Multivariate pattern classification and prediction offers an alternative to standard univariate analysis techniques, and has recently been applied in MR imaging using support vector machines (SVM), and used to attain real-time feedback. The standard approach has been to use reconstructed image magnitude data. However, information is also present in the image phase data, and in the k-space data itself. Further, multi-echo imaging offers possibilities of increased functional sensitivity and quantitative imaging. In this study, we explore applying SVM techniques to complex and multi-echo fMRI data, using both phase information and earlier echo-times for prediction.
{"title":"Multivariate Classification of Complex and Multi-echo fMRI Data","authors":"S. Peltier, D. Noll, J. Lisinski, S. LaConte","doi":"10.1109/PRNI.2013.65","DOIUrl":"https://doi.org/10.1109/PRNI.2013.65","url":null,"abstract":"Multivariate pattern classification and prediction offers an alternative to standard univariate analysis techniques, and has recently been applied in MR imaging using support vector machines (SVM), and used to attain real-time feedback. The standard approach has been to use reconstructed image magnitude data. However, information is also present in the image phase data, and in the k-space data itself. Further, multi-echo imaging offers possibilities of increased functional sensitivity and quantitative imaging. In this study, we explore applying SVM techniques to complex and multi-echo fMRI data, using both phase information and earlier echo-times for prediction.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114603294","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
F. Yanez, A. Fan, B. Bilgiç, C. Milovic, E. Adalsteinsson, P. Irarrazaval
Quantitative susceptibility mapping (QSM) is a last decade new concept which allows to determine the magnetic susceptibility distribution of tissue in-vivo. Nowadays it has several applications such as venous blood oxygenation and iron concentration quantification. To reconstruct high quality maps, a regularized scheme must be used to solve this ill-posed problem, due to the dipole kernel under sampling k-space. A widely used regularization penalty is Total Variation (TV), however, we can find stair casing artifacts in reconstructions due to the assumption that images are piecewise constant, not always true in MRI. In this sense, we propose a less restrictive functional, to avoid this problem and to improve QSM quality. A second order Total Generalized Variation (TGV) does not assume piecewise constancy in the images and is equivalent to TV in terms of edge preservation and noise removal. This work describes how TGV penalty addresses an increase in imaging efficiency in magnetic susceptibility maps from numerical phantom and in-vivo data. Currently, we report higher specificity with the proposed regularization. Moreover, the robustness of TGV suggest that is a possible alternative to tissue susceptibility mapping.
{"title":"Quantitative Susceptibility Map Reconstruction via a Total Generalized Variation Regularization","authors":"F. Yanez, A. Fan, B. Bilgiç, C. Milovic, E. Adalsteinsson, P. Irarrazaval","doi":"10.1109/PRNI.2013.59","DOIUrl":"https://doi.org/10.1109/PRNI.2013.59","url":null,"abstract":"Quantitative susceptibility mapping (QSM) is a last decade new concept which allows to determine the magnetic susceptibility distribution of tissue in-vivo. Nowadays it has several applications such as venous blood oxygenation and iron concentration quantification. To reconstruct high quality maps, a regularized scheme must be used to solve this ill-posed problem, due to the dipole kernel under sampling k-space. A widely used regularization penalty is Total Variation (TV), however, we can find stair casing artifacts in reconstructions due to the assumption that images are piecewise constant, not always true in MRI. In this sense, we propose a less restrictive functional, to avoid this problem and to improve QSM quality. A second order Total Generalized Variation (TGV) does not assume piecewise constancy in the images and is equivalent to TV in terms of edge preservation and noise removal. This work describes how TGV penalty addresses an increase in imaging efficiency in magnetic susceptibility maps from numerical phantom and in-vivo data. Currently, we report higher specificity with the proposed regularization. Moreover, the robustness of TGV suggest that is a possible alternative to tissue susceptibility mapping.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126104519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Virgile Fritsch, Benoit Da Mota, G. Varoquaux, V. Frouin, E. Loth, J. Poline, B. Thirion
Gene-neuroimaging studies involve high-dimensional data that have a complex statistical structure and that are likely to be contaminated with outliers. Robust, outlier-resistant methods are an alternative to prior outliers removal, which is a difficult task under high-dimensional unsupervised settings. In this work, we consider robust regression and its application to neuroimaging through an example gene-neuroimaging study on a large cohort of 300 subjects. We use randomized brain parcellation to sample a set of adapted low-dimensional spatial models to analyse the data. Combining this approach with robust regression in an analysis method that we show is outperforming state-of-the-art neuroimaging analysis methods.
{"title":"Robust Group-Level Inference in Neuroimaging Genetic Studies","authors":"Virgile Fritsch, Benoit Da Mota, G. Varoquaux, V. Frouin, E. Loth, J. Poline, B. Thirion","doi":"10.1109/PRNI.2013.15","DOIUrl":"https://doi.org/10.1109/PRNI.2013.15","url":null,"abstract":"Gene-neuroimaging studies involve high-dimensional data that have a complex statistical structure and that are likely to be contaminated with outliers. Robust, outlier-resistant methods are an alternative to prior outliers removal, which is a difficult task under high-dimensional unsupervised settings. In this work, we consider robust regression and its application to neuroimaging through an example gene-neuroimaging study on a large cohort of 300 subjects. We use randomized brain parcellation to sample a set of adapted low-dimensional spatial models to analyse the data. Combining this approach with robust regression in an analysis method that we show is outperforming state-of-the-art neuroimaging analysis methods.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122521917","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Inflammatory lesions form in the brain and spinal cord of patients with multiple sclerosis (MS). In many active MS lesions, blood flows abnormally into the white matter of the brain due to breakdown of the blood-brain barrier (BBB), which is know to be associated with morbidity and disability. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) allows quantitative study of blood flow and permeability dynamics throughout the brain. In our study, we observe 15 patients who undergo DCE-MRI periodically throughout a year. In this paper, we design and study spatiotemporal parameters of interest that cannot be obtained by visual inspection. Examples of such parameters are the rate and maximum intensity observed in regions of interest. We develop semi parametric methods for this quantification of BBB disruption at each visit.
{"title":"Information Criteria for Dynamic Contrast-Enhanced Magnetic Resonance Imaging","authors":"R. Shinohara, C. Crainiceanu, B. Caffo, D. Reich","doi":"10.1109/PRNI.2013.19","DOIUrl":"https://doi.org/10.1109/PRNI.2013.19","url":null,"abstract":"Inflammatory lesions form in the brain and spinal cord of patients with multiple sclerosis (MS). In many active MS lesions, blood flows abnormally into the white matter of the brain due to breakdown of the blood-brain barrier (BBB), which is know to be associated with morbidity and disability. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) allows quantitative study of blood flow and permeability dynamics throughout the brain. In our study, we observe 15 patients who undergo DCE-MRI periodically throughout a year. In this paper, we design and study spatiotemporal parameters of interest that cannot be obtained by visual inspection. Examples of such parameters are the rate and maximum intensity observed in regions of interest. We develop semi parametric methods for this quantification of BBB disruption at each visit.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125689778","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Meriem El Azami, A. Hammers, N. Costes, C. Lartizien
We designed a machine learning system based on a one-class support vector machine (OC-SVM) classifier in view of the detection of abnormalities in magnetic resonance images (MRIs) of patients with intractable epilepsy. This system performs a voxelwise analysis and outputs clusters of detected voxels ranked by size and suspicion degree. Features correspond to a combination of six maps: three tissue probabilities (grey matter, white matter and cerebrospinal fluid), cortical thickness, grey matter extension, and greywhite matter junction. The OC-SVM is trained using 29 controls, and tested on two patients with histologically proven focal cortical dysplasia (FCD). To assess the performance of the OC-SVM classifier, the classifier was compared with a statistical parametric mapping (SPM) single subject analysis using junction and extension maps only. The identified regions were also visually evaluated by an expert and compared to other data such as FDG-positron Emission tomography (PET) and magneto encephalography (MEG). For the two patients, both analyses agreed with the visually determined localization of the FCD lesions. No match was found for the other detected regions. The OC-SVM classifier was more specific in region localization and generated fewer false positive detections than the mass-univariate SPM approach.
{"title":"Computer Aided Diagnosis of Intractable Epilepsy with MRI Imaging Based on Textural Information","authors":"Meriem El Azami, A. Hammers, N. Costes, C. Lartizien","doi":"10.1109/PRNI.2013.32","DOIUrl":"https://doi.org/10.1109/PRNI.2013.32","url":null,"abstract":"We designed a machine learning system based on a one-class support vector machine (OC-SVM) classifier in view of the detection of abnormalities in magnetic resonance images (MRIs) of patients with intractable epilepsy. This system performs a voxelwise analysis and outputs clusters of detected voxels ranked by size and suspicion degree. Features correspond to a combination of six maps: three tissue probabilities (grey matter, white matter and cerebrospinal fluid), cortical thickness, grey matter extension, and greywhite matter junction. The OC-SVM is trained using 29 controls, and tested on two patients with histologically proven focal cortical dysplasia (FCD). To assess the performance of the OC-SVM classifier, the classifier was compared with a statistical parametric mapping (SPM) single subject analysis using junction and extension maps only. The identified regions were also visually evaluated by an expert and compared to other data such as FDG-positron Emission tomography (PET) and magneto encephalography (MEG). For the two patients, both analyses agreed with the visually determined localization of the FCD lesions. No match was found for the other detected regions. The OC-SVM classifier was more specific in region localization and generated fewer false positive detections than the mass-univariate SPM approach.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116728345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Anne-Laure Fouque, C. Fischer, V. Frouin, P. Ciuciu, E. Duchesnay
The aim of this paper is to identify the relevant features that improve the identification of associations between structural (T1-weighted) MRI and a group (clinical status) of each subject. We compare 5 features derived from grey matter and deformation, on both simulated and experimental data. With voxel-based analysis we compare sensitivity of detection of anatomical differences, with pattern recognition approaches, we compare the accuracies of group prediction. The best results on our data are achieved by a multivariate representation of the deformation, the strain tensor, that can be associated with grey matter.
{"title":"Comparison of Features for Voxel-Based Analysis and Classification of Anatomical Neuroimaging Data","authors":"Anne-Laure Fouque, C. Fischer, V. Frouin, P. Ciuciu, E. Duchesnay","doi":"10.1109/PRNI.2013.55","DOIUrl":"https://doi.org/10.1109/PRNI.2013.55","url":null,"abstract":"The aim of this paper is to identify the relevant features that improve the identification of associations between structural (T1-weighted) MRI and a group (clinical status) of each subject. We compare 5 features derived from grey matter and deformation, on both simulated and experimental data. With voxel-based analysis we compare sensitivity of detection of anatomical differences, with pattern recognition approaches, we compare the accuracies of group prediction. The best results on our data are achieved by a multivariate representation of the deformation, the strain tensor, that can be associated with grey matter.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117219243","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
E. Olivetti, Danilo Benozzo, S. M. Kia, Marta Ellero, T. Hartmann
Assessing whether the patterns of brain activity systematically differ when the subject is presented with different sets of stimuli is called "brain decoding". The most common solution to this problem is based on testing whether a classifier can accurately predict the type of stimulus from brain data. In this work we present a novel approach to the brain decoding problem which does not require any classifier. The proposed method is based on a high-dimensional two-sample test recently proposed in the machine learning literature. The test tries to determine whether the set of brain recordings related to one kind of stimulus, i.e. the first sample, and the ones related to the other kind of stimulus, i.e. the second sample, are drawn from the same probability distribution or not. In this work we illustrate the advantages of this novel approach together with experimental evidence of its efficacy on magneto encephalographic (MEG) data from a Face, House and Body discrimination task.
{"title":"The Kernel Two-Sample Test vs. Brain Decoding","authors":"E. Olivetti, Danilo Benozzo, S. M. Kia, Marta Ellero, T. Hartmann","doi":"10.1109/PRNI.2013.41","DOIUrl":"https://doi.org/10.1109/PRNI.2013.41","url":null,"abstract":"Assessing whether the patterns of brain activity systematically differ when the subject is presented with different sets of stimuli is called \"brain decoding\". The most common solution to this problem is based on testing whether a classifier can accurately predict the type of stimulus from brain data. In this work we present a novel approach to the brain decoding problem which does not require any classifier. The proposed method is based on a high-dimensional two-sample test recently proposed in the machine learning literature. The test tries to determine whether the set of brain recordings related to one kind of stimulus, i.e. the first sample, and the ones related to the other kind of stimulus, i.e. the second sample, are drawn from the same probability distribution or not. In this work we illustrate the advantages of this novel approach together with experimental evidence of its efficacy on magneto encephalographic (MEG) data from a Face, House and Body discrimination task.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132603320","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Krzysztof J. Gorgolewski, P. Bazin, Haakon G. Engen, D. Margulies
To increase the power of neuroimaging analyses, it is common practice to reduce the whole-brain search space to subset of hypothesis-driven regions-of-interest (ROIs). Rather than strictly constrain analyses, we propose to incorporate prior knowledge using probabilistic ROIs (pROIs) using a hierarchical Bayesian framework. Each voxel prior probability of being "of-interest" or "of-non-interest" is used to perform a weighted fit of a mixture model. We demonstrate the utility of this approach through simulations with various pROIs, and the applicability using a prior based on the NeuroSynth database search term "emotion" for thresholding the fMRI results of an emotion processing task. The modular structure of pROI correction facilitates the inclusion of other innovations in Bayesian mixture modeling, and offers a foundation for balancing between exploratory analyses without neglecting prior knowledge.
{"title":"Fifty Shades of Gray, Matter: Using Bayesian Priors to Improve the Power of Whole-Brain Voxel- and Connexelwise Inferences","authors":"Krzysztof J. Gorgolewski, P. Bazin, Haakon G. Engen, D. Margulies","doi":"10.1109/PRNI.2013.57","DOIUrl":"https://doi.org/10.1109/PRNI.2013.57","url":null,"abstract":"To increase the power of neuroimaging analyses, it is common practice to reduce the whole-brain search space to subset of hypothesis-driven regions-of-interest (ROIs). Rather than strictly constrain analyses, we propose to incorporate prior knowledge using probabilistic ROIs (pROIs) using a hierarchical Bayesian framework. Each voxel prior probability of being \"of-interest\" or \"of-non-interest\" is used to perform a weighted fit of a mixture model. We demonstrate the utility of this approach through simulations with various pROIs, and the applicability using a prior based on the NeuroSynth database search term \"emotion\" for thresholding the fMRI results of an emotion processing task. The modular structure of pROI correction facilitates the inclusion of other innovations in Bayesian mixture modeling, and offers a foundation for balancing between exploratory analyses without neglecting prior knowledge.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122686751","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
F. Segovia, C. Bastin, E. Salmon, J. Górriz, J. Ramírez, C. Phillips
In recent years, several approaches to develop computer aided diagnosis systems for dementia have been proposed. The purpose of this work is to measure the advantages of using not only brain images as data source for those systems but also some psychological scores. To this aim, we compared the accuracy rates achieved by systems that use psychological scores beside the image data in the classification step and systems that use only the image data. The experiments show that the formers achieve higher accuracy rates regardless of the procedure carried out to analyze the image data.
{"title":"Automatic Differentiation between Alzheimer's Disease and Mild Cognitive Impairment Combining PET Data and Psychological Scores","authors":"F. Segovia, C. Bastin, E. Salmon, J. Górriz, J. Ramírez, C. Phillips","doi":"10.1109/PRNI.2013.45","DOIUrl":"https://doi.org/10.1109/PRNI.2013.45","url":null,"abstract":"In recent years, several approaches to develop computer aided diagnosis systems for dementia have been proposed. The purpose of this work is to measure the advantages of using not only brain images as data source for those systems but also some psychological scores. To this aim, we compared the accuracy rates achieved by systems that use psychological scores beside the image data in the classification step and systems that use only the image data. The experiments show that the formers achieve higher accuracy rates regardless of the procedure carried out to analyze the image data.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115933459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this study, we propose the discrete cosine transform coefficients as a new and effective set of features for recognizing patterns of brain activity in MEG recording. We claim that computing DCT coefficients on the time-frequency representation of MEG signals is an efficient technique to reduce the dimensionality of feature space without losing discriminative power in brain decoding tasks. Our classification results on single-trial MEG decoding suggest that DCT is a viable method comparing to standard methods and it improves decoding accuracy by preserving the dynamic patterns of signal in time, frequency and space domains.
{"title":"Discrete Cosine Transform for MEG Signal Decoding","authors":"S. M. Kia, E. Olivetti, P. Avesani","doi":"10.1109/PRNI.2013.42","DOIUrl":"https://doi.org/10.1109/PRNI.2013.42","url":null,"abstract":"In this study, we propose the discrete cosine transform coefficients as a new and effective set of features for recognizing patterns of brain activity in MEG recording. We claim that computing DCT coefficients on the time-frequency representation of MEG signals is an efficient technique to reduce the dimensionality of feature space without losing discriminative power in brain decoding tasks. Our classification results on single-trial MEG decoding suggest that DCT is a viable method comparing to standard methods and it improves decoding accuracy by preserving the dynamic patterns of signal in time, frequency and space domains.","PeriodicalId":144007,"journal":{"name":"2013 International Workshop on Pattern Recognition in Neuroimaging","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2013-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124376875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：