CN113298038B - Construction method of multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD - Google Patents
Construction method of multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD Download PDFInfo
- Publication number
- CN113298038B CN113298038B CN202110684023.9A CN202110684023A CN113298038B CN 113298038 B CN113298038 B CN 113298038B CN 202110684023 A CN202110684023 A CN 202110684023A CN 113298038 B CN113298038 B CN 113298038B
- Authority
- CN
- China
- Prior art keywords
- network
- sub
- frequency
- brain
- networks
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 210000004556 brain Anatomy 0.000 title claims abstract description 117
- 238000003745 diagnosis Methods 0.000 title claims abstract description 35
- 238000010276 construction Methods 0.000 title abstract description 9
- 238000002599 functional magnetic resonance imaging Methods 0.000 claims abstract description 33
- 238000005065 mining Methods 0.000 claims abstract description 8
- 230000003925 brain function Effects 0.000 claims description 57
- 238000000034 method Methods 0.000 claims description 42
- 238000012545 processing Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 238000011282 treatment Methods 0.000 claims description 3
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 2
- 238000012217 deletion Methods 0.000 claims description 2
- 230000037430 deletion Effects 0.000 claims description 2
- 230000006870 function Effects 0.000 claims 1
- 208000024827 Alzheimer disease Diseases 0.000 description 53
- 230000007177 brain activity Effects 0.000 description 8
- 239000008280 blood Substances 0.000 description 4
- 210000004369 blood Anatomy 0.000 description 4
- 238000012937 correction Methods 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000000284 resting effect Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 3
- 238000009499 grossing Methods 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 229910052760 oxygen Inorganic materials 0.000 description 3
- 239000001301 oxygen Substances 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 239000010410 layer Substances 0.000 description 2
- 238000002595 magnetic resonance imaging Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000001638 cerebellum Anatomy 0.000 description 1
- 230000003920 cognitive function Effects 0.000 description 1
- 230000001054 cortical effect Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000004886 head movement Effects 0.000 description 1
- 239000011229 interlayer Substances 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005415 magnetization Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000002610 neuroimaging Methods 0.000 description 1
- 238000006213 oxygenation reaction Methods 0.000 description 1
- 238000010827 pathological analysis Methods 0.000 description 1
- 230000035479 physiological effects, processes and functions Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000002269 spontaneous effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Signal Processing (AREA)
- Probability & Statistics with Applications (AREA)
- Multimedia (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The application discloses a construction method of a multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD, and belongs to the technical field of computer auxiliary diagnosis. Dividing time series signals of each region of interest of the brain in the fMRI data set containing AD and NC into a plurality of frequency bands; calculating correlation coefficients of any two regions of interest in the same frequency band; thresholding the correlation coefficient to construct each tested multi-frequency brain network, obtaining AD and NC two multi-frequency brain network data sets, and performing frequency sub-network mining to obtain two multi-frequency frequent sub-network sets; respectively calculating the distinguishing capability of each sub-network in the two multi-frequency frequent sub-networks, and respectively taking the first plurality of sub-networks with the strongest capability from the two sub-networks to combine and construct a sub-network pair of each frequency band, namely a distinguishing sub-network pair; and calculating the difference degree of the sub-network pairs, sorting the top k sub-network pairs with the largest difference degree of each frequency band according to the descending order of the difference degree, and selecting the top k' sub-network pairs as the sub-network pairs for auxiliary diagnosis of AD.
Description
Technical Field
The application designs a construction method of a multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD, belonging to the technical field of computer auxiliary diagnosis.
Background
Functional magnetic resonance imaging (functional Magnetic Resonance Imaging, fMRI) technology is an emerging neuroimaging method based on blood oxygen level dependent signals. The fMRI technology has the advantages of no wound and no radiation, provides superior spatial resolution and coverage, can effectively locate an activation region of brain functions, can generate a digitized MRI signal by a blood oxygen level dependent signal, can further perform spatial calibration, normalization, smoothing and other treatments, and is an ideal tool for researching the whole brain activity mode and relation, unlike a magnetization signal which is sampled into discrete data points.
The brain function magnetic resonance imaging in the resting state is to study spontaneous brain activities of people in the resting state of being awake and without specific tasks, and based on the resting brain function network artificially constructed, the brain activity state and interaction among various neurons or brain areas can be well described, some characteristics of a brain default network are studied and analyzed, the system shows functional connection of the whole brain, and the brain network foundation of normal cognitive functions of the human brain is revealed. The functional connectivity analysis and graph theory method can also be used to detect changes in brain function network topology of a brain region or whole brain region caused by certain diseases such as Alzheimer's Disease (AD) based on the established resting brain function network.
The brain function network constructed by the traditional method is mostly constructed based on time domain signals, so that detail differences of brain activity information at different frequencies are covered, namely, the network constructed under different frequency division scales is analyzed, and different conclusions can be obtained. As in the low frequency band, bilaterally homologous brain regions tend to be tightly connected and functional connectivity is generally greater, and the correlation of cortical networks is generally concentrated in the ultra-low frequency (0.01-0.06 Hz) while the connections of edge networks are distributed over a wider frequency range (0.01-0.14 Hz). These all demonstrate that the fMRI signal is frequency specific. Therefore, the analysis method of the brain function network constructed based on the single frequency cannot describe the frequency domain information of the brain activity signal more accurately. On the other hand, the traditional AD classification auxiliary diagnosis method based on the brain function network is mainly based on the analysis between AD and normal contrast groups, a series of features such as node degree, cluster coefficient, path length, centrality and the like are selected from network topology attributes, and are connected in series to form feature vectors, and then the feature vectors are used for subsequent classification by combining a machine learning method. However, such feature selection may lose some finer topology information in the network, such as the network's own topology and the common topology among the networks, which in turn affects the accuracy of the subsequent classification.
Disclosure of Invention
Aiming at the problems of the prior art, the application provides a construction method of a multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD, aiming at fully revealing detail differences of brain activity information at different frequencies through the constructed multi-frequency brain network region molecular network pair, describing frequency domain information of brain activity signals more accurately, extracting finer brain network topology structure information, and being capable of being used for auxiliary diagnosis of AD and other symptoms so as to help improve diagnosis accuracy of AD and other symptoms.
The technical scheme of the application is as follows:
a method for constructing a multi-frequency brain network region molecular network pair for aided diagnosis of AD, comprising the steps of:
step 1: obtaining fMRI original data, and processing the fMRI original data to obtain time sequence signals of each interested region of each tested brain; the fMRI original data comprise AD data and NC data of normal control;
step 2: frequency division processing is carried out on the time series signals of each region of interest, each time series signal is divided into a plurality of frequency bands, and frequency division time series signals of the frequency bands are obtained;
step 3: for each frequency band, calculating the correlation coefficient between any two interested areas, and representing the correlation between two interested brain area nodes in the same frequency band through the correlation coefficient;
step 4: thresholding the correlation coefficient to construct brain function networks of each tested frequency band to obtain each tested multi-frequency brain network, and further obtain AD multi-frequency brain network data sets and NC multi-frequency brain network data sets;
step 5: aiming at brain function networks of each frequency band in the AD multi-frequency brain network data set and the NC multi-frequency brain network data set, frequent sub-network mining is carried out to obtain frequent sub-networks of each frequency band in each data set, and then an AD multi-frequency frequent sub-network set and an NC multi-frequency frequent sub-network set are obtained;
step 6: aiming at each frequency band, respectively calculating the respective 2-classification distinguishing capability of all the sub-networks in the AD multi-frequency sub-network set and the NC multi-frequency sub-network set, namely the capability of distinguishing the AD and NC brain function networks, and respectively taking the first plurality of sub-networks with the strongest distinguishing capability in the two frequent sub-network sets to combine and construct a sub-network pair of each frequency band, namely a distinguishing sub-network pair;
step 7: calculating the difference degree of two sub-networks in each sub-network pair, sorting the sub-network pairs of each frequency band according to the sequence from big to small, and selecting the first k sub-network pairs from the sorting sequence of the sub-network pairs of each frequency band;
step 8: and sequencing the difference degrees of the first k sub-network pairs of the all frequency bands according to the sequence from large to small, and selecting the first k' sub-network pairs as final sub-network pairs for carrying out category prediction on a given brain network.
Further, according to the construction method of the multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD, the time series signal of each region of interest is subjected to frequency division processing by using a multi-scale wavelet transformation method.
Further, according to the construction method of the multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD, the correlation coefficient is a Pearson correlation coefficient.
Further, according to the method for constructing the multi-frequency brain network region sub-network pair for auxiliary diagnosis of AD, the method for mining frequent sub-networks for all brain function networks of each frequency band j in the multi-frequency brain network data set comprises:
(a) Counting the occurrence frequencies of all sides in all brain function networks of each frequency band j in the multi-frequency brain network data set, selecting sides with the frequency larger than a preset support degree, taking each side as a candidate sub-network, and obtaining a candidate sub-network set to be marked as S j 1 ；
(b) For set S j 1 Each candidate sub-network in (a) is expanded for K times at most in the form of adding one edge each time, and the formed set is respectively marked as S j 2 ，S j 3 ，…，S j K And during the expansion, the set S j 2 ，S j 3 ，…，S j K Repeating the sub-network deletion in the middle;
(c) Calculate set S j 2 ，S j 3 ，……，S j K The occurrence frequency of each sub-network in the multi-frequency brain network data set is selected, and the sub-networks with the occurrence frequency larger than the preset support degree are respectively marked as S j 2’ ，S j 3’ ，…，S j K’ Then get the frequent sub-network set s= { S of each frequency band j 1 ,S j 2’ ,S j 3’ ,…,S j K’ }。
Further, according to the method for constructing the multi-frequency brain network region molecular network pair for aiding diagnosis of AD, the support degree is defined as: for a given set of brain function networks P and sub-networks g therein s Sub-network g s Support in network set P (g s P) is defined as sub-network g s The number of occurrences in the network set P is a percentage of the total number of networks N in the network set P, namely:
wherein ,is sonNetwork g s Support set of G i Representing the networks in the network set P, N represents the number of networks contained in P.
Further, according to the method for constructing the multi-frequency brain network area sub-network pair for auxiliary diagnosis of AD, the method for calculating the respective 2-classification distinguishing capability of all sub-networks in the AD multi-frequency frequent sub-network set and all sub-networks in the NC multi-frequency frequent sub-network set for each frequency band respectively includes: suppose G + G is a brain function network set of a certain frequency band in the AD multi-frequency brain network data set - A brain function network set of a certain frequency band in NC multi-frequency brain network data set; from G + A certain frequent sub-network mined in (1) is g + From G - A certain frequent sub-network mined in (1) is g - The method comprises the steps of carrying out a first treatment on the surface of the If subnetwork g + (g - ) At G + (G - ) Is much more frequent than it is at G - (G + ) The sub-network is a distinguishing sub-network with a classification of 2 distinguishing capability, and the distinguishing capability of the sub-network on the AD and NC brain function networks is calculated according to a formula (5):
similarly, subnetwork g - The discrimination capability for the two types of brain function networks is calculated according to equation (6):
further, according to the method for constructing the multi-frequency brain network region sub-network pairs for auxiliary diagnosis of AD, the difference degree calculation formula of two sub-networks in each sub-network pair is as follows:
diff(g 1 ,g 2 )＝1-corr(g 1 ,g 2 ) (7)
wherein corr (g 1 ,g 2 ) For a pair of subnetworks (g 1 ,g 2 ) Sub-network g of (3) 1 ，g 2 The correlation between the two is calculated according to the formula (8):
corr(g 1 ,g 2 )＝μsimN(g 1 ,g 2 )+(1-μ)simS(g 1 ,g 2 ) (8)
in the above formula, simN (g 1 ,g 2 ) For two sub-networks g 1 ，g 2 The structural similarity of (2) is calculated according to the formula (9); simS (g) 1 ,g 2 ) For two sub-networks g 1 ，g 2 The similarity of the support sets of (2) is calculated according to the formula (10);
wherein ,E(g1 ) Representing a subnetwork g 1 Is a set of edges; e (g) 2 ) Representing a subnetwork g 2 Is a set of edges;
in the above, cov (g) 1 ，G + ) For sub-network g 1 In network set G + Support of (3); cov (g) 2 ，G - ) For sub-network g 2 In network set G - Support of (3).
Further, according to the method for constructing the multi-frequency brain network region sub-network pair for auxiliary diagnosis of AD, the method for predicting the category of the given brain network by utilizing the sub-network region sub-network pair comprises the following steps: using each of k' distinguished subnetwork pairs according to a decision ruleFor a given tested brain function network t i Class prediction is performed, and the result is recorded as m r Where r=1, 2, …, k' is predicted by calculation of the prediction m r The number of sub-network pairs being "AD" or "NC" is the ratio of the sub-network pairs to the total number k' to obtain pairs t i Final prediction of (2)Results;
the judging rule is as follows:
(1) If it isAnd->Then m is r ＝‘AD’；
(2) If it isAnd->Then m is r ＝‘NC’；
(3) If it isAnd->Then m is r ＝null；
(4) If it isAnd->Then m is r ＝null。
Compared with the prior art, the application has the following beneficial effects: based on the frequency division of the time sequence signals, a multi-frequency brain function network is established, the topological attribute difference of the brain function network under different frequency bands is fully considered, and the information of brain activities can be better reserved; the intra-group commonality and inter-group specificity of the tested brain function network set are fully characterized by constructing the sub-network pairs by excavating frequent sub-networks, so that the medical auxiliary diagnosis service can be better provided.
Drawings
FIG. 1 is a schematic flow chart of a method for constructing a multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD;
FIG. 2 is a flow chart of a method for constructing a multi-frequency brain network region molecular network pair for aided diagnosis of AD according to the present application;
FIG. 3 is a schematic diagram of a method for mining frequent subnetworks according to an embodiment of the present application;
FIG. 4 is a flow chart of a method for constructing regional sub-network pairs using frequent sub-networks in accordance with the present application.
Detailed Description
In order that the application may be readily understood, a more complete description of the application will be rendered by reference to the appended drawings. The drawings illustrate preferred embodiments of the application. This application may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
Fig. 1 is a flow chart of a construction method of a multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD according to the present application, and fig. 2 is a specific flow chart of a construction method of a multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD according to the present application. The method for constructing the brain network multi-frequency region molecular network pair comprises the following steps:
step 1: acquiring an ADNI public data set and processing fMRI original data in the ADNI data set to obtain time sequence signals of each interested region of each tested brain; the fMRI original data comprises two types of data, namely AD and Normal Control (NC);
the specific contents include: preprocessing fMRI original data in an ADNI data set to obtain a required fMRI sample set, wherein the required fMRI sample set comprises an AD subsampleset composed of AD fMRI samples and an NC subsampleset composed of NC fMRI samples; then, for one sample, obtaining fMRI data of different brain areas by using a brain area dividing template, so that fMRI data of different brain areas of each sample can be obtained; and finally, taking the average value of the data of each time point of one region of interest of one sample as the value of the time point of the region, and forming a time sequence signal of each region of interest of the brain by the values of a plurality of time points of the same region, so that the time sequence signals of each region of interest of the brain of the sample can be obtained, and further, the time sequence signals of each region of interest of the brain of all samples can be obtained according to the same method.
Those of ordinary skill in the art will readily recognize that preprocessing the raw data of fMRI is a conventional processing step, and the goal of preprocessing is generally to eliminate extraneous information in the image, recover useful real information, enhance the detectability of the relevant information, and minimize data, thereby preparing qualified or optimized fMRI image data for subsequent steps of the method. In a preferred embodiment, the preprocessing of fMRI raw data includes a spatio-temporal correction process, a normalization process, and a smoothing filter process. Specifically, first removing fMRI images corresponding to unstable time points in the fMRI raw dataset; correcting fMRI images of all layers obtained by adopting a interlayer scanning method to the same time point, so as to eliminate time phase difference between layers caused by different acquisition time during scanning; then, performing head motion correction on the fMRI image subjected to time correction to eliminate the influence of tested fatigue and tested physiology caused by long acquisition time and ensure the usability of data; then, unified standardization is carried out on all tested brain fMRI images subjected to head movement correction, and the brain fMRI images are matched to the same space so as to facilitate subsequent statistics and space positioning; and finally, smoothing and filtering the standardized image to eliminate image noise.
In a preferred embodiment, the brain region segmentation template is a AAL (Anatomical Automatic Labeling) template. Specifically, the preprocessed fMRI sample is matched with the AAL template, 26 regions related to cerebellum in the matching result are removed, the remaining 90 brain regions are used as 90 nodes of the brain function network, and the average value of all voxel BOLD (Blood Oxygenation Level Dependent, blood oxygen level dependence) signal intensities in the brain regions is used as the representative of the activity condition of neurons in the brain regions, namely the value of the nodes.
Step 2: frequency division processing is carried out on the time series signals of each region of interest, each time series signal is divided into a plurality of frequency bands, and frequency division time series signals of the frequency bands are obtained;
in a preferred embodiment, the time series signals of each region of the brain are subjected to frequency division processing by using a multi-scale wavelet transform method, and each time series signal is divided into a plurality of frequency bands to obtain frequency division time series signals of the plurality of frequency bands.
The wavelet multi-scale transformation gradually performs multi-scale refinement on the signal through expansion and translation transformation, so that frequency subdivision at low frequency and time subdivision at high frequency are achieved. Assuming that the fMRI time-series signal is x (n), x (n) is respectively passed through a high-pass filter and a low-pass filter, and after x (n) is decomposed by wavelet transform, a result of one frequency band, for example, the e-th frequency band is:
in the above formula, n represents the number of time points included after pretreatment in one fMRI scan; z represents the number of divided frequency bands; z is a count variable; x is x e,H (n) represents the high frequency part of the e-th frequency band signal decomposition, and is also the frequency division result of the frequency band; x is x e,L (n) represents a low frequency part of the e-th frequency band signal decomposition, which is also an input signal for the next cascade decomposition; g (·) and h (·) are a high-pass filter and a low-pass filter, respectively.
Step 3: for each frequency band, calculating the correlation coefficient between any two interested areas, and representing the correlation between two interested brain area nodes in the same frequency band through the correlation coefficient;
the connection of the brain function network refers to the dynamic coordination of the activities between the neurons of the different brain regions. The weight of the connection represents the strength of the interaction relationship between the two brain regions, and the larger the weight is, the stronger the coupling between the brain regions is. After the frequency-divided time series signal is obtained, the relevance between two brain area nodes in the same frequency band can be represented by a correlation coefficient.
In a preferred embodiment, the correlation between two brain area nodes within the same frequency band is characterized by pearson correlation coefficients. The pearson correlation coefficient, as a linear correlation coefficient, may be used to reflect the degree of linear correlation of two variables. The preferred embodiment calculates the pearson correlation coefficient between any two brain region time series signals within the same frequency band according to equation (3).
wherein ,rij A pearson correlation coefficient between the time-series signal representing the ith brain region and the time-series signal representing the jth brain region; x is x i (i＝1,2,…,n)，y j (j=1, 2, …, n) respectively representing the time-series signal of the i-th brain region and the time-series signal of the j-th brain region;representing the mean value of the ith brain region time series signal and the mean value of the jth brain region time series signal; n represents the number of time points included after pretreatment in one scan.
Step 4: thresholding the correlation coefficient, constructing brain function networks of each tested frequency band, obtaining each tested multi-frequency brain network, and further obtaining an AD multi-frequency brain network data set consisting of all AD tested multi-frequency brain networks in the ADNI data set and an NC multi-frequency brain network data set consisting of all NC tested multi-frequency brain networks in the ADNI data set;
those skilled in the art will readily know that the range of the correlation coefficient value is (-1, 1), when thresholding is performed, it is assumed that the threshold value is set as T, the absolute value of the calculated correlation coefficient corresponding to each brain region in the same frequency band is compared with the threshold value T, and when the absolute value of the correlation coefficient between nodes is greater than the threshold value T, the connection between nodes is considered to be valid, and the value is set as 1; in contrast, when the absolute value of the correlation coefficient between the nodes is smaller than the threshold value, the connection between the nodes is considered invalid, and the value is set to 0. Thus, the brain function network of each frequency band can be constructed, and the multi-frequency brain network can be obtained.
Step 5: for an AD multi-frequency brain network data set and an NC multi-frequency brain network data set, aiming at brain function networks of all frequency bands, frequent sub-networks are mined, frequent sub-networks of all frequency bands in each data set are obtained, and then an AD multi-frequency frequent sub-network set formed by the frequent sub-networks of all frequency bands in the AD multi-frequency brain network data set and an NC multi-frequency frequent sub-network set formed by the frequent sub-networks of all frequency bands in the NC multi-frequency brain network data set are obtained;
the goal of frequent subnetwork mining of brain function networks is to mine frequently occurring subnetwork patterns in a given brain function network dataset to characterize the topology common to these brain function networks. Specifically, in the brain function network set under each frequency band, searching all the minimum supporters S meeting the preset min Is a frequent sub-network of (c). For a given set of brain function networks P and sub-networks g therein s Sub-network g s Support in network set P (g s P) is defined as sub-network g s The number of occurrences in the network set P is a percentage of the total number of networks N in the network set P, namely:
wherein ,is the sub-network g s Support set of G i Representing the networks in the network set P, N represents the number of networks contained in P.
In this embodiment, a specific method for mining frequent subnetworks for all brain function networks in one frequency band, such as the jth frequency band, in the multi-frequency brain network data set includes:
(a) Counting the occurrence frequencies of all sides in all brain function networks of the jth frequency band in the multi-frequency brain network data set, selecting the side with the frequency larger than the preset support degree,and each edge is used as a candidate sub-network, and the set of the candidate sub-networks is denoted as S j 1 ；
(b) For set S j 1 Each candidate sub-network in (a) is expanded for K times at most in the form of adding one edge each time, and the formed set is respectively marked as S j 2 ，S j 3 ，…，S j K . During the expansion process, the set S j 2 ，S j 3 ，…，S j K Is deleted by the repeated sub-network.
(c) Calculate set S j 2 ，S j 3 ，……，S j K The occurrence frequency of each sub-network in the multi-frequency brain network data set is selected, and the sub-networks with the occurrence frequency larger than the preset support degree are respectively marked as S j 2’ ，S j 3’ ，…，S j K’ Then a frequent sub-network set s= { S of the jth frequency band can be obtained j 1 ,S j 2’ ,S j 3’ ,…,S j K’ }。
Thus, frequent subnetworks can be respectively mined from all brain function networks of each frequency band in the AD multi-frequency brain network data set and the NC multi-frequency brain network data set to obtain frequent subnetworks of each frequency band in each data set, the frequent subnetworks of each frequency band in the AD multi-frequency brain network data set form an AD multi-frequency frequent subnetwork set, and the frequent subnetworks of each frequency band in the NC multi-frequency brain network data set form an NC multi-frequency frequent subnetwork set.
In a preferred embodiment, as shown in fig. 3, DFS encoding is performed on the candidate sub-network first, and then an edge is added to the L-edge frequent sub-network by using the rightmost path extension mode, so as to obtain the l+1-edge frequent sub-network. In building the depth-first search space tree, there may be multiple different DFS codes for the same candidate sub-network, but where only the smallest DFS code may uniquely identify a brain network, i.e., the other non-smallest DFS codes of the candidate sub-network are redundant. Redundant DFS codes and all their child nodes are tailored using Apriori properties. And calculating the occurrence frequency of the generated candidate sub-network in the brain function network data set, and deleting the candidate sub-network which does not meet the preset support degree condition, so that the frequent sub-network of each frequency band in each data set can be obtained.
Step 6: as shown in fig. 4, for each frequency band, the respective 2-class distinguishing capability of all the subnetworks in the AD multi-frequency subnetwork set and all the subnetworks in the NC multi-frequency subnetwork set, that is, the capability of distinguishing the AD and NC two brain function networks, is calculated, and the first several subnetworks with the strongest distinguishing capability in the two frequent subnetworks are taken and combined to construct a subnetwork pair, that is, a distinguishing subnetwork pair, of each frequency band.
In the present embodiment, G is assumed + G is a brain function network set of a certain frequency band in the AD multi-frequency brain network data set - Is a brain function network set of a certain frequency band in NC multi-frequency brain network data set. From G + A certain frequent sub-network mined in (1) is g + From G - A certain frequent sub-network mined in (1) is g - . If subnetwork g + (g - ) At G + (G - ) Is much more frequent than it is at G - (G + ) The sub-network is a distinguishing sub-network with 2 classification distinguishing capability, and the distinguishing capability of the sub-network on the AD and NC brain function networks can be calculated according to a formula (5):
similarly, the sub-network g may be calculated according to equation (6) - Discrimination capability for two types of brain function networks:
step 7: calculating the difference degree of two sub-networks in each sub-network pair, sequencing the sub-networks according to the sequence from the big difference degree to the small difference degree, and selecting the first k sub-network pairs from the sequencing sequence;
in the present embodiment, the sub-network pair (g 1 ,g 2 ) Calculating two sub-networks g in a sub-network pair 1 ，g 2 The expression of the degree of difference of (c) is as follows:
diff(g 1 ,g 2 )＝1-corr(g 1 ,g 2 ) (7)
wherein corr (g 1 ,g 2 ) For two sub-networks g 1 ，g 2 The correlation between the two is calculated according to the formula (8):
corr(g 1 ,g 2 )＝μsimN(g 1 ,g 2 )+(1-μ)simS(g 1 ,g 2 ) (8)
in the above formula, simN (g 1 ,g 2 ) For two sub-networks g 1 ，g 2 The structural similarity of (2) is calculated according to the formula (9); simS (g) 1 ,g 2 ) For two sub-networks g 1 ，g 2 Is calculated according to equation (10).
Where E (g) represents the edge set of sub-network g.
In the above, cov (g) 1 ，G + ) For sub-network g 1 In network set G + Support of (3); cov (g) 2 ，G - ) For sub-network g 2 In network set G - Support of (3).
Step 8: the difference degree of the first k sub-network pairs of all the frequency bands is ordered according to the order from big to small, the first k' sub-network pairs are selected as final sub-network pairs for carrying out 2-class distinction on a given brain network, and whether the given brain network is an AD brain function network or an NC brain function network is distinguished, so that the auxiliary diagnosis of AD is realized.
In the present embodiment, a brain function network t of an arbitrary subject is given i Utilizing the obtained k' differentiated sub-network pairs t i And performing category prediction, wherein the prediction result is used for auxiliary diagnosis of AD.
Using each of the k' sub-network pairs obtained by step 6 according to the determination rules described in (1) to (4) belowFor a given tested brain function network t i Class prediction is performed, and the result is recorded as m r Where r=1, 2, …, k':
(1) If it isAnd->Then m is r ＝‘AD’；
(2) If it isAnd->Then m is r ＝‘NC’；
(3) If it isAnd->Then m is r ＝null；
(4) If it isAnd->Then m is r ＝null；
According to the rule, the k' sub-network pairs are utilized to respectively pair t i Predicting and calculating a prediction result m r The ratio of the number of the sub-network pairs which are AD or NC to the total sub-network pair number k' is given to the pair t i If for t i And the final prediction result of the test pathological diagnosis label i If the brain function networks are consistent, the method of the application is demonstrated for the given tested brain function network t i The prediction is effective and can provide effective service for the auxiliary diagnosis of AD.
It will be appreciated by those skilled in the art in light of the present teachings that various modifications and changes can be made in light of the above teachings without departing from the spirit and scope of the application.
Claims (8)
1. A method for constructing a multi-frequency brain network region molecular network pair for aided diagnosis of AD, comprising the steps of:
step 1: obtaining fMRI original data, and processing the fMRI original data to obtain time sequence signals of each interested region of each tested brain; the fMRI original data comprise AD data and NC data of normal control;
step 2: frequency division processing is carried out on the time series signals of each region of interest, each time series signal is divided into a plurality of frequency bands, and frequency division time series signals of the frequency bands are obtained;
step 3: for each frequency band, calculating the correlation coefficient between any two interested areas, and representing the correlation between two interested brain area nodes in the same frequency band through the correlation coefficient;
step 4: thresholding the correlation coefficient to construct brain function networks of each tested frequency band to obtain each tested multi-frequency brain network, and further obtain AD multi-frequency brain network data sets and NC multi-frequency brain network data sets;
step 5: aiming at brain function networks of each frequency band in the AD multi-frequency brain network data set and the NC multi-frequency brain network data set, frequent sub-network mining is carried out to obtain frequent sub-networks of each frequency band in each data set, and then an AD multi-frequency frequent sub-network set and an NC multi-frequency frequent sub-network set are obtained;
step 6: for each frequency band, respectively calculating the respective capacity of distinguishing AD and NC brain function networks of all the sub-networks in the AD multi-frequency sub-network set and the NC multi-frequency sub-network set, and respectively combining a plurality of sub-networks with the strongest distinguishing capacity in the AD multi-frequency sub-network set and the NC multi-frequency sub-network set to construct a distinguishing sub-network pair of each frequency band;
step 7: calculating the difference degree of two sub-networks in each sub-network pair, sorting the sub-network pairs of each frequency band according to the sequence from big to small, and selecting the first k sub-network pairs from the sorting sequence of the sub-network pairs of each frequency band;
step 8: and sequencing the difference degrees of the first k sub-network pairs of the all frequency bands according to the sequence from large to small, and selecting the first k' sub-network pairs as final sub-network pairs for carrying out category prediction on a given brain network.
2. The method for constructing a multi-frequency brain network area molecular network pair for aiding diagnosis of AD according to claim 1, wherein the time-series signal of each region of interest is subjected to frequency division processing by using a multi-scale wavelet transform method.
3. The method for constructing a multi-frequency brain network region molecular network pair for aiding in the diagnosis of AD according to claim 1, wherein the correlation coefficient is a pearson correlation coefficient.
4. The method for constructing a multi-frequency brain network region sub-network pair for aiding diagnosis of AD according to claim 1, wherein the method for mining frequent sub-networks for all brain function networks of each frequency band j in the multi-frequency brain network data set comprises:
(a) Counting all of the brain function networks of each frequency band j in the multi-frequency brain network data setThe frequency of the edge occurrence, selecting the edge with the frequency larger than the preset support degree, taking each edge as a candidate sub-network, and obtaining a candidate sub-network set to be S j 1 ；
(b) For set S j 1 Each candidate sub-network in (a) is expanded for K times at most in the form of adding one edge each time, and the formed set is respectively marked as S j 2 ，S j 3 ，…，S j K And during the expansion, the set S j 2 ，S j 3 ，…，S j K Repeating the sub-network deletion in the middle;
(c) Calculate set S j 2 ，S j 3 ，……，S j K The occurrence frequency of each sub-network in the multi-frequency brain network data set is selected, and the sub-networks with the occurrence frequency larger than the preset support degree are respectively marked as S j 2’ ，S j 3’ ，…，S j K’ Then get the frequent sub-network set s= { S of each frequency band j 1 ,S j 2’ ,S j 3’ ,…,S j K’ }。
5. The method for constructing a multi-frequency brain network segment molecular network pair for aiding in the diagnosis of AD according to claim 4, wherein the degree of support is defined as: for a given set of brain function networks P and sub-networks g therein s Sub-network g s Support in network set P (g s P) is defined as sub-network g s The number of occurrences in the network set P is a percentage of the total number of networks N in the network set P, namely:
wherein ,is the sub-network g s Support set of G i Representing the networks in the network set P, N represents the number of networks contained in P.
6. The method for constructing a multi-frequency brain network area sub-network pair for auxiliary diagnosis of AD according to claim 1, wherein the method for calculating the respective 2-class distinguishing capacities of all sub-networks in the AD multi-frequency frequent sub-network set and all sub-networks in the NC multi-frequency frequent sub-network set for each frequency band respectively is as follows: suppose G + G is a brain function network set of a certain frequency band in the AD multi-frequency brain network data set - A brain function network set of a certain frequency band in NC multi-frequency brain network data set; from G + A certain frequent sub-network mined in (1) is g + From G - A certain frequent sub-network mined in (1) is g - The method comprises the steps of carrying out a first treatment on the surface of the If subnetwork g + (g - ) At G + (G - ) Is much more frequent than it is at G - (G + ) The sub-network is a distinguishing sub-network with a classification of 2 distinguishing capability, and the distinguishing capability of the sub-network on the AD and NC brain function networks is calculated according to a formula (5):
similarly, subnetwork g - The discrimination capability for the two types of brain function networks is calculated according to equation (6):
7. the method for constructing a multi-frequency brain network segment sub-network pair for aiding diagnosis of AD according to claim 5, wherein the difference degree calculation formula of two sub-networks in each segment sub-network pair is as follows:
diff(g 1 ,g 2 )＝1-corr(g 1 ,g 2 ) (7)
wherein corr (g 1 ,g 2 ) For a pair of subnetworks (g 1 ,g 2 ) Sub-network g of (3) 1 ，g 2 The correlation between the two is calculated according to the formula (8):
corr(g 1 ,g 2 )＝μsimN(g 1 ,g 2 )+(1-μ)simS(g 1 ,g 2 ) (8)
in the above formula, simN (g 1 ,g 2 ) For two sub-networks g 1 ，g 2 The structural similarity of (2) is calculated according to the formula (9); simS (g) 1 ,g 2 ) For two sub-networks g 1 ，g 2 The similarity of the support sets of (2) is calculated according to the formula (10);
wherein ,E(g1 ) Representing a subnetwork g 1 Is a set of edges; e (g) 2 ) Representing a subnetwork g 2 Is a set of edges;
in the above, cov (g) 1 ，G + ) For sub-network g 1 In network set G + Support of (3); cov (g) 2 ，G - ) For sub-network g 2 In network set G - Support of (3).
8. The method for constructing a multi-frequency brain network regional molecular network pair for aiding diagnosis of AD according to claim 1, wherein the method for performing category prediction for a given brain network using the regional molecular network pair is: using each of k' distinguished subnetwork pairs according to a decision ruleFor a given testBrain function network t i Class prediction is performed, and the result is recorded as m r Where r=1, 2, …, k' is predicted by calculation of the prediction m r The number of sub-network pairs being "AD" or "NC" is the ratio of the sub-network pairs to the total number k' to obtain pairs t i Final prediction results of (2);
the judging rule is as follows:
(1) If it isAnd->Then m is r ＝‘AD’；
(2) If it isAnd->Then m is r ＝‘NC’；
(3) If it isAnd->Then m is r ＝null；
(4) If it isAnd->Then m is r ＝null。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110684023.9A CN113298038B (en) | 2021-06-21 | 2021-06-21 | Construction method of multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110684023.9A CN113298038B (en) | 2021-06-21 | 2021-06-21 | Construction method of multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113298038A CN113298038A (en) | 2021-08-24 |
| CN113298038B true CN113298038B (en) | 2023-09-19 |
Family
ID=77328858
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110684023.9A Active CN113298038B (en) | 2021-06-21 | 2021-06-21 | Construction method of multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113298038B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108830827A (en) * | 2017-05-02 | 2018-11-16 | 通用电气公司 | Neural metwork training image generation system |
| CN109034263A (en) * | 2018-08-15 | 2018-12-18 | 东北大学 | The Alzheimer disease auxiliary diagnostic equipment and method of the brain network multi-frequency fusion kernel of graph |
| WO2021026400A1 (en) * | 2019-08-06 | 2021-02-11 | Neuroenhancement Lab, LLC | System and method for communicating brain activity to an imaging device |
| EP3816853A1 (en) * | 2019-10-31 | 2021-05-05 | NVIDIA Corporation | Gaze determination using one or more neural networks |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11228501B2 (en) * | 2019-06-11 | 2022-01-18 | At&T Intellectual Property I, L.P. | Apparatus and method for object classification based on imagery |
-
2021
- 2021-06-21 CN CN202110684023.9A patent/CN113298038B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108830827A (en) * | 2017-05-02 | 2018-11-16 | 通用电气公司 | Neural metwork training image generation system |
| CN109034263A (en) * | 2018-08-15 | 2018-12-18 | 东北大学 | The Alzheimer disease auxiliary diagnostic equipment and method of the brain network multi-frequency fusion kernel of graph |
| WO2021026400A1 (en) * | 2019-08-06 | 2021-02-11 | Neuroenhancement Lab, LLC | System and method for communicating brain activity to an imaging device |
| EP3816853A1 (en) * | 2019-10-31 | 2021-05-05 | NVIDIA Corporation | Gaze determination using one or more neural networks |
Non-Patent Citations (1)
| Title |
|---|
| 基于脑功能网络的阿尔茨海默症生物学标记物研究;陈金义;《中国优秀硕士学位论文全文数据库基础科学辑》(第5期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113298038A (en) | 2021-08-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107909117B (en) | Classification device for early and late mild cognitive impairment based on brain function network characteristics | |
| CN109409416B (en) | Feature vector dimension reduction method, medical image identification method, device and storage medium | |
| CN110188836B (en) | Brain function network classification method based on variational self-encoder | |
| Rajendran et al. | Hybrid medical image classification using association rule mining with decision tree algorithm | |
| CN113616184A (en) | Brain network modeling and individual prediction method based on multi-mode magnetic resonance image | |
| CN111967495A (en) | Classification recognition model construction method | |
| CN112348833B (en) | Dynamic connection-based brain function network variation identification method and system | |
| CN112634214A (en) | Brain network classification method combining node attributes and multilevel topology | |
| Liao et al. | Classify autism and control based on deep learning and community structure on resting-state fMRI | |
| CN112348090A (en) | Neighbor anomaly detection system based on neighbor self-encoder | |
| CN113947157A (en) | Dynamic brain effect connection network generation method based on hierarchical clustering and structural equation model | |
| CN116433590A (en) | Mental disorder population whole brain function analysis method and system based on functional magnetic resonance image | |
| CN113298038B (en) | Construction method of multi-frequency brain network region molecular network pair for auxiliary diagnosis of AD | |
| CN112883905B (en) | Human behavior recognition method based on coarse-grained time-frequency features and multi-layer fusion learning | |
| Saneipour et al. | Improvement of MRI brain image segmentation using Fuzzy unsupervised learning | |
| Preetha et al. | Normalization and deep learning based attention deficit hyperactivity disorder classification | |
| CN115760785A (en) | Brain somatotropin morphology high-order feature extraction method | |
| CN114287910A (en) | Brain function connection classification method based on multi-stage graph convolution fusion | |
| Shekerbek et al. | APPLICATION OF MATHEMATICAL METHODS AND MACHINE LEARNING ALGORITHMS FOR CLASSIFICATION OF X-RAY IMAGES. | |
| Vallaboju et al. | Bioinformatics image based decision support system for bone cancer detection | |
| Naveen et al. | A Novel Layer Based Logical Approach (LLA) Clustering Method for Performance Analysis in Medical Images | |
| CN116030941B (en) | Alzheimer's disease diagnosis method based on edge-centric effect connection network | |
| Emmanuel | Improved approach to feature dimension reduction for efficient diagnostic classification of breast cancer | |
| CN116070120B (en) | Automatic identification method and system for multi-tag time sequence electrophysiological signals | |
| CN114305451B (en) | Construction method of children brain electrical function connection map of entropy stability criterion |
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 |