Brain oxygen signal processing method based on complete ensemble empirical mode decomposition

Abstract

The brain oxygen signal processing method based on complete ensemble empirical mode decomposition performs N-order complete ensemble empirical mode decomposition on the spatial differential optical density signal in each optical band and outputs each modal component; the permutation entropy is calculated for each modal component; the modal component with permutation entropy less than the permutation entropy threshold is used to reconstruct the signal to obtain the spatial differential optical density signal ΔOD λ Reconstructed signal; for each complete set of empirical mode decomposition, dynamically update the permutation entropy threshold; calculate the local absorption of oxygenated hemoglobin and the local absorption of reduced hemoglobin ΔC based on the reconstructed signal Hb N screens out components with PE less than a set threshold for signal reconstruction. The signal quality after reconstruction is significantly improved, the signal-to-noise ratio is significantly improved, and the calculation of brain tissue oxygen saturation based on the reconstructed signal is more accurate.

Description

Technical Field

[0001] The present invention relates to the field of bioelectronic technology, and in particular to a method and device for processing brain oxygen saturation signals based on complete set empirical mode decomposition. Background Art

[0002] Brain tissue oxygen detection based on near-infrared spectroscopy is a non-invasive detection technology that is of great significance for observing brain oxygen changes in patients during the perioperative period and is also widely used in brain function research.

[0003] In the prior art, due to the large amount of red light and infrared light in the environment, it is difficult to avoid environmental interference in oxygen saturation measurement, and it is also difficult to measure the concentration of tissue oxygenated hemoglobin and reduced hemoglobin. Even if it can be measured, it is difficult to ensure the accuracy of the measurement. At the same time, in the prior art, when filtering the photoelectric signal based on empirical mode decomposition combined with permutation entropy, a fixed threshold is selected, resulting in large limitations in signal reconstruction. When a fixed permutation entropy threshold is used to reconstruct the signal, the signal will change with the body position and physiological state of the subject, and the fixed threshold cannot adapt to this change in time; therefore, this patent proposes an adaptive signal reconstruction method based on complete set empirical mode decomposition, which realizes adaptive signal reconstruction through dynamic permutation entropy threshold, which can effectively reduce the impact of these changes on the monitoring signal, improve the signal-to-noise ratio of the signal and the universality of the calculation.

[0004] Glossary:

[0005] CEEMDAN, the full name of which is "Complete Ensemble Empirical Decomposition with Adaptive Noise", means: Complete Ensemble Empirical Mode Decomposition

[0006] EMD, the full name of which is "Complete Ensemble Empirical Decomposition with Adaptive Noise", means: Empirical Mode Decomposition

[0007] PE, the full name of English is "Permutation Entropy", the Chinese meaning is permutation entropy

[0008] SNR, the full name of which is "Signal to Interference plus Noise Ratio", means signal to noise ratio in Chinese.

[0009] IMF, the full English name is "Intrinsic Mode Function", and its Chinese meaning is inherent mode function, also known as eigenmode function. Summary of the Invention

[0010] The technical problem to be solved by the technical solution in this application is to propose an adaptive threshold method, which combines the CEEMDAN, PE, and SNR methods. In the brain tissue oxygen saturation signal processing calculation, the signal can be reconstructed according to time changes, and the threshold changes adaptively, which solves the problem of insufficient calculation accuracy and low signal-to-noise ratio caused by the use of fixed thresholds in the existing technology, and calculates blood oxygen saturation based on the measured concentrations of tissue oxygenated hemoglobin and reduced hemoglobin.

[0011] The technical solution of the present application to solve the above-mentioned problem is a brain oxygen signal processing method based on complete set empirical mode decomposition, which includes the following steps: step S1: using two receivers to respectively obtain the photoelectric signals of at least W light bands after the light source passes through the brain tissue; W is a natural number greater than or equal to 2; step S2: based on the two photoelectric signals B1 and B2 in the same light band obtained by the two receivers, calculate the spatial differential optical density signal in each light band in is the light intensity at receiver 1 at wavelength λ, is the light intensity of the receiver 2 at wavelength λ; Step S3: spatial differential optical density signal ΔOD over a period of time λ Perform N-order complete set empirical mode decomposition and output each modal component; Step S4: calculate the permutation entropy of each modal component output in Step S3; set the permutation entropy threshold, and reconstruct the modal components with permutation entropy less than the permutation entropy threshold to obtain the spatial differential optical density signal ΔOD λ The reconstructed signal A1; dynamically update the permutation entropy threshold according to the current permutation entropy threshold, which is used as the basis for selecting the modal component in the next period of time; S5: the spatial differential optical density signal ΔOD obtained in step S4 λ The reconstructed signal A1 is used to calculate the local absorption of oxygenated hemoglobin according to the following formulas: and the local absorption of reduced hemoglobin ΔC hb .

[0012] After step S5, the process further includes step S6: calculating the brain tissue oxygen saturation under the two-band combination according to the differential absorption of oxygenated hemoglobin and reduced hemoglobin according to the following formula:

[0013] In step S4, permutation entropy calculation is performed on each modal component output in step S3, using a dynamic permutation entropy threshold; first, an initial permutation entropy threshold is set, and the modal component whose permutation entropy is less than the initial permutation entropy threshold is used to perform signal reconstruction to obtain a reconstructed signal A1 of the spatial differential optical density signal; the permutation entropy threshold updating rule is as follows: after the permutation entropy calculation is performed on each modal component, the Jth modal component closest to the initial permutation entropy threshold is taken as the center, and its two adjacent modal components, i.e., the J-1th modal component and the J+1th modal component, are selected, and the signal-to-noise ratios (SNRs) corresponding to the Jth modal component, the J-1th modal component, and the J+1th modal component are calculated respectively, and the permutation entropy value corresponding to the modal component with the smallest signal-to-noise ratio (SNR) is used as the next permutation entropy threshold.

[0014] In step S4: after calculating the permutation entropy of each modal component output in step S3, the permutation entropy value is normalized, and the normalized permutation entropy value is used as the permutation entropy.

[0015] The initial threshold of permutation entropy was selected in the range of 0.5-0.7 including 0.6.

[0016] In step S2, based on the two photoelectric signals B1 and B2 in the same optical band obtained by the two receivers, conventional filtering and denoising are performed on the photoelectric signals B1 and B2 respectively, and the spatial differential optical density signal ΔOD in each optical band is calculated using the filtered and denoised photoelectric signals B1 and B2. λ In step S5, the local absorption of oxygenated hemoglobin and the local absorption of reduced hemoglobin ΔC hb The calculation formula is as follows:

[0017]

[0018]

[0019] in Δr is the center distance between receiver 1 and receiver 2,

[0020] The brain oxygen signal processing method based on complete set empirical mode decomposition includes the steps of selecting a wavelength combination, where W is 2, i.e., two bands of infrared light, and calculating the spatial differential optical density signal ΔOD under the two light bands. λ For subsequent calculations;

[0021] For each spatial differential optical density signal ΔOD λ Perform steps S3 and S4 to reconstruct the signal, and select any reconstructed spatial differential optical density signal ΔOD λ Go to step S5.

[0022] The brain oxygen signal processing method based on complete set empirical mode decomposition includes the step of selecting a wavelength combination, where W is 3, i.e., three bands of infrared light, and the wavelength combinations include (1, 2), (1, 3), and (2, 3); any multiple wavelength combinations can be selected to perform spatial differential optical density signal ΔOD λ Calculate the spatial differential optical density signal ΔOD λ Perform steps S3 and S4 to reconstruct the signal, and select any reconstructed spatial differential optical density signal ΔOD λ Go to step S5.

[0023] After step S6, step S7 is also included: according to the formula The brain oxygen saturation is calculated, where A, B, C, and D are the correction coefficients and correction values ​​corresponding to brain tissue oxygen, respectively.

[0024] The technical solution of the present application to solve the above-mentioned problem can also be a readable storage medium having a computer program stored thereon, characterized in that when the program is executed by a processor, the brain oxygen signal processing method based on complete set empirical mode decomposition as described in any one of claims 1 to 9 is implemented.

[0025] Compared with the prior art, one of the beneficial effects of the present invention is to improve the spatial differential optical density signal ΔOD λ Performing N-order complete set empirical mode decomposition, the spatial differential optical density signal itself is the ratio of the photoelectric signals output by two detectors for the same light source through different tissue paths. Since some noise introduced by the light source and detector is eliminated, the signal consistency is better.

[0026] Compared with the prior art, one of the beneficial effects of the present invention is that the N-order complete set empirical mode decomposition is more suitable for brain oxygen signal processing scenarios. Pulse blood oxygen calculation is based on at least one light absorption signal under each pulse wave fluctuation state. The brain oxygen signal is a light absorption signal for brain tissue and does not have an obvious pulse rhythm. Therefore, the signal tends to be low-frequency, more difficult to extract, and more easily submerged by various noises. The complete set empirical mode decomposition can play a role in this scenario and is particularly suitable for extracting signal processing scenarios where such signal characteristics are not particularly obvious. After the N-order complete set empirical mode decomposition, the components with PE less than the set threshold are screened out for signal reconstruction. The signal quality after reconstruction is greatly improved, the SNR is significantly improved, and the calculation of brain tissue oxygen saturation based on the reconstructed signal is also more accurate.

[0027] Compared with the prior art, one of the beneficial effects of the present invention is that the dynamic permutation entropy threshold can update the permutation entropy threshold of signal reconstruction in real time, making the signal reconstruction process more adaptable.

[0028] Compared with the prior art, one of the beneficial effects of the present invention is that the dynamic permutation entropy threshold uses the two adjacent modal components of the Jth modal component closest to the initial permutation entropy threshold, and screens them according to the signal-to-noise ratio, and uses the permutation entropy value corresponding to the modal component with the smallest signal-to-noise ratio SNR as the permutation entropy threshold for the next signal reconstruction, so that the permutation entropy threshold is updated in the direction of smaller and smaller signal-to-noise ratio.

[0029] Compared with the prior art, one of the beneficial effects of the present invention is that the spatial differential optical density signal ΔOD after reconstruction is λ , calculate the local absorption of oxygenated hemoglobin and the local absorption of reduced hemoglobin ΔC hb , can obtain a variety of brain tissue oxygen saturation related parameters, which is convenient for subsequent calculation of more parameters.

[0030] Compared to existing technologies, one of the benefits of the present invention is that it calculates brain tissue oxygen saturation (rSO2) based on the differential absorption of oxygenated and reduced hemoglobin, using a combination of two wavelengths. This allows calculation of brain tissue oxygen saturation using a variety of different tissues, and further allows for multiple combinations or averaging to obtain more realistic brain tissue oxygen saturation data. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a schematic diagram of the positions of the light source and receiver of the brain tissue oxygen saturation test device;

[0032] Figure 2 It is the absorption spectrum of oxygenated hemoglobin HbO2, reduced hemoglobin Hb and water in human tissue;

[0033] Figure 3 This is a schematic diagram of the overall steps of the brain oxygen signal processing method;

[0034] Figure 4 This is a schematic diagram of one of the steps of the brain oxygen signal processing method based on complete set empirical mode decomposition;

[0035] Figure 5 This is a schematic diagram of one of the steps of the brain oxygen signal processing method based on complete set empirical mode decomposition;

[0036] Figure 6 This is the second step diagram of the brain oxygen signal processing method based on complete set empirical mode decomposition;

[0037] Figure 7 It is the photoelectric detection signal of three bands obtained on the detector far away from the light source. The signal amplitude is small and the effective information is almost submerged in the white noise.

[0038] Figure 8It is the photoelectric detection signal of three bands obtained on the detector closer to the light source. The signal amplitude is larger than that of the detector at the far end, and the effective information representation is also better. Figure 5 The situation is better;

[0039] Figure 9 is based on Figure 7 and Figure 8 Schematic diagram of the photoelectric detection signals obtained on the two detectors in the three bands, and the calculated spatial differential optical density signals in the three optical bands. DETAILED DESCRIPTION

[0040] The present invention will be further described below in conjunction with the accompanying drawings.

[0041] It should be noted that the following is a description of the preferred embodiments of the present application and does not constitute any limitation to the present application. The description of the preferred embodiments of the present application is merely an illustration of the general principles of the present application. The embodiments described in this application are only a part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.

[0042] In the description of this application, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings and are only for the convenience of describing this application and simplifying the description. They do not indicate or imply that the devices or components referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore should not be understood as limiting this application. In addition, the terms "first" and "second", as well as technical features numbered with Arabic numerals 1, 2, 3, etc., and numbers such as "A" and "B", are used for descriptive purposes only and are for the convenience of explanation only. They do not represent a temporal or spatial order relationship and should not be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, features defined as "first" and "second", as well as features numbered with Arabic numerals 1, 2, 3, etc., may explicitly or implicitly include one or more of such features. In the description of this application, “several” means two or more, unless otherwise clearly and specifically defined.

[0043] like Figure 1Schematic diagram of the light source and receiver positions in a brain tissue oxygen saturation test device. Because brain tissue oxygen saturation measures oxygen saturation in tissue, unlike pulse wave-based arterial oxygen saturation (SPO2), brain tissue oxygen saturation measurement requires at least two photodetectors or photodetectors in different locations to capture the light signals from light sources of different wavelengths absorbed by the tissue. Because the signal must be acquired through a reference method, the combination of multiple wavelengths and multiple detectors is necessary. However, the combination of multiple photodetectors and multiple light sources can result in significant variations in background noise due to device consistency and spatial arrangement issues. Furthermore, the effective signal at the remote photodetector is typically much weaker, posing a significant challenge to signal processing. Tissue oxygen saturation signals themselves are weak, making acquiring effective signals under these measurement conditions extremely challenging.

[0044] like Figure 7 The photoelectric detection signals of the three bands obtained on the detector far away from the light source have small signal amplitudes, and the effective information is almost submerged in the white noise. Figure 8 The photoelectric detection signals of the three bands obtained on the detector closer to the light source have larger signal amplitudes than those of the detector at the far end, and the effective information representation is also better. Figure 7 The situation is better; Figure 9 is based on Figure 7 and Figure 8 Figure 1. The photoelectric detection signals obtained from two detectors in the three optical bands are shown, along with the calculated spatial differential optical density signals for the three optical bands. Although the spatial differential optical density signal is the ratio of the signals obtained on the detectors, and some noise is removed after the ratio is calculated, the influence of noise still exists, and this noise can affect the subsequent measurement of tissue oxygen saturation.

[0045] like Figure 2 , the absorption spectra of oxygenated hemoglobin HbO2, reduced hemoglobin Hb and water in human tissue. Figure 3 Schematic diagram of the overall steps of the brain oxygen signal processing method. The photoelectric signal acquired by the detector is first subjected to a first-level denoising process, which can be any signal processing method known in the art. During the second-level denoising process, the brain oxygen signal processing method based on complete set empirical mode decomposition (CEMD) described in this application is employed. Brain tissue oxygen saturation is then calculated based on the processed signal, resulting in a more accurate and realistic result.

[0046] like Figures 4 to 6 In an embodiment of a brain oxygen signal processing method based on complete ensemble empirical mode decomposition, the method includes the following steps:

[0047] S0: includes the steps of selecting a wavelength combination, W is 2, that is, two bands of infrared light, and calculating the spatial differential optical density signal ΔOD under the two light bands λ Used for subsequent calculations.

[0048] S1: Use two receivers to obtain the photoelectric signals of the two light bands after the light source passes through the brain tissue. S2: Based on the two photoelectric signals B1 and B2 in the same light band obtained by the two receivers, calculate the spatial difference optical density signal in the two light bands:

[0049]

[0050] in is the light intensity at receiver 1 at wavelength λ1, is the light intensity of receiver 2 at wavelength λ1;

[0051]

[0052] in is the light intensity of receiver 1 at wavelength λ2, is the light intensity at receiver 2 at wavelength λ2.

[0053] S3: Spatial differential optical density signal ΔOD λ1 and ΔOD λ2 Perform N-order complete set empirical mode decomposition and output the corresponding modal components.

[0054] The specific process of complete set empirical mode decomposition is as follows:

[0055] Step A1: Select the original brain tissue oxygen signal within a period of time, and add N different Gaussian white noises x to the optical density difference ΔOD(t) in the noisy original brain tissue oxygen signal. i (t), construct a new signal ΔOD i (t)=ΔOD(t)+σ i x i (t), where σ i is the standard deviation of white noise, x i (t) obeys normal distribution. ΔOD(t) is the digital signal, and t is the sampling time.

[0056] Step A2: Using EMD to calculate ΔOD i (t) is decomposed to obtain the first IMF component and the average value is calculated Calculate the first residual component

[0057] The standard deviation of white noise used in this application is 0.2, generally 0-1; the number of times white noise is added is 100, and the range of the number of times white noise is added can be 50-100; EMD is used to calculate ΔODi (t) The number of IMF components to be decomposed is 10. In practical applications, the number of times white noise is added and the number of IMF components can be adjusted according to actual needs.

[0058] Step A3: Calculate the average value of the kth IMF component, as shown in Eq. As shown, where rk(t)=rk-1(t)-IMFk(tt), CEk(·) is the kth IMF component after the signal is decomposed by EMD.

[0059] Repeat step A3 until the remaining components do not meet the EMD decomposition conditions. The EMD decomposition conditions include: Condition 1: The number of local extreme points and zero crossing points in the signal must be equal or differ by at most one over the entire time range; Condition 2: At any time point, the upper envelope of the signal's local maximum (i.e., the signal amplitude greater than zero) and the lower envelope of the local minimum (i.e., the signal amplitude less than zero) average to zero.

[0060] S4: Calculate the permutation entropy for each modal component output in step S3 and perform threshold iteration. The permutation entropy method is mainly used to analyze the complexity of dynamic time domain signals. Suppose there is a set of time series {x(s), s=1,2,3,...,n}, which is reconstructed into a new space to obtain: Where m is the embedding dimension, τ is the time delay, and the value of s is 1≤s≤n-(m-1)τ. Rearrangement will result in K=m! permutations, and P(h) represents the probability of each permutation occurring: Normalization processing: In this patent, the embedding dimension is selected as 3 and the time delay is selected as 1.

[0061] In this application, step S4 is used to iterate the permutation entropy threshold: the initial permutation entropy threshold TH1 is selected as 0.6, and the PE value obtained is the series where λ i is the wavelength of near-infrared light, 1≤j≤n, n is the wavelength λ in S3 i The number of modal components obtained by the complete set of empirical mode decomposition, select The modal components of the signal are reconstructed to obtain a new reconstructed signal. The modal component with the PE value closest to the initial permutation entropy threshold TH1 is selected as the main modal component PEf. The first-order modal components are selected before and after PEf, and the signal-to-noise ratio values ​​of these three modal components are calculated. The PE value corresponding to the modal component with the largest signal-to-noise ratio is selected as the permutation entropy threshold for the next modal decomposition.

[0062] That is, in step S4, the permutation entropy is calculated for each modal component output in step S3, and the permutation entropy threshold is dynamically updated; first, the permutation entropy initial threshold is set, and the modal component whose permutation entropy is less than the permutation entropy initial threshold is used to perform signal reconstruction to obtain the reconstructed signal A1 of the spatial differential optical density signal; the permutation entropy threshold updating rule is as follows: after the permutation entropy is calculated for each modal component, the J-th modal component closest to the initial permutation entropy threshold is taken as the center, and its two adjacent modal components, i.e., the J-1 modal component and the J+1 modal component, are selected, and the signal-to-noise ratios (SNRs) corresponding to the J-th modal component, the J-1 modal component, and the J+1 modal component are calculated respectively, and the permutation entropy value corresponding to the modal component with the smallest signal-to-noise ratio (SNR) is used as the next permutation entropy threshold.

[0063] Set the initial threshold of permutation entropy; the selection range of the initial threshold of permutation entropy is 0.5-0.7 including 0.6. Based on the permutation entropy threshold and the selected modal component, perform signal reconstruction to obtain the spatial differential optical density signal ΔOD λ1 The reconstructed signal A1 of wavelength λ1;

[0064] S5: Reconstruct the signal using the spatial differential optical density signal obtained in step S4 and calculate the local absorption of oxygenated hemoglobin and the local absorption of reduced hemoglobin ΔC Hb Local absorption of oxygenated hemoglobin and the local absorption of reduced hemoglobin ΔC Hb The calculation formula is as follows:

[0065]

[0066] Where x and y are at least one selected wavelength combination;

[0067] After step S5, the process further includes step S6: calculating the brain tissue oxygen saturation under the two-band combination according to the differential absorption of oxygenated hemoglobin and reduced hemoglobin according to the following formula:

[0068] In an embodiment of a brain oxygen signal processing method based on complete set empirical mode decomposition, the method includes the steps of selecting a wavelength combination, where W is 2, i.e., two bands of infrared light, and calculating the spatial difference optical density signal ΔOD under the two light bands. λ For subsequent calculations; for each spatial differential optical density signal ΔOD λ Perform steps S3 and S4 to reconstruct the signal, and select any reconstructed spatial differential optical density signal ΔOD λ Go to step S5.

[0069] In another embodiment of a brain oxygen signal processing method based on complete set empirical mode decomposition, the method includes the step of selecting a wavelength combination, where W is 3, i.e., three-band infrared light, and the wavelength combinations include (1, 2), (1, 3), and (2, 3); any multiple wavelength combinations can be selected to generate a spatial differential optical density signal ΔOD. λ Calculate and convert the spatial differential optical density signal ΔOD λ Used for subsequent calculations.

[0070] S1: Use two receivers to obtain the photoelectric signals of the three light bands after the light source passes through the brain tissue. S2: Based on the two photoelectric signals B1 and B2 in the same light band obtained by the two receivers, calculate the spatial difference optical density signal in the two light bands:

[0071]

[0072] in is the light intensity at receiver 1 at wavelength λ1, is the light intensity of receiver 2 at wavelength λ1;

[0073]

[0074] in is the light intensity of receiver 1 at wavelength λ2, is the light intensity of receiver 2 at wavelength λ2;

[0075]

[0076] in is the light intensity of receiver 1 at wavelength λ3, is the light intensity of receiver 2 at wavelength λ3;

[0077] S3: Spatial differential optical density signal ΔOD λ1 , ΔOD λ2 and ΔOD λ3 Perform N-order complete set empirical mode decomposition and output the corresponding modal components;

[0078] Step S4: Calculate the permutation entropy of each modal component output from step S3; set a permutation entropy threshold, and reconstruct the modal components whose permutation entropy is less than the permutation entropy threshold to obtain the spatial differential optical density signal ΔOD λ The reconstructed signal A1;

[0079] S5: The spatial differential optical density signal ΔOD obtained in step S4 λ The reconstructed signal is used to calculate the local absorption of oxygenated hemoglobin. and the local absorption of reduced hemoglobin ΔCHb Local absorption of oxygenated hemoglobin and the local absorption of reduced hemoglobin ΔC Hb The calculation formula is as follows:

[0080]

[0081] Where x and y are at least one selected wavelength combination;

[0082] After step S65, the process further includes step S6: calculating the brain tissue oxygen saturation under the two-band combination according to the differential absorption of oxygenated hemoglobin and reduced hemoglobin according to the following formula:

[0083] After step S6,

[0084] Also includes step S7: According to the formula The brain oxygen saturation is calculated, where A, B, C, and D are the correction coefficients and correction values ​​corresponding to brain tissue oxygen, respectively.

[0085] In an embodiment of a brain tissue blood oxygen saturation acquisition device not shown in the accompanying drawings, the brain oxygen signal processing method based on the above-mentioned complete set empirical mode decomposition is used.

[0086] In an embodiment of a readable storage medium not shown in the drawings, a computer program is stored thereon, and when the program is executed by a processor, the above-mentioned brain oxygen signal processing method based on complete set empirical mode decomposition is implemented.

[0087] Four segments of signals were cut out from a section of brain oxygen data, each signal segment was 8s, 1000 data points, and the sampling rate was 125Hz. The PE fixed values ​​of 0.7 and 0.5 were set for each segment of data, and the spatial differential optical density signal was reconstructed using the patented method to compare the signal-to-noise ratio of the reconstructed signal. Tables 1 to 4 are the signal-to-noise ratio comparisons of the four segments of data. The three groups of spatial differential optical density signals were processed using conventional signal processing methods, CEEMDAN method with a fixed permutation entropy threshold of 0.7, CEEMDAN method with a fixed permutation entropy threshold of 0.5, and CEEMDAN method based on dynamic permutation entropy threshold. Conventional signal processing methods include conventional signal processing methods based on Fourier transform or wavelet transform, and conventional methods also include conventional filtering methods such as median filtering, 30Hz low-pass filtering, and smoothing filtering. In the conventional methods in Tables 1 to 4, median filtering, 30Hz low-pass filtering, and smoothing filtering were performed in sequence. As shown in Tables 1 through 4, the dynamic threshold CEEMDAN method is well suited for processing brain tissue oxygen saturation signals, significantly reducing the signal-to-noise ratio (SNR) compared to conventional methods. Furthermore, the dynamic permutation entropy threshold CEEMDAN method achieves an even lower SNR than the fixed threshold CEEMDAN method. This improvement in SNR significantly improves the accuracy of subsequent parameter calculations.

[0088] It can be seen from the data in Tables 1 to 4 that the CEEMDAN method with a fixed permutation entropy threshold does not demonstrate the advantages of the CEEMDAN method compared to conventional signal processing methods. However, the CEEMDAN method based on a dynamic permutation entropy threshold in this application can achieve a better signal-to-noise ratio than conventional signal processing methods. The method in this application significantly improves the spatial differential optical density signal value.

[0089] Table 1

[0090] Conventional methods Permutation entropy threshold 0.7 Permutation entropy threshold 0.5 Dynamic Permutation Entropy Threshold ΔOD1 1.2436 1.4743 15.5391 30.5556 ΔOD2 18.0387 18.4527 35.3778 48.0009 ΔOD3 3.1681 3.2210 3.2978 19.7932

[0091] Table 2

[0092] Conventional methods Permutation entropy threshold 0.7 Permutation entropy threshold 0.5 Dynamic Permutation Entropy Threshold ΔOD1 11.7474 11.7556 22.2417 47.8360 ΔOD2 11.6492 12.4138 23.6245 79.2458 ΔOD3 3.2114 3.5403 33.4771 48.7996

[0093] Table 3

[0094] Conventional methods Permutation entropy threshold 0.7 Permutation entropy threshold 0.5 Dynamic Permutation Entropy Threshold ΔOD1 0.4355 0.7543 2.1277 24.0778 ΔOD2 1.5347 1.7819 2.1120 30.3207 ΔOD3 4.8457 19.1508 53.2274 58.3720

[0095] Table 4

[0096] Conventional methods Permutation entropy threshold 0.7 Permutation entropy threshold 0.5 Dynamic Permutation Entropy Threshold ΔOD1 18.9286 24.6177 30.1338 30.1338 ΔOD2 16.7667 17.9927 35.0705 17.9927 ΔOD3 4.8457 5.4097 13.1199 29.9339

[0097] Brain oxygen signal processing method based on complete set empirical mode decomposition is used to analyze the spatial differential optical density signals in each optical band. Perform N-order complete set empirical mode decomposition and output each modal component; calculate the permutation entropy of each modal component; reconstruct the signal to obtain the spatial differential optical density signal ΔOD based on the modal component whose permutation entropy is less than the permutation entropy threshold λ reconstructed signal; dynamically update the permutation entropy threshold for each complete ensemble empirical mode decomposition; calculate the local absorption of oxygenated hemoglobin based on the reconstructed signal and the local absorption of reduced hemoglobin ΔC Hb N screens out components with PE less than a set threshold for signal reconstruction. The signal quality after reconstruction is significantly improved, the signal-to-noise ratio is significantly improved, and the calculation of brain tissue oxygen saturation based on the reconstructed signal is more accurate.

[0098] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A brain oxygen signal processing method based on complete set empirical mode decomposition, characterized in that: The following steps are included: Step S1: using two receivers to respectively acquire photoelectric signals of at least W light bands after light sources pass through brain tissue; W is a natural number greater than or equal to 2; Step S2: Based on the two photoelectric signals B1 and B2 in the same optical band obtained by the two receivers, calculate the spatial differential optical density signal in each optical band in is the light intensity at receiver 1 at wavelength λ, is the light intensity at receiver 2 at wavelength λ; Step S3: Calculate the spatial differential optical density signal ΔOD within a period of time λ Perform N-order complete set empirical mode decomposition and output each modal component; Step S4: Calculate the permutation entropy of each modal component output from step S3; set a permutation entropy threshold, and reconstruct the modal components whose permutation entropy is less than the permutation entropy threshold to obtain the spatial differential optical density signal ΔOD λ The reconstructed signal A1 is obtained by dynamically updating the permutation entropy threshold value according to the current permutation entropy threshold value, and used as the basis for selecting the modal component in the next period of time; The permutation entropy threshold update rule is as follows: After calculating the permutation entropy of each modal component, take the Jth modal component closest to the initial permutation entropy threshold as the center, select its two adjacent modal components, i.e., the J-1th modal component and the J+1th modal component, and calculate the signal-to-noise ratio (SNR) corresponding to the Jth modal component, the J-1th modal component, and the J+1th modal component respectively. The permutation entropy value corresponding to the modal component with the smallest SNR is used as the next permutation entropy threshold. S5: The spatial differential optical density signal ΔOD obtained in step S4 λ The reconstructed signal A1 is used to calculate the local absorption of oxygenated hemoglobin ΔC according to the following formulas: HbO2 and the local absorption of reduced hemoglobin ΔC Hb ; in Δr is the center distance between receiver 1 and receiver 2, 2. The brain oxygen signal processing method based on complete set empirical mode decomposition according to claim 1, characterized in that: After step S5, the process further includes step S6: calculating the brain tissue oxygen saturation under the two-band combination according to the differential absorption of oxygenated hemoglobin and reduced hemoglobin according to the following formula:

3. The brain oxygen signal processing method based on complete set empirical mode decomposition according to claim 1, characterized in that: In step S4, the permutation entropy is calculated for each modal component output in step S3, using a dynamic permutation entropy threshold. An initial permutation entropy threshold is first set, and the modal components whose permutation entropy is less than the initial permutation entropy threshold are used to perform signal reconstruction to obtain a reconstructed signal A1 of the spatial differential optical density signal.

4. The brain oxygen signal processing method based on complete set empirical mode decomposition according to claim 1, characterized in that: In step S4: after calculating the permutation entropy of each modal component output in step S3, the permutation entropy value is normalized, and the normalized permutation entropy value is used as the permutation entropy.

5. The brain oxygen signal processing method based on complete set empirical mode decomposition according to claim 4, characterized in that: The initial threshold of permutation entropy was selected in the range of 0.5-0.7 including 0.

6.

6. The brain oxygen signal processing method based on complete set empirical mode decomposition according to claim 1, characterized in that: In step S2, based on the two photoelectric signals B1 and B2 in the same optical band obtained by the two receivers, conventional filtering and denoising are performed on the photoelectric signals B1 and B2 respectively, and the spatial differential optical density signal ΔOD in each optical band is calculated using the filtered and denoised photoelectric signals B1 and B2. λ ; 7. The brain oxygen signal processing method based on complete set empirical mode decomposition according to claim 6, characterized in that: The method includes the steps of selecting a wavelength combination, where W is 2, i.e. two bands of infrared light, and calculating the spatial differential optical density signal ΔOD under the two light bands. λ For subsequent calculations; For each spatial differential optical density signal ΔOD λ Perform steps S3 and S4 to reconstruct the signal, and select any reconstructed spatial differential optical density signal ΔOD λ Go to step S5.

8. The brain oxygen signal processing method based on complete set empirical mode decomposition according to claim 1, characterized in that: The method includes the steps of selecting a wavelength combination, where W is 3, i.e., three-band infrared light, and the wavelength combinations include (1, 2), (1, 3), and (2, 3); any number of wavelength combinations can be selected to generate the spatial differential optical density signal ΔOD. λ Calculate the spatial differential optical density signal ΔOD λ Perform steps S3 and S4 to reconstruct the signal, and select any reconstructed spatial differential optical density signal ΔOD λ Go to step S5.

9. The brain oxygen signal processing method based on complete set empirical mode decomposition according to claim 8, characterized in that: After step S6, Also includes step S7: According to the formula The brain oxygen saturation is calculated, where A, B, C, and D are the correction coefficients and correction values ​​corresponding to brain tissue oxygen, respectively.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the brain oxygen signal processing method based on complete set empirical mode decomposition as claimed in any one of claims 1 to 9 is implemented.

Images