart rate frequency. .

[0091] 2. Calculate the bandpass power: using Calculate a preset narrow frequency band centered on the target (e.g., ...). The power spectrum integral value within ±0.1 Hz is used as the second parameter. .

[0092] Related technical principle: The signal energy of the heart rate main frequency and its adjacent frequency bands mainly originates from the periodic blood volume changes caused by the mechanical pulsation of the heart. Therefore, the power of this frequency band is highly correlated with the local tissue perfusion level and can be used as an effective estimator of the local gain G(t).

[0093] Step S3, Association Generation Step: The association generation step involves using the second parameter to perform an association operation on the first parameter to suppress the local interference represented by the gain factor G(t), thereby correcting and generating a stable autonomic nerve function assessment index. The assessment index is a quantitative representation of the vascular tension oscillation component N(t), and its value has excluded the interference of the local gain factor G(t).

[0094] Method C: Gain estimation based on frame-by-frame envelope energy, as an alternative to Example 2:

[0095] As another preferred embodiment, the second parameter Alternatively, a continuous-time amplitude-modulated signal can be obtained by extracting the cardiac cycle-related envelope from the PPG signal, and the temporal statistics of its squared amplitude (such as mean or mean square value) can be calculated as a second parameter. This method directly reflects the local gain factor through envelope modeling. The energy level is suitable for high temporal resolution scenarios. For example... Figure 7 The diagram shows a flowchart illustrating envelope extraction and perfusion gain estimation based on photoplethysmography (PPG) signals. Specifically, it includes:

[0096] Envelope extraction is performed, including bandpass filtering (e.g., 0.7-5.0 Hz) of the preprocessed PPG signal to highlight the cardiac carrier component. Then, the systolic peak and diastolic trough are detected, and the upper envelope (peak sequence) and lower envelope (trough sequence) are constructed by interpolation.

[0097] The envelope energy calculation involves calculating the square of the envelope amplitude (the difference between the upper and lower envelopes) and averaging it over a time window, as shown in the following expression:

[0098] ;

[0099] in, For time-varying gain factor, This is the envelope signal after low-pass smoothing.

[0100] By providing continuous gain estimation through method C, the discrete sampling error of method A can be avoided, and it is insensitive to motion artifacts.

[0101] Furthermore, this includes performing gain factor analysis. The steps for energy estimation are described in detail below:

[0102] 1. Perform signal preprocessing on the PPG signal. Bandpass filtering (0.7-5.0 Hz) was performed to extract the cardiac carrier component. ;

[0103] 2. Construct the envelope, including the following steps:

[0104] Detection The peak (systolic) and trough (diastolic) values ​​are denoted as peaks and valleys, respectively.

[0105] The peak and valley sequences are converted into continuous-time signals using linear interpolation to obtain the upper envelope. and lower envelope .

[0106] The envelope amplitude signal is calculated using the following expression:

[0107] ;

[0108] 3. Smoothing and energy statistics, including the following processes:

[0109] right A low-pass filter with a cutoff frequency of 0.3 Hz is applied to obtain a smooth envelope. .

[0110] Calculate the envelope energy parameters:

[0111] ;

[0112] Method C via envelope energy It directly corresponds to the square mean of local injection gain, complementing Method A (discrete amplitude statistics) and Method B (frequency domain power), and is particularly suitable for dynamic evaluation.

[0113] To obtain stable evaluation metrics, the first parameter needs to be correlated and corrected. The purpose of this correction is to use the second parameter to counteract local interference. The specific correction path can be selected based on the application scenario.

[0114] One approach is to directly calculate the ratio of the two parameters to obtain a dimensionless index. Another approach is to establish a functional model between the two and then use the model predictions or residuals as an index. A third approach is to intervene at an earlier stage, that is, first use the second parameter to standardize the amplitude of the original PPG signal, and then extract the first parameter from the processed signal.

[0115] Regardless of the path used to generate the initial indicators, subsequent processing can be introduced as needed, such as numerical mapping, nonlinear transformation, or scaling, to ultimately optimize the statistical properties and interpretability of the indicators.

[0116] In the feature extraction step, extracting the first parameter includes: performing power spectral density analysis on the PPG signal;

[0117] In the power spectral density spectrum, the power value within the preset autonomic nervous system modulation-related frequency band is calculated by integration and used as the first parameter;

[0118] The autonomic nervous system modulation-related frequency bands include at least one of the Mayer wave frequency band and the respiratory wave frequency band.

[0119] The Mayer wave frequency band is 0.04Hz to 0.15Hz, and the respiratory wave frequency band is 0.15Hz to 0.4Hz.

[0120] In the signal acquisition step, to ensure the accuracy of frequency domain analysis, the original signal undergoes the following preprocessing:

[0121] To ensure the effectiveness of frequency domain analysis, the raw PPG signal needs to undergo standardization preprocessing during signal acquisition. This process mainly includes:

[0122] Baseline drift removal and filtering: Eliminate extremely low frequency baseline drift components in the signal (e.g., filter out extremely low frequency components below 0.01Hz) and high frequency noise (including artifacts and power frequency interference), thereby completing the basic purification of the signal.

[0123] Power spectral density (PSD) conversion: The purified time-domain signal is converted into power spectral density through spectrum estimation, forming the frequency domain representation required for subsequent feature extraction.

[0124] For power spectral density estimation, the Welch average periodogram method is recommended. Its key parameters (such as window function type, window length, and overlap rate) should be rationally configured based on the actual signal sampling rate and specific physiological analysis requirements, aiming to achieve the optimal balance between spectral resolution and estimation variance suitable for physiological signal analysis.

[0125] In the feature extraction step, extracting the second parameter includes at least one of the following methods:

[0126] Method A: Based on time-domain amplitude statistics, calculate the statistics of the pulse wave amplitude sequence of the PPG signal in the time domain, and use it as the second parameter;

[0127] Method B: Based on the frequency domain heart rate band power, the power value in the frequency band where the heart rate main frequency is located is calculated by integration in the power spectral density spectrum of the PPG signal, and used as the second parameter;

[0128] Method C: Based on gain estimation of envelope energy per beat, the amplitude envelope signal is constructed by extracting the envelope of the photoplethysmography pulse wave signal; the square of the amplitude of the envelope signal is statistically calculated within a preset time window, and the obtained statistical value is defined as the envelope energy parameter, and this parameter is used as the second parameter.

[0129] In the association generation step, the association operation includes at least one of the following methods:

[0130] Ratio calculation: Calculate the ratio of the first parameter to the second parameter to generate a dimensionless index;

[0131] Functional relationship modeling: Construct a functional relationship model with the second parameter as the independent variable and the first parameter as the dependent variable, and use the model output or residual as the evaluation index;

[0132] Amplitude standardization: Before extracting the first parameter, the PPG signal is subjected to amplitude standardization using the second parameter or its functional form.

[0133] The association generation step further includes numerical mapping, nonlinear transformation, or scaling of the evaluation index to improve its statistical distribution characteristics.

[0134] It also includes a dynamic evaluation step:

[0135] The PPG signal is divided into multiple continuous sub-segments;

[0136] For each sub-segment, the signal acquisition step, feature extraction step, and association generation step are performed to generate an evaluation index sequence for the sub-segment;

[0137] Based on the evaluation index sequence, statistical or dynamic characteristics are calculated to quantify the short-term dynamic characteristics of autonomic nervous system regulation.

[0138] The statistical or dynamic characteristics include at least one of standard deviation, coefficient of variation, adjacent fluctuation intensity, trend of change, or dynamic pattern.

[0139] Example 2: Figure 4 As shown, this invention also proposes an autonomic nervous system function assessment system based on photoplethysmography (PPG) to suppress local interference, used to implement any of the above-mentioned autonomic nervous system function assessment methods based on PPG to suppress local interference, including:

[0140] The signal acquisition module 100 is configured to acquire the photoplethysmography (PPG) signal S(t) of the target object.

[0141] Processing module 200, connected to the signal acquisition module, is configured to process the photoplethysmography (PPG) signal S(t), wherein the processing module includes:

[0142] The feature extraction unit 210 is configured to extract a first parameter and a second parameter from the signal S(t), wherein the first parameter corresponds to the intensity characterization of the vascular tension oscillation component N(t) driven by the central nervous system in the signal S(t), and the second parameter corresponds to the estimated value of the time-varying gain factor G(t) affected by local physiological and physical factors.

[0143] The signal correction unit 220 is configured to receive the first parameter and the second parameter, and to suppress the local interference represented by the gain factor G(t) by performing correlation operation on the first parameter and the second parameter, thereby correcting and generating a stable autonomic nerve function assessment index. The assessment index is a quantitative representation of the vascular tension oscillation component N(t), and its value has excluded the interference of the local gain factor G(t).

[0144] The output module 300 is connected to the processing module and is configured to output the stabilized autonomic nervous function assessment index.

[0145] The signal acquisition module 100 includes a photoelectric sensor and a signal conditioning circuit. The processing module includes a processor and a memory. The memory stores program instructions. When the program instructions are executed by the processor, the method steps are implemented.

[0146] This invention is based on the following biophysical model: the observed PPG signal S(t) is the product of the central nervous system-driven vascular tension oscillation N(t) and the local physical / physiological transfer function G(t), i.e., S(t) ≈ G(t) × N(t). Here, N(t) represents the true vascular tension oscillation component driven by central nervous system activity, and G(t) represents the time-varying gain factor, which comprehensively reflects the influence of local blood flow, ambient light intensity, sensor sensitivity, etc. Therefore, directly calculating the power of S(t) actually yields the integral of G(t)² × N(t)² in the frequency domain, and the result is severely contaminated by fluctuations in the interference gain G(t), leading to extremely unstable evaluation indicators and huge errors. Therefore, the key technical problem of this invention is: how to effectively remove or suppress the influence of this time-varying gain from the observation signal S(t), which is susceptible to interference from G(t), thereby reliably correcting the true neural modulation component N(t).

[0147] To address the aforementioned issues, this invention departs from the traditional approach of relying on additional reference signals (such as ECG). Instead, it creatively proposes extracting two feature parameters with a clear physiological correspondence from a single PPG signal S(t), and achieving self-calibration through correlation operations between the two parameters.

[0148] Extraction of the first parameter: Extract the first parameter from the signal S(t) to characterize the intensity of N(t). For example, by performing spectral analysis on S(t), calculate the power in its Mayer band (0.04-0.15 Hz), which is essentially the result of N(t)² modulated by G(t)².

[0149] Extraction of the second parameter: A second parameter is extracted from the same signal S(t) to estimate the intensity of the time-varying gain G(t). This parameter is a direct or indirect reflection of the intensity of G(t), such as the average amplitude of the pulse wave of the PPG signal, or the band power near the dominant heart rate frequency.

[0150] Association generation step: The first parameter is correlated using the second parameter to suppress the local interference represented by the gain factor G(t), thereby correcting and generating a stable autonomic nerve function assessment index. The assessment index is a quantitative representation of the vascular tension oscillation component N(t), and its value has excluded the interference of the local gain factor G(t).

[0151] Advanced Example: Short-Term Dynamic Assessment of the Autonomic Nervous System

[0152] To assess the short-term dynamic characteristics of autonomic nervous system function, this advanced embodiment further includes the following steps based on the aforementioned method:

[0153] 1. Signal segmentation and index sequence generation

[0154] The acquired PPG signals were divided into multiple continuous and non-overlapping sub-segments in chronological order. The aforementioned signal acquisition, feature extraction, and association generation steps were performed on each sub-segment to calculate the corresponding stable physiological assessment index. This resulted in a time-series sequence of assessment indices arranged chronologically.

[0155] 2. Dynamic Feature Extraction and Analysis

[0156] Based on the time series of the aforementioned evaluation indicators, their statistical and dynamic characteristics are calculated to quantify the short-term dynamic properties of autonomic nervous system regulation. Extractable features include, but are not limited to:

[0157] Static dispersion: such as the standard deviation or coefficient of variation of a series, is used to describe the overall fluctuation level of the evaluation index during the observation period.

[0158] Adjacent fluctuation intensity: Calculates the root mean square of continuous differences in the sequence to quantify the intensity of rapid changes in the evaluation index between adjacent sub-segments, reflecting the instantaneous volatility of autonomic nervous regulation.

[0159] Trend of change: By performing linear or nonlinear fitting on the sequence, the overall slope of change is obtained to assess the trend of increasing or decreasing autonomic nerve tension during the observation period.

[0160] Dynamic pattern: Analyze the recovery curve pattern of the sequence before and after a specific intervention (such as postural change, cold stimulation, etc.), for example, fit an exponential decay function and calculate its time constant, or calculate its response amplitude and rate to the stimulus, in order to assess the stress responsiveness and recovery ability of the autonomic nervous system.

[0161] By extracting the aforementioned dynamic features, this embodiment can provide quantitative information on the transient volatility, recovery ability, and stress response of autonomic nervous regulation based on a single point-in-time assessment, thereby more comprehensively and dynamically characterizing an individual's physiological state and neural regulatory function.

[0162] Detailed Implementation (Supplementary Examples)

[0163] To fully disclose the present invention, a preferred embodiment is described in detail below with reference to specific experimental data and parameters. Those skilled in the art can reproduce the present invention based on this description, but this does not constitute a limitation on the scope of protection.

[0164] 1. Use resting-state data from publicly available PPG databases (such as the BIDMC PPG and Respiration Dataset). Select a 5-minute segment of PPG signal S_raw(t) with a sampling rate of 125 Hz. First, perform preprocessing: use a first-order high-pass filter (cutoff frequency 0.5 Hz) to remove baseline drift, and then use a fourth-order Butterworth bandpass filter (passband 0.5-10 Hz) to suppress high-frequency noise and power frequency interference, obtaining the preprocessed signal S(t).

[0165] 2. Feature extraction in step S2

[0166] (a) Extracting the first parameter (Specifically, Mayer wave power)

[0167] The power spectral density (PSD) of S(t) was calculated using the Welch method with the following parameters: Hamming window, window length of 256 points (approximately 2 seconds), and overlap of 50%. The power value was obtained by integrating the PSD spectrum over the frequency band from 0.04 Hz to 0.15 Hz. For example, the signal of a certain subject was calculated as follows: = 15.6 (au).

[0168] (b) Extract the second parameter (Average amplitude of pulse wave)

[0169] Peak and trough values ​​of S(t) are detected in the time domain. The pulse wave amplitude (peak-trough value) of 10 consecutive cardiac cycles is calculated, and its arithmetic mean is taken as the peak value. The signal calculated for the same subject yielded: = 102.4 (au).

[0170] 3. Correlation Correction and Indicator Generation (Step S3)

[0171] The ratio method was used for correlation correction to generate a stable evaluation index, namely the autonomic nerve drive intensity ANS_Index corresponding to each unit of local perfusion gain.

[0172] The calculation formula is: ;

[0173] Substitute the values ​​from the example above: .

[0174] 4. Stability Verification (Effective Examples)

[0175] To verify the effectiveness of this invention in suppressing local interference, a control experiment was designed as follows:

[0176] I. The verification experiments for method A and method B are as follows:

[0177] Condition I (normal temperature):

[0178] Measure the PPG of the subject's right hand and calculate it. .

[0179] Condition II (Heat Interference): After applying a 40°C heat compress to the left hand of the same subject for 5 minutes, the PPG of the left hand was measured and calculated. .

[0180] Comparison method: Using existing technology (US6805673B2), the absolute power of the Mayer wave band before and after heat application was directly calculated. (Uncorrected) With Power_II.

[0181] This represents the absolute power of the Mayer wave directly calculated from the PPG signal measured from the subject's right hand under condition I (normal temperature, no interference).

[0182] Power_II represents the absolute power of the Mayer wave directly calculated from the PPG signal measured from the left hand (after heat treatment) of the same subject under condition II (heat interference).

[0183] Results: Theoretically, central nervous system activity remains stable over a short period, and ideal assessment indicators should be essentially consistent under conditions I and II. Experimental data show:

[0184] Invention Specifications: = 0.150, = 0.155, the deviation is approximately 3.3%.

[0185] Current technical specifications: = 15.6, Power_II = 25.8, with a deviation as high as 65.4%.

[0186] The results show that Power_II was used as a benchmark for comparing the effectiveness with the method of this invention. Their significant numerical difference (65.4%) vividly demonstrates the sensitivity and instability of the prior art to local interference, thus highlighting how the method of this invention (with a deviation of only 3.3%) effectively suppresses interference through correlation correction, achieving much more stable evaluation results. This invention, through correlation correction, significantly suppresses measurement bias caused by localized heat application interference, resulting in a substantial improvement in the stability of the evaluation results.

[0187] As another embodiment, the second parameter It can also be extracted from the frequency domain. First, calculate the PSD to locate the heart rate frequency. (e.g., how many Hz), then calculate Integral power within the frequency band as Then use the formula The same stabilization assessment can be achieved by performing calculations.

[0188] This embodiment demonstrates that the method provided by the present invention can effectively isolate local interference and achieve stable and reliable autonomic nervous function assessment.

[0189] II. The verification experiments for Method C, Method A, and Method B are as follows:

[0190] Conditions: Compare PPG signals under hot compress (left hand) and normal temperature (right hand).

[0191] Result: Extracted by method C The gain is 120.5 au under heat and 88.3 au at room temperature, with a ratio of 1.36 (close to the ideal value of 1.0), while the original Mayer wave power ratio reaches 2.76. This indicates that mode C can more stably characterize gain changes.

[0192] Conclusion: Method C effectively suppresses local perfusion interference and improves the robustness of evaluation indicators through envelope energy estimation.

[0193] like Figure 5 The diagram shows the technical route of the autonomic nervous function assessment method of the present invention. Figure 6 This is a schematic diagram of a platform implementation embodiment of the autonomic nervous function assessment system of the present invention.

[0194] In summary, through the aforementioned "extraction-association-correction generation" technical means, this invention achieves the following outstanding technical effects:

[0195] 1) Effective suppression of interference is achieved: The correlation generation step can significantly suppress various local interferences characterized by the gain factor G(t), and solve the problem of invalid fluctuations in evaluation results caused by fluctuations in G(t) in traditional methods.

[0196] 2) Corrected the true physiological information: The final evaluation index is essentially a stable assessment of the vascular tension oscillation component N(t), which more realistically and accurately reflects the central autonomic nervous system regulation function of the target object, rather than the physical and physiological gain changes in the signal acquisition process.

[0197] 3) A complete innovation closed loop is constructed: This invention follows a complete logic of identifying the problem (G(t) interference) → proposing a means (intrinsic parameter correlation correction) → achieving the effect (stable evaluation of N(t)). This "problem-means-effect" closed loop fundamentally distinguishes this invention from the crude processing method of directly calculating S(t)² (i.e., G(t)²×N(t)²) in existing technologies. It provides a novel technical path that does not rely on multimodal signal fusion and can achieve highly robust autonomic neural function evaluation from within a single PPG signal, possessing outstanding substantive characteristics and significant progress.

[0198] Example 3: A non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the autonomic nervous function assessment method based on photoplethysmography (PPG) suppression of local interference according to Example 1 of the present invention. The present invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate methods for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0199] Example 4: An electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the autonomic nervous function assessment method based on photoplethysmography (PPG) suppression of local interference according to Example 1 of the present invention. These computer program instructions may also be stored in a computer-readable storage medium capable of directing a computer or other programmable data processing device to operate in a specific manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a flow... Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0200] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0201] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0202] This invention is not limited to the embodiments described above. The above description of specific embodiments is intended to illustrate and explain the technical solutions of this invention. The specific embodiments described above are merely illustrative and not restrictive. Without departing from the spirit and scope of the claims, those skilled in the art can make many specific modifications based on the teachings of this invention, and these modifications all fall within the scope of protection of this invention.

Claims

1. A method for assessing autonomic nervous system function based on photoplethysmography (PPG) to suppress local interference, characterized in that, Includes the following steps: Signal acquisition steps: Acquire the photoplethysmography (PPG) signal S(t) of the target object; Feature extraction steps: Extract a first parameter and a second parameter from the photoplethysmography (PPG) signal S(t), wherein the first parameter characterizes the intensity of the vascular tension oscillation component N(t) driven by the central nervous system in the signal S(t), and the second parameter characterizes the intensity of the gain factor G(t) affected by local physiological and physical factors, including at least one of local tissue perfusion level, ambient light interference, or device photoelectric gain level; Association generation step: The first parameter is correlated using the second parameter to suppress the local interference represented by the gain factor G(t), thereby correcting and generating a stable autonomic nerve function assessment index. The assessment index is a quantitative representation of the vascular tension oscillation component N(t), and its value has excluded the interference of the local gain factor G(t).

2. The method for assessing autonomic nervous function based on photoplethysmography (PPG) to suppress local interference as described in claim 1, characterized in that, In the feature extraction step, extracting the first parameter includes: Power spectral density analysis was performed on the PPG signal; In the power spectral density spectrum, the power value within the preset autonomic nervous system modulation-related frequency band is calculated by integration and used as the first parameter; The autonomic nervous system modulation-related frequency bands include at least one of the Mayer wave frequency band and the respiratory wave frequency band.

3. The method for assessing autonomic nervous function based on photoplethysmography (PPG) to suppress local interference as described in claim 3, characterized in that, The Mayer wave frequency band is 0.04Hz to 0.15Hz, and the respiratory wave frequency band is 0.15Hz to 0.4Hz.

4. The method for assessing autonomic nervous function based on suppressing local interference using photoplethysmography (PPG) waves as described in claim 1, characterized in that, In the feature extraction step, extracting the second parameter includes at least one of the following methods: Method A: Based on time-domain amplitude statistics, calculate the statistics of the pulse wave amplitude sequence of the PPG signal in the time domain, and use it as the second parameter; Method B: Based on the frequency domain heart rate band power, the power value in the frequency band where the heart rate main frequency is located is calculated by integration in the power spectral density spectrum of the PPG signal, and used as the second parameter; Method C: Based on gain estimation of the frame-by-frame envelope energy, the amplitude envelope signal is constructed by extracting the envelope of the photoplethysmography pulse wave signal. The squared amplitude of the envelope signal is statistically calculated within a preset time window, and the obtained statistical value is defined as the envelope energy parameter, which is then used as the second parameter.

5. The method for assessing autonomic nervous function based on suppressing local interference using photoplethysmography (PPG) waves as described in claim 4, characterized in that... Method A specifically includes: Identify the systolic peak and diastolic trough of the PPG signal; Calculate the pulse wave amplitude for each cardiac cycle to form an amplitude sequence; The mean square value or functional form of the amplitude sequence is calculated as the second parameter; wherein the functional form includes the perfusion index.

6. The method for assessing autonomic nervous function based on suppressing local interference using photoplethysmography (PPG) waves as described in claim 1, characterized in that, In the association generation step, the association operation includes at least one of the following methods: Ratio calculation: Calculate the ratio of the first parameter to the second parameter to generate a dimensionless index; Functional relationship modeling: Construct a functional relationship model with the second parameter as the independent variable and the first parameter as the dependent variable, and use the model output or residual as the evaluation index; Amplitude standardization: Before extracting the first parameter, the PPG signal is subjected to amplitude standardization using the second parameter or its functional form.

7. The method for assessing autonomic nervous function based on suppressing local interference using photoplethysmography (PPG) waves as described in claim 1, characterized in that, The association generation step further includes numerical mapping, nonlinear transformation, or scaling of the evaluation index to improve its statistical distribution characteristics.

8. The method for assessing autonomic nervous function based on suppressing local interference using photoplethysmography (PPG) waves as described in claim 1, characterized in that, It also includes a dynamic evaluation step: The PPG signal is divided into multiple continuous sub-segments; For each sub-segment, the signal acquisition step, feature extraction step, and association generation step are performed to generate an evaluation index sequence for the sub-segment; Based on the evaluation index sequence, statistical or dynamic characteristics are calculated to quantify the short-term dynamic characteristics of autonomic nervous system regulation. The statistical or dynamic characteristics include at least one of standard deviation, coefficient of variation, adjacent fluctuation intensity, trend of change, or dynamic pattern.

9. An autonomic nervous system function assessment system based on photoplethysmography (PPG) to suppress local interference, characterized in that, include: The signal acquisition module is configured to acquire the photoplethysmography (PPG) signal S(t) of the target object. A processing module, connected to the signal acquisition module, is configured to process the photoplethysmography (PPG) signal S(t), wherein the processing module includes: The feature extraction unit is configured to extract a first parameter and a second parameter from the signal S(t), wherein the first parameter corresponds to the intensity characterization of the vascular tension oscillation component N(t) driven by the central nervous system in the signal S(t), and the second parameter corresponds to the estimated value of the time-varying gain factor G(t) affected by local physiological and physical factors. A signal correction unit is configured to receive the first parameter and the second parameter, and to perform correlation operations between the first parameter and the second parameter to suppress the local interference represented by the gain factor G(t), thereby correcting and generating a stable autonomic nerve function assessment index. The assessment index is a quantitative representation of the vascular tension oscillation component N(t), and its value has excluded the interference of the local gain factor G(t). An output module, connected to the processing module, is configured to output the stabilized autonomic nervous function assessment indicators.

10. The autonomic nervous function assessment system based on photoplethysmography (PPG) to suppress local interference as described in claim 9, characterized in that, The signal acquisition module includes a photoelectric sensor and a signal conditioning circuit, and the processing module includes a processor and a memory. The memory stores program instructions, and when the program instructions are executed by the processor, the method steps are implemented.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the autonomic nervous function assessment method based on photoplethysmography suppression of local interference as described in any one of claims 1 to 8.

12. An electronic device, comprising: The memory, the processor, and the computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the autonomic nervous function assessment method based on photoplethysmography suppression of local interference as described in any one of claims 1 to 8.

Images