CN116030941B - Alzheimer's disease diagnosis method based on edge-centric effect connection network - Google Patents
Alzheimer's disease diagnosis method based on edge-centric effect connection network Download PDFInfo
- Publication number
- CN116030941B CN116030941B CN202310329085.7A CN202310329085A CN116030941B CN 116030941 B CN116030941 B CN 116030941B CN 202310329085 A CN202310329085 A CN 202310329085A CN 116030941 B CN116030941 B CN 116030941B
- Authority
- CN
- China
- Prior art keywords
- connection network
- edge
- brain
- effect
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000000694 effects Effects 0.000 title claims abstract description 75
- 238000000034 method Methods 0.000 title claims abstract description 46
- 208000024827 Alzheimer disease Diseases 0.000 title claims abstract description 43
- 238000003745 diagnosis Methods 0.000 title abstract description 17
- 210000004556 brain Anatomy 0.000 claims abstract description 126
- 230000005764 inhibitory process Effects 0.000 claims abstract description 9
- 239000008280 blood Substances 0.000 claims abstract description 5
- 210000004369 blood Anatomy 0.000 claims abstract description 5
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims abstract description 3
- 230000001419 dependent effect Effects 0.000 claims abstract description 3
- 229910052760 oxygen Inorganic materials 0.000 claims abstract description 3
- 239000001301 oxygen Substances 0.000 claims abstract description 3
- 230000001364 causal effect Effects 0.000 claims description 21
- 239000011159 matrix material Substances 0.000 claims description 14
- 238000012937 correction Methods 0.000 claims description 12
- 238000002598 diffusion tensor imaging Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 11
- 238000002599 functional magnetic resonance imaging Methods 0.000 claims description 10
- 238000002595 magnetic resonance imaging Methods 0.000 claims description 10
- 239000000835 fiber Substances 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 6
- 210000004885 white matter Anatomy 0.000 claims description 6
- 238000011282 treatment Methods 0.000 claims description 5
- 230000004886 head movement Effects 0.000 claims description 4
- 239000006185 dispersion Substances 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 3
- 210000003625 skull Anatomy 0.000 claims description 3
- 238000012706 support-vector machine Methods 0.000 claims description 3
- 108010076504 Protein Sorting Signals Proteins 0.000 claims description 2
- 230000000670 limiting effect Effects 0.000 claims description 2
- 230000033001 locomotion Effects 0.000 claims description 2
- 239000012636 effector Substances 0.000 claims 4
- 241001280442 Boldia Species 0.000 claims 1
- 210000004907 gland Anatomy 0.000 claims 1
- 238000002610 neuroimaging Methods 0.000 claims 1
- 230000002123 temporal effect Effects 0.000 claims 1
- 238000013399 early diagnosis Methods 0.000 abstract description 5
- 230000009850 completed effect Effects 0.000 abstract description 2
- 238000006213 oxygenation reaction Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 230000015654 memory Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 208000014644 Brain disease Diseases 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 2
- 208000010877 cognitive disease Diseases 0.000 description 2
- 230000000295 complement effect Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000010410 layer Substances 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 208000027061 mild cognitive impairment Diseases 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 206010012289 Dementia Diseases 0.000 description 1
- 208000025174 PANDAS Diseases 0.000 description 1
- 208000021155 Paediatric autoimmune neuropsychiatric disorders associated with streptococcal infection Diseases 0.000 description 1
- 240000000220 Panda oleosa Species 0.000 description 1
- 235000016496 Panda oleosa Nutrition 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000003925 brain function Effects 0.000 description 1
- 230000008309 brain mechanism Effects 0.000 description 1
- 230000001684 chronic effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 210000004884 grey matter Anatomy 0.000 description 1
- 230000000971 hippocampal effect Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000001404 mediated effect Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000004770 neurodegeneration Effects 0.000 description 1
- 208000015122 neurodegenerative disease Diseases 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 239000002344 surface layer Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses an Alzheimer disease diagnosis method based on an effect connection network with edges as centers, which comprises the following steps: acquiring multi-mode brain image data of a tested person, and based on the multi-mode brain image data, obtaining a time sequence, a functional connection network and a structural connection network of the tested person (Blood Oxygenation Level Dependent, blood oxygen level dependence); based on the inhibition relation between the structural connection network and the functional connection network, improving the conditional gracile cause and effect algorithm to obtain an improved gracile cause and effect algorithm, and based on the BOLD time sequence and the structural connection network, constructing an effect connection network with edges as centers by using the improved gracile cause and effect algorithm; based on the established edge-centered effect connection network, the identification of the Alzheimer's disease is completed by using a preset classifier. The invention is helpful for solving the problem of difficult early diagnosis of Alzheimer's disease.
Description
Technical Field
The invention relates to the technical field of brain image processing and brain science, in particular to an Alzheimer disease diagnosis method based on an effect connection network with edges as centers.
Background
Alzheimer's disease is a chronic neurodegenerative disease, the most common type of dementia, usually begins to slow and worsen over time, most commonly in elderly people over 65 years. At present, research on Alzheimer's disease has been urgent.
In recent years, non-invasive images such as magnetic resonance imaging (Magnetic Resonance Image, MRI), functional magnetic resonance imaging (fMRI), and diffusion tensor imaging (Diffusion Tensor Image, DTI) allow a doctor to intuitively understand the condition of a patient. The method widely applied to Alzheimer disease diagnosis at present is based on data analysis of brain images. Various machine learning algorithms are applied to image segmentation and other treatments on brain image data by researchers, and features such as cortex thickness, hippocampal volume, gray matter volume and the like in brain images are extracted as classification standards, so that Alzheimer disease is identified and diagnosed. However, such methods often only extract image surface layer information and do not correspond well to human brain mechanisms.
The brain network research method can break through the limitations of the traditional analysis technology in various aspects. The structure connection network and the brain function connection network contain rich brain structure information and function information. It has been found that the structural and functional connection networks constructed based on brain images of Alzheimer's patients undergo abnormal topological changes. However, the structural connection network and the functional connection network can only indicate whether or not there is a connection relationship between brain regions and the strength of the connection relationship, and have no directivity.
The effect connection network is used as a directed graph model, can display causal effect information of the connection relation between brain nerve signals, and display the information transmission mode between brain regions, and is more biological and interpretable. The high-level features of the multi-mode medical image fusion MRI, fMRI, DTI in recent years can provide complementary information of different modes, and avoid the information which is only contained in single-mode data and is related to abnormal parts. However, the current effect connection networks are centered on nodes and cannot reflect the common fluctuation mode between the connected edges, and no researcher has proposed a brain network capable of reflecting the causal relationship between the brain network connections.
Disclosure of Invention
The invention provides a diagnosis method of Alzheimer's disease based on an effect connection network with edges as centers, which aims to solve the technical problems that early diagnosis of Alzheimer's disease is difficult and diagnosis indexes are difficult to find.
In order to solve the technical problems, the invention provides the following technical scheme:
in one aspect, the present invention provides a method for diagnosing alzheimer's disease based on an edge-centric effect connection network, the method comprising:
acquiring multi-mode brain image data of a tested person, and obtaining a tested blood oxygen level dependence (Blood Oxygenation Level Dependent, BOLD) time sequence, a functional connection network and a structural connection network based on the multi-mode brain image data;
based on the inhibition relation between the structural connection network and the functional connection network, improving the conditional gracile cause and effect algorithm to obtain an improved gracile cause and effect algorithm, and based on the BOLD time sequence and the structural connection network, constructing an effect connection network with edges as the center by utilizing the improved gracile cause and effect algorithm;
and (3) connecting a network based on the effect taking the edge as the center, and completing Alzheimer disease identification by using a preset classifier.
Further, the multi-modality brain image data includes MRI data, fMRI data, and DTI data.
Further, based on the multi-mode brain image data, obtaining a BOLD time sequence, a functional connection network and a structural connection network to be tested, including:
performing time layer correction, head motion correction, spatial standardization, smooth filtering and brain region division operation based on an anatomic automatic labeling (Anatomical Automatic Labeling, AAL) template on the tested MRI data and fMRI data, dividing a human brain into a plurality of brain regions and obtaining a tested BOLD time sequence, and obtaining a functional connection network by calculating a Pearson correlation coefficient of each brain region;
performing skull removal, head movement correction, vortex correction, spatial standardization, smooth filtering and brain region division operation based on an AAL template on the DTI data to be tested, and defining white matter fiber connection between brain regions through the solved anisotropy value; wherein, the physical meaning of the anisotropy value indicates the intensity of the dispersion degree, and the number and the density of the white matter fibers can be reflected by the magnitude of the anisotropy value; and removing irrelevant anisotropic values by using a probabilistic fiber tracking algorithm to obtain a tested structural connection network.
Further, the functional connection network and the structural connection network are brain network expression forms centering on nodes, and the functional connection network and the structural connection network both take brain areas as nodes, wherein the structural connection network takes structural connection strength between brain areas as edges, and the functional connection network takes functional connection strength between brain areas as edges.
Further, the method for improving the conditional glaring causal algorithm based on the inhibition relation between the structural connection network and the functional connection network to obtain an improved glaring causal algorithm, and constructing an edge-centered effect connection network based on the BOLD time sequence and the structural connection network by using the improved glaring causal algorithm comprises the following steps:
constructing an edge time sequence based on the BOLD time sequence;
constructing an edge structure connection network based on the structure connection network;
based on the inhibition relation between the structural connection network and the functional connection network, introducing the structural connection network as a parameter matrix into the Grangel causal algorithm to obtain an improved Grangel causal algorithm;
substituting the constructed edge time sequence and the edge structure connection network into an improved Grangel causal algorithm to obtain an effect connection network taking edges as centers; wherein the nodes of the edge-centric effect connection network represent connections between two brain regions, and the edges of the edge-centric effect connection network represent causal relationships between the connections between two brain regions and the connections between two other brain regions, which are directional.
Further, the method for completing Alzheimer's disease identification by using a preset classifier based on the edge-centered effect connection network comprises the following steps:
taking the edge weight of the constructed edge-centered effect connection network as a characteristic, and selecting the characteristic;
and sending the selected characteristics into a preset classifier to finish the identification of the Alzheimer disease.
Further, the characterizing the edge weights of the constructed edge-centric effect connection network, and performing feature selection, includes:
and taking the edge weight of the constructed edge-centered effect connection network as a characteristic, and selecting the characteristic based on a maximum correlation minimum redundancy algorithm.
Further, the preset classifier is a support vector machine (Support Vector Machine, SVM) classifier with radial basis function.
In yet another aspect, the present invention also provides an electronic device including a processor and a memory; wherein the memory stores at least one instruction that is loaded and executed by the processor to implement the above-described method.
In yet another aspect, the present invention also provides a computer readable storage medium having at least one instruction stored therein, the instruction being loaded and executed by a processor to implement the above method.
The technical scheme provided by the invention has the beneficial effects that at least:
the invention adopts the multi-mode brain image data, overcomes the defect brought by single data, enables the constructed brain network to be more accurate, and the effect connection network which is constructed by taking the edge as the center can better reflect the causal relationship between brain network connection, has strong biological interpretability, can provide a certain thought for the research of the pathological mechanism of the Alzheimer's disease by applying the multi-mode brain image data to the diagnosis and identification task of the Alzheimer's disease, greatly improves the screening accuracy, is beneficial to solving the problems of difficult early diagnosis and difficult searching of diagnosis indexes of the Alzheimer's disease, and has promotion effect on searching the diagnosis standard of the Alzheimer's disease. The invention enriches the knowledge in the field of brain network construction and provides an important thought for brain network research application in treating difficult brain diseases.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of an alzheimer's disease diagnosis method based on an edge-centric effect connection network according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the embodiments of the present invention will be described in further detail with reference to the accompanying drawings.
First embodiment
Aiming at the problems that the early diagnosis of the Alzheimer's disease is difficult and the diagnosis index is difficult to find, the embodiment establishes an effect connection network with the center by taking the edge as the center with higher biological interpretability by preprocessing the brain image data of a tested person, further provides an Alzheimer's disease diagnosis method based on the effect connection network with the center by taking the edge, and greatly improves the screening accuracy by using a machine learning method to classify and identify the Alzheimer's disease, thereby being beneficial to solving the problems that the early diagnosis of the Alzheimer's disease is difficult and the diagnosis index is difficult to find and being beneficial to the development of the brain network field. The method may be implemented by an electronic device, which may be a terminal or a server. The execution flow of the method is shown in fig. 1, and comprises the following steps:
1. a brain image data preprocessing step, comprising:
acquiring multi-mode brain image data of a tested, and acquiring a BOLD time sequence, a functional connection network and a structural connection network of the tested based on the multi-mode brain image data;
specifically, in this embodiment, the multi-modal brain image data includes MRI data, fMRI data, and DTI data, and in the brain image data preprocessing step, the acquired multi-modal brain images of the four stages of the alzheimer's disease test, the normal person, the early mild cognitive impairment, the late mild cognitive impairment, and the alzheimer's disease test need to be processed. The format of the multi-modal brain images is first converted to the nifi format. The MRI and fMRI data to be tested are processed, and the data processing is carried out on a DPARSF kit, and the main steps are as follows: the human brain can be divided into 116 brain areas and a BOLD time sequence tested in each stage can be obtained by performing time layer correction, head movement correction, spatial standardization, smooth filtering and brain area division operation based on an AAL template, and a functional connection network can be obtained by calculating the Pearson correlation coefficient of each brain area. Then, the tested DTI data is processed, and the processing process is completed in a PANDA tool kit, and the main steps are as follows: performing skull removal, head movement correction, vortex correction, spatial standardization, smooth filtering and brain region division operation based on an AAL template on the tested DTI, defining white matter fiber connection between brain regions through solved anisotropy values FA, wherein the physical meaning of the FA values represents the intensity of dispersion degree, the number and density of the white matter fibers can be reflected through the size of the FA values, and irrelevant anisotropy values are removed through a probabilistic fiber tracking algorithm to obtain the tested structural connection network.
The functional connection network and the structural connection network are brain network expression forms taking nodes as centers, brain areas are taken as nodes, structural connection strength or functional connection strength between the brain areas is taken as an edge, and the brain network is a traditional brain network and is widely used at present.
2. An edge-centric effect connection network construction step comprising:
based on the inhibition relation between the structural connection network and the functional connection network, improving the conditional gracile cause and effect algorithm to obtain an improved gracile cause and effect algorithm, and based on the BOLD time sequence and the structural connection network, constructing an effect connection network with edges as the center by utilizing the improved gracile cause and effect algorithm;
specifically, in this embodiment, in the edge-centric effect connection network building step, in order to build an edge-centric brain network, an edge time sequence needs to be built first, specifically as follows:
let the number of subjects be M and the BOLD time series of the mth subject be described asWherein N and T represent the number of brain regions and the length of the time signal sequence, respectively. Let->I and j represent the ith and jth brain regions, s, respectively, in the brain network i Time series representing the ith brain region in the mth test BOLD time series, s j Time series representing the jth brain region in the mth test BOLD time series, s i (T) represents a T-th time point in the time series in which an i-th brain region is located in the m-th subject BOLD time series, t=1, 2, T; s is(s) j (t) represents the t-th time point of the time series in which the j-th brain region in the m-th subject BOLD time series is located. Then we obtained R after treatment with z-score i and Rj And calculates their pairwise products:
wherein , Ri (t) represents the t-th time point at which the ith brain region of the BOLD time series after z-score treatment is located, R j (t) represents the t-th time point X at which the j-th brain region of the BOLD time series after z-score treatment is located ij For the time series of the product of the ith brain region and the jth brain region, referred to herein as the edge time series, X ij (t) the t time point X of the time series of the ith brain region and the jth brain region uv (t) isThe nth brain region and the nth brain region border time series. EFCN ij,uv Representing the edge function connection strength of the order pair of the ith and jth brain regions and the order pair of the ith and jth brain regions. From EFCN ij,uv The network is an edge-centric functional connectivity network EFCN which corresponds in structure to the edge-centric effect connectivity network set forth herein, of the size。
Then constructing an edge structure connection network, which is specifically as follows:
wherein , SCij Representing structural connection strength of ith brain region and jth brain region, SC uv The structural connection strength of the ith brain region and the jth brain region is represented, and i, j, u and v represent brain region index numbers.Represents the edge structure connection strength of the sequence pair of the ith and the jth brain regions and the sequence pair of the ith and the jth brain regions, which is composed of +.>The network is the edge structure connection network W corresponding to the edge-centered effect connection network, and W is a symmetric matrix. To avoid matrix irreversibility problems caused by singular matrices, a value of [10 ] is added to W -3 ,10 -4 ]Noise in the same, and symmetric processing is performed on W.
In order to fully utilize the complementary relation of the multi-mode brain images and construct a more accurate effect connection network, the embodiment creatively introduces the structure connection network as a parameter matrix into a glabellar causal algorithm based on the inhibition relation of the structure connection network to the functional connection network, wherein the specific algorithm is as follows:
let U be the multivariate stationary time series:
,
wherein ,Ut Representing a multiple stationary time sequence at time t, U t-i Representing a multi-plateau time sequence at time t-i,representing U t Auto-covariance matrix of A i Represents regression coefficients, q represents model order, +.>Representing the error. Since the effect connection network is essentially a functional connection network, the structure connection network still has a limiting effect on it, and the introduction of the edge structure connection network results in:
wherein ,representing a multiplex stationary time sequence U when i=0 t Is a self-covariance matrix of (1);
from the Yule-Walker equation we can get:
wherein ,auto-covariance matrix, A, each representing a multivariate stationary time series 1 …A q All are regression coefficients.
Namely:
specifically, let x, y, z be a subset of the multivariate stationary time series U:
wherein ,xt ，x t-i ，y t-i ，z t-i Are all stationary time sequences, q is the model order,for model coefficients +.>Is a random error term that is independent of time. The grange cause and effect from y to x under z conditions (given z) is as follows:
wherein ,representing the graininess causal relationship of y to x under the condition of z, var represents a vector autoregressive model, when var (ε 1t ) > var (ε 2t ) At this time, we consider that introducing y-variables under the condition of z leads to an improvement in the accuracy of the prediction of the x-value, at this time +.>Y versus x is the cause of glaring. When the cause and effect of Grandigo between x, y is mediated entirely by z, then +.>， var(ε 1t ) = var(ε 2t ) Thus->We consider that under the condition z, y does not have a glabellar cause for x. Finally, whether the causal relationship is significant or not is verified through F test.
Substituting the constructed edge time sequence and edge structure connection network into improved Granges causal algorithm to obtain edge-centered effect connection network, wherein the node represents the connection between two brain regions, the edge represents the causal relationship between the connection between two brain regions and the connection between the other two brain regions, the obtained edge-centered effect connection network has a size of。
3. A feature extraction and selection step comprising:
taking the edge weight of the constructed edge-centered effect connection network as a characteristic, and selecting the characteristic; specifically, in this embodiment, the feature selection is performed by using a maximum correlation minimum redundancy algorithm, which is specifically as follows:
taking the edge weight of the constructed edge-centered effect connection network as the characteristic, selecting the number of brain areas as N, and the number of nodes in the edge-centered effect connection network should be N (N-1)/2, sharingAnd (5) a strip edge. To solve the problem of high-dimensional small samples, the present embodiment performs feature selection based on a maximum correlation minimum redundancy algorithm (mRMR). mRMR can use mutual information as a metric to address the tradeoff between feature redundancy and relevance. Mutual information between variables l1, l2The method comprises the following steps:
in the formula ,、/>probability densities of l1, l2, respectively,/->For the joint probability density of l1, l2, the resulting value is a number between 0 and 1, the larger this value is, the more information the two variables contain in common, the smaller the mutual information is, the smaller the information intersection between the two variables is.
Maximum correlationThe definition is as follows:
here the number of the elements is the number,representing a feature set with m features, D being the mutual information values between the attribute subsets, c being the category, +.>Representing the ith feature, ++>Representation->And->Mutual information between them.
Maximum correlationThe result is a value between 0 and 1, the greater the result, the more closely the relationship between the feature and the class label, and conversely the lower the degree of correlation.
Minimum redundancyThe definition is as follows:
wherein ,representation feature->Mutual information between them. The result of the minimum redundancy is a certain value between 0 and 1, the larger the result, the more relevant the two features, the greater the redundancy between them, and vice versa the smaller the redundancy. Finally, selecting a feature subset with the largest correlation metric and the smallest redundancy metric, wherein the feature subset is as follows:
4. an alzheimer's disease diagnosis step comprising:
and sending the selected characteristics into a preset classifier to finish the identification of the Alzheimer disease.
Specifically, in this embodiment, in the step of diagnosing alzheimer's disease, the features selected by the step of feature extraction and feature selection are sent to a preset classifier to perform training optimization, so as to be used for classifying alzheimer's disease. The classifier adopted in the embodiment is an SVM model with a radial basis function, a 10-fold cross validation strategy is adopted to find optimal parameters, in order to reduce experimental errors, 200 repeated experiments are carried out on each test, and an average result is used as a final result to verify the effectiveness of the method.
In summary, the present embodiment designs a diagnosis method for alzheimer's disease based on an edge-centric effect connection network, which can help doctors to assist diagnosis, and discover diseases earlier and more accurately. Meanwhile, from the perspective of neuroscience, the edge-centered effect connection network provided by the embodiment can be regarded as extension and expansion of a traditional brain network, so that the effect connection network more reflects how communication between different areas of the brain evolves and researches on causal relations among the areas, and has good inspired significance for excavating etiology and mechanism of difficult brain diseases such as Alzheimer disease and overcoming the problem of brain black boxes.
Second embodiment
The embodiment provides an electronic device, which comprises a processor and a memory; wherein the memory stores at least one instruction that is loaded and executed by the processor to implement the method of the first embodiment.
The electronic device may vary considerably in configuration or performance and may include one or more processors (central processing units, CPU) and one or more memories having at least one instruction stored therein that is loaded by the processors and performs the methods described above.
Third embodiment
The present embodiment provides a computer-readable storage medium having stored therein at least one instruction that is loaded and executed by a processor to implement the method of the first embodiment described above. The computer readable storage medium may be, among other things, ROM, random access memory, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc. The instructions stored therein may be loaded by a processor in the terminal and perform the methods described above.
Furthermore, it should be noted that the present invention can be provided as a method, an apparatus, or a computer program product. Accordingly, embodiments of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the invention may take the form of a computer program product on one or more computer-usable storage media having computer-usable program code embodied therein.
Embodiments of the present invention are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It should also be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or terminal device comprising the element.
It is finally pointed out that the above description of the preferred embodiments of the invention, it being understood that although preferred embodiments of the invention have been described, it will be obvious to those skilled in the art that, once the basic inventive concepts of the invention are known, several modifications and adaptations can be made without departing from the principles of the invention, and these modifications and adaptations are intended to be within the scope of the invention. It is therefore intended that the following claims be interpreted as including the preferred embodiment and all such alterations and modifications as fall within the scope of the embodiments of the invention.
Claims (7)
1. A method for diagnosing alzheimer's disease based on an edge-centric effect connection network, the method comprising:
acquiring multi-mode brain image data of a tested person, and acquiring a BOLD time sequence, a functional connection network and a structural connection network of the tested person based on the multi-mode brain image data;
based on the inhibition relation between the structural connection network and the functional connection network, improving the conditional gracile cause and effect algorithm to obtain an improved gracile cause and effect algorithm, and based on the BOLD time sequence and the structural connection network, constructing an effect connection network with edges as the center by utilizing the improved gracile cause and effect algorithm;
based on the effect connection network with the edges as the center, completing Alzheimer disease identification by using a preset classifier;
the method for improving the conditional gracile cause and effect algorithm based on the inhibition relation between the structural connection network and the functional connection network to obtain an improved gracile cause and effect algorithm, based on the BOLD time sequence and the structural connection network, utilizes the improved gracile cause and effect algorithm to construct an effect connection network with edges as centers, and comprises the following steps:
constructing an edge time sequence based on the BOLD time sequence; the method comprises the following steps:
let the tested quantity beMFirst, themThe individual BOLD time series tested are described as:，m=1, 2, …, M; wherein,NandTrepresenting the number of brain regions and the length of the time signal sequence, respectively; order the, s i Represent the firstmThe first time series of the tested BOLDiA time series of individual brain regions,s j represent the firstmThe first time series of the tested BOLDjA time series of individual brain regions,s i (t) Represent the firstmThe first time series of the tested BOLDiTime series of the brain regiontThe time point at which the time point is the same,t=1,2,...,T；s j (t) Represent the firstmThe first time series of the tested BOLDjTime series of the brain regiontTime points; then, after treatment with z-score, obtainR i AndR j and calculates their pairwise products:
;
；
wherein ,R i (t) Representing time series of BOLD after z-score processingiThe first brain regiontThe time point at which the time point is the same,R j (t) Representing time series of BOLD after z-score processingjThe first brain regiontAt each time point ,X ij Is the firstiBrain region and thjThe product of the individual brain regions is used to determine the time series, i.e. the edge time series,X ij (t) is the firstiBrain region and thjThe t-th time point of the temporal sequence of the sides of the brain region,X uv (t) is the firstuBrain region and thvT time points of the brain region border time sequence; EFCN ij,uv Represent the firstiAnd (d)jThe sequence of the brain regions is the same as the sequence of the brain regionsuAnd (d)vThe edge function connection strength of the sequence pair where the brain regions are located; from EFCN ij,uv The network is an edge-centric functional connection network EFCN which corresponds in structure to the edge-centric effect connection network and has a size of；
Constructing an edge structure connection network based on the structure connection network; the method comprises the following steps:
；
wherein ,SC ij represent the firstiBrain region and thjStructural connection strength of individual brain regions，SC uv Represent the firstuBrain region and thvStructural connection strength of the individual brain regions,i, j, u, vall represent brain region index numbers;represent the firstiAnd (d)jThe sequence of the brain regions is the same as the sequence of the brain regionsuAnd (d)vThe connection strength of the edge structure of the sequence pair of the brain regions is defined by +.>The network is a side structure connection network corresponding to the effect connection network centered on the sideWThe edge structure connection network is a symmetrical matrix; in order to avoid matrix irreversibility caused by singular matrix, inWIs added with a value of [10 ] -3 ,10 -4 ]Noise in between, and toWPerforming alignmentWeighing;
based on the inhibition relation between the structural connection network and the functional connection network, introducing the structural connection network as a parameter matrix into the Grangel causal algorithm to obtain an improved Grangel causal algorithm; the specific algorithm is as follows:
is provided withUIs a multiple stationary time sequence:
；
wherein ,U t represent the firsttA multi-plateau time series of time points,U t-i represent the firstt-iA multi-plateau time series of time points,showing theU t Is a matrix of auto-covariance of (c),A i the regression coefficient is represented as a function of the regression coefficient,qrepresenting model order, +.> t Representing the error; since the effect connection network is essentially a functional connection network, the structure connection network still has a limiting effect on it, and the introduction of the edge structure connection network results in:
；
；
wherein ,representation ofiWhen=0, the multiplex stationary time sequenceU t Is a self-covariance matrix of (1);
according to the Yule-Walker equation, we get:
；
wherein ,each representing an autocovariance matrix of the multivariate stationary time series,A 1 …A q are regression coefficients;
namely:；
；
specifically, it is provided withx, y, zIs a multiple stationary time seriesUA subset of (a):
；
；
；
wherein ,x t ，x t-i ，y t-i ，z t-i are all in a steady time sequence, and the time sequence,qfor the order of the model, the number of the model,as the coefficients of the model,，/>is a random error term independent of time; at the position ofzFrom under conditions ofyTo the point ofxThe gracile cause and effect of (c) is as follows:
；
wherein ,is shown inzUnder the condition of (2)，yFor the followingxIs a Grangel causal relationship, var represents a vector autoregressive model, when var is @ε t1 ) > var (ε t2 ) At the time, consider to bezIs introduced under the condition ofyThe variables being such that the pairxThe prediction accuracy of the value is improved, in this case +.>，yFor a pair ofxThere is a gland's cause; when (when)x, yThe cause and effect of the Granges are completely determined byzAt the time of mediation, at this time->，var(ε t1 ) = var(ε t2 ) Thus->Under the condition ofzThe lower part of the upper part is provided with a lower part,yno pair is presentxIs of the Grangel cause;
substituting the constructed edge time sequence and the edge structure connection network into an improved Grangel causal algorithm to obtain an effect connection network taking edges as centers; wherein the nodes of the edge-centric effect connection network represent connections between two brain regions, and the edges of the edge-centric effect connection network represent causal relationships between the connections between two brain regions and the connections between two other brain regions, which are directional.
2. The method for diagnosing alzheimer's disease based on edge-centric effect connected networks of claim 1, wherein said multi-modality brain imaging data includes magnetic resonance imaging MRI data, functional magnetic resonance imaging fMRI data, and diffusion tensor imaging DTI data.
3. The method for diagnosing alzheimer's disease based on edge-centric effector connectivity network according to claim 2, wherein obtaining a blood oxygen level dependent BOLD time series, functional connectivity network and structural connectivity network of the subject based on the multi-modal brain image data comprises:
performing time layer correction, head motion correction, spatial standardization, smooth filtering and brain region division operation based on an anatomical automatic labeling AAL template on the tested MRI data and fMRI data, dividing a human brain into a plurality of brain regions, obtaining a tested BOLD time sequence, and obtaining a functional connection network by calculating a pearson correlation coefficient of each brain region;
performing skull removal, head movement correction, vortex correction, spatial standardization, smooth filtering and brain region division operation based on an AAL template on the DTI data to be tested, and defining white matter fiber connection between brain regions through the solved anisotropy value; wherein, the physical meaning of the anisotropy value indicates the intensity of the dispersion degree, and the number and the density of the white matter fibers can be reflected by the magnitude of the anisotropy value; and removing irrelevant anisotropic values by using a probabilistic fiber tracking algorithm to obtain a tested structural connection network.
4. The method for diagnosing alzheimer's disease based on edge-centric effector connectivity network according to claim 1, wherein said functional connectivity network and structural connectivity network are both node-centric brain network representations, said functional connectivity network and structural connectivity network each having brain regions as nodes, wherein said structural connectivity network has structural connection strength between brain regions as edges, and said functional connectivity network has functional connection strength between brain regions as edges.
5. The method for diagnosing alzheimer's disease based on edge-centric effector connection network of claim 1, wherein said edge-centric effector connection network based on utilizes a predetermined classifier to accomplish alzheimer's disease identification, comprising:
taking the edge weight of the constructed edge-centered effect connection network as a characteristic, and selecting the characteristic;
and sending the selected characteristics into a preset classifier to finish the identification of the Alzheimer disease.
6. The method for diagnosing alzheimer's disease based on edge-centric effect connected networks of claim 5, wherein said characterizing the edge weights of the structured edge-centric effect connected networks and selecting features comprises:
and taking the edge weight of the constructed edge-centered effect connection network as a characteristic, and selecting the characteristic based on a maximum correlation minimum redundancy algorithm.
7. The method for diagnosing alzheimer's disease based on edge-centric effect connected networks of claim 5, wherein said predetermined classifier is a support vector machine SVM classifier with radial basis function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310329085.7A CN116030941B (en) | 2023-03-30 | 2023-03-30 | Alzheimer's disease diagnosis method based on edge-centric effect connection network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310329085.7A CN116030941B (en) | 2023-03-30 | 2023-03-30 | Alzheimer's disease diagnosis method based on edge-centric effect connection network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116030941A CN116030941A (en) | 2023-04-28 |
| CN116030941B true CN116030941B (en) | 2023-08-25 |
Family
ID=86070792
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310329085.7A Active CN116030941B (en) | 2023-03-30 | 2023-03-30 | Alzheimer's disease diagnosis method based on edge-centric effect connection network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116030941B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9107595B1 (en) * | 2014-09-29 | 2015-08-18 | The United States Of America As Represented By The Secretary Of The Army | Node excitation driving function measures for cerebral cortex network analysis of electroencephalograms |
| CN109034360A (en) * | 2018-07-13 | 2018-12-18 | 北京工业大学 | A kind of ant colony method constructing brain effective connectivity network from fMRI and DTI data |
| CN113947157A (en) * | 2021-10-26 | 2022-01-18 | 北京工业大学 | Dynamic brain effect connection network generation method based on hierarchical clustering and structural equation model |
| CN114036736A (en) * | 2021-11-05 | 2022-02-11 | 大连理工大学 | Cause and effect network learning method based on local granger cause and effect analysis |
| CN114188013A (en) * | 2021-09-01 | 2022-03-15 | 北京智精灵科技有限公司 | Cognitive and brain image data integration evaluation method for Alzheimer's disease |
| CN115005798A (en) * | 2022-06-02 | 2022-09-06 | 四川大学 | Brain image feature extraction method based on edge connection function connection |
| CN115778323A (en) * | 2022-11-21 | 2023-03-14 | 山东大学 | User rehabilitation level assessment system with multi-source data fusion |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016200952A1 (en) * | 2015-06-09 | 2016-12-15 | Children's Medical Center Corporation | Method and system for locating seizure focus from interictal data |
-
2023
- 2023-03-30 CN CN202310329085.7A patent/CN116030941B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9107595B1 (en) * | 2014-09-29 | 2015-08-18 | The United States Of America As Represented By The Secretary Of The Army | Node excitation driving function measures for cerebral cortex network analysis of electroencephalograms |
| CN109034360A (en) * | 2018-07-13 | 2018-12-18 | 北京工业大学 | A kind of ant colony method constructing brain effective connectivity network from fMRI and DTI data |
| CN114188013A (en) * | 2021-09-01 | 2022-03-15 | 北京智精灵科技有限公司 | Cognitive and brain image data integration evaluation method for Alzheimer's disease |
| CN113947157A (en) * | 2021-10-26 | 2022-01-18 | 北京工业大学 | Dynamic brain effect connection network generation method based on hierarchical clustering and structural equation model |
| CN114036736A (en) * | 2021-11-05 | 2022-02-11 | 大连理工大学 | Cause and effect network learning method based on local granger cause and effect analysis |
| CN115005798A (en) * | 2022-06-02 | 2022-09-06 | 四川大学 | Brain image feature extraction method based on edge connection function connection |
| CN115778323A (en) * | 2022-11-21 | 2023-03-14 | 山东大学 | User rehabilitation level assessment system with multi-source data fusion |
Non-Patent Citations (1)
| Title |
|---|
| 尚志刚 ; 沈晓阳 ; 李蒙蒙 ; 万红 ; .基于格兰杰因果的效应性连接分析方法综述.郑州大学学报(工学版).2020,41(第03期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116030941A (en) | 2023-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yao et al. | A mutual multi-scale triplet graph convolutional network for classification of brain disorders using functional or structural connectivity | |
| Zeng et al. | A new deep belief network-based multi-task learning for diagnosis of Alzheimer’s disease | |
| CN113616184B (en) | Brain network modeling and individual prediction method based on multi-mode magnetic resonance image | |
| Amico et al. | Mapping hybrid functional-structural connectivity traits in the human connectome | |
| CN109528197B (en) | Individual prediction method and system for mental diseases based on brain function map | |
| CN111863244B (en) | Functional connection mental disease classification method and system based on sparse pooling graph convolution | |
| Liu et al. | An enhanced multi-modal brain graph network for classifying neuropsychiatric disorders | |
| CN116503680B (en) | Brain image structured analysis and brain disease classification system based on brain atlas | |
| CN109344889A (en) | A kind of cerebral disease classification method, device and user terminal | |
| CN113936172B (en) | Disease classification method and device based on integrated learning fusion multi-mode features | |
| Chen et al. | Recent advances of deep learning in psychiatric disorders | |
| CN115272295A (en) | Dynamic brain function network analysis method and system based on time domain-space domain combined state | |
| CN113786185A (en) | Static brain network feature extraction method and system based on convolutional neural network | |
| CN116959714A (en) | Autism spectrum disorder classification method, electronic equipment and storage medium | |
| CN116843667A (en) | Brain image classification and brain cognition scoring prediction method based on multitask learning | |
| Alharthi et al. | Do it the transformer way: a comprehensive review of brain and vision transformers for autism spectrum disorder diagnosis and classification | |
| CN111488934A (en) | Brain image data processing method, storage medium, computer device and apparatus | |
| Xi et al. | Brain Functional Networks with Dynamic Hypergraph Manifold Regularization for Classification of End-Stage Renal Disease Associated with Mild Cognitive Impairment. | |
| CN116759096B (en) | Disease prediction system based on multiple comparison function connection matrix | |
| Patil et al. | Identification of brain white matter regions for diagnosis of Alzheimer using diffusion tensor imaging | |
| CN117612735A (en) | Structure-constrained white matter function network construction method and system | |
| CN117765530A (en) | Multi-mode brain network classification method, system, electronic equipment and medium | |
| Li et al. | Diagnosis of Alzheimer's disease by feature weighted-LSTM: a preliminary study of temporal features in brain resting-state fMRI | |
| CN116030941B (en) | Alzheimer's disease diagnosis method based on edge-centric effect connection network | |
| CN116468677A (en) | MCI conversion intelligent prediction method and system based on brain network and deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |