CN110797123B - Graph convolution neural network evolution method of dynamic brain structure - Google Patents
Graph convolution neural network evolution method of dynamic brain structure Download PDFInfo
- Publication number
- CN110797123B CN110797123B CN201911033766.9A CN201911033766A CN110797123B CN 110797123 B CN110797123 B CN 110797123B CN 201911033766 A CN201911033766 A CN 201911033766A CN 110797123 B CN110797123 B CN 110797123B
- Authority
- CN
- China
- Prior art keywords
- brain
- network
- evolution
- brain structure
- neural 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
- 210000004556 brain Anatomy 0.000 title claims abstract description 134
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 20
- 239000013598 vector Substances 0.000 claims abstract description 22
- 230000008569 process Effects 0.000 claims abstract description 15
- 238000003062 neural network model Methods 0.000 claims abstract description 10
- 239000011159 matrix material Substances 0.000 claims description 25
- 210000004884 grey matter Anatomy 0.000 claims description 10
- 230000000994 depressogenic effect Effects 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 238000009499 grossing Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000012300 Sequence Analysis Methods 0.000 claims description 3
- 210000001175 cerebrospinal fluid Anatomy 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 210000001519 tissue Anatomy 0.000 claims description 3
- 230000001737 promoting effect Effects 0.000 claims description 2
- 238000012549 training Methods 0.000 claims description 2
- 210000004027 cell Anatomy 0.000 claims 1
- 238000010219 correlation analysis Methods 0.000 claims 1
- 230000001054 cortical effect Effects 0.000 claims 1
- KJONHKAYOJNZEC-UHFFFAOYSA-N nitrazepam Chemical compound C12=CC([N+](=O)[O-])=CC=C2NC(=O)CN=C1C1=CC=CC=C1 KJONHKAYOJNZEC-UHFFFAOYSA-N 0.000 claims 1
- 238000013135 deep learning Methods 0.000 abstract description 2
- 238000005457 optimization Methods 0.000 abstract 1
- 208000024714 major depressive disease Diseases 0.000 description 6
- 230000006870 function Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000001575 pathological effect Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000010220 Pearson correlation analysis Methods 0.000 description 2
- 210000004885 white matter Anatomy 0.000 description 2
- 208000004547 Hallucinations Diseases 0.000 description 1
- 208000019022 Mood disease Diseases 0.000 description 1
- 206010065604 Suicidal behaviour Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003001 depressive effect Effects 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 230000036285 pathological change Effects 0.000 description 1
- 231100000915 pathological change Toxicity 0.000 description 1
- 201000000980 schizophrenia Diseases 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- General Engineering & Computer Science (AREA)
- Epidemiology (AREA)
- Pathology (AREA)
- Primary Health Care (AREA)
- Databases & Information Systems (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a graph convolution neural network evolution method of a dynamic brain structure, which adopts a graph convolution neural network to simulate the evolution process of a normal human brain structure network to be changed into depression. The direction vector is introduced in the evolution process, the vector not only contains brain structure network information of normal people, but also contains brain structure network information of depression patients, the characteristics of the normal people and the depression patients can be extracted simultaneously through graph convolution operation, and the evolution direction and the evolution degree can be controlled. The invention provides a graph convolution neural network model of brain structure network evolution, which utilizes a deep learning method based on a tensorf low framework to ensure that the network evolution always proceeds towards the direction of the brain network of a patient suffering from depression by calculating the cross entropy of a first evolution result and the brain network of a real patient suffering from depression and utilizing an optimization method of gradient descent. And finally outputting a brain structure network which is close to a real depression patient, and obtaining an evolution model which is more close to the real network.
Description
Technical Field
The invention relates to a network evolution method, in particular to a graph convolution neural network evolution method of a dynamic brain structure.
Background
Major depression is a serious illness next to schizophrenia, and patients may be pessimistic, hopeless, delusional hallucinations, hypofunction, accompanied by serious suicidal attempts, and even suicidal behavior, constituting a serious threat to human health. At present, the pathological mechanism of depression is not clear, and quantitative biological indexes are still lacking to clearly define the depression. However, an important sign of major depression is represented by a mood disorder caused by a change in functional connectivity between brain regions. Therefore, brain network analysis is of great importance for studying the pathological mechanism and diagnostic method of major depressive disorder. Specifically, the pathological change process of the brain is simulated by constructing a brain structure network and establishing a dynamic evolution model of the network. The method provides a new idea for exploring the pathological mechanism of depression.
At present, a plurality of domestic and foreign researchers discover that compared with normal people, the static topological characteristics (such as clustering coefficients, degree distribution and the like) of the functional connection network of the brain of the patient suffering from major depression are obviously changed by combining complex network theory and brain science knowledge. However, these changes only reveal the difference between healthier people with the patient at a specific moment, and the disease consequences are found but the whole disease occurrence process cannot be traced. The defect can be made up to a certain extent by researching the dynamic evolution process of the brain network of the patients suffering from major depression.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a graph convolution neural network evolution method capable of tracing the dynamic brain structure of the whole disease occurrence process.
In order to simulate the dynamic evolution process of the brain network of a patient suffering from major depressive disorder, the technical scheme of the invention is as follows: a method for evolving a graph convolution neural network of a dynamic brain structure constructs a graph convolution neural network model of brain structure network evolution, and learns the dynamic evolution process of the brain network by using a graph convolution neural network tool, wherein the method comprises the following specific steps:
A. data preprocessing
The brain structure magnetic resonance images were acquired using a simon 1.5T magnetic resonance scanner, an 8-channel head coil, and then preprocessed using a brain image data sequence analysis tool (SPM) in matrix laboratory (MATLAB).
A1, standardization: the brain structure magnetic resonance image of each individual is standardized to the same three-dimensional space, so that the same brain region of different individuals is positioned at the same position of the standard space.
A2, segmentation: according to the prior probability distribution template, the brain structure magnetic resonance image is divided into three parts of gray matter, white matter and cerebrospinal fluid, and images representing the three tissue types are generated, and the volumes of the gray matter and the white matter of the brain are quantitatively represented.
A3, smoothing: the gray image generated by the segmentation is subjected to spatial smoothing, so that noise and errors generated in the standardization and segmentation process are reduced, and the signal-to-noise ratio is improved, thereby improving the effectiveness of statistical analysis.
B. Construction of brain structure network
Based on the brain structure magnetic resonance image, 90 brain regions in the automatic anatomical label template AAL are selected as nodes in the network, and the interrelationship or anatomical connection among the 90 brain regions is defined as an edge in the network. According to the calibrated 90 brain regions, extracting gray matter volume of the brain regions, and then evaluating the interrelation between two nodes by using Pearson correlation analysis, constructing a gray matter brain network, wherein the steps are as follows:
b1, extracting gray matter volumes of brain areas of 90 automatic anatomical label templates of each tested according to the gray matter images obtained by preprocessing and dividing the data.
And B2, estimating the mutual relation between the two brain areas by using a Pearson correlation analysis, and obtaining a correlation coefficient r between every two brain areas by calculating the Pearson correlation coefficient, thereby obtaining a 90 x 90 relation matrix A. The pearson correlation coefficient, also known as pearson product moment correlation coefficient, is a method for calculating a linear correlation, calculated by equation (1):
wherein ,rq,f Representing the correlation coefficient between brain regions q and f, x q(t) and respectively representing time series value and time series mean value of brain region q, x f(t) and />The time series value and the time series mean value of the brain region f are respectively represented, and n represents the number of the scanned brain cortex slices.
B3, determining the continuous edge. The value of r is between-1 and 1, the absolute value of r represents the strength of the correlation between every two brain areas, and the larger the value is, the larger the correlation degree is represented: r > 0.8 indicates a strong correlation between two brain regions; r < 0.3 indicates that there is a weak correlation between the two brain regions; r=0 indicates that the two brain regions are uncorrelated. The positive and negative of r means the direction of correlation between brain regions, i.e. r > 0 indicates that the two variables are in a mutually promoting relationship, or positive correlation; r < 0 indicates that the two variables are in a mutually resisting relationship, or are inversely related.
C. Graph convolution neural network model for establishing brain structure network evolution
The graph convolution neural network model for establishing brain structure network evolution is as follows: f (X, A) is a two-layer graph roll-up neural network, where A is the relationship matrix of the brain structure network and X is a direction vector representing the difference between each brain region of the brains of normal and depressed patients.
C1, respectively extracting gray matter volumes of 90 brain areas of normal people and depression patients through a data preprocessing step to respectively form column vectors X of 90X 1 1 and X2 Then define a direction vector x=x 1 -X 2 I.e. the difference in brain structure between normal and depressed patients, and is normalized as shown in formula (2):
where a is the minimum value of the elements in the direction vector X and b is the maximum value of the elements in the direction vector X.
And C2, carrying out the following normalization processing on the brain structure network relation matrix A, wherein the normalization processing is shown in a formula (3):
wherein ,is a relation matrix added with self-connection of brain structure network, I N Is an identity matrix> Representing the degree of each node in the brain structure network.
And C3, training a brain structure evolution model based on a two-layer graph convolution neural network, wherein the feedforward neural network model uses the following form:
wherein ,W(0) ∈R 90*90 For the weight matrix of the input layer to the hidden layer, W (1) ∈R 90*90 Is a weight matrix between the hidden layer and the output layer. softmax is an activation function, defined as wherein Z=∑i exp(x i )。
C4, evaluating the cross entropy loss function by using the relation matrix of all depression patients as a label sample, as shown in formula (5):
wherein ,yi Representing the resulting vector of the evolving network,a network vector representing a truly depressed patient.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention adopts the graph convolution neural network to simulate the evolution process of the normal human brain structural network to be changed into depression. The direction vector is introduced in the evolution process, the vector not only contains brain structure network information of normal people, but also contains brain structure network information of depression patients, the characteristics of the normal people and the depression patients can be extracted simultaneously through graph convolution operation, and the evolution direction and the evolution degree can be controlled.
2. The invention provides a graph convolution neural network model of brain structure network evolution, which uses a deep learning method based on a tensorsurface framework to normalize a correlation matrix and a direction vector of a normal human brain structure network as input of a first layer neural network, obtains an intermediate matrix through one layer of graph convolution, represents first-order adjacent node information of brain structure network nodes, and obtains output of the first layer convolution network through an activation function Relu. And then taking the output of the first layer as the input of a second layer neural network, further obtaining second-order neighbor node information of the current node through graph convolution, and outputting a first evolution result of the brain network. By calculating the cross entropy of the first evolution result and the brain network of the actual depression patient, the evolution of the network is always carried out towards the direction of the brain network of the depression patient by using the gradient descent optimizing method. The brain network after each evolution reveals the trend of illness from normal people to depression patients, and finally an evolution model which is closer to a real network is obtained. In the current related method, the realization is simpler, the data acquisition and processing are clear and concise, and the practicability is strong.
Drawings
The invention is illustrated in fig. 5, in which:
fig. 1 is a process of brain structure network construction.
Fig. 2 is a network diagram of a normal human brain structure.
Fig. 3 is a network diagram of brain structures of a depressive patient.
Fig. 4 is a network diagram of the resulting brain structure of the evolution model.
Fig. 5 is a flow chart of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The brain structure network evolution model of the invention is composed of a two-layer graph convolution neural network. As shown in fig. 1, brain structure magnetic resonance images are first preprocessed using matrix laboratory (MATLAB) brain image data sequence analysis tools (SPMs). Based on the gray matter volumes of brain areas of 90 automatic anatomical label templates obtained by data preprocessing, the pearson correlation coefficient between every two is calculated to obtain a brain structure network relation matrix A. The direction vector X between the brain structure networks of normal and depressed patients is constructed according to the gray matter volumes of 90 brain areas of normal and depressed patients. Next, a network of brain structure network evolution is obtained according to the flow shown in fig. 5. The correlation matrix and the direction vector of the normal human brain structure network (as shown in figure 2) are normalized and then used as the input of the first-layer graph convolution neural network. And obtaining an intermediate matrix through one layer of graph convolution, wherein the intermediate matrix represents first-order adjacent node information of brain structure network nodes, and then obtaining output of the first layer of convolution network through an activation function Relu. And then taking the output of the first layer as the input of a second layer neural network, further obtaining second-order neighbor node information of the current node through graph convolution, and outputting a first evolution result of the brain network. Finally, the model result and a true brain network (as shown in figure 3) of the depressed patient are subjected to cross entropy verification, and parameters W in the network are continuously optimized through back propagation (0) and W(1) The evolution of the network is always towards the patients with depressionThe direction of the brain network proceeds, and the final output approximates to the brain structure network of the real depression patient (as shown in fig. 4), and an evolution model which approximates to the real network is obtained.
The present invention is not limited to the present embodiment, and any equivalent concept or modification within the technical scope of the present invention is listed as the protection scope of the present invention.
Claims (1)
1. The method for evolving a graph convolution neural network of a dynamic brain structure constructs a graph convolution neural network model of brain structure network evolution, learns the dynamic evolution process of the brain network by using a graph convolution neural network tool, and is characterized in that: the method specifically comprises the following steps:
A. data preprocessing
Acquiring brain structure magnetic resonance images by using a Simon 1.5T magnetic resonance scanner and an 8-channel magnetic head coil, and preprocessing the brain structure magnetic resonance images by using a brain image data sequence analysis tool SPM in matrix laboratory MATLAB;
a1, standardization: normalizing the brain structure magnetic resonance image of each individual to the same three-dimensional space, so that the same brain region of different individuals is positioned at the same position of the standard space;
a2, segmentation: dividing the brain structure magnetic resonance image into three tissue types of grey brain matter, white brain matter and cerebrospinal fluid according to a priori probability distribution template, generating images representing the three tissue types, namely grey brain matter images, white brain matter images and cerebrospinal fluid images, and quantitatively representing the volumes of grey brain matter and white brain matter;
a3, smoothing: performing spatial smoothing on the brain gray image generated by segmentation;
B. construction of brain structure network
Based on the brain structure magnetic resonance image, selecting 90 brain regions in an automatic anatomical label template AAL as nodes in a network, and defining the interrelationship or anatomical connection among the 90 brain regions as edges in the network; according to the calibrated 90 brain regions, extracting the grey brain volume of the brain regions, evaluating the interrelation between two nodes by using Pelson correlation analysis, and constructing a grey brain network, wherein the steps are as follows:
b1, extracting the grey brain matter volumes of the 90 brain areas of each individual according to the grey brain matter images obtained by preprocessing the data;
b2, calculating the Pearson correlation coefficient through the formula (1) to obtain the correlation coefficient between every two brain areasObtaining a singleRelation matrix of brain structure network->:
wherein ,representing brain region-> and />Correlation coefficient between-> and />Respectively represent brain region->Time series value and time series mean value, +.> and />Respectively represent brain region->Time series value and time series mean value, +.>Representing the number of scanned cortical slices;
b3, determining the continuous edge;the value of (2) is->Between (I)>The absolute value of (2) represents the intensity of the correlation between brain regions, and the larger the value is, the greater the degree of correlation is: />Representing strong correlation between two brain regions;indicating that there is a weak correlation between the two brain regions; />Indicating that the two brain regions are uncorrelated; />Positive and negative meaning the direction of correlation between brain regions, i.e. +.>Representing that the two variables are in a mutually promoting relationship or positive correlation; />Indicating that the two variables are in a mutually resisting relationship or negative correlation;
C. graph convolution neural network model for establishing brain structure network evolution
The graph convolution neural network model for establishing brain structure network evolution is as follows:is a two-layer graph roll-up neural network, wherein +.>Is a relational matrix of the brain structure network, +.>Is a direction vector representing the difference between each brain region of the brains of normal and depressed patients;
c1, respectively extracting gray matter volumes of the 90 brain areas of normal people and depression patients through a data preprocessing step to respectively formColumn vector +.> and />Define a direction vector +.>I.e. the difference in brain structure between normal and depressed patients, and for said direction vector +.>The following normalization process is performed as shown in formula (2):
wherein ,is a direction vector +>Minimum value of element->Is a direction vector +>Maximum value of the element;
c2, brain structure network relation matrixThe following normalization process is performed as shown in formula (3):
wherein ,is a relation matrix added with self-connection of brain structure network, < >>Is an identity matrix of the unit cell,representing the degree of each node in the brain structure network;
and C3, training a brain structure evolution model based on a two-layer graph convolution neural network, wherein the feedforward neural network model uses the following form:
wherein ,for the weight matrix of the input layer to the hidden layer, < +.>Is a weight matrix between the hidden layer and the output layer;softmaxis an activation function, defined as +.>, wherein ;
C4, evaluating the cross entropy loss function by using the relation matrix of all depression patients as a label sample, as shown in formula (5):
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911033766.9A CN110797123B (en) | 2019-10-28 | 2019-10-28 | Graph convolution neural network evolution method of dynamic brain structure |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911033766.9A CN110797123B (en) | 2019-10-28 | 2019-10-28 | Graph convolution neural network evolution method of dynamic brain structure |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110797123A CN110797123A (en) | 2020-02-14 |
CN110797123B true CN110797123B (en) | 2023-05-26 |
Family
ID=69441609
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911033766.9A Active CN110797123B (en) | 2019-10-28 | 2019-10-28 | Graph convolution neural network evolution method of dynamic brain structure |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110797123B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112258952B (en) * | 2020-09-17 | 2022-08-09 | 顺德职业技术学院 | Tissue inflammation simulation display method and system |
TWI768483B (en) * | 2020-09-28 | 2022-06-21 | 臺北醫學大學 | Method and apparatus for identifying white matter hyperintensities |
CN112820396B (en) * | 2020-12-30 | 2023-10-17 | 电子科技大学 | Automatic diagnosis device for realizing depression based on network fragility |
CN112862751B (en) * | 2020-12-30 | 2022-05-31 | 电子科技大学 | Automatic diagnosis device for autism |
CN113096127A (en) * | 2021-06-04 | 2021-07-09 | 壹药网科技(上海)股份有限公司 | System and method for generating brain network evolution model |
CN113786185B (en) * | 2021-09-18 | 2024-05-07 | 安徽师范大学 | Static brain network feature extraction method and system based on convolutional neural network |
CN114209320B (en) * | 2021-11-17 | 2024-05-07 | 山东师范大学 | Depression patient brain electricity identification system based on graph mutual information maximization |
CN114628036B (en) * | 2022-05-17 | 2022-08-02 | 中南大学湘雅医院 | Brain ischemia risk prediction platform based on neural network |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108427929A (en) * | 2018-03-19 | 2018-08-21 | 兰州大学 | A kind of depressed discriminance analysis system based on tranquillization state brain network |
CN109192298A (en) * | 2018-07-27 | 2019-01-11 | 南京航空航天大学 | Deep brain medical diagnosis on disease algorithm based on brain network |
CN110265148A (en) * | 2019-06-20 | 2019-09-20 | 上海海事大学 | A kind of dynamic function pattern learning method that fMRI brain network mechanism inspires |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
IL274548B1 (en) * | 2017-11-10 | 2024-08-01 | Lvis Corp | Efficacy and/or therapeutic parameter recommendation using individual patient data and therapeutic brain network maps |
-
2019
- 2019-10-28 CN CN201911033766.9A patent/CN110797123B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108427929A (en) * | 2018-03-19 | 2018-08-21 | 兰州大学 | A kind of depressed discriminance analysis system based on tranquillization state brain network |
CN109192298A (en) * | 2018-07-27 | 2019-01-11 | 南京航空航天大学 | Deep brain medical diagnosis on disease algorithm based on brain network |
CN110265148A (en) * | 2019-06-20 | 2019-09-20 | 上海海事大学 | A kind of dynamic function pattern learning method that fMRI brain network mechanism inspires |
Non-Patent Citations (2)
Title |
---|
Yang L et.al..Dynamic Brain Network Evolution in Major Depressive Disorder.2019,全文. * |
齐倩蕊.抑郁症患者大脑结构差异及脑网络动力学研究.2018,全文. * |
Also Published As
Publication number | Publication date |
---|---|
CN110797123A (en) | 2020-02-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110797123B (en) | Graph convolution neural network evolution method of dynamic brain structure | |
CN109376751B (en) | Human brain function network classification method based on convolutional neural network | |
CN113040715B (en) | Human brain function network classification method based on convolutional neural network | |
CN103886328B (en) | Based on the functional magnetic resonance imaging data classification method of brain mixed-media network modules mixed-media architectural feature | |
CN113616184A (en) | Brain network modeling and individual prediction method based on multi-mode magnetic resonance image | |
CN110188836B (en) | Brain function network classification method based on variational self-encoder | |
CN110070935B (en) | Medical image synthesis method, classification method and device based on antagonistic neural network | |
CN110598793B (en) | Brain function network feature classification method | |
CN112418337B (en) | Multi-feature fusion data classification method based on brain function hyper-network model | |
CN111090764B (en) | Image classification method and device based on multitask learning and graph convolution neural network | |
CN103020653B (en) | Structure and function magnetic resonance image united classification method based on network analysis | |
CN108921233A (en) | A kind of Raman spectrum data classification method based on autoencoder network | |
CN113693563A (en) | Brain function network classification method based on hypergraph attention network | |
CN110211671B (en) | Thresholding method based on weight distribution | |
CN113947157B (en) | Dynamic brain effect connection network generation method based on hierarchical clustering and structural equation model | |
CN117172294B (en) | Method, system, equipment and storage medium for constructing sparse brain network | |
Roewer-Despres et al. | Towards finite element simulation using deep learning | |
CN112950631A (en) | Age estimation method based on saliency map constraint and X-ray head skull positioning lateral image | |
CN114748053A (en) | fMRI high-dimensional time sequence-based signal classification method and device | |
Mouches et al. | Unifying brain age prediction and age-conditioned template generation with a deterministic autoencoder | |
Goutham et al. | Brain tumor classification using EfficientNet-B0 model | |
CN113255789A (en) | Video quality evaluation method based on confrontation network and multi-tested electroencephalogram signals | |
CN113143247A (en) | Method for constructing brain function hyper-network | |
CN115760785A (en) | Brain somatotropin morphology high-order feature extraction method | |
Lange et al. | A comparison between neural and fuzzy cluster analysis techniques for functional MRI |
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 |