CN105931281B - The method that brain function network is quantitatively portrayed based on network characterization entropy - Google Patents
The method that brain function network is quantitatively portrayed based on network characterization entropy Download PDFInfo
- Publication number
- CN105931281B CN105931281B CN201610230029.8A CN201610230029A CN105931281B CN 105931281 B CN105931281 B CN 105931281B CN 201610230029 A CN201610230029 A CN 201610230029A CN 105931281 B CN105931281 B CN 105931281B
- Authority
- CN
- China
- Prior art keywords
- mrow
- brain
- node
- network
- msub
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/003—Reconstruction from projections, e.g. tomography
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves
- A61B5/055—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/41—Medical
Landscapes
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Engineering & Computer Science (AREA)
- Pathology (AREA)
- Heart & Thoracic Surgery (AREA)
- Radiology & Medical Imaging (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- High Energy & Nuclear Physics (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses a kind of method that brain function network is quantitatively portrayed based on network characterization entropy, step includes：Gather the fMRI data under quiescent condition and pre-processed, selection includes the brain template of cerebral cortex function division, single brain area in brain template is regarded as to a node in brain function network, average time sequence signal of the abstraction function magnetic resonance imaging data in each node, function connects between calculate node, obtain function connects matrix, function connects matrix progress binaryzation is obtained into adjacency matrix, form brain function network, calculate the eigenvector centrality degree and encircled energy of each node in brain function network, brain network Characteristic Entropy is calculated according to the encircled energy of each node in brain function network.It is of the invention quantitatively to portray brain function network to disclose brain development and aging rule by brain network Characteristic Entropy, have the advantages that principle is simple, it is convenient to realize, the interior result of subject is stable, different classes of sample sensitivity is good.
Description
Technical field
The present invention relates to the brain network research field based on functional mri (fMRI) data, and in particular to a kind of
The method that brain function network is quantitatively portrayed based on network characterization entropy.
Background technology
The principle of functional mri be by magnetic shake radiography measure neuron activity caused by it is hemodynamic change.
For example：When stimulating some sense organs, the specific some brain domain activation of cerebral cortex, the brain area of these activation sends neural letter
Number, using magnetic shake radiography obtain nerve signal, obtain a series of images.It is various multiple that people perform motion, sensation, higher cognitive etc.
During miscellaneous task, some nerve signals of intracerebral can be evoked, the brain activity of this physiological causes local cerebral blood flow, brain blood to hold,
The change of energetic supersession, so as to cause the change of tissue magnetic.FMRI is as a kind of live body cerebral function imaging technology, because it has
The advantages of Noninvasive, preferable spatial and temporal resolution, rapidly become the most frequently used brain imaging technique of cranial nerve scientific research.
Either field of biology, technical field, social field or other scientific domains all generate similarly phase
Mutual associated data set.In the past few decades, in order to portray these data sets, Complex Networks Analysis is arisen at the historic moment.It is complicated
Network analysis mainly quantitatively portrays the topological property of given network by graph theory.Brain connection data refer to functional connection or knot
Structure follows the trail of the network that connection is formed, and it has common feature with biological or physical system, while researcher has found brain from micro-
Small neuron so utilizes Complex Networks Analysis technique study to could set up network model on the different level such as brain area
Brain network attribute recognizes and studied brain function activity and provides a kind of brand-new Research Thinking on the whole for people.
2006, Achard Sophi et al. portrayed function connects using discrete wavelet change conversion fMRI data estimations
Frequency dependence matrix, it is concluded that：Low-frequency oscillation data set has worldlet (small world) property in fMRI data.
Van den Heuvel MP in 2008 et al. are based on voxel (voxel- according to tranquillization function nuclear magnetic resonance (rs-fMRI) data
Wise) method, builds the function connects between voxel, and then the statistical attribute of calculating network figure finds the cluster of brain function network
Coefficient ratio corresponds to the big of random network, and average path length is shorter than random network, indicates the small of brain function network
World's characteristic；In addition, the power law distribution that power rate exponential approximation is equal to 2, table are obeyed in the distribution of function connects between voxel and voxel
The uncalibrated visual servo characteristic (scale free network) of bright brain function network.But, these topologicaies property derived above are only carved
A feature of brain function network is drawn, these are characterized in isolated, and its application is limited to.Such as we are explored by very
The brain development of multifactor impact, it is necessary to which these characteristics are carried out into comprehensive analysis, extracts the structure trend feature of network, how to lead to
Cross comprehensive analysis quantitatively to portray brain function network to explore network characterization entropy and brain development, aging rule, have become one
Key technical problem urgently to be resolved hurrily.
The content of the invention
The technical problem to be solved in the present invention：For the above mentioned problem of prior art, there is provided pass through brain network Characteristic Entropy
Announcement brain development and aging rule, with principle is simple, realization is easy, the interior result stability of subject is high, different classes of sensitive
The good method that brain function network is quantitatively portrayed based on network characterization entropy of property.
In order to solve the above-mentioned technical problem, the technical solution adopted by the present invention is：
A kind of method that brain function network is quantitatively portrayed based on network characterization entropy, step is included：
1) fMRI data under collection quiescent condition；
2) pre-processed for the fMRI data collected；
3) selection includes the brain template of cerebral cortex function division；
4) pretreated fMRI data is directed to, the Different brain region in brain template is regarded as brain function network
In node, so as to extract the coefficient correlation between the average time sequence signal of each node, calculate node, obtain function company
Matrix is connect, function connects matrix progress binaryzation is obtained into adjacency matrix, so that brain function network is formed, the adjacent square
Formation into brain function network feature node interphase interaction relation, reflect the topological attribute of brain function network；
5) the eigenvector centrality degree and encircled energy of each node in brain function network are calculated；
6) brain network Characteristic Entropy is calculated according to the encircled energy of each node in brain function network.
Preferably, the step 4) detailed step include：
4.1) pretreated fMRI data is directed to, the Different brain region in brain template is regarded as brain function net
Node in network, extracts the average time sequence signal of each node；
4.2) using the coefficient correlation between formula (1) calculate node, the time series that a symmetrical size is n*n is obtained
Correlation matrix A (aij), wherein n represents brain function network node number, aijRepresent the average time sequence signal x of node iiWith section
Point j average time sequence signal xjBetween coefficient correlation, the function connects between node, therefore institute are represented with coefficient correlation
State time series correlation matrix A (aij) illustrate in brain function network function connects, referred to as function connects matrix between each node；
In formula (1), E (xixj) represent xixjMathematic expectaion, E (xi) represent xiMathematic expectaion, E (xj) represent xjMathematics
Expect,Represent xi 2Mathematic expectaion, E2(xi) represent xiSquare of mathematic expectaion,Represent xj 2Mathematic expectaion,
E2(xj) represent xjSquare of mathematic expectaion；
4.3) the function connects matrix is directed to, first according to the corresponding statistical value p of function connects between any two node,
Corrected by Bonferroni, function connects correspondence statistical value p being more than between the node pair of default statistical value are set to 0, will
The function connects that correspondence statistical value p is less than or equal between the node pair of default statistical value are set to 1, secondly by the function connects
The element for being less than 0 in matrix is set to 0, so that the function connects matrix binaryzation is obtained into adjacency matrix formation brain function net
Network, the brain function network features node interphase interaction relation, reflects the topological attribute of brain function network.
Preferably, the step 4.3) in default statistical value be 0.05.
Preferably, the step 5) detailed step include：
5.1) i-th of component conduct of the eigenvalue of maximum character pair vector of the adjacency matrix is calculated according to formula (2)
The eigenvector centrality degree of i-th of node in adjacency matrix；
In formula (2), eiRepresent the eigenvector centrality degree of i-th of node in adjacency matrix, aijRepresent i-th in adjacency matrix
Function connects between individual node and j-th of node, λ represents the corresponding eigenvalue of maximum of adjacency matrix, ejRepresent the corresponding spies of λ
Vector is levied, n represents brain function network node number；
5.2) encircled energy of each node in adjacency matrix is calculated according to formula (3)；
In formula (3), IiRepresent the encircled energy of i-th of node in adjacency matrix, eiRepresent i-th of section in adjacency matrix
The eigenvector centrality degree of point, n represents brain function network node number.
Preferably, the step 6) in calculate brain network Characteristic Entropy function expression such as formula (4) shown in；
In formula (4), E represents brain network Characteristic Entropy, IiRepresent the encircled energy of i-th of node in adjacency matrix, n tables
Show brain function network node number.
Preferably, the step 2) detailed step include：
2.1) image of the preceding specified number of frames of remove function magnetic resonance imaging data；
2.2) subject that amplitude exceedes specified threshold is moved in removing function magnetic resonance imaging data on head；
2.3) interlayer time adjustment is carried out for fMRI data；
2.4) the dynamic correction of wardrobe is entered for fMRI data；
2.5) Spatial normalization processing is carried out for fMRI data；
2.6) fMRI data is carried out at space smoothing using the gaussian kernel function of specified size halfwidth
Reason；
2.7) time-domain filtering is carried out to fMRI data using assigned frequency is interval；
2.8) for fMRI data carry out return full brain average signal, white matter, hydrocrania average signal and
Dynamic parameter.
Preferably, the step 2.6) in specify size halfwidth specifically refer to 6mm half-breadth it is high.
Preferably, the step 2.7) in specify frequency separation specifically refer to 0.01~0.08Hz.
The present invention is had the advantage that based on the method tool that network characterization entropy quantitatively portrays brain function network：
1st, the present invention regards the single brain area in brain template as the node in network, is total to by extracting function magnetic after pretreatment
Shake average time sequence signal of the imaging data in each node, the function connects between calculate node, obtains function connects square
Battle array, obtains adjacency matrix by function connects matrix progress binaryzation, brain function network is formed, for the adjoining square of brain function network
Battle array, calculates the eigenvector centrality degree and encircled energy of each node respectively, has excavated and the essential topological attribute of network
Related feature.
2nd, the present invention is theoretical by existing shannon entropy, and brain net is calculated according to the encircled energy of each node
Network Characteristic Entropy, it is possible thereby to calculate the brain network of different developmental phases by comparing, explores brain development and aging rule, leads to
Cross and repeat experimental verification stability of the network characterization entropy on individual, to explore brain development, aging using network characterization entropy
Rule offer method and thinking, with principle is simple, realization is easy, the interior result stability of subject is high, different classes of sensitivity is good
Advantage.
Brief description of the drawings
Fig. 1 is the basic procedure schematic diagram of present invention method.
Fig. 2 is subject age distribution figure in the embodiment of the present invention.
Embodiment
As shown in figure 1, the present embodiment is included the step of the method for brain function network is quantitatively portrayed based on network characterization entropy：
1) fMRI data under collection quiescent condition；
2) pre-processed for the fMRI data collected；
3) selection includes the brain template of cerebral cortex function division；
4) pretreated fMRI data is directed to, the Different brain region in brain template is regarded as brain function network
In node, so as to extract the coefficient correlation between the average time sequence signal of each node, calculate node, obtain function company
Matrix is connect, function connects matrix progress binaryzation is obtained into adjacency matrix, so that brain function network is formed, the adjacent square
Formation into brain function network feature node interphase interaction relation and reflect the topological attribute of brain function network；
5) the eigenvector centrality degree and encircled energy of each node in brain function network are calculated；
6) brain network Characteristic Entropy is calculated according to the encircled energy of each node in brain function network.
The present embodiment step 1) in, gather the functional mri under quiescent condition (resting state)
During (functional magnetic resonance imaging, abbreviation fMRI) data, the equipment used is swept for nuclear magnetic resonance
Instrument, and selection fast gradient echo mode are retouched, needs setting scan period, thickness, the number of plies, scan matrix, scanning to regard before scanning
Require that subject keeps head still, closes eyes and loosens in the sweep parameters such as open country, scanning process, but keep clear-headed.
In the present embodiment, step 2) detailed step include：
2.1) image of the preceding specified number of frames of remove function magnetic resonance imaging data, for eliminate magnetic field saturation to number
It is specially to remove preceding 5 two field picture according to the inaccurate influence brought, in the present embodiment；
2.2) subject that amplitude exceedes specified threshold is moved in removing function magnetic resonance imaging data on head；
2.3) interlayer time adjustment is carried out for fMRI data；
2.4) the dynamic correction of wardrobe is entered for fMRI data；
2.5) Spatial normalization processing is carried out for fMRI data；
2.6) fMRI data is carried out at space smoothing using the gaussian kernel function of specified size halfwidth
Reason；In the present embodiment, step 2.6) in specify size halfwidth specifically refer to 6mm half-breadth it is high；
2.7) time-domain filtering is carried out to fMRI data using assigned frequency is interval；In the present embodiment, step
2.6) frequency separation is specified to specifically refer to 0.01~0.08Hz in；
2.8) for fMRI data carry out return full brain average signal, white matter, hydrocrania average signal and
Dynamic parameter, to reduce hardware noise.
In the present embodiment, step 2.2)~2.8) specifically use Statistical Parametric Mapping softwares
(SPM8) realized under Matlab operations, pass through above-mentioned processing, it is possible to increase the degree of accuracy of fMRI data.
Brain template is the cerebral cortex function division that forefathers are determined by methods such as task state active region detections, the present embodiment
Step 3) in, brain template of the selection comprising cerebral cortex function division is specially Montreal Neurological
Anatomical Automatic Labeling (AAL) template that Institute (MNI) mechanism is provided, it includes altogether 116
Individual brain area；In addition it is also possible to select other to generally acknowledge relatively stable brain template as needed.
In the present embodiment, step 4) detailed step include：
The step 4) detailed step include：
4.1) pretreated fMRI data is directed to, the Different brain region in brain template is regarded as brain function net
Node in network, extracts the average time sequence signal of each node；
4.2) using the coefficient correlation between formula (1) calculate node, the time series that a symmetrical size is n*n is obtained
Correlation matrix A (aij), wherein n represents brain function network node number, aijRepresent the average time sequence signal x of node iiWith section
Point j average time sequence signal xjBetween coefficient correlation, the function connects between node, therefore institute are represented with coefficient correlation
State time series correlation matrix A (aij) illustrate in brain function network function connects, referred to as function connects matrix between each node；
In formula (1), E (xixj) represent xixjMathematic expectaion, E (xi) represent xiMathematic expectaion, E (xj) represent xjMathematics
Expect,Represent xi 2Mathematic expectaion, E2(xi) represent xiSquare of mathematic expectaion,Represent xj 2Mathematic expectaion,
E2(xj) represent xjSquare of mathematic expectaion；
4.3) the function connects matrix is directed to, first according to the corresponding statistical value p of function connects between any two node,
Corrected by Bonferroni, function connects correspondence statistical value p being more than between the node pair of default statistical value are set to 0, will
The function connects that correspondence statistical value p is less than or equal between the node pair of default statistical value are set to 1, secondly by the function connects
The element for being less than 0 in matrix is set to 0, so that the function connects matrix binaryzation is obtained into adjacency matrix formation brain function net
Network, the brain function network features node interphase interaction relation, reflects the topological attribute of brain function network.
In the present embodiment, step 4.3) in default statistical value be 0.05.
In the present embodiment, step 5) detailed step include：
5.1) i-th of component conduct of the eigenvalue of maximum character pair vector of the adjacency matrix is calculated according to formula (2)
The eigenvector centrality degree (eigenvector centrality) of i-th of node in adjacency matrix；
In formula (2), eiRepresent the eigenvector centrality degree of i-th of node in adjacency matrix, aijRepresent i-th in adjacency matrix
Function connects between individual node and j-th of node, λ represents the corresponding eigenvalue of maximum of adjacency matrix, ejRepresent the corresponding spies of λ
Vector is levied, n represents brain function network node number；
5.2) encircled energy of each node in adjacency matrix is calculated according to formula (3)；
In formula (3), IiRepresent the encircled energy of i-th of node in adjacency matrix, eiRepresent i-th of section in adjacency matrix
The eigenvector centrality degree of point, n represents brain function network node number；
In the present embodiment, step 6) in calculate shown in the function expression such as formula (4) of brain network Characteristic Entropy；
In formula (4), E represents brain network Characteristic Entropy, IiRepresent the encircled energy of i-th of node in adjacency matrix, n tables
Show brain function network node number.Function expression shown in formula (4) is the information entropy theory brain net by existing Shannon
The entropy of network, brain function network is quantitatively portrayed based on network characterization entropy, can be used for exploring brain development, aging rule.
In order to assess performance of the brain network Characteristic Entropy in brain function network application, two factors must be considered：1st, again
Survey reliability (test-retest reliability)：It is big that the height of test reflects network Characteristic Entropy changeability in subject
It is small, weigh stability of the network characterization entropy in time domain, 2, sensitivity：Examined by double sample t or non-parametric test is examined
It is different classes of (for example：Between men and women) subject between difference it is whether notable.The present embodiment matches totally 97 people (male for sex
54, women 43) same a collection of subject, by two Scan fMRI data obtaining collecting in the case of different time (Scan it
Between be separated by one week), the test and sensitivity of calculating network Characteristic Entropy.It is 27.83 years old to be tested average age, and variance is 5.09
In year, age distribution is as shown in Figure 2.
Test refers to same tranquillization state functional MRI experiment in the different time to the testing of same subject colony
The degree of consistency of resulting measurement result.In function connects group, interclass correlation coefficient intra-class is commonly used
The test of one measurement of correlation (ICC) quantitative predication.Survey article is delivered with reference to Zuo Xi nian within 2014
test-retest reliabilities of resting-state FMRI measurements inhuman brain
functional connectomics:Test is divided into five in A systems neuroscience perspective
Individual grade：It is relatively low：0<ICC≤0.2, typically：0.2<ICC≤0.4, it is medium：0.4<ICC≤0.6, significantly：0.6<ICC≤0.8,
It is higher：0.8<ICC≤1.In the present embodiment, the step of calculating the test of brain network Characteristic Entropy using interclass correlation coefficient
Including：
A1 two Scan fMRI data (Scan1 and Scan2)) are respectively adopted into the present embodiment based on network characterization entropy to determine
The method that amount portrays brain function network calculates brain network Characteristic Entropy, is tested number n=97, pendulous frequency d=2, network characterization
Entropy is E, Eij(i=1 ... d；J=1 ..., n) refer to the network characterization entropy that j-th of subject is obtained when ith is measured.It will calculate
Obtained brain network Characteristic Entropy is modeled using linear mixed model Linear mixed models (LMM), linear mixed model
Expression formula be Eij=λ0j+eij,λ0j=μ00+p0j, wherein μ00It is a fixed parameter, it refers to all d measurements of subject
Network characterization entropy average, λ0jFor intermediate parameters, p0jAnd eijIt is that obedience average is that 0 variance is respectivelyWithIndependence just
State is distributed random error, p0jRefer to the error between subject, eijRefer to measurement error.If μ00Notable is not 0, then brain
Mean coefficient ICC (E) definition is as shown in formula (5) in the group of network characterization entropy.
In formula (5), MSb(E) it is the summation of variance between sample two-by-two, MSw(E) it is mean square error in all subjects,WithFor variance.
A2 SS) is calculatedpAnd SSe。
According to linear mixed model Eij, in order to estimate interclass correlation coefficient ICC, SS is calculated according to formula (6) and (7)pWith
SSe。
In formula (6) and (7), SSpFor poor quadratic sum, SS between groupeFor overall error quadratic sum,It is j-th of sample in ith
The network characterization entropy obtained during measurement, n is number of samples, whereinSample average between i-th of sample group,Sample in jth group group
This average,The expression formula of total sample average three refers to formula (8).
In formula (8), d is pendulous frequency, and remaining meaning of parameters is identical with formula (6) and (7), will not be repeated here.
A3) estimation ICC (E).
MSb=SSp/ (n-1) and MSw=SSe/ (n-1) (d-1) mathematic expectaion is respectively
Therefore,WithUnbiased esti-mator such as formula (9) shown in；
In formula (9),For varianceUnbiased esti-mator,For varianceUnbiased esti-mator.
Shown in so ICC (E) estimation such as formula (10)；
In formula (10),For varianceUnbiased esti-mator,For varianceUnbiased esti-mator.In the present embodiment, final meter
The estimate for calculating the test ICC (E) for obtaining brain network Characteristic Entropy is 0.96, and the present embodiment is quantitative based on network characterization entropy
The test value for portraying the method for brain function network belongs to higher level, and the present embodiment quantitatively portrays brain based on network characterization entropy
Changeability is small in this measurement subject of brain network Characteristic Entropy in the method for functional network, and time-domain stability is high.
Obtained by sensitivity refers to same tranquillization state functional MRI experiment to different classes of subject colony testing
The otherness degree of measurement result.In statistics, we examine the difference of two groups of different groups frequently with double sample t
Property, conspicuousness p value is set to 0.05, if p<0.05, then two groups of different groups there is significant difference, otherwise two under the measurement
Group different groups are not significantly different under the measurement.In the present embodiment, the step of determining sensitivity includes：
B1) same subject scan1 and scan2 time serieses are combined, and calculate the network characterization entropy being each tested.
B2 linear regression model (LRM), E=β) are set up0+β1× age, wherein E represent that observable variable refers specifically to network
Characteristic Entropy, obtained residual error is used for subsequent analysis, wherein β0、β1Regression coefficient is represented, age is independent variable.
B3 obtained residual error) is divided into two groups according to sex,
B4) examined using double sample t, examine the difference of two groups of subject network characterization entropys of men and women, calculating obtains p=0.024,
P is known according to significance p=0.05<0.05.
In summary, the research at present for brain network is generally qualitative analysis, and the present embodiment is fixed based on network characterization entropy
Amount portrays the definition of the method combining information entropy of brain function network, probes into the definition of network characterization entropy, brain net can be portrayed by providing
The entropy definition of network complexity, and measure intraindividual stability (test) and the other sensitivity checking of region class from investigating
The reliability of network characterization entropy, and then comparative development, the network characterization entropy of aging different times brain, seek network characterization entropy
The relation existed between brain development, aging rule, and development, rate of ageing, so as to study brain development, aging rule
Rule.
Described above is only the preferred embodiment of the present invention, and protection scope of the present invention is not limited merely to above-mentioned implementation
Example, all technical schemes belonged under thinking of the present invention belong to protection scope of the present invention.It should be pointed out that for the art
Those of ordinary skill for, some improvements and modifications without departing from the principles of the present invention, these improvements and modifications
It should be regarded as protection scope of the present invention.
Claims (7)
1. a kind of method that brain function network is quantitatively portrayed based on network characterization entropy, it is characterised in that step includes：
1) fMRI data under collection quiescent condition；
2) pre-processed for the fMRI data collected；
3) selection includes the brain template of cerebral cortex function division；
4) pretreated fMRI data is directed to, the Different brain region in brain template is regarded as in brain function network
Node, so as to extract the coefficient correlation between the average time sequence signal of each node, calculate node, obtains function connects square
Battle array, adjacency matrix is obtained by function connects matrix progress binaryzation, so that brain function network is formed, the adjacency matrix shape
Into brain function network feature node interphase interaction relation, reflect the topological attribute of brain function network；
5) the eigenvector centrality degree and encircled energy of each node in brain function network are calculated；
6) brain network Characteristic Entropy is calculated according to the encircled energy of each node in brain function network；
The step 4) detailed step include：
4.1) pretreated fMRI data is directed to, the Different brain region in brain template is regarded as in brain function network
Node, extract the average time sequence signal of each node；
4.2) using the coefficient correlation between formula (1) calculate node, the time series that a symmetrical size is n*n is obtained related
Matrix A (aij), wherein n represents brain function network node number, aijRepresent the average time sequence signal x of node iiWith node j
Average time sequence signal xjBetween coefficient correlation, the function connects between node are represented with coefficient correlation, therefore described
Time series correlation matrix A (aij) illustrate in brain function network function connects, referred to as function connects matrix between each node；
<mrow>
<msub>
<mi>a</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mi>E</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>E</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mi>E</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msqrt>
<mrow>
<mi>E</mi>
<mrow>
<mo>(</mo>
<msubsup>
<mi>x</mi>
<mi>i</mi>
<mn>2</mn>
</msubsup>
<mo>)</mo>
</mrow>
<mo>-</mo>
<msup>
<mi>E</mi>
<mn>2</mn>
</msup>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msqrt>
<msqrt>
<mrow>
<mi>E</mi>
<mrow>
<mo>(</mo>
<msubsup>
<mi>x</mi>
<mi>j</mi>
<mn>2</mn>
</msubsup>
<mo>)</mo>
</mrow>
<mo>-</mo>
<msup>
<mi>E</mi>
<mn>2</mn>
</msup>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msqrt>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
In formula (1), E (xixj) represent xixjMathematic expectaion, E (xi) represent xiMathematic expectaion, E (xj) represent xjMathematic expectaion,RepresentMathematic expectaion, E2(xi) represent xiSquare of mathematic expectaion,Represent xj 2Mathematic expectaion, E2(xj)
Represent xjSquare of mathematic expectaion；
4.3) the function connects matrix is directed to, first according to the corresponding statistical value p of function connects between any two node, is passed through
Bonferroni is corrected, and function connects correspondence statistical value p being more than between the node pair of default statistical value are set to 0, by correspondence
The function connects that statistical value p is less than or equal between the node pair of default statistical value are set to 1, secondly by the function connects matrix
In be less than 0 element be set to 0 so that by the function connects matrix binaryzation obtain adjacency matrix formation brain function network, institute
State brain function network and feature node interphase interaction relation, reflect the topological attribute of brain function network.
2. the method according to claim 1 that brain function network is quantitatively portrayed based on network characterization entropy, it is characterised in that institute
State step 4.3) in default statistical value be 0.05.
3. the method according to claim 1 that brain function network is quantitatively portrayed based on network characterization entropy, it is characterised in that institute
State step 5) detailed step include：
5.1) i-th of component that the eigenvalue of maximum character pair vector of the adjacency matrix is calculated according to formula (2) is used as adjoining
The eigenvector centrality degree of i-th of node in matrix；
<mrow>
<msub>
<mi>e</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>&lambda;</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mi>j</mi>
<mo>&NotEqual;</mo>
<mi>i</mi>
</mrow>
<mi>n</mi>
</munderover>
<msub>
<mi>a</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<msub>
<mi>e</mi>
<mi>j</mi>
</msub>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
In formula (2), eiRepresent the eigenvector centrality degree of i-th of node in adjacency matrix, aijRepresent i-th of section in adjacency matrix
Function connects between point and j-th of node, λ represents the corresponding eigenvalue of maximum of adjacency matrix, ejRepresent the corresponding features of λ to
Amount, n represents brain function network node number；
5.2) encircled energy of each node in adjacency matrix is calculated according to formula (3)；
<mrow>
<msub>
<mi>I</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>e</mi>
<mi>i</mi>
</msub>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msub>
<mi>e</mi>
<mi>i</mi>
</msub>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
In formula (3), IiRepresent the encircled energy of i-th of node in adjacency matrix, eiRepresent i-th node in adjacency matrix
Eigenvector centrality degree, n represents brain function network node number.
4. the method according to claim 1 that brain function network is quantitatively portrayed based on network characterization entropy, it is characterised in that institute
State step 6) in calculate brain network Characteristic Entropy function expression such as formula (4) shown in；
<mrow>
<mi>E</mi>
<mo>=</mo>
<mo>-</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msub>
<mi>I</mi>
<mi>i</mi>
</msub>
<mi>ln</mi>
<mi> </mi>
<msub>
<mi>I</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
In formula (4), E represents brain network Characteristic Entropy, IiThe encircled energy of i-th of node in adjacency matrix is represented, n represents brain
Functional network node number.
5. the method according to claim 1 that brain function network is quantitatively portrayed based on network characterization entropy, it is characterised in that institute
State step 2) detailed step include：
2.1) image of the preceding specified number of frames of remove function magnetic resonance imaging data；
2.2) subject that amplitude exceedes specified threshold is moved on head in removing function magnetic resonance imaging data；
2.3) interlayer time adjustment is carried out for fMRI data；
2.4) the dynamic correction of wardrobe is entered for fMRI data；
2.5) Spatial normalization processing is carried out for fMRI data；
2.6) space smoothing processing is carried out to fMRI data using the gaussian kernel function of specified size halfwidth；
2.7) time-domain filtering is carried out to fMRI data using assigned frequency is interval；
2.8) full brain average signal, white matter, hydrocrania average signal and head is returned for fMRI data progress to move
Parameter.
6. the method according to claim 5 that brain function network is quantitatively portrayed based on network characterization entropy, it is characterised in that institute
State step 2.6) in specify size halfwidth specifically refer to 6mm half-breadth it is high.
7. the method according to claim 5 that brain function network is quantitatively portrayed based on network characterization entropy, it is characterised in that institute
State step 2.7) in specify frequency separation specifically refer to 0.01~0.08Hz.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610230029.8A CN105931281B (en) | 2016-04-14 | 2016-04-14 | The method that brain function network is quantitatively portrayed based on network characterization entropy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610230029.8A CN105931281B (en) | 2016-04-14 | 2016-04-14 | The method that brain function network is quantitatively portrayed based on network characterization entropy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105931281A CN105931281A (en) | 2016-09-07 |
| CN105931281B true CN105931281B (en) | 2017-10-27 |
Family
ID=56838052
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610230029.8A Active CN105931281B (en) | 2016-04-14 | 2016-04-14 | The method that brain function network is quantitatively portrayed based on network characterization entropy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105931281B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845137A (en) * | 2017-02-27 | 2017-06-13 | 浙江工业大学 | A kind of sacred disease analysis method based on brain network structure |
| CN108354605A (en) * | 2017-12-20 | 2018-08-03 | 佛山科学技术学院 | The extraordinary teenager's brain function of mathematics based on tranquillization state EEG signal connects network analysis method |
| CN110827271A (en) * | 2019-11-13 | 2020-02-21 | 上海交通大学 | Method for constructing brain function network with improved individual identification capability based on functional image data |
| CN110838173B (en) * | 2019-11-15 | 2023-06-02 | 天津医科大学 | Individualized brain co-transformation network construction method based on three-dimensional texture features |
| CN114048848B (en) * | 2022-01-13 | 2022-05-10 | 深圳市永达电子信息股份有限公司 | Brain-like computing method and system based on memory mechanism |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9451883B2 (en) * | 2009-03-04 | 2016-09-27 | The Regents Of The University Of California | Apparatus and method for decoding sensory and cognitive information from brain activity |
| CN102366323B (en) * | 2011-09-30 | 2013-09-11 | 中国科学院自动化研究所 | Detection method for causal connection strength of magnetic resonance brain imaging based on PCA (Principal component analysis) and GCA (Granger causality analysis) |
| CN102722727B (en) * | 2012-06-11 | 2014-03-05 | 杭州电子科技大学 | Electroencephalogram feature extracting method based on brain function network adjacent matrix decomposition |
| CN104050394B (en) * | 2014-07-04 | 2017-10-17 | 北京师范大学 | The modeling method and its topological attribute analysis method of colony's brain network |
-
2016
- 2016-04-14 CN CN201610230029.8A patent/CN105931281B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105931281A (en) | 2016-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105931281B (en) | The method that brain function network is quantitatively portrayed based on network characterization entropy | |
| Qi et al. | The influence of construction methodology on structural brain network measures: A review | |
| Fukushima et al. | Fluctuations between high-and low-modularity topology in time-resolved functional connectivity | |
| Graham et al. | Criteria for analysis of multicomponent tissue T2 relaxation data | |
| EP1946701B1 (en) | Brain function analysis method and brain function analysis program | |
| Alexander | An introduction to computational diffusion MRI: the diffusion tensor and beyond | |
| Zhu et al. | Statistical analysis of diffusion tensors in diffusion-weighted magnetic resonance imaging data | |
| US6845342B1 (en) | Determination of an empirical statistical distribution of the diffusion tensor in MRI | |
| DSouza et al. | Mutual connectivity analysis of resting-state functional MRI data with local models | |
| CN107240125B (en) | Diffusion weighted imaging method | |
| CN104379057A (en) | Magnetic resonance imaging device | |
| Beall et al. | The non-separability of physiologic noise in functional connectivity MRI with spatial ICA at 3 T | |
| Sourbron et al. | Choice of the regularization parameter for perfusion quantification with MRI | |
| CN100348152C (en) | Method for eliminating functional magnetic resonance data noise based on independent component space relativity | |
| Frank et al. | Dynamic multiscale modes of resting state brain activity detected by entropy field decomposition | |
| WO2022083258A1 (en) | Electrical impedance imaging method, system, storage medium, and electronic device | |
| CN108294754B (en) | Magnetic resonance multi-parameter quantitative imaging method and device | |
| Lee et al. | Group sparse dictionary learning and inference for resting-state fMRI analysis of Alzheimer's disease | |
| JP4271873B2 (en) | Computer-readable medium and apparatus for analyzing diffusion tensor magnetic resonance signals | |
| Nagler et al. | Maximum likelihood estimation of cardiac fiber bundle orientation from arbitrarily spaced diffusion weighted images | |
| Hickey et al. | A method for the clinical measurement of relaxation times in magnetic resonance imaging | |
| WO2006091983A2 (en) | Variable diffusion-time magnetic resonance-based system and method | |
| CN113391250B (en) | Tissue attribute multi-parameter quantitative test system and method thereof | |
| McVeigh et al. | Optimization of MR protocols: A statistical decision analysis approach | |
| Luo et al. | Analysis of fMRI data with drift: Modified general linear model and Bayesian estimator |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |