CN111612746A - Dynamic detection method of functional brain network central node based on graph theory - Google Patents

Dynamic detection method of functional brain network central node based on graph theory Download PDF

Info

Publication number
CN111612746A
CN111612746A CN202010364805.XA CN202010364805A CN111612746A CN 111612746 A CN111612746 A CN 111612746A CN 202010364805 A CN202010364805 A CN 202010364805A CN 111612746 A CN111612746 A CN 111612746A
Authority
CN
China
Prior art keywords
node
matrix
time
nodes
brain
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.)
Granted
Application number
CN202010364805.XA
Other languages
Chinese (zh)
Other versions
CN111612746B (en
Inventor
颜成钢
陈安琪
朱嘉凯
孙垚棋
张继勇
张勇东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou Dianzi University
Original Assignee
Hangzhou Dianzi University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hangzhou Dianzi University filed Critical Hangzhou Dianzi University
Priority to CN202010364805.XA priority Critical patent/CN111612746B/en
Publication of CN111612746A publication Critical patent/CN111612746A/en
Application granted granted Critical
Publication of CN111612746B publication Critical patent/CN111612746B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10088—Magnetic resonance imaging [MRI]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20072—Graph-based image processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30016—Brain

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Magnetic Resonance Imaging Apparatus (AREA)

Abstract

The invention discloses a dynamic detection method of a functional brain network central node based on graph theory. The invention is improved on the basis of a multivariable central pivot node detection method, so that the central pivot node which is more reliable and accords with the cognitive activity of neuroscience can be detected. First, the blood oxygen signal is divided into several segments with time as dimension by using the sliding window technique. And detecting the pivot node in the corresponding time window in the sliding window of each period of time, thereby obtaining a change track of the pivot node along with the movement of the sliding window. Finally, the change track is used as a constraint to act on a multivariate detection method, so that more reliable and accurate central node can be dynamically detected.

Description

Dynamic detection method of functional brain network central node based on graph theory
Technical Field
The invention relates to the field of neuroscience brain network research, in particular to a dynamic detection method of a functional brain network central node based on graph theory.
Background
Resting state magnetic resonance imaging (fMRI) provides a non-invasive method to measure changes in brain blood oxygenation. In the resting state, the subject does not perform any specific task, and the subject spontaneously develops neural activity, the fluctuations of which are related to the changes in the blood oxygen concentration signal. The blood oxygen concentration signal is used to calculate the connectivity between brain regions, thereby constructing a functional brain network.
The brain network can be divided into a plurality of modules, some of which are responsible for vision, some of which are responsible for hearing and the like, and the modular structure enables people to distinguish different roles and statuses of brain area nodes more carefully. For example, some nodes are important in the module in which they are located, but not necessarily important for the entire network, and are called regional core nodes (provisonalhub), while other nodes, although having limited functionality in their own module, are connected to different modules and maintain connectivity of the entire network, and are called hub nodes (connectorhuub). The central node is connected with different functional modules in a brain network, and plays a very important role in connection in the brain due to the high centrality of the central node, such as information integration and participation in various cognitive activities, and the characteristics show that the central node is more easily influenced on the brain when suffering from disease attack. Recently, a consensus has been reached in the neuroimaging field that the brain's functional network changes throughout the scan time, even in an unprocessed environment. Many studies have shown that dynamic patterns are more relevant to certain cranial nerve diseases. Therefore, if we can dynamically detect central nodes more accurately, we can help us to analyze and understand the pathological mechanism of these diseases better, and can also be used for assisting the early diagnosis and treatment of diseases.
At present, methods for detecting a central node are designed for a static functional brain network, and are difficult to capture dynamic changes of the central node along with time, so that the detection result does not have time consistency and the change of the central node cannot be guaranteed to be consistent with cognitive changes presented in the functional brain network.
Disclosure of Invention
The invention provides a dynamic detection method for a functional brain network central node based on graph theory. Aiming at the defects of the static detection method, the multivariate central node detection method is improved, so that the central node which is more reliable and accords with the cognitive activity of neuroscience can be detected.
First, the blood oxygen signal is divided into several segments with time as dimension by using the sliding window technique. In the sliding window of each period of time, the hub node in the corresponding time window is detected, so as to obtain a variation track (as shown in fig. 1) of the hub node moving along with the sliding window. Finally, the change track is used as a constraint to act on a multivariate detection method, so that more reliable and accurate central node can be dynamically detected.
The technical scheme adopted by the invention for solving the technical problem comprises the following steps:
step 1, constructing a functional brain network in each sliding window by using Pearson correlation;
basic data format and preliminary knowledge, G ═ V, W, to represent a functional brain network, where V represents the set of N brain region nodes,
Figure BDA0002476193790000021
is an N × N adjacency matrix wijRepresenting the elements of the ith row and the jth column in the adjacent matrix W, and particularly representing the association degree of the ith brain area node and the jth brain area node; next, a Laplace matrix L ═ D-W is calculated, where D is a diagonal element of
Figure BDA0002476193790000022
The degree matrix of (c).
In graph theory, graph G is formed by a set of orthogonal vectors Φ obtained by eigen-decomposing a laplacian matrix L, i.e. L ═ ΦTΛ Φ, diagonal matrix Λ ═ diag [ λ [ ]1，λ2，...，λN]And λ1，λ2，...，λNRepresenting the eigenvalues sorted in ascending order.
The eigenvalues correspond to the orthogonal vectors one by one and are eigenvalues of a Laplace matrix L;
if all brain region nodes are connected in a functional brain network without separate parts, then the minimum eigenvalue λ is1Is zero. The number of zero eigenvalues equals the number of subgraphs. I.e. when lambda1And λ2The number of subgraphs is 2 when all values are zero.
Step 2: improved multivariate central node detection
2-1. Each brain region node v in a functional brain networkiAre all associated with a binary flag s in a selection vectoriIs associated with where s i0 denotes the brain region node viBeing a hub node, si1 means not a hub node. And judging whether the central node is the expected central node or not by removing the damage degree of the selected brain area node to the functional brain network, wherein the damage degree is judged by the quantity condition of zero eigenvalues of the residual functional brain network after the brain area node is removed.
The judgment criteria are as follows: and if the number of the zero eigenvalues of the residual functional brain networks is increased more than that of other nodes after the selected nodes are deleted, the node is taken as a candidate central node.
2-2. since multiple pivot nodes need to be selected at one time, the pair selection vector s ═ s1，s2，…，sn]Is an NP-hard problem. According to the KyFan theorem, the problem is converted into the sum of the minimum K characteristic values before minimization,
if there are K central nodes in the N brain area nodes of the functional brain network, the calculation is as follows:
Figure BDA0002476193790000031
then further derivation yields the final objective function:
Figure BDA0002476193790000032
wherein λ isiRepresenting the ith characteristic value;
Figure BDA0002476193790000033
representing a matrix;
2-3. place each element of the selection vector S on the diagonal of the diagonal matrix S. Wherein L iss＝D-STWS represents the laplace matrix of the remaining functional brain network, with subscript s being the marker for differentiation;
Figure BDA0002476193790000036
a K-dimensional matrix representing nodes of each brain region in the remaining functional brain network, and is subjected to FTF ═ I orthogonal constraint, so the solution needs to be optimized for F and S:
and F, optimizing: fixing the diagonal matrix S to obtain a closed form solution of F, the closed form solution being LsK orthogonal vectors corresponding to the first K minimum eigenvalues of (1).
And (4) optimizing S: fixing F, the objective function of the optimized diagonal matrix S becomes:
Figure BDA0002476193790000034
wherein
Figure BDA0002476193790000035
Is an element aij＝wij||fi-fj||2N × N, and fiAnd fjRespectively representing the ith row vector and the jth row vector of the F matrix.
2-4, because the optimized objective function is not strictly convex, and s is a binary selection vector which is not beneficial to optimization, an auxiliary vector is introduced
Figure BDA0002476193790000041
And simplifying the optimized objective function as follows:
Figure BDA0002476193790000042
where P is another diagonal matrix derived from the auxiliary vector P. Intuitively, the selection vector s is the result of the binarization of the auxiliary vector p.
Also in equation (4), P and S are solved alternately:
and F, optimizing: fixing the diagonal matrix S to obtain a closed form solution of F, the closed form solution being LsK orthogonal vectors corresponding to the first K minimum eigenvalues of (1).
Optimizing S and P: and F is fixed, and the diagonal matrix S and the auxiliary matrix P are optimized and solved by utilizing an augmented Lagrange multiplier method.
And step 3: dynamic detection of functional brain network hub nodes
The algorithm framework for backbone node detection versus dynamic function network of the present invention is shown in FIG. 2. Specifically, the following is shown:
and 3-1, segmenting the whole blood oxygen signal into T groups of overlapped sliding windows. At each time t, each brain region node v is estimatediCorresponding pi，piIs the ith element in the auxiliary vector p.
3-2. connection of piForm a track
Figure BDA0002476193790000043
Wherein
Figure BDA0002476193790000044
Is a continuous function pi(τ) at a time point τtDiscrete sampling of (2). Thus a continuous function pi(τ) the value of τ at any other point in time can be calculated using Radial Basis Functions (RBFs):
Figure BDA0002476193790000045
where the parameter sigma is used to control the strength of the trajectory smoothing,
Figure BDA0002476193790000046
representing the weight of the auxiliary vector of the ith brain region node at time t, i.e. the weight
Figure BDA0002476193790000047
Represents piThe weight at time t; tau-tautRepresenting the time difference.
3-3. given the trajectory
Figure BDA0002476193790000048
Set of radial basis functions
Figure BDA0002476193790000049
Calculated by the following method:
Figure BDA00024761937900000410
wherein the parameter λ is used to control the continuous function pi(τ) intensity of time dependence.
3-4. initializing a diagonal matrix S consisting of the selection vectors S for each sliding window using the improved diagonal matrix P consisting of auxiliary vectors, as indicated by the arrows in fig. 2.
And 3-5, substituting the improved diagonal matrix P formed by the auxiliary vectors into the step 2-3, and repeating the optimization solution until the target converges.
The invention has the following beneficial effects:
the output of the method of the invention is the central nodes which change along with the change of the brain network, and the central nodes are more accurate than the traditional method at each time point, and the correlation information between the time points is also kept.
Drawings
FIG. 1 is an overall framework of a hub node dynamic detection method;
FIG. 2 is an algorithm framework for a hub node dynamic detection method;
FIG. 3 is an analysis graph of the stability of the conventional method and the present method;
FIG. 4 is a graph of the visualization of the detection of normal human pivot nodes by the conventional method and the present method;
FIG. 5 is a graph of the visualization of the central node detection of a obsessive-compulsive patient using the conventional method and the present method;
Detailed Description
The present invention will be described in further detail below with reference to the accompanying drawings and experiments.
The invention utilizes the sliding window technology to divide the blood oxygen signal into several segments by taking time as a dimension. In the sliding window of each period of time, the hub node in the corresponding time window is detected, so as to obtain a variation track (as shown in fig. 1) of the hub node moving along with the sliding window. Finally, the change track is used as a constraint to act on a multivariate detection method, so that more reliable and accurate central node can be dynamically detected.
The embodiment of the invention mainly comprises the following steps:
step (1) data of 63 normal persons and 62 obsessive-compulsive patients were selected for the experiment, each subject had T1 weighted magnetic resonance image (specific parameters TR 8 ms, TE 1.7ms, flip angle 20 °, resolution 1.0 × 1.0 × 1.0.0 mm2) And resting state functional magnetic resonance data (specific parameters are TR 2s, TE 60ms, flip angle 90 degrees, resolution 3.0 × 3.0.0 3.0 × 4.0.0 mm2) Each subject contained 230 detection time points.
Step (2) all these experimental data were registered into an AAL template, divided into 116 brain regions. And calculating the correlation among the brain areas to obtain a corresponding functional brain network connection matrix W as the input of the experiment.
And (3) setting experimental parameters.
The size of the sliding window is set to 10% of the size of the whole time point; the detection number of the pivot nodes is set to be 12; the parameter σ of the radial basis RBF is set to 0.7; the optimal lambda parameter of the objective function (2) is found to be 0.6 by the grid search method.
And (4): finally, the objective function is aligned
Figure BDA0002476193790000061
And sequentially optimizing F and S until convergence, and obtaining a final result.
And (3) analyzing an experimental result:
for each individual, counting the number of the pivot nodes detected in each sliding window, constructing a corresponding histogram, and finally calculating a corresponding entropy value.
The lower the entropy shows that the change of the pivot node in the whole detection process is less, and as shown in fig. 3, the entropy values are all above the diagonal, which shows that the entropy values detected by the method are lower and more stable.
We further visualize the change results of the pivot nodes in a period of time, as shown in fig. 4 and fig. 5, the pivot nodes detected by our method have no change within 3 TRs (about 6 seconds), while the pivot nodes detected by the conventional method all have a jump at the second TR moment and return to the previous state at the next moment, which is unusual and not in accordance with the law of cognitive change. Thus, it can be seen that our method is more stable and reliable than the conventional method.

Claims (5)

1. A dynamic detection method for a functional brain network central node based on graph theory is characterized in that a sliding window technology is utilized to averagely divide blood oxygen signals into a plurality of segments by taking time as a dimension; detecting a pivot node in the corresponding time window in the sliding window of each period of time, thereby obtaining a change track of the pivot node along with the movement of the sliding window; and finally, the change track is used as a constraint to act on a multivariate detection method, so that more reliable and accurate central node can be dynamically detected.
2. The dynamic detection method for a graph theory-based functional brain network hub node according to claim 1, comprising the steps of:
step 1, constructing a functional brain network in each sliding window by using Pearson correlation;
a functional brain network is denoted by G ═ (V, W), where V denotes the set of N brain region nodes,
Figure FDA0002476193780000011
is an N × N adjacency matrix, wijRepresenting the elements of the ith row and the jth column in the adjacent matrix W, and particularly representing the association degree of the ith brain area node and the jth brain area node; next, a Laplace matrix L ═ D-W is calculated, where D is a diagonal element of
Figure FDA0002476193780000012
A degree matrix of (c);
in graph theory, graph G is formed by a set of orthogonal vectors Φ obtained by eigen-decomposing a laplacian matrix L, i.e. L ═ ΦTΛ Φ, diagonal matrix Λ ═ diag [ λ [ ]1，λ2，...，λN]And λ1，λ2，...，λNRepresenting eigenvalues sorted in ascending order;
step 2: improved multivariate central node detection
2-1. Each brain region node v in a functional brain networkiAre all associated with a binary flag s in a selection vectoriIs associated with where si denotes the brain region node viBeing a hub node, si1 means not a hub node; judging whether the central node is the expected central node or not by removing the damage degree of the selected brain area node to the functional brain network, wherein the damage degree is judged by the quantity condition of zero eigenvalues of the residual functional brain network after removing the brain area node;
2-2. since multiple pivot nodes need to be selected at one time, the pair selection vector s ═ s1，s2，…，sn]Is an NP-hard problem; converting the problem into the sum of K minimum eigenvalues before minimization according to the Ky Fan theorem;
if there are K central nodes in the N brain area nodes of the functional brain network, the calculation is as follows:
Figure FDA0002476193780000021
then further derivation yields the final objective function:
Figure FDA0002476193780000022
wherein λ isiRepresenting the ith characteristic value;
Figure FDA0002476193780000023
representing a matrix;
2-3, putting each element of the selection vector S on the diagonal of the diagonal matrix S; wherein L iss＝D-STWS represents the laplace matrix of the remaining functional brain network, with subscript s being the marker for differentiation;
Figure FDA0002476193780000024
a K-dimensional matrix representing nodes of each brain region in the remaining functional brain network, and is subjected to FTF ═ I orthogonal constraint, so the solution needs to be optimized for F and S:
and F, optimizing: fixing the diagonal matrix S to obtain a closed form solution of F, the closed form solution being LsK orthogonal vectors corresponding to the first K minimum eigenvalues;
and (4) optimizing S: fixing F, the objective function of the optimized diagonal matrix S becomes:
Figure FDA0002476193780000025
wherein
Figure FDA0002476193780000026
Is an element aij＝wij||fi-fj||2N × N, and fiAnd fjRespectively representing ith row vector and jth row vector of the F matrix;
2-4, because the optimized objective function is not strictly convex, and s is a binary selection vector which is not beneficial to optimization, an auxiliary vector is introduced
Figure FDA0002476193780000027
And simplifying the optimized objective function as follows:
Figure FDA0002476193780000028
where P is another diagonal matrix derived from the auxiliary vector P; intuitively, the selection vector s is the binarization result of the auxiliary vector p;
also in equation (4), P and S are solved alternately:
and F, optimizing: fixing the diagonal matrix S to obtain a closed form solution of F, the closed form solution being LsK orthogonal vectors corresponding to the first K minimum eigenvalues;
optimizing S and P: fixing F, and optimally solving a diagonal matrix S and an auxiliary matrix P by using an augmented Lagrange multiplier method;
and step 3: dynamic detection of functional brain network hub nodes
3-1, segmenting the whole blood oxygen signal into T groups of overlapped sliding windows; at each time t, each brain region node v is estimatediCorresponding pi，piIs the ith element in the auxiliary vector p;
3-2. connection of piForm a track
Figure FDA0002476193780000031
Wherein
Figure FDA0002476193780000032
Is a continuous function pi(τ) at a point in time { τt-discrete sampling of { right left over }; thus a continuous function pi(τ) the value of τ at any other point in time can be calculated using Radial Basis Functions (RBFs):
Figure FDA0002476193780000033
where the parameter sigma is used to control the strength of the trajectory smoothing,
Figure FDA0002476193780000034
representing the weight of the auxiliary vector of the ith brain region node at time t, i.e. the weight
Figure FDA0002476193780000035
Represents piThe weight at time t; tau-tautRepresents a time difference;
3-3. given the trajectory
Figure FDA0002476193780000036
Set of radial basis functions
Figure FDA0002476193780000037
Calculated by the following method:
Figure FDA0002476193780000038
wherein the parameter λ is used to control the continuous function pi(τ) intensity of the time dependence;
3-4, initializing a diagonal matrix S consisting of selection vectors S of each sliding window by using the diagonal matrix P consisting of the improved auxiliary vectors, as shown by an arrow in FIG. 2;
and 3-5, substituting the improved diagonal matrix P formed by the auxiliary vectors into the step 2-3, and repeating the optimization solution until the target converges.
3. The method according to claim 2, wherein the eigenvalues are in one-to-one correspondence with orthogonal vectors and are eigenvalues of a laplacian matrix L.
4. A method according to claim 2 or 3, wherein if all brain area nodes are connected in a functional brain network without separate parts, the minimum eigenvalue λ is1Is zero; the number of zero eigenvalues equals the number of subgraphs; i.e. when lambda1And λ2All have values ofAt zero, the number of subgraphs is 2.
5. The method according to claim 4, wherein the criterion is as follows: and if the number of the zero eigenvalues of the residual functional brain networks is increased more than that of other nodes after the selected nodes are deleted, the node is taken as a candidate central node.
CN202010364805.XA 2020-04-30 2020-04-30 Dynamic detection method for functional brain network central node based on graph theory Active CN111612746B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010364805.XA CN111612746B (en) 2020-04-30 2020-04-30 Dynamic detection method for functional brain network central node based on graph theory

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010364805.XA CN111612746B (en) 2020-04-30 2020-04-30 Dynamic detection method for functional brain network central node based on graph theory

Publications (2)

Publication Number Publication Date
CN111612746A true CN111612746A (en) 2020-09-01
CN111612746B CN111612746B (en) 2023-10-13

Family

ID=72196615

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010364805.XA Active CN111612746B (en) 2020-04-30 2020-04-30 Dynamic detection method for functional brain network central node based on graph theory

Country Status (1)

Country Link
CN (1) CN111612746B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117357132A (en) * 2023-12-06 2024-01-09 之江实验室 Task execution method and device based on multi-layer brain network node participation coefficient

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101243974A (en) * 2008-03-28 2008-08-20 天津和德脑象图技术开发研究有限公司 Method and apparatus for generating brain phase image detection and analysis with electroencephalogram
CN103325119A (en) * 2013-06-27 2013-09-25 中国科学院自动化研究所 Default state brain network center node detecting method based on modality fusion
US20150142714A1 (en) * 2013-11-18 2015-05-21 Sargon Partners Dynamic lighting system
CN106650231A (en) * 2016-11-10 2017-05-10 深圳市元征软件开发有限公司 Method and device for processing wireless body area network data
CN108519819A (en) * 2018-03-30 2018-09-11 北京金山安全软件有限公司 Intelligent device processing method and device, intelligent device and medium
CN109522894A (en) * 2018-11-12 2019-03-26 电子科技大学 A method of detection fMRI brain network dynamic covariant
US20190090749A1 (en) * 2017-09-26 2019-03-28 Washington University Supervised classifier for optimizing target for neuromodulation, implant localization, and ablation
CN109893126A (en) * 2019-03-21 2019-06-18 杭州电子科技大学 Epileptic seizure prediction method based on brain function network characterization

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101243974A (en) * 2008-03-28 2008-08-20 天津和德脑象图技术开发研究有限公司 Method and apparatus for generating brain phase image detection and analysis with electroencephalogram
CN103325119A (en) * 2013-06-27 2013-09-25 中国科学院自动化研究所 Default state brain network center node detecting method based on modality fusion
US20150142714A1 (en) * 2013-11-18 2015-05-21 Sargon Partners Dynamic lighting system
CN106650231A (en) * 2016-11-10 2017-05-10 深圳市元征软件开发有限公司 Method and device for processing wireless body area network data
US20190090749A1 (en) * 2017-09-26 2019-03-28 Washington University Supervised classifier for optimizing target for neuromodulation, implant localization, and ablation
CN108519819A (en) * 2018-03-30 2018-09-11 北京金山安全软件有限公司 Intelligent device processing method and device, intelligent device and medium
CN109522894A (en) * 2018-11-12 2019-03-26 电子科技大学 A method of detection fMRI brain network dynamic covariant
CN109893126A (en) * 2019-03-21 2019-06-18 杭州电子科技大学 Epileptic seizure prediction method based on brain function network characterization

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
THI MAI PHUONG NGUYEN 等: "Estimation of Brain Dynamics Under Visuomotor Task using Functional Connectivity Analysis Based on Graph Theory" *
耿广磊: "基于磁共振的脑结构功能网络分析方法应用研究" *
靳聪 等: "DTI脑网络中枢节点识别方法的比较" *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117357132A (en) * 2023-12-06 2024-01-09 之江实验室 Task execution method and device based on multi-layer brain network node participation coefficient
CN117357132B (en) * 2023-12-06 2024-03-01 之江实验室 Task execution method and device based on multi-layer brain network node participation coefficient

Also Published As

Publication number Publication date
CN111612746B (en) 2023-10-13

Similar Documents

Publication Publication Date Title
CN113616184B (en) Brain network modeling and individual prediction method based on multi-mode magnetic resonance image
CN107633522B (en) Brain image segmentation method and system based on local similarity active contour model
CN109528197B (en) Individual prediction method and system for mental diseases based on brain function map
Dey et al. Healthy and unhealthy rat hippocampus cells classification: A neural based automated system for Alzheimer disease classification
CN112418337B (en) Multi-feature fusion data classification method based on brain function hyper-network model
CN109480833A (en) The pretreatment and recognition methods of epileptic's EEG signals based on artificial intelligence
CN109645990A (en) A kind of CRT technology method of epileptic's EEG signals
CN109935321B (en) Risk prediction system for converting depression patient into bipolar affective disorder based on functional nuclear magnetic resonance image data
CN113255728A (en) Depression classification method based on map embedding and multi-modal brain network
KR101687217B1 (en) Robust face recognition pattern classifying method using interval type-2 rbf neural networks based on cencus transform method and system for executing the same
CN111402278B (en) Segmentation model training method, image labeling method and related devices
JP2020036633A (en) Abnormality determination program, abnormality determination method and abnormality determination device
CN113947157B (en) Dynamic brain effect connection network generation method based on hierarchical clustering and structural equation model
Chao et al. Multi-modality image fusion based on enhanced fuzzy radial basis function neural networks
Wang et al. Understanding the relationship between human brain structure and function by predicting the structural connectivity from functional connectivity
CN115969369A (en) Brain task load identification method, application and equipment
Ji et al. Convolutional neural network with graphical lasso to extract sparse topological features for brain disease classification
Li et al. Cell population tracking and lineage construction using multiple-model dynamics filters and spatiotemporal optimization
CN115985488A (en) Brain disease diagnosis method, device and system based on dynamic hyper-network
CN117671463B (en) Multi-mode medical data quality calibration method
CN111612746A (en) Dynamic detection method of functional brain network central node based on graph theory
Ma et al. A universal texture segmentation and representation scheme based on ant colony optimization for iris image processing
CN114155952A (en) Senile dementia illness auxiliary analysis system for elderly people
CN117137435B (en) Rehabilitation action recognition method and system based on multi-mode information fusion
Lohar et al. Automatic classification of autism spectrum disorder (ASD) from brain MR images based on feature optimization and machine 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