CN111785379B - Brain function connection prediction method, device, computer equipment and storage medium - Google Patents
Brain function connection prediction method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111785379B CN111785379B CN202010573592.1A CN202010573592A CN111785379B CN 111785379 B CN111785379 B CN 111785379B CN 202010573592 A CN202010573592 A CN 202010573592A CN 111785379 B CN111785379 B CN 111785379B
- Authority
- CN
- China
- Prior art keywords
- brain function
- function connection
- brain
- covariance matrix
- prediction model
- 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
- 230000003925 brain function Effects 0.000 title claims abstract description 106
- 238000000034 method Methods 0.000 title claims abstract description 58
- 239000011159 matrix material Substances 0.000 claims abstract description 80
- 210000004556 brain Anatomy 0.000 claims abstract description 43
- 238000010586 diagram Methods 0.000 claims abstract description 21
- 230000002123 temporal effect Effects 0.000 claims abstract description 16
- 238000004590 computer program Methods 0.000 claims description 19
- 238000002599 functional magnetic resonance imaging Methods 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 11
- 238000011480 coordinate descent method Methods 0.000 claims description 7
- 230000006870 function Effects 0.000 description 26
- 238000004422 calculation algorithm Methods 0.000 description 16
- 238000004088 simulation Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 8
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 4
- 239000008280 blood Substances 0.000 description 4
- 210000004369 blood Anatomy 0.000 description 4
- 230000001419 dependent effect Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 229910052760 oxygen Inorganic materials 0.000 description 4
- 239000001301 oxygen Substances 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 210000001175 cerebrospinal fluid Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000004886 head movement Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000000284 resting effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002195 synergetic effect Effects 0.000 description 1
- 210000004885 white matter Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The embodiment of the invention discloses a brain function connection prediction method, a device, computer equipment and a storage medium, and relates to the technical field of brain connection prediction. The method comprises the following steps: acquiring a covariance matrix of BOLD signal sample data of a preset brain region; inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model; and solving the brain function connection prediction model to obtain a prediction result of brain function connection. Because two L0 normal form regular terms for respectively controlling and estimating the space sparsity and the time homogeneity of the connection matrix are adopted, the error of the brain function connection prediction model is small, the sparsity is strong and the accuracy is high.
Description
Technical Field
The present invention relates to the field of brain connection prediction technology, and in particular, to a brain function connection prediction method, device, computer apparatus, and storage medium.
Background
The synergic action of different brain regions of the brain constructs a high-efficiency brain function network, thereby realizing the behavior cognition function of people. In order to more clearly understand the operation mechanism of brain functions, more and more researchers are focusing on the study of brain Function Connections (FCs).
Many studies in recent years have shown that the brain function network exhibits significant dynamic behavior, even in a resting state. Many important and significant achievements are achieved for the research of the dynamic change of the brain network in time. Functional magnetic resonance imaging (fMRI) is commonly used as a non-invasive imaging technique for brain functional studies. To estimate the time-varying connected network of the brain from BOLD (blood oxygen level dependent, blood oxygen dependent level, derived from fMRI, which can indirectly reflect the activity level of neurons) signals, researchers have proposed many methods of estimating dynamic brain networks, with sliding window based methods being most commonly used.
The current method based on the sliding window and Gao Situ model achieves good effect and also obtains wide acceptance of researchers. Many studies have shown that the brain network structure has sparsity, i.e. there are only a few nodes connected in the network, thereby completing more efficient information interaction. Thus, in order to make the estimated functional network conform to the sparsity feature, researchers typically implement the method by adding L1 regularization terms to the estimated model (e.g., SINGLE algorithm, TVGL algorithm). When solving the optimization problem, the regularization term of the L1 normal form can ensure that the objective function is a convex function, so that the optimal solution can be conveniently obtained. At the same time, however, the disadvantages of the regularization term of the L1 paradigm are also apparent. The method can lead to shrinkage of estimation parameters, estimation errors are introduced, and meanwhile, in the problem of strong sparsity, the L1 regularization term cannot obtain enough sparsity.
Disclosure of Invention
The embodiment of the invention provides a brain function connection prediction method, a device, computer equipment and a storage medium, which aim to solve the problems of large error, poor sparsity and low accuracy of the existing brain function connection prediction model based on regular terms of an L1 paradigm.
In a first aspect, an embodiment of the present invention provides a brain function connection prediction method, including:
acquiring a covariance matrix of sample data of a BOLD signal of a preset brain region;
inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model;
and solving the brain function connection prediction model to obtain a prediction result of brain function connection.
In a second aspect, an embodiment of the present invention further provides a brain function connection prediction apparatus, including:
an acquisition unit for acquiring a covariance matrix of sample data of a BOLD signal of a preset brain region;
the input unit is used for inputting the covariance matrix into a preset brain function connection prediction model, and the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the space sparsity and time homogeneity of the connection matrix on the basis of a Gaussian diagram model;
and the prediction unit is used for solving the brain function connection prediction model to obtain a prediction result of brain function connection.
In a third aspect, an embodiment of the present invention further provides a computer device, which includes a memory and a processor, where the memory stores a computer program, and the processor implements the method when executing the computer program.
In a fourth aspect, embodiments of the present invention also provide a computer readable storage medium storing a computer program which, when executed by a processor, implements the above method.
By applying the technical scheme of the embodiment of the invention, the covariance matrix of the sample data of the BOLD signal of the preset brain region is obtained; inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model; and solving the brain function connection prediction model to obtain a prediction result of brain function connection. Because two L0 normal form regular terms for respectively controlling and estimating the space sparsity and the time homogeneity of the connection matrix are adopted, the error of the brain function connection prediction model is small, the sparsity is strong and the accuracy is high.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a brain function connection prediction method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a multi-channel simulation result in a small world network connection mode according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a multi-channel simulation result in a non-scale network connection mode according to an embodiment of the present invention;
fig. 4 is a schematic diagram of a multi-channel simulation result in a random network connection mode according to an embodiment of the present invention;
FIG. 5 is a schematic block diagram of a brain function connection prediction device according to an embodiment of the present invention;
FIG. 6 is a schematic block diagram of a prediction unit of a brain function connection prediction device according to an embodiment of the present invention;
FIG. 7 is a schematic block diagram of a brain function connection prediction apparatus according to another embodiment of the present invention;
fig. 8 is a schematic block diagram of a computer device according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
It should be understood that the terms "comprises" and "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should be further understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.
As used in this specification and the appended claims, the term "if" may be interpreted as "when..once" or "in response to a determination" or "in response to detection" depending on the context. Similarly, the phrase "if a determination" or "if a [ described condition or event ] is detected" may be interpreted in the context of meaning "upon determination" or "in response to determination" or "upon detection of a [ described condition or event ]" or "in response to detection of a [ described condition or event ]".
The meanings of the symbols appearing herein are shown in Table 1 below.
TABLE 1 meaning of symbols commonly used herein
Referring to fig. 1, fig. 1 is a flowchart of a brain function connection prediction method according to an embodiment of the invention. As shown, the method includes the following steps S1-S3.
S1, acquiring a covariance matrix of sample data of a BOLD signal of a preset brain region
In a specific implementation, the sample data of the BOLD signal of the preset brain region is the data obtained by preprocessing fMRI data. The sample data is typically a multidimensional vector. fMRI data refers to functional magnetic resonance data.
The preset brain region is a brain region preset by a user, namely, a brain region of interest to the user, and the brain region which the user wants to study.
Note that, the BOLD (blood oxygen level dependent, blood oxygen dependent level) signal of the brain region is obtained from fMRI data, and can indirectly reflect the activity level of neurons.
The covariance matrix of the sample data is typically calculated using a sliding window based approach.
In an embodiment, before step S1, the brain function connection prediction method further includes: preprocessing fMRI data to obtain sample data of BOLD signals of the preset brain region.
The specific pretreatment process is as follows:
preprocessing of fMRI data is accomplished based on SPM12 and DPABI kits of MATLAB platform. Before pretreatment, the data from the first few time points should be removed in order to avoid T1 balance effects. In the preprocessing process, slice time correction is firstly carried out to eliminate time errors between adjacent slices. Then, correction of head movements is performed, because some head shaking is unavoidable during the experiment, which will cause errors in the scanned data. These functional images are then registered with the subject's high resolution T1 image, segmented and normalized into standard MNI space to facilitate comparative analysis of different brain data. And then carrying out space smoothing processing to eliminate the influence of interference signals generated by hardware instability and physiological motion and improve the signal to noise ratio. The linear trend and other interfering signals (including Friston's 24 motion parameters, cerebrospinal fluid, white matter and global signals) were then regressed. In the frequency domain, the above processed signal is bandpass filtered, leaving a signal of the relevant frequency range (the frequency range of the BOLD signal is generally considered to be 0.01-0.1 Hz). The preprocessing is completed, so that a time series signal of each voxel after preprocessing can be obtained. But more times, time series signals of all voxels are not required, and then signals of brain regions of interest per se need to be extracted according to research needs. Before extracting the brain region signal of interest, the brain region of interest needs to be predefined, and can be defined manually by the user, or some commonly used templates can be selected, such as an AAL template, a Harvard-Oxford template and the like. After selecting the brain regions of interest, the extraction of these specific brain region signals can be performed to obtain a sequence signal of T x P, where T is the length of the time point and P is the number of brain regions of interest.
S2, inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0-norm regularization terms for respectively controlling and estimating the spatial sparsity and time homogeneity of the connection matrix on the basis of a Gaussian diagram model.
In specific implementation, the covariance matrix is input into a preset brain function connection prediction model, and the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model. For convenience of description, the above brain function connection prediction model is denoted as L0-SSR model.
Note that the mathematical expression of the Gao Situ model is as follows:
-logdetX t +tr(S t X t )， (1)
in one embodiment, the mathematical expression of the brain function connection prediction model (L0-SSR model) is as follows:
wherein S is t Covariance matrix of sample data at time t; x is X t The inverse covariance matrix at the moment t is an estimated connection matrix of brain function connection at the moment t; x is X t-1 An inverse covariance matrix at the time t-1, namely an estimated connection matrix for brain function connection at the time t-1; lambda (lambda) 1 Coefficients for controlling spatial sparsity; lambda (lambda) 2 To control the coefficient of temporal homogeneity.
By solving for X t And obtaining the prediction result of the brain function connection at the time t.
And S3, solving the brain function connection prediction model to obtain a prediction result of brain function connection.
In specific implementation, the brain function connection prediction model is solved to obtain a prediction result of brain function connection.
Specifically, in one embodiment, the brain function connection prediction model is solved by an iterative method based on a coordinate descent method.
The specific solving process is as follows:
an iterative method based on a coordinate descent method is introduced, wherein only one element in a matrix is optimized in each iteration, and other elements are fixed. Once the selected minimum is found, matrix X t Will be updated and continue to the optimization of the next element of the matrix in the next iteration. This update process may be expressed as:
wherein,representing updated element x ij The latter matrix, delta (·) function is defined as +.> The superscript k indicates the number of iterations. For simplicity, the time point indicator variable t is omitted in (3), and the following derivation process is omitted as well, if not necessary. Thus, for any i and j, the minimized objective function can be generally expressed as:
wherein s is ij Is the element of the ith row and jth column of the sample covariance matrix S. D if i=j ij Equal to 1, d if i +.j ij Equal to 2.
Minimizing ψ when i=j
If i=j, then the selected element to be updated is located on the diagonal of the matrix, in which case the optimization problem can be expressed as:
if x ii，t-1 Within the definition domain of the objective function, that is to sayThen the function is atWill be discontinuous. Because the last item of the function->Will be equal to 0, which will result in a mutation in the function value.
At the same time, if x ii，t-1 Not within the domain of the function, ψ (·) will be contiguous.
First, whenDefining a continuous function ψ (·) as:
thus, the objective function in equation (5) using equation (6) can be expressed as:
it can now be seen that the extreme point of the function ψ (·) is the extreme point or extreme point of the function c () (i.e. of the function (6))Since function (6) is a convex function and is derivative, the optimal solution can be obtained by making its inverse equal to 0.
The Sherman-Morrison-Woodbury formula is utilized to obtain:
here e i Is a vector (see table 1) with the i element of 1 and the rest of 0, θ is a constant, y=x -1 ,y ij Representing the elements of the ith row and jth column of matrix Y, Y [i] Is the ith column of matrix Y. Order theThe method comprises the following steps:
next, substituting equation (10) into equation (8) and solving for x, the extreme points are:
on the other hand, ifThen x ii,t-1 If not within the definition domain of the objective function, then the extreme point of ψ ()' may be defined by +.>Given.
The above discussion is summarized as follows:
·When
·When
minimizing ψ when i+.j
Next, a more complex case will be discussed, when i+.j, the selected element to be updated is in the off-diagonal position of the matrix. Similar to equation (5), an objective function can be obtained:
in this case, ifAnd->The L0-SSR problem has a break point at each of these two points, if within the definition domain.
As discussed in section 2.3.1, consider first the case where ψ (·) is continuous:
by making the derivative equal to 0, there is
Again by using the shaerman-Morrison-Woodbury formula, there are:
where I is an identity matrix, U ij The definition of (c) is given in table 1,will-> Substituting formula (17) can result in:
when s is ij When=0, the following equation is obtained by substituting equation (18) into equation (16) and then solving for x.
When s is ij When not equal to 0, the minimum point is:
thus, if two discontinuities are not within the definition of ψ (-), i.e. detZ (0). Ltoreq.0 and detZ (x) ij,t-1 ) Less than or equal to 0, the following expression of the minimum point can be obtained by the formulas (19) and (20):
further, cases of extreme points of the objective function in the following two cases are discussed, respectively: (1) When x ij,t-1 ＝0and(2)when x ij,t-1 ≠0.
Case 1 if x ij,t-1 ＝0
In this case, the objective function may be written as:
thus, the minimum point is the minimum point of c (&) or is atWhere it is located. To make the formulation more compact, let a=ψ (0), let->The optimal solution can thus be summarized in this case as follows:
when detZ (0). Ltoreq.0, the optimal solution can be obtained from (21)
When detZ (0) >0:
case 2 if x ij,t-1 ≠0
Unlike the previous case, in this case the objective function can be written as:
the minimum point of the objective function in this case is the minimum point at c (,) orOr alternativelyOne of which is a solid. The optimal xsolution that minimizes the objective function is summarized as follows:
when detZ (0) is less than or equal to 0 and detZ (x) ij,t-1 ) And (3) obtaining the solution of x from the step (21).
When detZ (0)>0 and detZ (x) ij,t-1 ) And (3) obtaining the solution of x from the step (23).
When detZ (0) is less than or equal to 0 and detZ (x) ij,t-1 )>0:
When detZ (0)>0 and detZ (x) ij，t-1 )>0:
Here a and B are respectively:
A＝Ψ(0)＝-logdet(Z(0))+2λ 2 ， (27)
B＝Ψ(x ij，t-1 )＝-logdet(Z(x ij，t-1 ))+2s ij x ij，t-1 +2λ 1 . (28)
further, update Y k
Y also needs to be updated in real time as in (11) and (21) when each element of the matrix X is calculated and updated. Order theAccording to the Sherman-Morrison-Woodbury formula, there are:
assume that
Solving for updatable Y k
In one embodiment, lambda is determined by a predetermined Extended Bayesian Information Criterion (EBIC) 1 Lambda of 2 。
In a specific implementation, an Extended Bayesian Information Criterion (EBIC) method is used for parameter selection. For any lambda 1 Lambda of 2 EBIC is defined as:
where α ε [0,1], whose typical value is 0.5, n is the number of samples, P is the dimension of the covariance matrix, and K (t) represents the number of non-zero points estimated at time point t, which can be calculated from the following equation:
lambda can be determined by solving 1 Lambda of 2 。
For ease of understanding, the present invention summarizes the above-described solving process in table 2 below.
TABLE 2 summary of solving process
By applying the technical scheme of the embodiment of the invention, the covariance matrix of the sample data of the BOLD signal of the preset brain region is obtained; inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model; and solving the brain function connection prediction model to obtain a prediction result of brain function connection. Because two L0 normal form regular terms for respectively controlling and estimating the space sparsity and the time homogeneity of the connection matrix are adopted, the error of the brain function connection prediction model is small, the sparsity is strong and the accuracy is high.
In order to verify the superiority of the method provided by the embodiment of the invention, a series of simulation experiments are carried out, and by contrast, two other algorithms based on L1 regularization are also used in the simulation experiments.
The simulation experiment is carried out 3 times in total, the simulation data in different simulation experiments correspond to different network connection modes (3 connection modes are respectively small world, non-scale and random network connection modes), each connection mode data comprises 3 time periods in time, each time period comprises 10 nodes and 100 time points, and corresponds to a randomly generated network structure, and the total time is 300 time points. And then, estimating the connection network among the nodes by using an L0-SSR, SINGLE and TVGL algorithm respectively, wherein the simulation experiment under each connection mode is carried out 100 times.
Performance evaluation criteria:
whether functional connection exists between different brain areas or not is an embodiment of good performance of the algorithm, so that whether zero elements and non-zero elements in a real connection matrix are accurately judged by observing an accuracy matrix estimated by the algorithm is hoped to evaluate the performance of the algorithm.
The accuracy (precision) of the first measurement algorithm can be expressed as:
wherein P is t Representing the accuracy, X t And T t Representing the accuracy matrix and the true matrix, respectively, estimated at the t-th point in time. The above equation represents the ratio of the number of non-zero elements that the estimated precision matrix correctly estimates to the total number of non-zero elements estimated.
The recall rate (recall) of an algorithm can be expressed as:
it represents the ratio of the number of non-zero elements that the estimated precision matrix correctly estimates to the total number of true non-zero elements.
It is desirable to integrate precision and recovery to represent the estimated performance of the algorithm with their harmonic mean F1. F1 can be expressed as:
the value range of F1 is [0,1], and the value can be used for indicating the performance of the algorithm.
Experimental results as shown in fig. 2-4, fig. 2-4 show the multi-channel simulation results for the small world, scaleless, and random networks, respectively. The average F1 score of 100 simulation experiments is used as a measurement index to evaluate the performance of different methods. It can be seen from FIGS. 2 and 3 that the proposed L0-SSR model works best for all methods. In fig. 4, the algorithm proposed by the present invention has substantially the same estimation performance as that shown by the SINGLE algorithm, but is still superior to the TVGL method. In these three results, it is clearly noted that the estimated performance is significantly reduced around time points 100 and 200. The method is characterized in that the network structure changes at the two moments, and when the network connection at the current moment is estimated by utilizing a sliding window method, data of the network structure at the adjacent moment, which is different from the current moment, are utilized, so that certain estimation deviation is caused. Although all three algorithms can track network changes shortly after network connection changes occur, the proposed L0-SSR method is less affected and can maintain higher performance for longer than the other two methods. Meanwhile, when the network structure changes, the estimation can be quickly and accurately recovered. These results demonstrate the effectiveness of the L0-SSR algorithm in identifying time-varying brain connections.
Fig. 5 is a schematic block diagram of a brain function connection prediction apparatus 90 according to an embodiment of the present invention. As shown in fig. 5, the present invention also provides a brain function connection prediction apparatus 90 corresponding to the above brain function connection prediction method. The brain function connection predicting device 90 includes a unit for performing the above brain function connection predicting method, and the brain function connection predicting device 90 may be configured in a desktop computer, a tablet computer, a portable computer, or the like. Specifically, referring to fig. 5, the brain function connection prediction apparatus 90 includes an acquisition unit 91, an input unit 92, and a prediction unit 93.
An obtaining unit 91, configured to obtain a covariance matrix of sample data of a BOLD signal of a preset brain region;
the input unit 92 is configured to input the covariance matrix into a preset brain function connection prediction model, where the brain function connection prediction model is obtained by adding two L0-norm regularization terms for respectively controlling and estimating the spatial sparsity and temporal homogeneity of the connection matrix on the basis of a gaussian graph model;
and the prediction unit 93 is used for solving the brain function connection prediction model to obtain a prediction result of brain function connection.
In one embodiment, the mathematical expression of the brain function connection pre-model is as follows:
wherein S is t Covariance matrix of sample data at time t, X t Is the inverse covariance matrix of the moment t, X t-1 As an inverse covariance matrix at time t-1, lambda 1 Coefficients for controlling spatial sparsity; lambda (lambda) 2 To control the coefficient of temporal homogeneity.
In one embodiment, as shown in fig. 6, the prediction unit 93 includes a solving unit 931.
And a solving unit 931 for solving the brain function connection prediction model through an iterative method based on a coordinate descent method.
In an embodiment, as shown in fig. 7, the brain function connection prediction device 90 further includes a preprocessing unit 94.
A preprocessing unit 94, configured to preprocess fMRI data to obtain BOLD signal sample data of the preset brain region. .
In one embodiment, lambda is determined by a pre-set extended Bayesian information criterion 1 Lambda of 2 。
It should be noted that, as will be clearly understood by those skilled in the art, the specific implementation process of the brain function connection prediction device 90 and each unit may refer to the corresponding description in the foregoing method embodiment, and for convenience and brevity of description, the description is omitted here.
The brain function connection predicting means described above may be implemented in the form of a computer program which can be run on a computer device as shown in fig. 8.
Referring to fig. 8, fig. 8 is a schematic block diagram of a computer device according to an embodiment of the present application. The computer device 500 may be a terminal or a server, where the terminal may be an electronic device with a communication function, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, a personal digital assistant, and a wearable device. The server may be an independent server or a server cluster formed by a plurality of servers.
With reference to FIG. 8, the computer device 500 includes a processor 502, memory, and a network interface 505 connected by a system bus 501, where the memory may include a non-volatile storage medium 503 and an internal memory 504.
The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032, when executed, may cause the processor 502 to perform a brain function connection prediction method.
The processor 502 is used to provide computing and control capabilities to support the operation of the overall computer device 500.
The internal memory 504 provides an environment for the execution of a computer program 5032 in the non-volatile storage medium 503, which computer program 5032, when executed by the processor 502, causes the processor 502 to perform a brain function connection prediction method.
The network interface 505 is used for network communication with other devices. Those skilled in the art will appreciate that the architecture shown in fig. 8 is merely a block diagram of a portion of the architecture in connection with the present application and is not intended to limit the computer device 500 to which the present application is applied, and that a particular computer device 500 may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
Wherein the processor 502 is configured to execute a computer program 5032 stored in a memory to implement the steps of:
acquiring a covariance matrix of sample data of a BOLD signal of a preset brain region;
inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model;
and solving the brain function connection prediction model to obtain a prediction result of brain function connection.
In one embodiment, when implementing the step of solving the brain function connection prediction model to obtain a prediction result of a brain function connection, the processor 502 specifically implements the following steps:
and solving the brain function connection prediction model by an iteration method based on a coordinate descent method.
In an embodiment, before implementing the step of acquiring the covariance matrix of the sample data of the BOLD signal of the preset brain region, the processor 502 further implements the following steps:
preprocessing fMRI data to obtain sample data of BOLD signals of the preset brain region.
In one embodiment, the mathematical expression of the brain function connection pre-model is as follows: />
wherein S is t Covariance matrix of sample data at time t, X t Is the inverse covariance matrix of the moment t, X t-1 As an inverse covariance matrix at time t-1, lambda 1 Coefficients for controlling spatial sparsity; lambda (lambda) 2 To control the coefficient of temporal homogeneity.
In one embodiment, lambda is determined by a pre-set extended Bayesian information criterion 1 Lambda of 2 。
It should be appreciated that in embodiments of the present application, the processor 502 may be a central processing unit (Central Processing Unit, CPU), the processor 502 may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf Programmable gate arrays (FPGAs) or other Programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. Wherein the general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Those skilled in the art will appreciate that all or part of the flow in a method embodying the above described embodiments may be accomplished by computer programs instructing the relevant hardware. The computer program may be stored in a storage medium that is a computer readable storage medium. The computer program is executed by at least one processor in the computer system to implement the flow steps of the embodiments of the method described above.
Accordingly, the present invention also provides a storage medium. The storage medium may be a computer readable storage medium. The storage medium stores a computer program. The computer program, when executed by a processor, causes the processor to perform the steps of:
acquiring a covariance matrix of sample data of a BOLD signal of a preset brain region;
inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model;
and solving the brain function connection prediction model to obtain a prediction result of brain function connection.
In one embodiment, when the processor executes the computer program to implement the step of solving the brain function connection prediction model to obtain a prediction result of the brain function connection, the processor specifically implements the following steps:
and solving the brain function connection prediction model by an iteration method based on a coordinate descent method.
In an embodiment, before executing the computer program to implement the step of obtaining the covariance matrix of the sample data of the BOLD signal of the preset brain region, the processor further implements the following steps:
preprocessing fMRI data to obtain sample data of BOLD signals of the preset brain region.
In one embodiment, the mathematical expression of the brain function connection pre-model is as follows:
wherein S is t Covariance matrix of sample data at time t, X t Is the inverse covariance matrix of the moment t, X t-1 As an inverse covariance matrix at time t-1, lambda 1 Coefficients for controlling spatial sparsity; lambda (lambda) 2 To control the coefficient of temporal homogeneity.
In one embodiment, lambda is determined by a pre-set extended Bayesian information criterion 1 Lambda of 2 。
The storage medium is a physical, non-transitory storage medium, and may be, for example, a U-disk, a removable hard disk, a Read-Only Memory (ROM), a magnetic disk, or an optical disk.
Those of ordinary skill in the art will appreciate that the elements and algorithm steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, in computer software, or in a combination of the two, and that the elements and steps of the examples have been generally described in terms of function in the foregoing description to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the device embodiments described above are merely illustrative. For example, the division of each unit is only one logic function division, and there may be another division manner in actual implementation. For example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed.
The steps in the method of the embodiment of the invention can be sequentially adjusted, combined and deleted according to actual needs. The units in the device of the embodiment of the invention can be combined, divided and deleted according to actual needs. In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.
The integrated unit may be stored in a storage medium if implemented in the form of a software functional unit and sold or used as a stand-alone product. Based on such understanding, the technical solution of the present invention is essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a terminal, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention.
In the foregoing embodiments, the descriptions of the embodiments are focused on, and for those portions of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
While the invention has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.
Claims (5)
1. A brain function connection prediction method, comprising:
acquiring a covariance matrix of BOLD signal sample data of a preset brain region;
inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model;
solving the brain function connection prediction model to obtain a prediction result of brain function connection;
the mathematical expression of the brain function connection prediction model is as follows:
wherein S is t Covariance matrix of sample data at time t, X t At tInverse covariance matrix of scale, X t-1 As an inverse covariance matrix at time t-1, lambda 1 Coefficients for controlling spatial sparsity; lambda (lambda) 2 Coefficients for controlling temporal homogeneity;
lambda determination by means of preset extended Bayesian information criteria 1 Lambda of 2 ；
The solving the brain function connection prediction model to obtain a prediction result of brain function connection comprises:
and solving the brain function connection prediction model by an iteration method based on a coordinate descent method.
2. The brain function connection prediction method according to claim 1, further comprising, before said acquiring the covariance matrix of the preset brain region BOLD signal sample data:
preprocessing fMRI data to obtain the preset brain region BOLD signal sample data.
3. A brain function connection prediction device for implementing the method of any one of claims 1-2, comprising:
the acquisition unit is used for acquiring a covariance matrix of the BOLD signal sample data of the preset brain region;
the input unit is used for inputting the covariance matrix into a preset brain function connection prediction model, wherein the brain function connection prediction model is obtained by adding two L0 normal form regular terms for respectively controlling and estimating the spatial sparsity and the temporal homogeneity of the connection matrix on the basis of a Gaussian diagram model, and the mathematical expression of the brain function connection prediction model is as follows:
wherein S is t Covariance matrix of sample data at time t, X t Is the inverse covariance matrix of the moment t, X t-1 As an inverse covariance matrix at time t-1, lambda 1 Coefficients for controlling spatial sparsity; lambda (lambda) 2 Coefficients for controlling temporal homogeneity;
lambda determination by means of preset extended Bayesian information criteria 1 Lambda of 2 ；
The prediction unit is used for solving the brain function connection prediction model to obtain a prediction result of brain function connection;
the prediction unit includes:
and the solving unit is used for solving the brain function connection prediction model through an iteration method based on a coordinate descent method.
4. A computer device, characterized in that it comprises a memory on which a computer program is stored and a processor which, when executing the computer program, implements the method according to any of claims 1-2.
5. A computer readable storage medium, characterized in that the storage medium stores a computer program which, when executed by a processor, implements the method according to any of claims 1-2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010573592.1A CN111785379B (en) | 2020-06-22 | 2020-06-22 | Brain function connection prediction method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010573592.1A CN111785379B (en) | 2020-06-22 | 2020-06-22 | Brain function connection prediction method, device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111785379A CN111785379A (en) | 2020-10-16 |
| CN111785379B true CN111785379B (en) | 2023-12-26 |
Family
ID=72756150
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010573592.1A Active CN111785379B (en) | 2020-06-22 | 2020-06-22 | Brain function connection prediction method, device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111785379B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116313131B (en) * | 2023-05-24 | 2023-09-15 | 山东大学 | Brain network difference recognition system, equipment and storage medium based on imitation variables |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103838943A (en) * | 2012-11-21 | 2014-06-04 | 大连灵动科技发展有限公司 | Brain interval covariance analysis method for functional magnetic resonance data processing |
| CN104605853A (en) * | 2015-01-21 | 2015-05-13 | 北京师范大学 | Real-time nerve decoding system based on brain function features |
| CN109065128A (en) * | 2018-09-28 | 2018-12-21 | 郑州大学 | A kind of sparse brain network establishing method of weighted graph regularization |
| CN109522894A (en) * | 2018-11-12 | 2019-03-26 | 电子科技大学 | A method of detection fMRI brain network dynamic covariant |
| CN110689029A (en) * | 2019-11-01 | 2020-01-14 | 中国计量大学 | Method for determining sparsity of fMRI brain function connection network |
| CN110689536A (en) * | 2019-09-30 | 2020-01-14 | 深圳大学 | Method and device for tracking grey matter and white matter of brain based on multi-modal magnetic resonance image |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10945622B2 (en) * | 2016-08-23 | 2021-03-16 | Regents Of The University Of Minnesota | System and method for assessing electrical activity using an iterative sparse technique |
-
2020
- 2020-06-22 CN CN202010573592.1A patent/CN111785379B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103838943A (en) * | 2012-11-21 | 2014-06-04 | 大连灵动科技发展有限公司 | Brain interval covariance analysis method for functional magnetic resonance data processing |
| CN104605853A (en) * | 2015-01-21 | 2015-05-13 | 北京师范大学 | Real-time nerve decoding system based on brain function features |
| CN109065128A (en) * | 2018-09-28 | 2018-12-21 | 郑州大学 | A kind of sparse brain network establishing method of weighted graph regularization |
| CN109522894A (en) * | 2018-11-12 | 2019-03-26 | 电子科技大学 | A method of detection fMRI brain network dynamic covariant |
| CN110689536A (en) * | 2019-09-30 | 2020-01-14 | 深圳大学 | Method and device for tracking grey matter and white matter of brain based on multi-modal magnetic resonance image |
| CN110689029A (en) * | 2019-11-01 | 2020-01-14 | 中国计量大学 | Method for determining sparsity of fMRI brain function connection network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111785379A (en) | 2020-10-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Allen et al. | A generalized least-square matrix decomposition | |
| KR101889722B1 (en) | Method and Apparatus for Diagnosing Malignant Tumor | |
| US20230301542A1 (en) | Brain atlas individualization method and system based on magnetic resonance and twin graph neural network | |
| WO2020114329A1 (en) | Fast magnetic resonance parametric imaging and device | |
| Polzehl et al. | Structural adaptive segmentation for statistical parametric mapping | |
| CN111863244B (en) | Functional connection mental disease classification method and system based on sparse pooling graph convolution | |
| KR101889724B1 (en) | Method and Apparatus for Diagnosing Malignant Tumor | |
| CN113256592B (en) | Training method, system and device of image feature extraction model | |
| KR101889723B1 (en) | Method and Apparatus for Diagnosing Malignant Tumor | |
| CN114708263B (en) | Individual brain functional region positioning method, device, equipment and storage medium | |
| Ma et al. | Partition-based vector filtering technique for suppression of noise in digital color images | |
| US20070021952A1 (en) | General graphical Gaussian modeling method and apparatus therefore | |
| CN111785379B (en) | Brain function connection prediction method, device, computer equipment and storage medium | |
| Ciarleglio et al. | Wavelet-based scalar-on-function finite mixture regression models | |
| Kook et al. | NPBayes-fMRI: Non-parametric Bayesian general linear models for single-and multi-subject fMRI data | |
| Mouches et al. | Unifying brain age prediction and age-conditioned template generation with a deterministic autoencoder | |
| Lu et al. | Going deeper into cardiac motion analysis to model fine spatio-temporal features | |
| Yue et al. | Two-step sparse boosting for high-dimensional longitudinal data with varying coefficients | |
| JP2011141866A (en) | Data classifying device, data classifying system, and program | |
| Ng | Recent developments in expectation‐maximization methods for analyzing complex data | |
| Sjöstrand et al. | A path algorithm for the support vector domain description and its application to medical imaging | |
| Chao et al. | Multivariate factorizable expectile regression with application to fMRI data | |
| Tao et al. | Nonlinear association criterion, nonlinear Granger causality and related issues with applications to neuroimage studies | |
| Abbas et al. | Tangent functional connectomes uncover more unique phenotypic traits | |
| Seghouane | A Kullback-Leibler methodology for HRF estimation in fMRI data |
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 |