A method for extracting metabolic adaptability based on waveform reconstruction

Abstract

The present invention provides a method for extracting metabolic adaptability based on waveform reconstruction. By means of waveform reconstruction, considering the overall change trend of the waveform, geometric figures with the highest similarity to the original waveform are cyclically searched for, so as to simplify and reduce noise by converting waveform features into graphic features, and the metabolic adaptability is extracted based on the extracted simplified figures. The measurement of metabolic adaptability is based on the respiratory quotient RQ extracted by indirect calorimetry (IC). Indirect calorimetry calculates the respiratory quotient RQ by collecting oxygen consumption VO2 and carbon dioxide VCO2. Starting from the overall waveform change trend, the present invention reconstructs the waveform into simple geometric figures, filters out noise while retaining the overall change trend, and has the advantages of good repeatability and strong anti-interference ability.

Description

Technical Field

[0001] The present invention relates to the field of metabolic feature detection, and particularly to a method for extracting metabolic adaptability based on waveform reconstruction. Background Art

[0002] The International Diabetes Federation's survey report (2019) shows that the prevalence of adult hypertension is 25.2%, and the number of diabetes patients reaches 116 million. Both hypertension and diabetes belong to the category of chronic diseases, and the base numbers of chronic disease patients and potential patients in China are huge. According to the Report on Nutrition and Chronic Diseases of Chinese Residents (2020), as the survival period of chronic disease patients continues to extend, coupled with the accelerating process of population aging, urbanization, industrialization and the impact of the prevalence of behavioral risk factors on the incidence of chronic diseases, the base number of chronic disease patients in China will continue to expand. The proportion of deaths due to chronic diseases will also continue to increase. The number of deaths due to cardiovascular and cerebrovascular diseases, cancer, and chronic respiratory diseases accounts for 80.7% of the total deaths, and the situation remains severe. In response to this situation, the National Health Commission has pointed out that it is necessary to "explore new models of chronic disease management", "promote early detection of chronic diseases and reduce the incidence risk of high-risk groups", and form a "three-level prevention" model for chronic diseases, and has pointed out that the key to the three-level prevention is "early".

[0003] As a concept proposed in the 1990s, the significance of applying the indicators of metabolic adaptability in clinical practice has been gradually discovered in recent years. Metabolic adaptability represents the ability of the human body to switch and utilize different metabolic substrates (mainly sugar and fat) under different states. For example, a healthy normal person mainly metabolizes sugar after a meal to quickly and stably maintain blood sugar levels, and mainly metabolizes fat on an empty stomach to prevent excessive fat accumulation. With the intensity of exercise varying, the body switches between sugar metabolism and fat metabolism. Metabolic adaptability represents the ability of a person to respond to external changes and changes in their own state, quickly and stably supply and demand energy, and reflects the "stability" of human metabolic health. The occurrence and development of most chronic diseases stem from metabolic imbalance. Therefore, in chronic diseases, when the human body transforms from a healthy state to a diseased state, metabolic adaptability often changes prior to the onset of the disease, which has important value for the "early warning and early intervention" of chronic diseases.

[0004] The method for extracting metabolic adaptability is to measure the respiratory quotient RQ through indirect calorimetry. Indirect calorimetry (IC) is a non-invasive and non-invasive method for calculating human metabolic characteristics by collecting respiratory gases based on a fixed ratio formula of energy consumption. RQ is the ratio of carbon dioxide production VCO2 to oxygen consumption VO2, which reflects the current ratio of sugar metabolism and lipid metabolism. The ratio of sugar and fat energy supply used by the human body will adaptively change under different states. Therefore, the difference ΔRQ between the RQ values in the basal state and the insulin response state and the time required for the switch, that is, the speed and amplitude of the sugar-lipid metabolism ratio switch, jointly reflect the strength of metabolic adaptability.

[0005] The traditional method for extracting ΔRQ is to obtain the RQ in the basal state and the RQ in the insulin response state separately, and then calculate the difference to get ΔRQ. The RQ in the basal state is equal to the average value of RQ over 5 minutes, and the RQ in the insulin response state is equal to the average value of the 1 minute before and after the RQ reaches its maximum value. However, due to the small theoretical value range of ΔRQ (0.1 - 0.3) and the volatility of RQ in actual measurement (the fluctuation is ±0.05), when the average window is too large, the change characteristics will be lost and the most discriminative ΔRQ cannot be obtained. When the average window is too small, the stability is poor. Moreover, factors affecting metabolic adaptability include but are not limited to metabolic non-periodic oscillations caused by the body's chemical feedback regulation mechanism (the oscillation time is between 30 seconds and 5 minutes), the ventilation ability of the subject, the psychological state of the subject, the sleep state of the subject the day before the test, etc., which will cause large fluctuations in the measured values of the same sample each time. Therefore, the traditional value-taking method will result in poor repeatability of the extracted ΔRQ, and the switching adjustment time will also change due to the fluctuation of the value-taking position. The measurement results are difficult to be well applied to clinical and precision health assessment. Summary of the Invention

[0006] To overcome the bottleneck of the prior art, the present invention proposes a method for extracting metabolic adaptability based on waveform reconstruction, aiming to solve the problems of poor measurement repeatability and weak anti-interference ability of the current method. Starting from the overall waveform change trend, the waveform is reconstructed into a simple geometric figure, filtering out noise while retaining the overall change trend. It has the advantages of good repeatability and strong anti-interference ability.

[0007] To achieve the above object, the present invention proposes a method for extracting metabolic adaptability based on waveform reconstruction. By means of waveform reconstruction, considering the overall change trend of the metabolic waveform before and after stimulation, the geometric figure with the highest similarity to the original waveform is cyclically searched for, so as to simplify and reduce noise by converting the waveform features into graphic features, and extract the metabolic adaptability based on the extracted simplified figure; the measurement of the metabolic adaptability is based on the respiratory quotient RQ extracted by indirect calorimetry, and the indirect calorimetry calculates the respiratory quotient RQ by collecting the oxygen consumption VO2 and the carbon dioxide volume VCO2.

[0008] Furthermore, the waveform reconstruction method starts from the change of the overall waveform before and after stimulation, determines a right-angled quadrilateral BPDC with the value-taking point B of the respiratory quotient RQ in the basic state, the peak point P of the respiratory quotient RQ in the insulin response state after stimulation, the descending inflection point D where the fluctuation gradually tends to decrease, and the intersection point C of the vertical line of the descending inflection point D and the horizontal line of the value-taking point B. And through the way of cyclic optimization, change the positions of the peak point P and the descending inflection point D to make the quadrilateral after cyclic optimization have the highest similarity to the original waveform, and calculate the change amplitude ΔRQ of the respiratory quotient of the metabolic adaptability based on the right-angled quadrilateral to measure the time required for the change.

[0009] Further, the way of cyclic optimization includes: setting the time at the initial position of the value point B of the respiratory quotient RQ in the basic state before the cycle to be 1.5 minutes before the stimulation, and the value of the respiratory quotient RQ to be the average value of the respiratory quotient RQ within 1 minute; the initial position of the peak point P after the stimulation is the first peak point after being stimulated, and the peak point P is found by the change of the slope. The slope before the peak point P is positive, and the slope after the peak point P is negative. To prevent misjudgment of the peak point, the peak point is defined as the point with the largest respiratory quotient RQ within 2 minutes centered on itself; the initial position of the descending inflection point D is the point where the respiratory quotient RQ drops rapidly, that is, (the respiratory quotient RQ of the value point - the respiratory quotient RQ of the next value point) / the respiratory quotient RQ of the next value point > 10%; the angle PBC reflects the fluctuation amplitude of the respiratory quotient RQ, and the length of the side BC reflects the required stabilization time. There are two nested loops in the cycle. The first loop as the outer loop is to fix the position of the value point B in the basic state, and each time the cycle reduces the angle of the angle PBC by 1 degree. When the angle PBC drops to 0 degree, the first loop ends, and the position of the peak point P is the intersection of the PB line and the original waveform; the second loop as the inner loop shortens the length of the side BC to shorten the time, each time shortening by 3 minutes. When the abscissa of the horizontal intersection point C coincides with the abscissa of the peak point P, the second loop ends; each time the first and second loops calculate the graphic similarity between the obtained right-angled quadrilateral and the original waveform, and finally select the right-angled quadrilateral with the highest correlation with the original waveform.

[0010] Further, the method for calculating the graphic similarity includes: through the positions of the value point B, the peak point P, and the descending inflection point D, calculate the linear functions of the sides BP and PD, y = k1x + b1, y = x2x + b2, (x represents time, y represents the RQ measurement value corresponding to the time, k1 and k2 are the slopes of the BP and PD sides calculated, and b1 and b2 are the intercepts of the BP and PD sides calculated), and obtain the evaluation value x of the respiratory quotient RQ of each sampling point through the linear function e sequence, and the measured value x of the respiratory quotient of the sampling point a are compared to obtain the mean square error where n represents the number of sampling points. The smaller the mean square error, the higher the graphic similarity.

[0011] Further, the change amplitude ΔRQ of the respiratory quotient for calculating metabolic adaptability based on the right-angled quadrilateral includes: for the right-angled quadrilateral BPDC, the value of the side BP * cos BPC, which is the height of the right-angled quadrilateral BPDC perpendicular to the side BC, represents the peak value of the change amplitude ΔRQ, and the side DC represents the regulation rate.

[0012] Further, when the graphic similarity between the extracted optimal simplified quadrilateral and the original waveform is less than 80%, it indicates that the waveform itself changes irregularly and has a large fluctuation amplitude, then it is necessary to verify whether the experimental process is standardized and whether the collected waveform is reliable and effective.

[0013] The present invention has the following advantages and effects:

[0014] (1) Based on the indirect calorimetry method, the present invention uses waveform reconstruction to simplify the waveform, filter out noise and retain the change trend. Since metabolic adaptability is affected by many factors and has a large variability, it is difficult to distinguish whether the source of the fluctuation is the variability of RQ itself or the interference caused by noise, which makes it difficult to filter out noise. From the perspective of simplifying the graphics, this method eliminates a lot of filtering work and noise interference.

[0015] (2) The extraction method based on waveform reconstruction of the present invention is different from the traditional method. It starts from the overall perspective and extracts metabolic adaptability based on geometric features. It has the advantages of good repeatability and strong anti-interference ability.

[0016] (3) The method of calculating graphic similarity based on Euclidean distance can be used to evaluate the quality of the experiment. When it is difficult to extract a simplified graphic with high similarity based on the waveform, it means that the waveform itself changes irregularly and the fluctuation range is large, which does not conform to the expected change law of metabolic adaptability. It is necessary to verify whether the experimental process is standardized and whether the collected waveform is reliable and valid. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic diagram of a metabolic adaptability extraction method based on waveform reconstruction of the present invention;

[0018] Figure 2 The present invention is a flowchart of a metabolic adaptability extraction method based on waveform reconstruction. DETAILED DESCRIPTION

[0019] In order to make the above-mentioned features, advantages and implementation schemes of the present invention more obvious and easy to understand, the implementation steps of the present invention are further described below in conjunction with the accompanying drawings and specific implementation schemes.

[0020] The present invention considers the overall change trend of the waveform through waveform reconstruction, cyclically searches for the geometric figure with the highest similarity to the original waveform, simplifies and reduces noise by converting waveform features into graphic features, and extracts metabolic adaptability based on the extracted simplified graphics; the measurement of the metabolic adaptability is based on the respiratory quotient RQ extracted by indirect calorimetry, and the indirect calorimetry calculates the respiratory quotient RQ by collecting oxygen consumption VO2 and carbon dioxide amount VCO2; RQ=VCO2 / VO2.

[0021] like Figure 1As shown in the figure, the metabolic adaptability extraction method based on waveform reconstruction of the present invention includes: collecting the oxygen consumption VO2 and carbon dioxide production VCO2 during each breath, calculating the respiratory quotient RQ through a fixed ratio formula, and obtaining the "RQ - time" graph under different states. Metabolic adaptability is measured by the change between the baseline state of the human body before stimulation and the insulin response state after stimulation. The greater the change amplitude and the faster the change speed, the higher the metabolic adaptability, indicating that the human body has a stronger ability to switch between different energy sources. The change amplitude ΔRQ and the change speed together reflect this switching ability. In the extraction method based on waveform reconstruction, the waveform is simplified to a right - angled quadrilateral, that is, a right - angled quadrilateral BPDC is determined by the value - taking point B1 of the respiratory quotient RQ in the basic state, the peak point P2 of the respiratory quotient RQ in the insulin response state after stimulation, the inflection point D3 where the fluctuation gradually tends to decline, and the intersection point C4 of the vertical line of the inflection point D3 and the horizontal line of the value - taking point B1. The angle PBC5 represents the fluctuation amplitude, and is used to calculate the side BP * cos BPC, that is, the height of the right - angled quadrilateral BPDC perpendicular to the side BC. The greater the height of the right - angled quadrilateral BPDC, the greater the fluctuation amplitude; the length of the side BC represents the time required for the change, and the shorter the side BC, the faster the change speed of metabolic adaptability.

[0022] As Figure 2 shown, the metabolic adaptability extraction method based on waveform reconstruction of the present invention specifically includes the following implementation steps:

[0023] Step 1: Collect the oxygen consumption VO2 and carbon dioxide production VCO2 once during each breath, calculate the respiratory quotient RQ of the human body through a fixed ratio formula (RQ = VCO2 / VO2), draw the "RQ - time" curve, mark the time when the stimulation occurs, and divide different states. The different states include the basic state before stimulation and the insulin response state after stimulation.

[0024] Step 2: Based on the method of waveform reconstruction, determine the initial positions of the four points of the graph. The time of the initial position of the value - taking point B1 of the respiratory quotient RQ in the basic state before the cycle is one and a half minutes before stimulation, and the respiratory quotient RQ of the value - taking point B1 is the average value of the respiratory quotient RQ within one minute; the initial position of the peak point P2 of the respiratory quotient RQ in the insulin response state is the first peak point after being stimulated. The peak point is found through the change in slope. The slope is positive before the peak point and negative after the peak point. To prevent misjudging the peak point, the peak point is defined as the point with the largest respiratory quotient RQ within 2 minutes centered on itself; the initial position of the inflection point D3 where the waveform drops after stimulation is the point where the respiratory quotient RQ drops rapidly, that is, (the respiratory quotient RQ of the value - taking point - the respiratory quotient RQ of the next value - taking point) / the respiratory quotient RQ of the next value - taking point > 10%; the intersection point C4 of the horizontal line is the intersection point of the vertical line of the inflection point D3 and the horizontal line of the value - taking point B1.

[0025] Step 3: Change the angle of ∠PBC in a loop-optimized manner - which represents the change amplitude, and the length of BC - which represents the time required for the change, to find the approximate right-angled quadrilateral with the highest similarity to the original metabolic waveform. The two nested loops are as follows. The first loop (outer loop) is used to fix the position of point B, the value-taking point of the basic state. Each time the loop shrinks the angle of ∠PBC by 1 degree. When the angle of ∠PBC drops to 0 degrees, the loop ends. The position of the peak point P is the intersection of the PB line and the original metabolic waveform. When there are multiple intersections, both the abscissa and ordinate of the peak point P are the averages of the multiple intersections. The second loop (inner loop) shortens the length of the line segment BC - the length represents time, and each time it is shortened by 3 minutes. When the abscissa of the horizontal line intersection point C coincides with the abscissa of the peak point P, the inner loop ends. The inner loop is nested within the outer loop. Each time the outer loop changes the angle, the inner loop loops once until both loops are completed. The graphical similarity between the obtained right-angled quadrilateral and the original metabolic waveform is calculated once for each inner and outer loop. Finally, the right-angled quadrilateral with the highest correlation with the original waveform is selected.

[0026] Based on the positions of points B, P, and D, calculate the linear functions of lines BP and PD, y = k1x + b1 and y = k2x + b2, where x represents time and y represents the RQ measurement value corresponding to the time. k1 and k2 are the slopes of sides BP and PD calculated, and b1 and b2 are the intercepts of sides BP and PD calculated. Obtain the evaluation value of the respiratory quotient RQ at each sampling point i through the linear function ei and the measured value at the corresponding sampling point i ai are compared to obtain the mean square error where n represents the number of sampling points. The smaller the mean square error, the higher the graphical similarity.

[0027] Step 4: Determine the change amplitude ΔRQ through BP * cos BPC, that is, the height of the right-angled quadrilateral BPDC perpendicular to side BC. The higher the height, the greater the change amplitude. Determine the time required for the change through the length of BC. The greater the change amplitude, the shorter the change time, and the better the metabolic adaptability.

[0028] Step 5: Record and output the highest similarity between the graph and the original metabolic waveform. When the graphical similarity between the extracted optimal simplified graph and the original metabolic waveform is less than 80%, it indicates that the waveform itself changes irregularly and has a large fluctuation amplitude. Then, it is necessary to verify whether the experimental process is standardized and whether the collected waveform is reliable and effective. Based on this, the reliability of the experimental data can be judged.

[0029] It is easy for those skilled in the art to understand that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for extracting metabolic adaptability based on waveform reconstruction, characterized in that: By means of waveform reconstruction, considering the overall change trend of the metabolic waveform before and after stimulation, circularly search for the geometric figure with the highest similarity to the original metabolic waveform, simplify and reduce noise by converting the waveform features into graphic features, and extract metabolic adaptability based on the extracted simplified figure; the measurement of the metabolic adaptability is based on the respiratory quotient RQ extracted by indirect calorimetry. Indirect calorimetry calculates the respiratory quotient RQ by collecting the oxygen consumption VO2 and the carbon dioxide amount VCO2, where RQ = VCO2 / VO2; The described waveform reconstruction method starts from the overall change of the metabolic waveform before and after stimulation. By the value point B of the respiratory quotient RQ in the basic state, the peak point P of the respiratory quotient RQ in the insulin response state after stimulation, the descending inflection point D where the fluctuation gradually tends to decrease, and the intersection point C of the vertical line of the descending inflection point D and the horizontal line of the value point B, a right-angled quadrilateral BPDC is determined. The right-angled quadrilateral BPDC reflects the change trend and characteristics of the curve after the metabolism is stimulated. The length of the side BP reflects the amplitude of the change, and the side DC reflects the length of time required for the change. And through the way of circular optimization, change the positions of the peak point P and the descending inflection point D to make the quadrilateral after circular optimization have the highest similarity to the original waveform, and calculate the change amplitude ΔRQ of the respiratory quotient of the metabolic adaptability based on the right-angled quadrilateral to measure the time required for the change; The described way of circular optimization includes: setting the initial position time of the value point B of the respiratory quotient RQ in the basic state before the cycle to 1.5 minutes before stimulation, and the value of the respiratory quotient RQ is the average value of the respiratory quotient RQ within 1 minute; the initial position of the peak point P after stimulation is the first peak point after being stimulated, and the peak point P is found through the slope change. The slope before the peak point P is positive, and the slope after the peak point P is negative. To prevent misjudging the peak point, the peak point is defined as the point with the largest respiratory quotient RQ within 2 minutes centered on itself; the initial position of the descending inflection point D is the point where the respiratory quotient RQ drops rapidly, that is, (the value point respiratory quotient RQ - the next value point RQ) / the next value point RQ > 10%; the angle PBC reflects the fluctuation amplitude of the respiratory quotient RQ, and the length of the side BC reflects the time required for stability. There are two nested loops during the cycle. The first loop as the outer loop is to fix the position of the value point B in the basic state. Each time the cycle reduces the angle of the angle PBC by 1 degree. When the angle PBC drops to 0 degree, the first loop ends, and the position of the peak point P is the intersection point of the PB line and the original metabolic waveform; the second loop as the inner loop shortens the length of the side BC to shorten the time. Each time it shortens by 3 minutes. When the horizontal intersection point C coincides with the abscissa of the peak point P, the second loop ends; calculate the graphic similarity between the obtained right-angled quadrilateral and the original waveform every time the first and second loops are performed, and finally select the right-angled quadrilateral with the highest correlation with the original waveform; The methods for calculating the similarity of graphs include: calculating the linear functions of sides BP and PD by the positions of the value-taking point B, the peak point P, and the descending inflection point D. where x represents time, y represents the RQ measurement value corresponding to the time, k1 and k2 are the slopes of sides BP and PD obtained by calculation, b1 and b2 are the intercepts of sides BP and PD obtained by calculation, and the evaluation value of the respiratory quotient RQ at each sampling point is obtained through the linear function. sequence, and the measured value of the respiratory quotient at the sampling point are compared to obtain the mean square error where n represents the number of sampling points. The smaller the mean square error, the higher the graph similarity. The change amplitude ΔRQ of the respiratory quotient for calculating metabolic adaptability based on the right-angled quadrilateral includes: for the right-angled quadrilateral BPDC, the value of the side BP The numerical value of cos BPC, which is the height of the right-angled quadrilateral BPDC perpendicular to the side BC, represents the peak value of the change amplitude ΔRQ, and the side DC represents the regulation rate.

2. The metabolic adaptability extraction method based on waveform reconstruction according to claim 1, characterized in that: When the graphic similarity between the extracted optimal simplified quadrilateral and the original waveform is less than 80%, it indicates that the waveform itself changes irregularly and has a large fluctuation amplitude. Then verify whether the experimental process is standardized and whether the collected waveform is reliable and effective.

Images