CN106667490B - Relationship analysis method between the data of individual difference between a kind of tested object based on magnetic resonance brain image - Google Patents
Relationship analysis method between the data of individual difference between a kind of tested object based on magnetic resonance brain image Download PDFInfo
- Publication number
- CN106667490B CN106667490B CN201710015079.9A CN201710015079A CN106667490B CN 106667490 B CN106667490 B CN 106667490B CN 201710015079 A CN201710015079 A CN 201710015079A CN 106667490 B CN106667490 B CN 106667490B
- Authority
- CN
- China
- Prior art keywords
- tested object
- brain
- data
- variable
- magnetic resonance
- 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
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0033—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room
- A61B5/004—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room adapted for image acquisition of a particular organ or body part
- A61B5/0042—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room adapted for image acquisition of a particular organ or body part for the brain
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2576/00—Medical imaging apparatus involving image processing or analysis
- A61B2576/02—Medical imaging apparatus involving image processing or analysis specially adapted for a particular organ or body part
- A61B2576/026—Medical imaging apparatus involving image processing or analysis specially adapted for a particular organ or body part for the brain
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30016—Brain
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Heart & Thoracic Surgery (AREA)
- Pathology (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Biophysics (AREA)
- Theoretical Computer Science (AREA)
- Veterinary Medicine (AREA)
- Biomedical Technology (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Signal Processing (AREA)
- Neurology (AREA)
- Physiology (AREA)
- Psychiatry (AREA)
- Quality & Reliability (AREA)
- High Energy & Nuclear Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention provides relationship analysis method between the data of individual difference between a kind of tested object based on magnetic resonance brain image, including：A, the magnetic resonance brain image of each tested object is obtained；B, at least one parametric variable corresponding at least one analyzed unit and each analyzed unit based on each tested object magnetic resonance brain image, the multidimensional variable of each tested object is built；C, according to the multidimensional variable of each tested object, the characteristic distance between the multidimensional variable of tested object two-by-two is calculated respectively, and build the fisrt feature distance matrix of each tested object accordingly；D, second of data of each tested object are obtained, calculate the characteristic distance between second of data of tested object two-by-two respectively；The second feature distance matrix of each tested object is built accordingly；E, according to the similarity size of first, second characteristic distance matrix, determine that the parametric variable of the analyzed unit of the tested object and the correlation of second of data are strong and weak.
Description
Technical field
The present invention relates to brain image field, individual between more particularly to a kind of tested object based on subject magnetic resonance brain image
Relationship analysis method between the data of difference.
Background technology
At present, the Neuscience analysis method based on image at this stage is typically to investigate the brain image of single analytic unit
The relation of information and other data in addition to the analyzed image information.For example, in the prior art in contrast mongolism
When with the brain image difference of normal person, for the complete brain area image of each tested object, it is assumed that by n brain area structure
Into only this n brain area in independent analysis brain image one by one is carried out after the independent brain area of each tested object brain image is selected
Comparative analysis one by one, to attempt to explore mongolism and the difference brain area of normal person.This mode have ignored image data not
With the relative information between analytic unit, i.e. only independently consider multiple brain areas, and have ignored in n brain area, each brain area it
Between relative information, i.e., do not consider the variation between each brain area and correlation.
Meanwhile the neuroimaging analysis method of nuclear magnetic resonance image data is used based on brain image initial data mostly at present
Group difference compares, be tested between correlation analysis etc., have ignored using the difference between being tested two-by-two as data elementary cell to deploy point
Analysis.
As can be seen here, prior art have ignored relative information between the multiple analytic units of tested individual brain image and two-by-two
Individual difference information between subject, partly hinder to tested object brain image information and in addition to the analyzed image information
The relation research of other data.
The content of the invention
In view of this, it is a primary object of the present invention to provide individual between a kind of tested object based on magnetic resonance brain image
Relationship analysis method between the data of difference, the phase between being tested two-by-two by using multidimensional variable measurement in nuclear magnetic resonance image data
Like property (i.e. feature space distance), and analyze similarity and other data in addition to the image data between the subject of image data
Subject between relation between similarity, with determine the tested object brain image multidimensional variable with except the analyzed image is believed
The correlation of other data outside breath.
Relationship analysis side between the data of individual difference between a kind of tested object based on magnetic resonance brain image of the application offer
Method, it the described method comprises the following steps：
A, the magnetic resonance brain image of each tested object is obtained；
B, at least one analyzed unit based on each tested object magnetic resonance brain image and each it is analyzed unit
Corresponding at least one parametric variable, build the multidimensional variable of each tested object；
C, according to the multidimensional variable of each tested object, calculate respectively tested object two-by-two the multidimensional variable it
Between characteristic distance, and build the fisrt feature distance matrix of each tested object accordingly；
D, second of data of each tested object are obtained, are calculated respectively between second of data of tested object two-by-two
Characteristic distance；The second feature distance matrix of each tested object is built accordingly；
E, according to the similarity size of first, second characteristic distance matrix, the described analyzed of the tested object is determined
The parametric variable of unit and the correlation of second of data are strong and weak.
By upper, similitude (the i.e. feature between being tested two-by-two by using multidimensional variable measurement in nuclear magnetic resonance image data
Space length), and analyze between the subject of image data phase between the subject of similarity and other data in addition to the image data
Like the relation between degree, to determine the tested object brain image multidimensional variable and its in addition to the analyzed image information
The correlation of his data.
Preferably, the multidimensional variable of each tested object includes：
The n1 that 1 analyzed unit based on the tested object, n1 parametric variable corresponding to the analyzed unit are formed
Tie up variable；Or
What 1 parametric variable corresponding to the m2 based on the tested object analyzed units, each analyzed unit was formed
M2 ties up variable；Or
What n3 parametric variable corresponding to the m3 based on the tested object analyzed units, each analyzed unit was formed
M3*n3 ties up variable；
Wherein：N1 is more than 1, m2 and is more than 1, m3 more than 2, n3 more than 2.
Preferably, the nuclear magnetic resonance image includes following any：Structure nuclear magnetic resonance image, disperse nuclear magnetic resonance image, task
State functional magnetic resonance imaging, tranquillization state functional magnetic resonance imaging.
Preferably, the analyzed unit comprises at least following one：Whole brain, cerebral hemisphere, by any individual element
Each area-of-interest of composition, the brain connection between each area-of-interest.
Preferably, the parametric variable is any quantifiable brain index；
Wherein, the brain index includes：Brain magnetic resonance signal intensity, ectocinerea index of correlation, white matter of brain index of correlation, brain
Function activation index of correlation, brain tranquillization state function index of correlation, brain network connection data, brain network topology attribute or brain laterality
Index.
Preferably, the ectocinerea index of correlation includes following any：Grey matter cortical surface product, grey matter volume, grey matter skin
Thickness degree, gray matter concentration；
The white matter of brain index of correlation includes following any：Average dispersion coefficient, fractional anisotropy, relatively respectively to different
Property, axial dispersion coefficient and radial direction dispersion coefficient；
The brain function activation index of correlation includes following any：Activation area, intensity of activation；
The brain tranquillization state function index of correlation includes following any：Locally coherence, low-frequency oscillation amplitude, fraction low frequency
Amplitude.
Preferably, each parametric variable includes：
The one-dimensional vector being made up of the related one or more data of same index；
By the p different indexs p dimensional vectors that related one or more data are formed respectively, p is more than 1.
Preferably, second of data of the tested object comprise at least following one：
Each quilt that other analyzed units or other specification variable based on tested object magnetic resonance brain image extract
Try multidimensional variable, environmental data, genetic data, clinical data, the behavioral data of object.
By upper, relation between the data of individual difference between a kind of tested object based on magnetic resonance brain image of the application offer
Analysis method, by using similitude of the subject of multidimensional variable in magnetic resonance brain image data between, and analyze the image number
According to subject between relation between the subject of similarity and other data in addition to the image data between similarity, to determine
State the multidimensional variable and the correlation of second of data of the brain image of tested object.
Brief description of the drawings
Between a kind of tested object based on magnetic resonance brain image that Fig. 1 provides for present invention implementation between the data of individual difference
The flow chart of relationship analysis method；
Fig. 2 is the image schematic diagram example of corresponding diagram 1, wherein (A) is the full brain of each tested object magnetic resonance brain image
The spatial distribution map of multidimensional variable；(B) for each tested object magnetic resonance brain image multidimensional variable schematic diagram；(C) it is to adopt
The schematic diagram of characteristic distance between the multidimensional variable of certain two tested object is measured with linear correlative method；(D) it is to count respectively
Calculate the characteristic distance matrix schematic diagram two-by-two between the multidimensional variable of tested object, wherein each matrix pixel correspond to it is each
Characteristic distance value.
The example schematic of Fig. 3 (A) characteristic distance matrixes between the subject based on nuclear magnetic resonance image；(B) it is based on base
Because of the first example schematic of characteristic distance matrix between the subject of sequence information；(C) between the subject based on gene sequence information
Second example schematic of characteristic distance matrix.
Embodiment
Below in conjunction with accompanying drawing 1-3, the embodiment of the present invention is described in detail.
Referring to Fig. 1, relation between the data of individual difference between a kind of tested object based on magnetic resonance brain image of the invention
Analysis method comprises the following steps：
S101, for s tested object, obtain the magnetic resonance brain image of each tested object；Wherein, the nuclear magnetic resonance image
It may include following any：Structure nuclear magnetic resonance image, disperse nuclear magnetic resonance image, task state functional magnetic resonance imaging, tranquillization state function
Nuclear magnetic resonance image.Wherein, the figure (A) in Fig. 2 examples show s=43.
S102, at least one analyzed unit based on each tested object magnetic resonance brain image and each analyzed
At least one parametric variable corresponding to unit, build the multidimensional variable of each tested object.
Wherein, the analyzed unit is the object being analyzed, and can be whole brain, cerebral hemisphere, by any
Brain connection between each area-of-interest or each area-of-interest of individual voxel composition.For example, based on magnetic resonance shadow
As data, cerebral cortex can be divided into different area-of-interests, i.e. Different brain region according to certain rule.In brain network point
In analysis, brain area is also referred to as the node of brain network, and these brain area/nodes can be as each analytic unit of the present invention.Similarly,
If calculating the correlation between each brain area, in brain network research, this correlation is referred to as brain connection.Each brain
Connection can also be as each analytic unit of the present invention.
Each analyzed unit can correspond at least one parameter and become there is at least one quantifiable brain desired value
Amount.Wherein, the brain index may include：Brain magnetic resonance signal intensity, ectocinerea index of correlation, white matter of brain index of correlation, brain work(
Index of correlation, brain tranquillization state function index of correlation, brain network connection data, brain network topology attribute or brain laterality can be activated to refer to
Mark etc..
And for above-mentioned each brain index, i.e., each parametric variable, it can also be made up of some subparameter variables, example
Such as：The ectocinerea index of correlation includes grey matter cortical surface product, grey matter volume, grey matter skin thickness, gray matter concentration etc.；It is described
White matter of brain index of correlation includes average dispersion coefficient, fractional anisotropy, relative anisotropies, axial dispersion coefficient and radial direction
Dispersion coefficient etc.；The brain function activation index of correlation includes activation area, intensity of activation etc.；The brain tranquillization state function is related
Index includes locally coherence, low-frequency oscillation amplitude, fraction low frequency amplitude etc..
Therefore for each tested object, it can be with according to its some analyzed unit, each analyzed list
(parametric variable and subparameter variable can for some subparameter variables that some parametric variables, each parametric variable corresponding to member include
It is referred to as parametric variable), the multidimensional variable of each tested object is constructed, is designated as t dimension variables herein, the figure (B) in Fig. 2 examples
It show t=90.Wherein, the t dimension variables of each tested object can be constructed as below：
Situation a：1 analyzed unit based on the tested object, n1 parametric variable corresponding to the analyzed unit, really
T=n1 is made, therefore builds n1 dimension variables, n1 is more than 1.Such as when each tested object uses left hemisphere as analyzed
Unit, using the n1 parameter such as the grey matter cortical surface product of left hemisphere, grey matter volume, grey matter skin thickness as n1 ginseng
Number variable, then each tested object be configured to n1 dimension variable；
Situation b：M2 based on the tested object analyzed units, each 1 parametric variable corresponding to analyzed unit,
T=m2 is determined, therefore builds m2 dimension variables, m2 is more than 1.Such as when each tested object uses m2 brain area as m2 quilt
Analytic unit, when each brain area uses this 1 parameter of brain network node efficiency as parametric variable, each tested object corresponds to m2
Tie up variable；
Situation c：M3 based on the tested object analyzed units, each n3 parameter corresponding to analyzed unit becomes
Amount, t=m3*n3 is determined, therefore build m3*n3 dimension variables, m3 is more than 2, n3 and is more than 2.Such as when each tested object uses
M3 brain area uses grey matter cortical surface product, grey matter volume, grey matter skin thickness etc. as m3 analyzed units, each brain area
When n3 parameter is as n3 parametric variable, each tested object corresponds to m3*m2 dimension variables.
It is described herein as, it is the above situation b as corresponding to Fig. 2 (A) (B) example, Fig. 2 (B) shows that abscissa is 43
Individual tested object label, ordinate are 90 nodal schemes, and the depth of each color lump corresponds to brain network node efficiency in coordinate, i.e.,
Represent：43 tested objects (i.e. s=43) are shared, each tested object, which employs 90 brain areas (or being node) and is used as, to be divided
Unit (i.e. m2=90) is analysed, each node is using the node efficiency as parametric variable.
Each parametric variable described in step S102, the parametric variable can be a value, or by same index phase
The one-dimensional vector that one or more data of pass are formed, or have the related one or more data institutes of multiple different indexs difference
The multi-C vector (such as the multi-C vector being made up of some subparameter variables) of composition.For example, when parametric variable is as shown in Figure 2
Brain network node efficiency when, each parametric variable be one value；And for example, the connection data when parametric variable between brain area
When (such as fiber number between brain area), then the parametric variable be the brain area and other multiple brain areas connection data (such as
Fiber number) one group of data (forming one-dimensional matrix) for being formed, i.e. one-dimensional vector；And for example, when the parametric variable also contains it
The value of his index (such as ectocinerea index of correlation), then together constitute multi-group data (forming multi-dimensional matrix), i.e. multidimensional by these
Vector.
Here, the principle of this step is further described below：Based on nuclear magnetic resonance image data, by cerebral cortex according to one
Fixed rule is divided into the node in different regions, i.e. brain area, also referred to as brain network.Afterwards, it is mutual between calculating brain area
Relation, this correlation are referred to as the side of brain network, and also referred to as brain connects.Brain bonding strength can be defined as form between brain area
Learn correlation (the i.e. ectocinerea structure of data (such as grey matter skin thickness, grey matter volume, grey matter cortical surface product, gray matter concentration)
Network), can also be defined as white matter fiber tract between two brain areas attribute (such as fiber number, fibre length, fibre density,
Average fractional anisotropy etc.) (i.e. white matter of brain structural network), the functional activity that can also be defined as between two brain areas when
Between linear or non-linear dependencies (i.e. brain tranquillization state functional network).Connected by calculating the brain between any two brain area,
The annexation between each node in network is obtained, this relation can be represented with adjacency matrix.By the analysis method of graph theory,
Calculate the topological attribute of the network.For example, the cerebral gray matter of each tested object is divided into 90 brain areas, (each brain area is one
Individual network node, namely an analytic unit), then calculating network nodal community, such as the network node attribute of calculating is net
Network node efficiency, so each tested object are achieved with 90 node efficiency values (i.e. parameter variable values).
S103, according to the multidimensional variable of each tested object magnetic resonance brain image, tested object two-by-two is calculated respectively
Characteristic distance (such as Pearson correlation coefficients) between the multidimensional variable of diencephalon image.
Above-mentioned s tested object is directed to, the t dimension variables of each object, tested object brain image t dimensions two-by-two is calculated and becomes
Characteristic distance between amount.Wherein, the algorithm of the characteristic distance of two groups of t dimension variables is calculated, existing algorithm or analysis can be used
Instrument, and non-invention focus, to this therefore are repeated no more.
In Fig. 2 examples, each tested object has 90 node efficiency values, this step is corresponded to, by calculating any two
The Pearson correlation coefficients of 90 node efficiency values between subject, the spy between tested object two-by-two is obtained according to the coefficient correlation
Levy distance.Fig. 2 (C) illustrates the feature between the data of the 26th tested object 90 dimension and the data of the 39th tested object 90 dimension
The measurement of distance, illustrate that using Pearson's measure of linear correlation characteristic distance in the figure.
S104, according to the characteristic distance between the multidimensional variable of the diencephalon image of tested object two-by-two, construct
The characteristic distance matrix of magnetic resonance brain image multidimensional variable, referred to herein as fisrt feature distance matrix between tested object.Such as Fig. 2
(D) each tested object characteristic distance matrix diagram is illustrated, wherein the multidimensional becomes between each pixel corresponds to tested object
The characteristic distance of amount, the depth of pixel grey scale represent the size of characteristic distance value.
S105, second of data of each tested object are obtained, calculate second of data of tested object two-by-two respectively
Between characteristic distance；The second feature distance matrix of each tested object is built accordingly.
Wherein, second of data of the tested object are the multidimensional variable in addition to the related brain image of the first data
Data.For example, the multidimensional variable data of other analytic units of the brain image of non-the first data correlation, environmental data, hereditary number
According to, clinical data, behavioral data.
S106, judge the fisrt feature distance matrix of the magnetic resonance brain image multidimensional variable and second of data
The similarity of second feature distance matrix, to determine each tested object brain image multidimensional variable, i.e., described tested object
The intensity of correlation between the parametric variable of the analyzed unit, with second of data described in each tested object.
Wherein, the fisrt feature distance matrix for judging the magnetic resonance brain image multidimensional variable and described second number
According to second feature distance matrix similarity method, using existing two matrix similarities determination methods, or analysis tool
To realize, repeat no more, such as following manner can be used：
Calculate the second feature distance matrix correlation of brain image feature fisrt feature distance matrix and second of data
(such as Pearson correlation coefficients etc.).Can to measure, whether two characteristic distance matrixes significantly correlated and phase using similar value
Pass degree.Pearson correlation coefficients analysis herein is one of which correlation analysis, and other may determine that the similar of the two
The correlation analysis of degree is equally applicable.
Below, specific characteristic distance matrix of the method for the invention with reference to shown in Fig. 3 obtained in actual experiment,
The principle of this step is explained further for example, Fig. 3 (A) is shown：22 mongolisms and 21 normal persons are taken respectively as quilt
Object is tried, the characteristic distance matrix schematic diagram of the multidimensional variable of the magnetic resonance brain image of this 43 tested objects, the figure shows two
The distribution of difference between two subjects.Observed for the ease of reader, 22 objects (need before mongolism tested object is used as in the figure
It is noted that herein only to facilitate reader, which is easier to diagramatic way directly perceived, understands the present invention just by mongolism subject pair
As showing the figure as preceding 22 objects, in fact, because the intensity of two matrix similarities of determination of this step can be by counting
Calculation machine performs judgement, and being need not 22 objects, 43 tested objects be arranged at random as before using mongolism tested object
Row), represent that characteristic distance is more and more remote by 0 to 1 in the diagram, the characteristic distance value illustrated from this can be substantially
Go out：The characteristic distance of 22 mongolisms between any two is nearer, is predominantly located in about 0.3；21 normal persons are between any two
Characteristic distance is nearer, is predominantly located in about 0.2；And the characteristic distance between mongolism and normal person is farther out, most of position
On 0.85.
Fig. 3 (B) is shown：Some or some hereditary feature numbers of above-mentioned 22 mongolism and 21 normal persons are taken respectively
Similarity measures two-by-two are carried out according to (can be one or more dimensions characteristic), hereditary feature data are using base in the example
Because of sequence information, the characteristic distance of the gene sequence information of tested individual two-by-two is calculated, obtains 43 tested objects
The gene sequence characteristic distance matrix, wherein, the order of tested object is identical with Fig. 3 (A).
When the similarity between two characteristic distance matrixes that Fig. 3 (A) and (B) are illustrated is higher, then it represents that, it is identified this
The multidimensional variable of 43 tested object magnetic resonance brain images, and the hereditary feature data, i.e. gene sequence information, between
Relationship degree it is stronger.
If the characteristic distance matrix of the hereditary feature data of 43 tested objects being obtained as described above is not such as Fig. 3 (B)
And the result as shown in Fig. 3 (C), that is, similarity is not high between representing the two characteristic distance matrixes, then it represents that identified
The multidimensional variable of this 43 tested object magnetic resonance brain images, the relationship degree between the gene sequence information be not high.
Wherein, Fig. 3 (C) is a kind of situation of hypothesis, and the figure actually obtained is Fig. 3 (B)；That is, this 43 tested objects
The multidimensional variable of magnetic resonance brain image, and the hereditary feature data, i.e. gene sequence information, between relationship degree compared with
By force, similarity is high.
It according to mode above, then can apply in data analysis process, such as can be used for judging in above-mentioned steps
Multidimensional variable used and the size of the relationship degree of second of data used in step S105 in S102, so as to for entering
The screening of row index.When for substantial amounts of tested object, have accumulated the multidimensional variable of a variety of brain indexs respectively, be also accumulated from
During polytype second of data, then method of the invention can be used, which kind of brain achievement data and any class second determined
The relationship degree of kind data is closer.
Further illustrate in addition, above-mentioned described image, refer to sensu lato image data, can be with image side
Formula exists or presents or exist in data format or present.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
God any modification, equivalent substitution and improvements made etc., should be included in the scope of the protection with principle.
Claims (8)
1. relationship analysis method, its feature exist between the data of individual difference between a kind of tested object based on magnetic resonance brain image
In, including step：
A, the magnetic resonance brain image of each tested object is obtained；
B, at least one analyzed unit and each analyzed unit based on each tested object magnetic resonance brain image are corresponding
At least one parametric variable, build the multidimensional variable of each tested object；
C, according to the multidimensional variable of each tested object, calculate respectively between the multidimensional variable of tested object two-by-two
Characteristic distance, and the fisrt feature distance matrix of each tested object is built accordingly；
D, second of data of each tested object are obtained, calculate the spy between second of data of tested object two-by-two respectively
Levy distance；The second feature distance matrix of each tested object is built accordingly；
E, according to the similarity size of first, second characteristic distance matrix, the analyzed unit of the tested object is determined
The parametric variable and second of data correlation it is strong and weak.
2. according to the method for claim 1, it is characterised in that the multidimensional variable of each tested object includes：
The n1 dimensions that 1 analyzed unit based on the tested object, n1 parametric variable corresponding to the analyzed unit are formed become
Amount；Or
The m2 dimensions that 1 parametric variable corresponding to m2 based on the tested object analyzed units, each analyzed unit is formed
Variable；Or
The m3* that n3 parametric variable corresponding to m3 based on the tested object analyzed units, each analyzed unit is formed
N3 ties up variable；
Wherein：N1 is more than 1, m2 and is more than 1, m3 more than 2, n3 more than 2.
3. method according to claim 1 or 2, it is characterised in that the magnetic resonance brain image includes following any：Structure
Magnetic resonance brain image, disperse magnetic resonance brain image, task state functional MRI brain image, tranquillization state functional MRI brain image.
4. method according to claim 1 or 2, it is characterised in that the analyzed unit comprises at least following one：It is whole
Brain connection between individual brain, cerebral hemisphere, each area-of-interest being made up of any individual element, each area-of-interest.
5. according to the method for claim 4, it is characterised in that the parametric variable is any quantifiable brain index；
Wherein, the brain index includes：Brain magnetic resonance signal intensity, ectocinerea index of correlation, white matter of brain index of correlation, brain function
Activation index of correlation, brain tranquillization state function index of correlation, brain network connection data, brain network topology attribute or brain laterality refer to
Mark.
6. according to the method for claim 5, it is characterised in that the ectocinerea index of correlation includes following any：Grey matter
Cortical surface product, grey matter volume, grey matter skin thickness, gray matter concentration；
The white matter of brain index of correlation includes following any：Average dispersion coefficient, fractional anisotropy, relative anisotropies, axle
To dispersion coefficient and radial direction dispersion coefficient；
The brain function activation index of correlation includes following any：Activation area, intensity of activation；
The brain tranquillization state function index of correlation includes following any：Locally coherence, low-frequency oscillation amplitude, fraction low frequency shake
Width.
7. according to the method for claim 5, it is characterised in that each parametric variable includes：
The one-dimensional vector being made up of the related one or more data of same index；
By the p different indexs p dimensional vectors that related one or more data are formed respectively, p is more than 1.
8. according to the method for claim 1, it is characterised in that second of data of the tested object comprise at least following
One：
Each subject pair that other analyzed units or other specification variable based on tested object magnetic resonance brain image extract
Multidimensional variable, environmental data, genetic data, clinical data, the behavioral data of elephant.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710015079.9A CN106667490B (en) | 2017-01-09 | 2017-01-09 | Relationship analysis method between the data of individual difference between a kind of tested object based on magnetic resonance brain image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710015079.9A CN106667490B (en) | 2017-01-09 | 2017-01-09 | Relationship analysis method between the data of individual difference between a kind of tested object based on magnetic resonance brain image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106667490A CN106667490A (en) | 2017-05-17 |
| CN106667490B true CN106667490B (en) | 2017-12-29 |
Family
ID=58850414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710015079.9A Active CN106667490B (en) | 2017-01-09 | 2017-01-09 | Relationship analysis method between the data of individual difference between a kind of tested object based on magnetic resonance brain image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106667490B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108345903B (en) * | 2018-01-25 | 2019-06-28 | 中南大学湘雅二医院 | A kind of multi-modal fusion image classification method based on mode distance restraint |
| CN108520283B (en) * | 2018-04-13 | 2021-04-27 | 山东管理学院 | Medical image classification method constructed based on brain region indirect relation network |
| CN109102492B (en) * | 2018-07-02 | 2020-06-09 | 重庆邮电大学 | Functional magnetic resonance image brain connection map construction method based on collaborative clustering |
| CN110148108A (en) * | 2019-03-27 | 2019-08-20 | 深圳市南山区人民医院 | Herpes zoster neuralgia curative effect prediction method and system based on functional MRI |
| CN110197729A (en) * | 2019-05-20 | 2019-09-03 | 华南理工大学 | Tranquillization state fMRI data classification method and device based on deep learning |
| CN111870808B (en) * | 2020-07-27 | 2024-08-09 | 杭州般意科技有限公司 | Intervention method for transcranial direct current stimulation |
| CN114300044B (en) * | 2021-12-31 | 2023-04-18 | 深圳华大医学检验实验室 | Gene evaluation method, gene evaluation device, storage medium, and computer device |
| CN115359305B (en) * | 2022-10-19 | 2023-01-10 | 之江实验室 | Accurate positioning system for abnormal area of cerebral fiber bundle |
| CN115937129B (en) * | 2022-12-01 | 2024-04-02 | 北京邮电大学 | Method and device for processing left and right half brain relations based on multi-mode magnetic resonance image |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9072496B2 (en) * | 2012-02-02 | 2015-07-07 | International Business Machines Corporation | Method and system for modeling and processing fMRI image data using a bag-of-words approach |
| US9265441B2 (en) * | 2013-07-12 | 2016-02-23 | Siemens Aktiengesellschaft | Assessment of traumatic brain injury |
| CN105046709B (en) * | 2015-07-14 | 2018-06-29 | 华南理工大学 | A kind of brain age analysis method based on nuclear magnetic resonance image |
-
2017
- 2017-01-09 CN CN201710015079.9A patent/CN106667490B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN106667490A (en) | 2017-05-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106667490B (en) | Relationship analysis method between the data of individual difference between a kind of tested object based on magnetic resonance brain image | |
| Amico et al. | Mapping hybrid functional-structural connectivity traits in the human connectome | |
| Abdelnour et al. | Functional brain connectivity is predictable from anatomic network's Laplacian eigen-structure | |
| Sloan et al. | Learning rigid image registration-utilizing convolutional neural networks for medical image registration | |
| CN101287410B (en) | Brain function analysis method and brain function analysis program | |
| Hagmann et al. | Mapping human whole-brain structural networks with diffusion MRI | |
| Cherukuri et al. | Deep MR brain image super-resolution using spatio-structural priors | |
| CN103886328A (en) | Functional magnetic resonance image data classification method based on brain network modular structure characteristics | |
| Knock et al. | The effects of physiologically plausible connectivity structure on local and global dynamics in large scale brain models | |
| Chung et al. | Exact topological inference for paired brain networks via persistent homology | |
| Kavitha et al. | Efficient DWT-based fusion techniques using genetic algorithm for optimal parameter estimation | |
| Chung et al. | Topological distances between brain networks | |
| Amico et al. | Centralized and distributed cognitive task processing in the human connectome | |
| CN110797123A (en) | Graph convolution neural network evolution method of dynamic brain structure | |
| CN109978872A (en) | White matter microstructure features screening system and method based on white matter fiber tract | |
| CN113476032B (en) | Brain structure and function coupling method based on directed graph harmonic analysis | |
| Tariq et al. | Why are deep representations good perceptual quality features? | |
| CN110443785A (en) | The feature extracting method of three-dimensional point cloud under a kind of lasting people having the same aspiration and interest | |
| CN104573745B (en) | Fruit-fly classified method based on magnetic resonance imaging | |
| Zhu et al. | Exploring high-order functional interactions via structurally-weighted lasso models | |
| Wang et al. | Topological inference on brain networks across subtypes of post-stroke aphasia | |
| Rafael-Patino et al. | The microstructural features of the diffusion-simulated connectivity (disco) dataset | |
| Anand et al. | Hodge-Decomposition of Brain Networks | |
| CN107240098A (en) | A kind of sparse Bayesian network of brain function linking parsing and the double constrained procedures of Granger | |
| Geng et al. | Image Fusion Adopting Guided Filter and Multi-resolution Singular Value Decomposition. |
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 |