A method for quantitatively evaluating the treatment effect of Alzheimer's disease based on electroencephalogram spectrum

Abstract

In the task state, the whole brain works in a coordinated manner. In the group with neurodegenerative diseases, during this process, significant changes will occur in the overall interval of EEG data. The physiological electrical activity components in the degenerative brain regions change, and the brain regions that compensate for the functions of the degenerative brain regions will also show component changes. The analysis of a single component cannot effectively evaluate the EEG data of AD patients in the task state. Therefore, the analysis of the overall EEG data may obtain higher effectiveness. In view of the problem that the treatment effect of AD disease is difficult to quantify, the present invention constructs a method for quantitatively evaluating the treatment effect of Alzheimer's disease based on EEG spectrum, which integrates the quantitative evaluation of various brain physiological electrical activity components that change due to neurodegeneration in the EEG of the subject at rest and the evaluation of the overall change of the EEG of the subject in the task state to quantify the treatment effect of AD disease, which brings certain assistance and guidance to the treatment of AD disease.

Description

Technical Field

[0001] The present invention belongs to the field of image processing, and particularly relates to a method for quantitatively evaluating the treatment effect of Alzheimer's disease based on electroencephalogram spectrum. Background Art

[0002] Alzheimer's disease (AD) is a common geriatric disease in which patients show prominent memory decline, reduced cognitive function, and accompanied by emotional and personality changes. As the disease progresses, the incidence and mortality of the affected population will increase accordingly. Currently, there are mature technologies for diagnosing this disease, and accurate diagnoses can be obtained in the early stage of the disease through various means, and certain measures can be taken to intervene and treat the disease.

[0003] However, the brain itself is an extremely complex neural organ, which contains a complex network of neural tissues working in parallel. When the disease occurs, the brain regions where the lesions occur show functional degenerative decline, and other regions will show compensatory replenishment, masking the functional degenerative phenomenon of the lesions. Moreover, after intervention and treatment, the recovery of the lesions may lead to a decline in the functions of other compensatory brain regions, causing certain interference to data analysts. Therefore, single-dimensional data analysis cannot meet the needs of patients with brain neurological diseases, especially the elderly population with degenerative brains, for rehabilitation quantification during the treatment process.

[0004] EEG is a brain information acquisition method that records the physiological electrical activities of the brain by placing electrodes on the cerebral cortex, and can provide data with extremely high time resolution. The EEG data contains various complex components, and different components are the physiological electrical activity mappings of different brain functions. Patients with neurodegenerative diseases will show significant changes in various aspects of EEG, which is significantly helpful for the evaluation of neurodegenerative diseases. The EEG in the brain resting state contains components of multiple frequency bands. Among them, the energy ratios of the α and β frequency bands in AD patients will show a certain decline, while the energy ratios of the δ and θ frequency bands will relatively increase. The changes of these two types of components will have significant changes through collaborative calculation, which is one of the important evaluation parameters in the diagnosis and treatment of AD.

[0005] In the task state, the whole brain will work collaboratively. In this process, for the population with neurodegenerative diseases, the overall interval of EEG data will show significant changes, the physiological electrical activity components of the degenerative brain regions will change, and the brain regions that compensate for the functions of the degenerative brain regions will also show component changes. The analysis of a single component cannot effectively evaluate the EEG data of AD patients in the task state. Therefore, the analysis of the overall EEG data may obtain higher effectiveness. Summary of the Invention

[0006] Aiming at the problem that the treatment effect of AD disease is difficult to quantify, a quantitative evaluation method for the treatment effect of AD disease is constructed based on EEG data. By integrating the quantitative evaluation of various electroencephalogram physiological electrical activity components that change due to neurodegeneration in the EEG of the subjects at rest and the data evaluation of the overall change of the EEG of the subjects in the task state, the treatment effect of AD disease is quantified, which brings certain assistance and guidance to the treatment of AD disease. Specifically, it includes:

[0007] Step 1: EEG data preprocessing

[0008] (1-1) Conduct enrollment screening to obtain the resting state and task state EEG data of AD patients before and after treatment and healthy control groups;

[0009] (1-2) Preprocess the obtained data through the Matlab toolbox eeglab;

[0010] Step 2: Resting state EEG evaluation

[0011] (2-1) Design α and β band filters, send the resting state EEG data into the filters, and screen the components with the decreased proportion of EEG energy in AD patients and healthy control groups;

[0012] (2-2) Calculate the absolute difference in the decreased proportion of energy of AD patients before and after treatment relative to the healthy control group;

[0013] (2-3) Construct δ and θ band filters, send the resting state EEG data into the filters, and screen the components with the increased proportion of EEG energy in AD patients and healthy control groups;

[0014] (2-4) Calculate the absolute difference in the increased proportion of energy of AD patients before and after treatment relative to the healthy control group;

[0015] (2-5) According to the results obtained in (2-2), calculate its proportion in the energy of the healthy control group and add them with a weight ratio of 2:1 to obtain the energy decrease evaluation result;

[0016] (2-6) According to the results obtained in (2-4), calculate its proportion in the energy of the healthy control group and add them with a weight ratio of 1:1 to obtain the energy increase evaluation result;

[0017] (2-7) According to the evaluation results of the decrease and increase of the resting state EEG energy of AD patients obtained in (2-5) and (2-6), add them in a 1:1 mode to obtain the resting state treatment evaluation result;

[0018] Step 3: Task state EEG evaluation

[0019] (3-1) With the set interval as the length, perform superposition averaging on the EEG data under the task-state image recognition task to obtain the superposition average data;

[0020] (3-2) Calculate the average data volatility of the EEG data in units of 1 ms;

[0021] (3-3) Calculate the absolute difference between the EEG data volatility of AD patients before and after treatment and that of healthy controls;

[0022] (3-4) Calculate the proportion distance of the absolute difference fluctuation of the values of AD patients before and after treatment in (3-3) in the healthy control group to obtain the volatility evaluation result;

[0023] (3-5) Locate the sampling points of the positive potential peak and the negative potential peak in the data of the healthy control group;

[0024] (3-6) And with the sampling points obtained in (3-5) as the center, perform superposition averaging on the data within the window with a fixed interval before and after as the window frame to obtain the positive and negative peak changes in the task state of the healthy control;

[0025] (3-7) According to the sampling point positions obtained in (3-5), calculate the positive and negative peak changes of AD patients before and after treatment to obtain the positive and negative peak changes in the task state of AD patients before and after treatment;

[0026] (3-8) Calculate the absolute difference between the superposition results of the positive and negative peaks of AD patients before and after treatment and those of the healthy control group;

[0027] (3-9) Calculate the proportion distance of the absolute difference before and after treatment obtained in (3-8) in the peaks of the healthy control group, and assign weights with a ratio of 1:1 to obtain the peak evaluation result;

[0028] (3-10) Add the results obtained in (3-4) and (3-9) with weights in a 1:1 mode to obtain the task-state treatment evaluation result;

[0029] Step 4: Evaluation of treatment effect

[0030] Fuse the scores of the resting-state and task-state treatment evaluation results with a weight ratio of 3:2 to obtain the final quantitative evaluation result of the treatment effect.

[0031] The beneficial effects of the present invention are as follows:

[0032] The present invention quantitatively evaluates the treatment effect in the clinical treatment of AD disease, and brings certain guidance and assistance to the accurate treatment of AD disease. Specific implementation manner

[0033] The present invention will be further described below in conjunction with specific embodiments, so that those skilled in the art can better understand the present invention. It should be noted that the following description is only for explaining the present invention and is not used to limit the present invention.

[0034] A method for treating and evaluating Alzheimer's disease based on electroencephalogram spectrum according to the present invention includes the following steps:

[0035] Step 1: Construct EEG data

[0036] (1) Comprehensively evaluate various biochemical indexes of clinical AD patients and healthy people, and accurately select AD patients and healthy controls, which is the data basis for high-quality evaluation of treatment effects.

[0037] (2) Collect resting-state EEG data X of AD patients before and after treatment and healthy control groups ri , where i takes values of 1, 2, and 3, representing AD patients before treatment, after treatment, and healthy control groups respectively, and among them, X ri (x ri1 , x ri2 , x ri3 ...... x rij ...... x rin ) represents the data of n subjects in each group of data. For example, x rij represents the experimental data of the j-th subject in the i-th group of resting-state EEG data.

[0038] (3) Design a standard AD detection "image recognition" EEG experiment, and collect task-state EEG data X of AD patients before and after treatment and healthy control groups ti , where X ti (x ti1 , x ti2 , x ti3 ...... x tij ...... x tin ) represents the data of n subjects in each group of data. For example, x tij represents the experimental data of the j-th subject in the i-th group of task-state EEG data.

[0039] (4) EEG data preprocessing: Based on the Matlab toolbox eeglab, preprocess the two types of data according to the same standard: low-frequency filtering below 0.1, whole-brain average re-reference, channel data inspection, independent component decomposition, and noise component removal, etc.

[0040] Step 2: Resting-state EEG evaluation

[0041] (1) Design filters for two frequency bands, and input the data X ri respectively to obtain the alpha and beta band data M of the EEG data riWith N ri ;

[0042] (2) Calculate the energy ratios P of the α and β bands before and after treatment for the healthy control group and AD patients ri and Q ri , which are P ri = 10·log 10 M ri 2 / 10·log 10 X ri 2 and Q ri = 10·log 10 N ri 2 / 10·log 10 X ri 2 ;

[0043] (3) Calculate the energy differences of the α and β bands between the healthy control group and AD patients before treatment: TP 1 = P r3 - P r1 and TP 2 = Q r3 - Q r1 ;

[0044] (4) Calculate the energy differences of the α and β bands between the healthy control group and AD patients after treatment: TP 1 ' = P r3 - P r2 and TP 2 ' = Q r3 - Q r2 ;

[0045] (5) Design filters for two frequency bands, δ (0.5 Hz - 3 Hz) and θ (3.5 Hz - 7.5 Hz), and input the data X ri to obtain the EEG data of the δ and θ bands, M' ri and N' ri ;

[0046] (6) Calculate the energy ratios P' ri and Q' ri of the δ and θ bands before and after treatment for the healthy control group and AD patients, respectively and

[0047] (7) Calculate the energy differences TP of the δ and θ bands between the healthy control group and AD patients 3 = P' r3 - P' r1 and TP 4 = Q' r3-Q′ r1 ;

[0048] (8) Calculate the energy difference TP′ between the δ and θ bands of healthy control group and AD patients after treatment 3 =P′ r3 -P′ r2 and TP 4 '=Q r ' 3 -Q r ' 2 ;

[0049] (9) Quantitative evaluation results of treatment effect based on resting-state EEG data:

[0050] Y 1 =0.5((2 / 3)*(1-(TP′ 1 -TP 1 ) / P r3 )+(1 / 3)*(1-(TP′ 2 -TP 2 ) / Q r3 ))+0.5(0.5(1-(TP′ 1 -TP 1 ) / P r3 )+0.5(1-(TP 4 -TP′ 4 ) / Q r3 ))

[0051] Step 3: Task-state EEG evaluation

[0052] (1) For the three groups of task-state EEG data X ti , where u = 1, 2, 3, representing AD patients before treatment, after treatment, and healthy control group, with a time window of 500 ms sampling points, perform superposition averaging on the task-state EEG data to obtain the superposition average data X t ' i ;

[0053] (2) Select the superposition average data X′ ti , and calculate its data volatility J i (j i1 ,j i2 ,j i3 ......j in ), where n represents the nth subject, and k represents the position of the sampling point;

[0054] (3) Calculate the difference J 1 ' and J' 2 between the changes in EEG volatility during the process of AD patients before and after treatment and that of the healthy control group, where J1 ' = abs(J 3 - J 1 ), J' 2 = abs(J 3 - J 2 ), abs represents the absolute value;

[0055] (4) Select the superimposed average data X of the healthy control group t ' 3 The sampling point position L where the maximum value is located t , and select the 5 points before and after this point for superimposed averaging to obtain O 3 ;

[0056] (5) Select the data X before and after the treatment of AD patients t ' 1 and X t ' 2 at the position L t , the sampling point and the 5 points before and after it are superimposed and averaged to obtain O 1 and O 2 , to obtain the positive peak change;

[0057] (6) Calculate the difference between the positive peak change of EEG during the task and the healthy control before and after the treatment of AD subjects O 1 ‘and O' 2 , where O 1 ' = abs(O 3 - O 1 ), O' 2 = abs(O 3 - O 2 );

[0058] (7) Select the superimposed average data X of the healthy control group t ' 3 The sampling point position L' where the minimum value is located t , and select the 5 points before and after this point for superimposed averaging to obtain U 3 ;

[0059] (8) Select the data X before and after the treatment of AD subjects t ' 1 and X t ' 2 at the position L' t , the sampling point and the 5 points before and after it are superimposed and averaged to obtain U 1 and U 2 , to obtain the negative peak change;

[0060] (9) Calculate the difference between the negative peak change of EEG during the task and the healthy control before and after the treatment of AD subjects U 1 ' and U'2 , where U 1 ' = abs(U 3 - U 1 ), U' 2 = abs(U 3 - U 2 );

[0061] (10) Quantitative evaluation of treatment effect based on task-state data:

[0062] Y 2 = (2 / 3)*(1 - abs(J' 2 - J′ 1 ) / J 3 ) + (1 / 3)*(0.5(1 - abs(O' 2 - O′ 1 ) / O 3 ) + 0.5(1 - abs(U' 2 - U′ 1 ) / U 3 ))

[0063] Step 4: Quantitative evaluation of treatment effect

[0064] Combined resting-state and task-state EEG evaluation result Y 1 and Y 2 , and output the treatment evaluation score Z = 0.6*Y 1 + 0.4*Y 2 .

[0065] The above content has introduced the present invention in detail. However, the description of the specific implementation manners is only used to explain the method and its core idea of the present invention, so as to facilitate those skilled in the art to understand the present invention. It should be clear that the present invention is not limited to the scope of the specific implementation manners. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are within the scope of protection.

Claims

1. A method for quantitatively evaluating the treatment effect of Alzheimer's disease based on electroencephalogram spectrum, characterized in that: It includes the following steps: Step 1: EEG data preprocessing (1-1) Conduct enrollment screening to obtain resting-state and task-state EEG data of AD patients before and after treatment and healthy control groups; (1-2) Preprocess the obtained data through the Matlab toolbox eeglab; Step 2: Resting-state EEG evaluation (2-1) Design α and β band filters, send the resting-state EEG data into the filters, and screen the components with a decrease in the EEG energy proportion of AD patients and healthy control groups; (2-2) Calculate the absolute difference in the proportion of the decrease in energy of AD patients before and after treatment relative to the healthy control group; (2-3) Construct δ and θ band filters, send the resting-state EEG data into the filters, and screen the components with an increase in the EEG energy proportion of AD patients and healthy control groups; (2-4) Calculate the absolute difference in the proportion of the increase in energy of AD patients before and after treatment relative to the healthy control group; (2-5) According to the results obtained in step (2-2), calculate its proportion in the energy of the healthy control group and add it with a weight ratio of 2:1 to obtain the energy decrease evaluation result; (2-6) According to the results obtained in step (2-4), calculate its proportion in the energy of the healthy control group and add it with a weight ratio of 1:1 to obtain the energy increase evaluation result; (2-7) According to the evaluation results of the decrease and increase in the resting-state EEG energy of AD patients obtained in steps (2-5) and (2-6), add them in a 1:1 mode to obtain the resting-state treatment evaluation result; Step 3: Task-state EEG evaluation (3-1) Take the set interval as the length, perform task-state image recognition on the task-state EEG data, and perform superposition averaging to obtain the superposition average data; (3-2) Calculate the average data volatility of the EEG data in units of 1 ms; (3-3) Calculate the absolute difference between the EEG data volatility of AD patients before and after treatment and that of the healthy control; (3-4) Calculate the proportion distance of the absolute difference in the values of AD patients before and after treatment in step (3-3) in the healthy control group to obtain the volatility evaluation result; (3-5) Locate the sampling points of the positive potential peak and negative potential peak in the healthy control group data; (3-6) And take the sampling points obtained in step (3-5) as the center, and perform superposition averaging on the data within the window with a fixed interval before and after to obtain the positive and negative peak changes in the healthy control task state; (3-7) According to the sampling point positions obtained in step (3-5), calculate the positive and negative peak changes of AD patients before and after treatment to obtain the positive and negative peak changes of AD patients in the task state before and after treatment; (3-8) Calculate the absolute difference between the superposition results of the positive and negative peaks of AD patients before and after treatment and those of the healthy control group; (3-9) Calculate the proportion distance of the absolute difference before and after treatment obtained in step (3-8) in the peaks of the healthy control group and allocate it with a weight of 1:1 to obtain the peak evaluation result; (3-10) Add the results obtained in steps (3-4) and (3-9) with a weight ratio of 1:1 to obtain the treatment evaluation result in the task state; Step 4: Treatment effect evaluation Fuse the scores of the resting state and task state treatment evaluation results with a weight of 3:2 to obtain the final quantitative evaluation result of the treatment effect.