CN109242816B - Brain glioma pathological grade imaging auxiliary judgment method based on resting state brain function contralateral positioning - Google Patents
Brain glioma pathological grade imaging auxiliary judgment method based on resting state brain function contralateral positioning Download PDFInfo
- Publication number
- CN109242816B CN109242816B CN201710551743.1A CN201710551743A CN109242816B CN 109242816 B CN109242816 B CN 109242816B CN 201710551743 A CN201710551743 A CN 201710551743A CN 109242816 B CN109242816 B CN 109242816B
- Authority
- CN
- China
- Prior art keywords
- brain
- positioning
- functional
- resting state
- region
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30016—Brain
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Abstract
The invention belongs to the field of medical image processing and application, relates to a non-invasive resting state magnetic resonance imaging auxiliary judgment method, and particularly relates to a non-invasive resting state magnetic resonance imaging auxiliary judgment method for brain glioma WHO pathological grade.
Description
Technical Field
The invention belongs to the field of medical image processing and application, relates to a non-invasive resting state magnetic resonance imaging auxiliary judgment method, in particular to a non-invasive resting state magnetic resonance imaging auxiliary judgment method for brain glioma WHO pathological grade, and particularly relates to a method for realizing high-accuracy high-low-grade tumor non-invasive imaging judgment by utilizing the brain functional region characteristic of resting state brain function contralateral positioning labeling.
Background
Brain gliomas are reported to be common central nervous system tumors with high morbidity, high mortality, and high recurrence rates. Clinical practice shows that the reasonable judgment of the pathological grade of brain glioma (WHO) is crucial to clinical diagnosis, determination of intervention treatment schemes and the like; according to the latest standards of the world health organization, brain gliomas are classified into low grade (grade I and II) and high grade (grade III and IV). The current gold standard for tumor grade remains biopsy and histopathological evaluation. However, biopsy also faces many challenges, especially local sampling analysis of tumors may cause high-grade tumors to be misdiagnosed as low-grade tumors, and cause intracranial hemorrhage, etc., and meanwhile, unlike tumors in other parts, tissue biopsy of brain tumors is very difficult; various non-invasive approaches other than biopsy have also been used for the assessment of malignancy of tumors. In recent years, morphological and metabolic indexes displayed by magnetic resonance and other nerve imaging technologies can play a certain indicating role, the accuracy mainly depends on experience accumulation and professional knowledge of a film reading doctor, quantifiable standard indexes are lacked, the prediction effect evaluation is not facilitated, and the standardization popularization and application are also not facilitated. Previously, imaging techniques such as Magnetic Resonance Spectroscopy (MRS) have been used to assess the malignancy of tumors based primarily on the metabolic features of tumor tissue as distinguished from normal tissue. The current method for maximizing sensitivity and specificity is to calculate the ratio of Cho (Choline) and NAA (N-acetyl-aspartate) in a tumor region, and the sensitivity of 0.80 and the specificity of 0.76 can be achieved. By using the resting state magnetic resonance technology, the research on high-level judgment of tumors by using indexes such as Signal Intensity Compensation (SIC) of resting state signals of tumor regions and the like is carried out, and the accuracy (the sensitivity is 0.82, and the specificity is 0.96) of 89% is realized on 18 high-level and 17 low-level brain glioma samples. However, no independent data verification is found in the brain glioma grading results, and the tumor region needs to be manually drawn on the magnetic resonance image, which is easily interfered by human factors. In fact, the influence of tumors on the brain function architecture, i.e., the function connection pattern of each brain region, has been reported in many documents, including causing function compensation, function network reconstruction, etc., and the research group of the present invention proposed a "resting state function magnetic resonance-based brain function region-to-side positioning method" (patent application No. 201710377307.7) in the early stage, in which an auxiliary motion region is used as a positioning target, and the resting state function connection pattern from the affected side to the healthy side is used to position the auxiliary motion region on the affected side. At present, no method for evaluating the brain glioma height classification by adopting a resting state functional connection network exists.
In order to overcome the difficulties in the prior art and the basis based on the prior art, the inventor of the application proposes an auxiliary judgment method for brain glioma pathological grade imaging based on resting state brain function contralateral positioning. The method is a non-invasive, objective and quantitative evaluation method only requiring the minimum degree of adaptability of a tested person, has the high-low-level glioma judgment precision of more than 90 percent, and is suitable for being popularized and applied to hospitals with magnetic resonance imaging equipment.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, provides a non-invasive resting state magnetic resonance imaging auxiliary judgment method, and particularly relates to a non-invasive resting state magnetic resonance imaging auxiliary judgment method for a brain glioma WHO pathological grade.
The method disclosed by the invention is based on the brain glioma pathological grade assessment of the contralateral positioning of the brain functional region of the resting state functional magnetic resonance data, can realize high-precision noninvasive brain glioma pathological grade based on imaging, and is particularly suitable for brain language and motor functional region glioma patients.
In order to achieve the above object, the method for determining pathology grade of brain glioma based on resting state brain function contralateral positioning in an auxiliary imaging manner of brain glioma of the present invention comprises:
firstly, the auxiliary motion area at the affected side of a patient with brain glioma is positioned functionally by using a method of 'a brain functional area contralateral positioning method based on resting state functional magnetic resonance' (patent application number 201710377307.7); secondly, constructing a Logistic model by using the size of the positioned auxiliary motion area (namely the number of voxels contained in the positioning area), wherein the Lotistic model takes the size of the positioned auxiliary motion area as input and takes high-level tumors as positive cases and low-level tumors as negative cases; by applying the model to patient data, the case level of the tumor it is suffering from can be determined.
In the invention, biopsy cases are used as a gold standard, and the logistic model achieves more than 90% of judgment accuracy on a training set (20 low levels versus 31 high levels, and the model accuracy is evaluated by one-out-of-one cross validation) and a testing set (7 low levels versus 6 high levels, and model parameters obtained on the training set are directly applied).
Specifically, the brain glioma pathological grade imaging auxiliary judgment method based on resting state brain function contralateral positioning is characterized by comprising the following steps of:
1) Preprocessing the acquired resting state functional image and high-precision structural image of the glioma patient in the brain functional region, comprising the following steps of: scanning layer time correction, head motion correction, mapping to a standardized space, trend item removal, band-pass filtering and scattering; while registering, removing the influence of the tumor MASK on standardization, and mapping the healthy brain which is not subjected to tumor images to a standard space;
2) performing functional positioning on an auxiliary motion area on the affected side of a patient with brain glioma by using a brain functional area contralateral positioning method based on resting state functional magnetic resonance;
3) constructing a Logistic model by using the size of the auxiliary motion area obtained by positioning, namely the number of voxels contained in the positioning area, wherein the output of the Logistic model is higher than the threshold value and is judged to be in a high level; otherwise, it is considered as a low level.
In step 2) of the invention, the auxiliary motion area which has a significant relation with the tumor grade is positioned.
In the invention, the resting state functional magnetic resonance-based brain functional region contralateral positioning method adopts the following technical scheme:
dividing functional sub-regions of a brain region by using resting state functional connection of the whole brain, training a Support Vector Machine (SVM) classifier for each functional sub-region, establishing specific contralateral semi-brain resting state functional connection for each functional sub-region, and training the classifier on big data of a healthy person to finally realize the positioning of each functional sub-region; the method comprises the following steps:
1. The method comprises the following steps of acquiring resting state functional images and high-precision structural images of glioma patients in a brain functional region within 5 minutes by adopting a resting state functional magnetic resonance technology, and adopting a plurality of standardized pretreatments: scanning layer time correction, head motion correction, mapping to a standardized space, trend item removal, band-pass filtering and scattering; then, the preprocessed brain data is divided into 45 brain areas (45 brain areas are symmetrical left and right) according to an AAL (automated atomic labeling) template provided by the Montreal neuroscience research center;
2. dividing each brain region into functional sub-regions by using resting state functional connection of the whole brain voxel level, and dividing the whole brain functional region into 218 functional sub-regions; in the embodiment, for each brain area, resting state signals of all voxels of the brain area are calculated, then correlation coefficients of the resting state signals and average signals of the remaining 88 brain areas of the whole brain are respectively calculated, and after a correlation coefficient matrix is obtained, a proper lambda opt is selected to carry out 0-1 conversion on the matrix:
then, classifying by utilizing an LM algorithm (a minimization network link algorithm based on local properties) so as to obtain sub-region division of each brain region; wherein the selected λ opt is obtained by 50 times of packet cross-checking, specifically by calculating standardized mutual information
The lambda which enables the NMI (X, Y) to be maximum obtains the most stable partitioning result, and the voxels in each sub-area have basically consistent resting state functional link characteristics;
3. training a Support Vector Machine (SVM) classifier for each functional subarea; in an embodiment, for each brain sub-region, focusing on all voxels in this region and its surrounding 6mm, the output of the trained classifier marks the voxels inside the target functional region as 1 and the voxels around the target functional region as 0, and then the input features of the classifier are the resting brain functional connectivity pattern of the contralateral half-brain, which is specifically defined as follows: judging whether a certain voxel belongs to a designated functional area, and calculating specific resting state functional connection of the contralateral semi-brain of the voxel, wherein the specific functional connection is given by comparing the resting state brain functional connection of internal and external voxels of the given functional area to all voxels of the contralateral semi-brain: calculating average signals of two groups of voxels inside and outside the functional region respectively, calculating functional connection strength to the contralateral half brain, performing comparison t-test between groups, identifying a brain region cluster with significant difference between groups after multiple comparison correction, and finding out the contralateral brain functional connection with significant difference inside and outside the functional region as the input characteristic of a classifier; establishing specific contralateral semi-brain resting state function connection for each functional sub-region, and training a classifier on big data of a healthy person to finally realize the positioning of each functional sub-region;
4. And combining the positioning results of all the functional sub-regions of the AAL template to obtain functional positioning maps of 45 brain regions.
More specifically, the resting state functional magnetic resonance-based brain functional region contralateral positioning method comprises the following steps:
4) establishing a map of the 218 sub-area of the whole brain by using resting state data through a big data sample of a healthy person;
5) respectively training an SVM classifier for each sub-region in the 218 sub-region map through a big data sample of a healthy person;
6) preprocessing the acquired resting state functional image and high-precision structural image of the glioma patient in the brain functional region, comprising the following steps of: scanning layer time correction, head motion correction, mapping to a standardized space, trend item removal, band-pass filtering and scattering; while registering, removing the influence of the tumor MASK on standardization, and mapping the healthy brain which is not subjected to tumor images to a standard space;
7) for each voxel of the target region on the affected side, calculating the correlation coefficient between the voxel and a plurality of feature region signals on the opposite side respectively, taking the correlation coefficient as the input of a Support Vector Machine (SVM) classifier, and outputting whether each voxel of the target region belongs to the brain sub-region or not;
8) Finally, all the positioning results are pieced together and mapped back to the individual space to form the positioning result of the whole affected side brain functional area.
In the step 1), functional connections of voxels in the functional region and the rest 88 brain regions of the whole brain are calculated simultaneously to obtain a connection matrix M, after a similar matrix N of the connection matrix M is calculated, binarization and classification are carried out on the connection matrix M, and a stable partitioning result is obtained by maximizing mutual information;
wherein, when the similar matrix N is binarized, 50 times of grouping cross check is adopted, and by calculating the standardized mutual information,
λ is obtained that maximizes NMI (X, Y), by which the similarity matrix N is binarized:
obtaining a map of a 218 sub-area of the whole brain in the step 1);
in the step 2), defining a target area: for each brain area, taking the brain area position positioned by the AAL template as the center, taking the area with 2 voxels (namely 6mm) expanded outwards as a target area, and dividing the functional area from the target area;
in step 2), calculating the functional connection from each brain subregion and surrounding voxels to each voxel of the contralateral half brain on a training set, performing comparison t-test between groups, and identifying a brain region cluster with significant difference between groups after multiple comparison correction;
In the step 2), a Support Vector Machine (SVM) classifier is trained for each sub-area by taking the connection of the average signal of the voxel in the cluster and the voxel calculation function of the target area as a characteristic.
In the step 3), a tumor part is manually animated on each layer of image through the T2 image, and the normalized weight of the part is set to 0 in the registration process, so that the influence of the tumor on the normalization is removed;
in the step 4), for each voxel in the target region on the affected side, calculating a correlation coefficient between the voxel and a plurality of feature region signals on the opposite side, taking the correlation coefficient as the input of a Support Vector Machine (SVM) classifier, and outputting whether each voxel in the target region belongs to a certain specific brain functional region;
in the step 5), the function positioning result in the standard space is mapped back to the individual space, and a brain function positioning map is drawn on the affected side of the semi-brain by positioning 45 functional brain regions one by one.
According to the brain glioma pathological grading method based on the imaging, the opposite side positioning of the auxiliary motion functional area is carried out through the resting state functional image and the high-precision structural image, the brain glioma pathological grade is judged through the Logistic model, the result is displayed, the high-precision noninvasive brain glioma pathological grading based on the imaging can be realized, and the brain glioma pathological grading method based on the imaging is particularly suitable for brain language and motion functional area glioma patients, and can be used for guiding the selection of a treatment scheme and the like.
The invention has the beneficial effects that:
1. the method evaluates the pathological grade of the brain glioma through non-invasive examination.
2. The method has the precision of assessing the pathology level of the brain glioma by more than 90 percent, and achieves the specificity of 95 percent and the sensitivity of 87 percent on a training set.
3. The method achieved 83% specificity and 100% sensitivity on the independent test set.
Detailed Description
Example 1
The method comprises the following steps of obtaining resting state functional images and high-precision structural images of glioma patients in a brain functional region within 5 minutes by adopting a resting state functional magnetic resonance technology, and adopting multiple pretreatment steps: scanning slice time correction, head motion correction, mapping to a standardized space, de-trending terms, band-pass filtering and scattering. While registering, removing the influence on standardization by tumor MASK, and mapping the brain (healthy side) of the healthy side of the patient, which is not subjected to tumor images, to a standard space;
then, the auxiliary motion area on the affected side is subjected to kinetic energy positioning by adopting the 'resting state-based auxiliary motion area function positioning method'; then, the size of the auxiliary movement function area obtained by positioning is taken as input and is brought into the Logistic model for tumor case classification evaluation established by the invention, and the probability that the case grade of the tumor is a high grade is estimated.
Because systematic deviation may exist in different machine or scanning parameters, when the model is used on a certain magnetic resonance scanner, a certain amount of data with a case grade gold standard of tissue biopsy is needed to be used as training, and a threshold value is set; the output of the Logistic model is higher than the threshold value and is judged to be high level; otherwise, it is considered as a low level.
In the embodiment of the invention, contralateral positioning of the auxiliary motion functional area is carried out through the resting state functional image and the high-precision structural image, the case grade of the tumor is judged through a Logistic model, the high-grade glioma judgment precision reaches more than 90 percent, and the method can be used for guiding the selection of an intervention treatment scheme and the like.
The above description is only exemplary of the present invention and should not be taken as limiting the invention, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (2)
1. The brain glioma pathological grade imaging auxiliary judgment method based on resting state brain function contralateral positioning is characterized by comprising the following steps of:
1) preprocessing the acquired resting state functional image and high-precision structural image of the glioma patient in the brain functional region, comprising the following steps of: scanning layer time correction, head motion correction, mapping to a standardized space, trend item removal, band-pass filtering and scattering; while registering, removing the influence of the tumor MASK on standardization, and mapping the healthy brain without the tumor influence to a standard space;
2) Performing functional positioning on an auxiliary motion area on the affected side of a patient with brain glioma by using a brain functional area contralateral positioning method based on resting state functional magnetic resonance; wherein, the brain functional region contralateral positioning based on the resting state functional magnetic resonance comprises the following steps:
(1) establishing a map of the 218 sub-area of the whole brain by using resting state data through a big data sample of a healthy person;
(2) respectively training an SVM classifier for each sub-region in the 218 sub-region map through a big data sample of a healthy person;
(3) for each voxel of the target region on the affected side, calculating the correlation coefficient between the voxel and a plurality of characteristic region signals on the opposite side respectively, taking the correlation coefficient as the input of a support vector machine classifier, and outputting whether each voxel of the target region belongs to the brain subregion;
(4) finally, all positioning results are pieced together and mapped back to an individual space to form a positioning result of the whole affected side brain functional area;
3) establishing a Logistic model by utilizing the size of the positioned auxiliary motion area, namely the number of voxels contained in the positioning area, wherein the output of the Logistic model is higher than a threshold value and is judged to be in a high level; otherwise, it is considered as a low level.
2. The method for judging the pathological grade of brain glioma through the imaging of the opposite side positioning based on the resting state brain function as claimed in claim 1, wherein in the step 2), the functional positioning of the auxiliary motor area on the affected side of the patient with brain glioma is to position the auxiliary motor area which has a significant relation with the tumor grade.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710551743.1A CN109242816B (en) | 2017-07-07 | 2017-07-07 | Brain glioma pathological grade imaging auxiliary judgment method based on resting state brain function contralateral positioning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710551743.1A CN109242816B (en) | 2017-07-07 | 2017-07-07 | Brain glioma pathological grade imaging auxiliary judgment method based on resting state brain function contralateral positioning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109242816A CN109242816A (en) | 2019-01-18 |
| CN109242816B true CN109242816B (en) | 2022-01-18 |
Family
ID=65082746
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710551743.1A Active CN109242816B (en) | 2017-07-07 | 2017-07-07 | Brain glioma pathological grade imaging auxiliary judgment method based on resting state brain function contralateral positioning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109242816B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114187227B (en) * | 2021-09-13 | 2023-01-24 | 北京银河方圆科技有限公司 | Method and device for determining functional area of brain tumor affected area |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106204562A (en) * | 2016-07-04 | 2016-12-07 | 西安交通大学 | A kind of method of the arched roof white matter segmentation merged based on fMRI Yu DTI |
| CN106485039A (en) * | 2015-08-24 | 2017-03-08 | 复旦大学附属华山医院 | The construction method of Butut distinguished in a kind of Chinese brain language |
| CN106683081A (en) * | 2016-12-17 | 2017-05-17 | 复旦大学 | Brain glioma molecular marker nondestructive prediction method and prediction system based on radiomics |
| CN106780515A (en) * | 2017-01-04 | 2017-05-31 | 南京审计大学 | Glioma method for extracting region in a kind of cerebral magnetic resonance image |
| US9984283B2 (en) * | 2015-02-14 | 2018-05-29 | The Trustees Of The University Of Pennsylvania | Methods, systems, and computer readable media for automated detection of abnormalities in medical images |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10909675B2 (en) * | 2015-10-09 | 2021-02-02 | Mayo Foundation For Medical Education And Research | System and method for tissue characterization based on texture information using multi-parametric MRI |
-
2017
- 2017-07-07 CN CN201710551743.1A patent/CN109242816B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9984283B2 (en) * | 2015-02-14 | 2018-05-29 | The Trustees Of The University Of Pennsylvania | Methods, systems, and computer readable media for automated detection of abnormalities in medical images |
| CN106485039A (en) * | 2015-08-24 | 2017-03-08 | 复旦大学附属华山医院 | The construction method of Butut distinguished in a kind of Chinese brain language |
| CN106204562A (en) * | 2016-07-04 | 2016-12-07 | 西安交通大学 | A kind of method of the arched roof white matter segmentation merged based on fMRI Yu DTI |
| CN106683081A (en) * | 2016-12-17 | 2017-05-17 | 复旦大学 | Brain glioma molecular marker nondestructive prediction method and prediction system based on radiomics |
| CN106780515A (en) * | 2017-01-04 | 2017-05-31 | 南京审计大学 | Glioma method for extracting region in a kind of cerebral magnetic resonance image |
Non-Patent Citations (1)
| Title |
|---|
| 神经导航系统中的高低级别胶质瘤的鉴别诊断以及基于多DTI参数诊断模型的建立;马琳;《中国博士学位论文全文数据库 医药卫生科技辑》;20150315;第[0049]-[0059]页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109242816A (en) | 2019-01-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220215563A1 (en) | Grading of structures for state determination | |
| Rana et al. | Regions-of-interest based automated diagnosis of Parkinson’s disease using T1-weighted MRI | |
| US8121375B2 (en) | Methods and devices for producing the parameters of the brain tissues and assessing data of the suitability for thrombolysis of a patient | |
| US8634614B2 (en) | System and method for volumetric analysis of medical images | |
| CN104414636B (en) | Cerebral microbleeds computer-aided detection system based on MRI | |
| CN110840468B (en) | Autism risk assessment method and device, terminal device and storage medium | |
| Coupé et al. | LesionBrain: an online tool for white matter lesion segmentation | |
| WO2012074039A1 (en) | Medical image processing device | |
| Klein et al. | Early diagnosis of dementia based on intersubject whole-brain dissimilarities | |
| KR102288078B1 (en) | device that provides information related to Parkinson's disease by analyzing neuromelanine related magnetic resonance images based on artificial intelligence | |
| Rueda et al. | Bag of features for automatic classification of Alzheimer’s disease in magnetic resonance images | |
| CN115393269A (en) | Extensible multi-level graph neural network model based on multi-modal image data | |
| WO2008138810A1 (en) | Method for analysing an image of the brain of a subject, computer program product for analysing such image and apparatus for implementing the method. | |
| CN112927187A (en) | Method for automatically identifying and positioning focal cortical dysplasia epileptic focus | |
| Perez-Gonzalez et al. | Mild cognitive impairment classification using combined structural and diffusion imaging biomarkers | |
| Yang et al. | Diagnosis of Parkinson’s disease based on 3D ResNet: The frontal lobe is crucial | |
| CN109242816B (en) | Brain glioma pathological grade imaging auxiliary judgment method based on resting state brain function contralateral positioning | |
| Somasundaram et al. | Brain extraction method for T1-weighted magnetic resonance scans | |
| Ahmad et al. | Brain tumor detection & features extraction from MR images using segmentation, image optimization & classification techniques | |
| CN114862799B (en) | Full-automatic brain volume segmentation method for FLAIR-MRI sequence | |
| CN114847922A (en) | Brain age prediction method based on automatic fiber bundle identification | |
| CN113080929A (en) | anti-NMDAR encephalitis image feature classification method based on machine learning | |
| CN113303782A (en) | Alzheimer disease SVM classification model construction method based on nuclear magnetic resonance spectrum | |
| CN115500794B (en) | Method for identifying subjective cognitive decline and electronic equipment | |
| González Villà | Automated brain structure segmentation in magnetic resonance images of multiple sclerosis patients |
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 |