Brain function activation level lateral detection method based on near-infrared image data

Abstract

The invention discloses a brain function activation level bias detection method based on near-infrared image data. The method comprises the following steps: acquiring the near-infrared image data and extracting at least one group of brain function activation indexes of left and right brain regions from the near-infrared image data; and obtaining optimal solutions of a, b and c calculated according to the bias calculation model: respectively comparing the extracted brain function activation indexes with 0, a, b and c according to a preset bias detection process, and outputting a bias detection result. Based on the near-infrared image data, by analyzing the near-infrared blood oxygen data characteristics of the time sequences of the left brain and the right brain, the size difference of the function activation level between the left brain and the right brain is accurately recognized, and help is provided for clinical departments to timely and effectively recognize the brain function activation characteristics of mental patients; meanwhile, a feasible nerve biomarker is provided for evaluating the treatment effect and tracking the prognosis of a patient.

Description

TECHNICAL FIELD

[0001] The present application relates to a method for detecting brain function activation level lateralization based on near-infrared image data. BACKGROUND

[0002] Near-infrared imaging technology is a non-invasive brain function imaging technology, which mainly uses the difference in absorption rate of near-infrared light of different wavelengths (600-900 nm) by oxyhemoglobin (HbO) and deoxyhemoglobin (HbR) in brain tissue to directly detect the hemodynamic activity of the cerebral cortex in real time, and indirectly assess the brain neural activity status by combining the neural vascular coupling rule. When using near-infrared imaging technology to measure brain function, a task or paradigm is usually combined to observe the brain blood flow activity pattern of the subject when performing a specific task, in order to judge whether the task can cause the brain function activation of the subject and the size of the activation level.

[0003] In recent years, many researchers have used near-infrared imaging technology to conduct a large number of studies on neurobiological markers and neuroimaging characteristics of mental diseases, and found that there are significant differences in brain function activation patterns between mental disease patients and healthy controls when performing VFT (Verbal Fluency Task, verbal fluency task). One of the more obvious differences is that some mental disease patients will show inconsistent activation levels between the left and right brains, while this feature called lateralization is less common in healthy controls. Therefore, the lateralization feature has great value for identifying mental disease patients. However, although this feature has been reported in many studies, there is currently no algorithm to detect whether lateralization exists.

[0004] Some researchers have proposed using a lateralization index as an indicator to study the lateralization feature, but this indicator still has some problems that make it difficult to be practically applied, including: 1. There is no uniform threshold to distinguish between healthy controls and mental disease patients for this indicator. There are not many cases in previous literature, and there is a lack of research based on large sample data mining methods; 2. In some cases, calculating the lateralization index has no practical significance, such as when the activation levels of the left and right brains are both negative, in which case there is no activation in either the left or right brain, so there is no left or right activation bias. Previous methods did not make such a distinction; 3. The lateralization index can misjudge, because brain activity can have a hemispheric advantage phenomenon. A slight lateralization in healthy controls is completely normal, but lateralization caused by disease is different. There is currently no algorithm to make this judgment. SUMMARY

[0005] To solve the problem of how to accurately and effectively identify the brain function activation level lateralization, the present application provides a brain function activation level lateralization detection method based on near-infrared image data, which is based on near-infrared image data, analyzes the near-infrared blood oxygen data characteristics of the left brain and right brain time sequence, accurately identifies the size difference of the function activation level between the left and right brains, provides help for the clinical departments to timely and effectively identify the brain function activation characteristics of mental illness patients, and also provides a feasible neurobiological marker for evaluating the treatment effect and tracking the prognosis of patients.

[0006] According to an aspect of the present application, a brain function activation level lateralization detection method based on near-infrared image data is provided, comprising: Obtaining near-infrared image data and extracting at least one set of brain function activation indicators of left and right brain regions therefrom; Obtaining the optimal solutions of a, b and c calculated according to the following lateralization calculation model:

[0007] Wherein a is the minimum activation difference, b is the minimum activation threshold, c is the lateralization index threshold, P(Y=1) is the probability of lateralization of the current data, and β0, β1, β2 and β3 are model parameters; Comparing the extracted at least one set of brain function activation indicators of the left and right brain regions with 0, a, b and c respectively according to a preset lateralization detection process, and outputting the lateralization detection result.

[0008] Preferably, the at least one set of brain function activation indicators of the left and right brain regions includes but is not limited to: task period oxyhemoglobin integral value, oxyhemoglobin beta value, oxyhemoglobin mean value or oxyhemoglobin peak value.

[0009] As a further technical solution, comparing the extracted at least one set of brain function activation indicators of the left and right brain regions with 0, a, b and c respectively according to a preset lateralization detection process, comprising: Determining whether the left brain region brain function activation indicator is less than 0; If the left brain region brain function activation indicator is less than 0, determining whether the right brain region brain function activation indicator is less than 0, and if so, outputting no lateralization.

[0010] As a further technical solution, when the left brain region brain function activation indicator is greater than or equal to 0, further comprising: Determining whether the left brain region brain function activation indicator is less than or equal to b, and if not, when the right brain region brain function activation indicator is less than 0, outputting left lateralization.

[0011] As a further technical solution, if the left brain area brain function activation index is greater than b, and the right brain area brain function activation index is greater than or equal to 0, or the left brain area brain function activation index is less than or equal to b, and the right brain area brain function activation index is greater than b, the detection result is given according to the calculation result of g(T) as follows: , , wherein IV 左 represents the left brain area brain function activation index, IV 右 represents the right brain area brain function activation index.

[0012] As a further technical solution, when the left brain area brain function activation index is less than or equal to b, it further comprises: determining whether the right brain area brain function activation index is less than 0, if not, when the right brain area brain function activation index is less than or equal to b, outputting no lateralization.

[0013] As a further technical solution, when the right brain area brain function activation index is less than 0, it further comprises: determining whether the difference between the left brain area brain function activation index and the right brain area brain function activation index is greater than or equal to a, if yes, outputting left lateralization, if not, outputting no lateralization.

[0014] As a further technical solution, when the left brain area brain function activation index is less than 0, and the right brain area brain function activation index is greater than or equal to 0, it further comprises: determining whether the difference between the right brain area brain function activation index and the left brain area brain function activation index is greater than or equal to a, if yes, outputting right lateralization, if not, outputting no lateralization.

[0015] As a further technical solution, the method further comprises: solving the lateralization calculation model using the simulated annealing algorithm based on the near-infrared image data sample set according to the constructed objective function and constraint conditions to obtain the optimal solution of a, b and c.

[0016] As a further technical solution, the constructed objective function is: , The constraint condition is: , wherein, , respectively represent the number of healthy people and patients, , respectively represent whether the data of healthy people and patients exist lateralization.

[0017] According to an aspect of the present application, a brain function activation level lateralization detection system based on near-infrared image data is provided for implementing the method, comprising: A first main module is configured to acquire near-infrared image data and extract at least one set of brain function activation indicators of left and right brain regions from the data; A second main module is configured to acquire the optimal solutions of a, b and c calculated according to the following lateralization calculation model: , Wherein, a is the minimum activation difference, b is the minimum activation threshold, c is the lateralization index threshold, P(Y=1) is the probability of the current data existing lateralization, and β0, β1, β2 and β3 are model parameters; A third main module is configured to compare the extracted at least one set of brain function activation indicators of the left and right brain regions with 0, a, b and c respectively according to a preset lateralization detection process, and output a lateralization detection result.

[0018] Compared with the prior art, the present application has the following advantages: The present application analyzes and mines the near-infrared image data of mental illness patients and healthy people, proposes a discrimination method for detecting whether there is a lateralization feature, can assist in identifying whether the brain function activation has lateralization, which is a neural biomarker of mental illness, and then assist clinicians to make objective diagnosis, and can also provide objective basis for doctors to develop subsequent treatment plan. The present application provides a feasible solution to the problem of excessive subjectivity caused by the diagnosis and treatment of clinicians relying on behavior observation, scale or experience judgment, which has great potential value and significance for the clinical work of related departments. BRIEF DESCRIPTION OF DRAWINGS

[0019] To make the technical solutions of the present application or the prior art clearer, the drawings used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0020] Figure 1 The flowchart of the brain function activation level lateralization detection method based on near-infrared image data provided by the present application is shown.

[0021] Figure 2 The judgment logic diagram of the brain function activation level lateralization detection based on near-infrared image data provided by the present application is shown.

[0022] Figure 3A result schematic diagram after processing of near-infrared image data of a mental illness patient provided by an embodiment of the present application is shown.

[0023] Figure 4 Another result schematic diagram after processing of near-infrared image data of a mental illness patient provided by an embodiment of the present application is shown.

[0024] Figure 5 A result schematic diagram after processing of near-infrared image data of a healthy person provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0025] The technical solutions of the embodiments of the present application will be described clearly and completely in combination with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of the present application.

[0026] The present application is based on large-sample norm database research and development, and describes a detection method for brain function activation level laterality. Figure 1 The embodiment of the present application provides a brain function activation level laterality detection method based on near-infrared image data. First, near-infrared image data is acquired and at least one set of brain function activation indexes of left and right brain regions is extracted therefrom. Then, the optimal solutions of a, b and c calculated according to a laterality calculation model are acquired. Subsequently, the extracted at least one set of brain function activation indexes of the left and right brain regions are compared with 0, a, b and c according to a preset laterality detection process, and a laterality detection result is output.

[0027] The brain function activation level laterality detection method based on near-infrared image data provided by the embodiment of the present application specifically includes the following steps: 1. Collecting near-infrared image data of a subject during execution of VFT.

[0028] 2. Processing the collected data, which follows a basic process of near-infrared image data, and the purpose is to extract the concentration change of HbO2 of the brain during execution of VFT.

[0029] 3. Calculate the task period HbO2 integral value, oxygenated hemoglobin beta value, oxygenated hemoglobin mean value or oxygenated hemoglobin peak value of each brain area. According to previous literature research, patients with mental illness will show abnormal activation function in the left frontal lobe, right frontal lobe, left temporal lobe and right temporal lobe. Therefore, the activation degree of the above four brain areas is mainly calculated in the embodiment of the present application. The task period HbO2 integral value is taken as the brain function activation index for illustration, the integral value is an index for measuring the intensity of cerebral hemodynamic response, the area under the curve of the change of HbO2 concentration with time during the task period is calculated, the size of the increase of HbO2 concentration during the task period is reflected, and then the activation degree of the brain area is reflected. The calculation formula of the integral value (IV) is as follows: , Wherein, t is the time point, f(t) is the HbO2 concentration at time t, and a and b represent the start time point and the end time point of the task period respectively.

[0030] It should be noted that the calculation of the oxygenated hemoglobin beta value, the oxygenated hemoglobin mean value or the oxygenated hemoglobin peak value can be realized by using the existing technology. The key point of the present application is that the preset process of outputting the lateralization result is based on the minimum activation difference (a), the minimum activation threshold (b) and the lateralization index threshold (c) and the brain function activation index, and the selection and calculation of the brain function activation index are not limited.

[0031] 4. Determine whether the lateralization of the left and right brain function activation levels of the subject exists according to the following algorithm. According to the principle of near-infrared detection, the embodiment of the present application distinguishes the different cases of positive and negative integral values of the left and right brain, and defines the following three parameters for quantitative judgment: minimum activation difference (a), minimum activation threshold (b) and lateralization index threshold (c). The minimum activation difference refers to that even if there is a certain difference between the left and right brain, but it does not reach a certain degree, at this time, there is still no lateralization. The minimum activation threshold refers to that even if the integral value is positive, but it does not reach a certain numerical range, at this time, there is still no activation. The lateralization index threshold refers to that when the condition is met, whether the lateralization exists is calculated by the formula g(T), the value needs to exceed the threshold value to be recognized as lateralization. a, b, c are all positive values.

[0032] Taking the task period HbO2 integral value as the brain function activation index as an example, the specific judgment implementation process is shown in Figure 2 , in which, IV 左 represents the integral value of the left brain area (i.e. the task period oxygenated hemoglobin integral value of the left brain area), and IV 右 represents the integral value of the right brain area (i.e. the task period oxygenated hemoglobin integral value of the right brain area). The calculation formula of g(T) is as follows: , Wherein, the calculation formula of T is as follows: .

[0033] The specific judgment process is as follows: First, it is judged whether IV 左 is less than 0, if yes, it is further judged whether IV 右 is less than 0.

[0034] If IV 右 is less than 0, the unbiased lateralization is outputted.

[0035] If IV 右 is greater than or equal to 0, it is judged whether IV 右 -IV 左 is greater than or equal to a, if yes, the right biased lateralization is outputted, if not, the unbiased lateralization is outputted.

[0036] Next, when IV 左 is greater than or equal to 0, it is further judged whether IV 左 is less than or equal to b.

[0037] If IV 左 is less than or equal to b, it is further judged whether IV 右 is less than 0, if yes, it is further judged whether IV 左 -IV 右 is greater than or equal to a, if yes, the left biased lateralization is outputted, if not, the unbiased lateralization is outputted.

[0038] When IV 左 is less than or equal to b, and IV 右 is greater than or equal to 0, it is judged whether IV 右 is less than or equal to b, if yes, the unbiased lateralization is outputted, if not, g(T) is calculated, and according to the calculation result, it is judged whether it is left biased lateralization, right biased lateralization or unbiased lateralization.

[0039] Further, if IV 左 is greater than b, it is further judged whether IV 右 is less than 0, if yes, the left biased lateralization is outputted, if not, g(T) is calculated, and according to the calculation result, it is judged whether it is left biased lateralization, right biased lateralization or unbiased lateralization.

[0040] The optimal solution solving process of the lateralization calculation model described in the embodiment of the application includes the following process: The collected large sample near-infrared image data is divided into a training set and a test set. The training set has 1552 cases, including 762 cases of patients with mental illness and 790 cases of healthy people. The test set has 4280 cases, including 2011 cases of patients with mental illness and 2269 cases of healthy people. The training set is used to study the characteristics of the image data, and the unsupervised learning method is used to optimize the parameter values and improve the accuracy of the algorithm. The test set is mainly used to verify the accuracy of the algorithm, determine the generalizability of the algorithm, and evaluate the performance of the algorithm. The training set and the test set are not touched during the model training and parameter tuning process, that is, they are two batches of relatively independent data. The a, b and c calculated from the training set are included in the lateralization calculation model, and then verified in the test set to determine whether it can be generalized to other conditions.

[0041] Specifically, the training set is first learned and arranged, and the lateralization characteristics of the brain function activation level in the training set data are integrated, and the three core parameters in the algorithm, i.e., the minimum activation difference (a), the minimum activation threshold (b), and the lateralization index threshold (c), are optimized.

[0042] The lateralization calculation model training process is as follows: First, according to the above algorithm process, whether a piece of near-infrared image data has lateralization is determined by the values of a, b and c, that is: , Where P(Y=1) is the probability of the data having lateralization, a, b and c are the minimum activation difference, the minimum activation threshold and the lateralization index threshold, respectively. The logstic regression model is used to construct the lateralization calculation model in the embodiment of the present application, wherein β0, β1, β2 and β3 are parameters of the logstic regression model.

[0043] According to the probability, the classification is as follows: , Wherein, That is, it is judged whether the data has lateralization. "1" means that there is lateralization, and "0" means that there is no lateralization.

[0044] Secondly, according to the meta-analysis research literature, the proportion of lateralization in the patient population with mental illness is more than 30%, and only less than 5% of the healthy population will have lateralization. Therefore, the following constraints exist: (1) , and the value is as large as possible.

[0045] (2) , and the value is as small as possible.

[0046] Finally, according to the literature research and clinical experience, the proportion of the lateralization in the patient group of mental illness is much larger than that in the healthy people, i.e. it is necessary to optimize a, b, c, so that under the condition of meeting the above constraints, there is: , According to the objective function and the constraint condition, the optimal solution S=(a, b, c) is obtained by using the simulated annealing algorithm. The simulated annealing algorithm starts from a higher initial temperature, and with the continuous decrease of the temperature parameter, the global optimal solution of the objective function is randomly searched in the solution space, i.e. the local optimal solution can be probabilistically jumped out and finally tends to be global optimal. The specific algorithm implementation process is as follows: 1. Initialization, set the initial temperature T, the initial solution state S, and the iteration number of each T value is L; 2. Perform steps 3 to 6 for k=1,....,L; 3. Generate a new solution S1 according to the current temperature and the predetermined condition; 4. Calculate the objective function increment ; 5. If , then accept S1 as the new current solution, otherwise accept S1 as the new current solution with a probability ; 6. If the termination condition is met, output the current solution as the optimal solution; 7. Reduce T, so that T gradually tends to 0, and then go to step 2.

[0047] The simulated annealing algorithm avoids falling into local optimization in the process of finding the solution by the probability jump feature, so as to find the global optimal solution of the objective function. In the process of generating a new solution S1, the size of will be used to dynamically adjust the generation of a suitable new solution, i.e. when is larger, the difference between the new solution and the old solution is larger, and vice versa. This facilitates the rapid convergence of the solution and reduces the time consumption.

[0048] As an implementation mode, as shown in Figure 3 , the result of the near-infrared image data analysis and processing of a mental illness patient is shown, the brain area is the temporal lobe, the horizontal axis is the time, and the vertical axis is the HbO2 concentration. After calculation, the integral value of the left temporal lobe is 76.0138, and the integral value of the right temporal lobe is 327.0274. According to the algorithm flow, the calculation condition is met, the T value calculation result is-0.6228, and the threshold judgment condition is met, which is determined as right lateralization.

[0049] As an implementation mode, as shown in Figure 4The result of the analysis and processing of the near-infrared image data of a mental illness patient is shown, the brain region is the temporal lobe, the horizontal axis is time, and the vertical axis is HbO2 concentration. After calculation, the integral value of the left temporal lobe is 288.6495, and the integral value of the right temporal lobe is -70.4879. According to the algorithm process, it can be directly determined that it is left lateralization.

[0050] As an embodiment, as Figure 5 The result of the analysis and processing of the near-infrared image data of a healthy person is shown, the brain region is the temporal lobe, the horizontal axis is time, and the vertical axis is HbO2 concentration. After calculation, the integral value of the left temporal lobe is 115.7382, and the integral value of the right temporal lobe is 168.4069. According to the algorithm process, it is determined that there is no lateralization.

[0051] Based on the same inventive concept as the foregoing method embodiments, the present embodiment also provides a brain function activation level lateralization detection system based on near-infrared image data, for realizing the method of any one of the foregoing embodiments.

[0052] The system comprises: a first main module for acquiring near-infrared image data and extracting at least one set of brain function activation indicators of left and right brain regions therefrom; a second main module for acquiring the optimal solutions of a, b and c calculated according to the following lateralization calculation model: , wherein a is the minimum activation difference, b is the minimum activation threshold, c is the lateralization index threshold, P(Y=1) is the probability of the presence of lateralization in the current data, β0, β1, β2 and β3 are model parameters; a third main module for comparing the extracted at least one set of brain function activation indicators of the left and right brain regions with 0, a, b and c respectively according to a preset lateralization detection process, and outputting a lateralization detection result.

[0053] The system of the present embodiment can assist in identifying whether the brain function activation has lateralization, which is a neurobiological marker of mental illness, thereby assisting clinicians to objectively make a diagnosis, and also providing an objective basis for doctors to develop subsequent treatment plans.

[0054] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent substitutions for part or all of the technical features; and these modifications or substitutions do not make the essence of the corresponding technical solution deviate from the technical solutions of the present embodiment.

Claims

1. A method for detecting the lateralization of brain functional activation level based on near-infrared image data, characterized in that, The method comprises the following steps: obtaining near-infrared image data and extracting at least one set of brain function activation indicators of left and right brain regions from the near-infrared image data; obtaining the optimal solutions of a, b and c calculated according to the following lateralization calculation model: ， wherein a is the minimum activation difference, b is the minimum activation threshold, c is the lateralization index threshold, P(Y=1) is the probability of the current data existing lateralization, and β0, β1, β2 and β3 are model parameters; comparing the extracted at least one set of brain function activation indicators of the left and right brain regions with 0, a, b and c respectively according to a preset lateralization detection process, and outputting a lateralization detection result.

2. The method of claim 1, wherein the method further comprises: comparing the extracted at least one set of brain function activation indicators of the left and right brain regions with 0, a, b and c respectively according to a preset lateralization detection process, comprising: determining whether the brain function activation indicator of the left brain region is less than 0; when the brain function activation indicator of the left brain region is less than 0, determining whether the brain function activation indicator of the right brain region is less than 0, and if yes, outputting no lateralization. 3.The method of claim 2, wherein, when the brain function activation indicator of the left brain region is greater than or equal to 0, further comprising: determining whether the brain function activation indicator of the left brain region is less than or equal to b, and if no, when the brain function activation indicator of the right brain region is less than 0, outputting left lateralization.

4. The method of claim 3, wherein the method further comprises: if the brain function activation indicator of the left brain region is greater than b and the brain function activation indicator of the right brain region is greater than or equal to 0, or the brain function activation indicator of the left brain region is less than or equal to b and the brain function activation indicator of the right brain region is greater than b, the detection result is given according to the calculation result of g(T) as follows: ， ， Wherein, IV 左 represents the brain function activation index of the left brain region, IV 右 represents the brain function activation index of the right brain region.

5. The method of claim 3, wherein the method further comprises: when the brain function activation indicator of the left brain region is less than or equal to b, further comprising: determining whether the brain function activation indicator of the right brain region is less than 0, and if no, when the brain function activation indicator of the right brain region is less than or equal to b, outputting no lateralization.

6. The method of claim 5, wherein the method further comprises: when the brain function activation indicator of the right brain region is less than 0, further comprising: determining whether the difference between the brain function activation indicator of the left brain region and the brain function activation indicator of the right brain region is greater than or equal to a, and if yes, outputting left lateralization, and if no, outputting no lateralization.

7. The method of claim 2, wherein the method further comprises: when the brain function activation indicator of the left brain region is less than 0 and the brain function activation indicator of the right brain region is greater than or equal to 0, further comprising: determining whether the difference between the brain function activation indicator of the right brain region and the brain function activation indicator of the left brain region is greater than or equal to a, and if yes, outputting right lateralization, and if no, outputting no lateralization. 8.The method of claim 1, wherein, The method further comprises: solving the lateralization calculation model based on the near-infrared image data sample set using a simulated annealing algorithm according to the constructed objective function and constraint condition, to obtain the optimal solutions of a, b and c.

9. The method of claim 8, wherein the method further comprises: The constructed objective function is: ， The constraint condition is: ， wherein, , respectively represent the number of healthy and patient, , respectively represent whether the data of healthy and patient are lateralized.

10. A system for detecting lateralization of brain functional activation level based on near-infrared image data, for implementing the method of any one of claims 1-9, characterized in that, comprising: a first main module for obtaining near-infrared image data and extracting at least one set of brain function activation indicators of left and right brain regions from the near-infrared image data; a second main module for obtaining the optimal solutions of a, b and c calculated according to the following lateralization calculation model: ， wherein a is the minimum activation difference, b is the minimum activation threshold, c is the lateralization index threshold, P(Y=1) is the probability of the current data existing lateralization, and β0, β1, β2 and β3 are model parameters; a third main module for comparing the extracted at least one set of brain function activation indicators of the left and right brain regions with 0, a, b and c respectively according to a preset lateralization detection process, and outputting a lateralization detection result.

Images