In this work, we explore the use of classification algorithms in predicting mental states from functional neuroimaging data. We train a linear support vector machine classifier to characterize spatial fMRI activation patterns. We employ a general linear model based feature extraction method and use the t-test for feature selection. We evaluate our method on a memory encoding task, using participants' subjective prediction about learning as a benchmark for our classifier. We show that the classifier achieves better than random predictions and the average accuracy is close to subject's own prediction performance. In addition, we validate our tool on a simple motor task where we demonstrate an average prediction accuracy of over 90%. Our experiments demonstrate that the classifier performance depends significantly on the complexity of the experimental design and the mental process of interest.
We explore unsupervised, hypothesis-free methods for fMRI analysis in two different types of experiments. First, we employ clustering to identify large-scale functionally homogeneous systems. We formulate a generative mixture model, derive the EM algorithm and apply it to delineate functional systems. We also investigate spectral clustering in application to this problem and demonstrate that both methods give rise to similar partitions of the brain based on resting state fMRI data. Second, we demonstrate how to extend this approach to include information about the experimental protocol. Specifically, we formulate a mixture model in the space of possible profiles of brain response to stimuli. In both applications, our methods confirm previously known results in brain mapping and point to new research directions for exploratory analysis of fMRI data.
We present iCluster, a fast and efficient algorithm that clusters a set of images while co-registering them using a parameterized, nonlinear transformation model. The output is a small number of template images that represent different modes in a population. This is in contrast with traditional approaches that assume a single template to construct atlases. We validate and explore the algorithm in two experiments. First, we employ iCluster to partition a data set of 416 whole brain MR volumes of subjects aged 18-96 years into three sub-groups, which mainly correspond to age groups. The templates reveal significant structural differences across these age groups that confirm previous findings in aging research. In the second experiment, we run iCluster on a group of 30 patients with dementia and 30 age-matched healthy controls. The algorithm produced three modes that mainly corresponded to a sub-population of healthy controls, a sub-population of patients with dementia and a mixture group that contained both types. These results suggest that the algorithm can be used to discover sub-populations that correspond to interesting structural or functional "modes".
We present a method for discovering patterns of activation observed through fMIRI in experiments with multiple stimuli/tasks. We introduce an explicit parameterization for the profiles of activation and represent fMRI time courses as such profiles using linear regression estimates. Working in the space of activation profiles, we design a mixture model that finds the major activation patterns along with their localization maps and derive an algorithm for fitting the model to the fMRI data. The method enables functional group analysis independent of spatial correspondence among subjects. We validate this model in the context of category selectivity in the visual cortex, demonstrating good agreement with prior findings based on hypothesis-driven methods.
We propose a novel l1l2-norm inverse solver for estimating the sources of EEG/MEG signals. Based on the standard l1-norm inverse solver, the proposed sparse distributed inverse solver integrates the l1-norm spatial model with a temporal model of the source signals in order to avoid unstable activation patterns and "spiky" reconstructed signals often produced by the original solvers. The joint spatio-temporal model leads to a cost function with an l1l2-norm regularizer whose minimization can be reduced to a convex second-order cone programming problem and efficiently solved using the interior-point method. Validation with simulated and real MEG data shows that the proposed solver yields source time course estimates qualitatively similar to those obtained through dipole fitting, but without the need to specify the number of dipole sources in advance. Furthermore, the l1l2-norm solver achieves fewer false positives and a better representation of the source locations than the conventional l2 minimum-norm estimates.
We describe a method for correcting the distortions present in echo planar images (EPI) and registering the EPI to structural MRI. A fieldmap is predicted from an air / tissue segmentation of the MRI using a perturbation method and subsequently used to unwarp the EPI data. Shim and other missing parameters are estimated by registration. We obtain results that are similar to those obtained using fieldmaps, however neither fieldmaps, nor knowledge of shim coefficients is required.
Several recent studies explored the use of unsupervised segmentation methods for segmenting thalamic nuclei from diffusion tensor images. These methods provide a plausible segmentation on individual subjects; however, they do not address the problem of consistently identifying the same functional areas in a population. The lack of correspondence between the segmented nuclei make it more difficult to use the results from the unsupervised segmentation tools for morphometry. In this paper we present a novel segmentation algorithm to automatically segment the gray matter nuclei while ensuring consistency between subjects in a population. This new algorithm, referred to as Consistency Clustering, finds correspondence between the nuclei as the segmentation is achieved through a single model for the whole population, similar to the brain atlases experts use to identify thalamic nuclei.
In this paper, we propose a natural framework that allows any region-based segmentation energy to be re-formulated in a local way. We consider local rather than global image statistics and evolve a contour based on local information. Localized contours are capable of segmenting objects with heterogeneous feature profiles that would be difficult to capture correctly using a standard global method. The presented technique is versatile enough to be used with any global region-based active contour energy and instill in it the benefits of localization. We describe this framework and demonstrate the localization of three well-known energies in order to illustrate how our framework can be applied to any energy. We then compare each localized energy to its global counterpart to show the improvements that can be achieved. Next, an in-depth study of the behaviors of these energies in response to the degree of localization is given. Finally, we show results on challenging images to illustrate the robust and accurate segmentations that are possible with this new class of active contour models.
Recent developments in MR data acquisition technology are starting to yield images that show anatomical features of the hippocampal formation at an unprecedented level of detail, providing the basis for hippocampal subfield measurement. Because of the role of the hippocampus in human memory and its implication in a variety of disorders and conditions, the ability to reliably and efficiently quantify its subfields through in vivo neuroimaging is of great interest to both basic neuroscience and clinical research. In this paper, we propose a fully-automated method for segmenting the hippocampal subfields in ultra-high resolution MRI data. Using a Bayesian approach, we build a computational model of how images around the hippocampal area are generated, and use this model to obtain automated segmentations. We validate the proposed technique by comparing our segmentation results with corresponding manual delineations in ultra-high resolution MRI scans of five individuals.
BACKGROUND: A reduction in interhemispheric connectivity is thought to contribute to the etiology of schizophrenia. Diffusion Tensor Imaging (DTI) measures the diffusion of water and can be used to describe the integrity of the corpus callosum white matter tracts, thereby providing information concerning possible interhemispheric connectivity abnormalities. Previous DTI studies in schizophrenia are inconsistent in reporting decreased Fractional Anisotropy (FA), a measure of anisotropic diffusion, within different portions of the corpus callosum. Moreover, none of these studies has investigated corpus callosum systematically, using anatomical subdivisions. METHODS: DTI and structural MRI scans were obtained from 32 chronic schizophrenic subjects and 42 controls. Corpus callosum cross sectional area and its probabilistic subdivisions were determined automatically from structural MRI scans using a model based deformable contour segmentation. These subdivisions employ a previously generated probabilistic subdivision atlas, based on fiber tractography and anatomical lobe subdivision. The structural scan was then co-registered with the DTI scan and the anatomical corpus callosum subdivisions were propagated to the associated FA map. RESULTS: Results revealed decreased FA within parts of the corpus interconnecting frontal regions in schizophrenia compared with controls, but no significant changes for callosal fibers interconnecting parietal and temporo-occipital brain regions. In addition, integrity of the anterior corpus was statistically significantly correlated with negative as well as positive symptoms, while posterior measures correlated with positive symptoms only. CONCLUSIONS: This study provides quantitative evidence for a reduction of interhemispheric brain connectivity in schizophrenia, involving corpus callosum, and further points to frontal connections as possibly disrupted in schizophrenia.
In this paper, we explore the use of over-complete spherical wavelets in shape analysis of closed 2D surfaces. Previous work has demonstrated, theoretically and practically, the advantages of overcomplete over bi-orthogonal spherical wavelets. Here we present a detailed formulation of over-complete wavelets, as well as shape analysis experiments of cortical folding development using them. Our experiments verify in a quantitative fashion existing qualitative theories of neuroanatomical development. Furthermore, the experiments reveal novel insights into neuro-anatomical development not previously documented.
We present the fast Spherical Demons algorithm for registering two spherical images. By exploiting spherical vector spline interpolation theory, we show that a large class of regularizers for the modified demons objective function can be efficiently implemented on the sphere using convolution. Based on the one parameter subgroups of diffeomorphisms, the resulting registration is diffeomorphic and fast - registration of two cortical mesh models with more than 100k nodes takes less than 5 minutes, comparable to the fastest surface registration algorithms. Moreover, the accuracy of our method compares favorably to the popular FreeSurfer registration algorithm. We validate the technique in two different settings: (1) parcellation in a set of in-vivo cortical surfaces and (2) Brodmann area localization in ex-vivo cortical surfaces.
We introduce a versatile framework for characterizing and extracting salient structures in three-dimensional symmetric second-order tensor fields. The key insight is that degenerate lines in tensor fields, as defined by the standard topological approach, are exactly crease (ridge and valley) lines of a particular tensor invariant called mode. This reformulation allows us to apply well-studied approaches from scientific visualization or computer vision to the extraction of topological lines in tensor fields. More generally, this main result suggests that other tensor invariants, such as anisotropy measures like fractional anisotropy (FA), can be used in the same framework in lieu of mode to identify important structural properties in tensor fields. Our implementation addresses the specific challenge posed by the non-linearity of the considered scalar measures and by the smoothness requirement of the crease manifold computation. We use a combination of smooth reconstruction kernels and adaptive refinement strategy that automatically adjust the resolution of the analysis to the spatial variation of the considered quantities. Together, these improvements allow for the robust application of existing ridge line extraction algorithms in the tensor context of our problem. Results are proposed for a diffusion tensor MRI dataset, and for a benchmark stress tensor field used in engineering research.
Richly labeled images representing several sub-structures of an organ occur quite frequently in medical images. For example, a typical brain image can be labeled into grey matter, white matter or cerebrospinal fluid, each of which may be subdivided further. Many manipulations such as interpolation, transformation, smoothing, or registration need to be performed on these images before they can be used in further analysis. In this work, we present a novel multi-shape representation and compare it with the existing representations to demonstrate certain advantages of using the proposed scheme. Specifically, we propose label space, a representation that is both flexible and well suited for coupled multi-shape analysis. Under this framework, object labels are mapped to vertices of a regular simplex, e.g. the unit interval for two labels, a triangle for three labels, a tetrahedron for four labels, etc. This forms the basis of a convex linear structure with the property that all labels are equally spaced. We will demonstrate that this representation has several desirable properties: algebraic operations may be performed directly, label uncertainty is expressed equivalently as a weighted mixture of labels or in a probabilistic manner, and interpolation is unbiased toward any label or the background. In order to demonstrate these properties, we compare label space to signed distance maps as well as other implicit representations in tasks such as smoothing, interpolation, registration, and principal component analysis.
A software strategy to provide intuitive navigation for MRI-guided robotic transperineal prostate therapy is presented. In the system, the robot control unit, the MRI scanner, and open-source navigation software are connected to one another via Ethernet to exchange commands, coordinates, and images. Six states of the system called "workphases" are defined based on the clinical scenario to synchronize behaviors of all components. The wizard-style user interface allows easy following of the clinical workflow. On top of this framework, the software provides features for intuitive needle guidance: interactive target planning; 3D image visualization with current needle position; treatment monitoring through real-time MRI. These features are supported by calibration of robot and image coordinates by the fiducial-based registration. The performance test shows that the registration error of the system was 2.6 mm in the prostate area, and it displayed real-time 2D image 1.7 s after the completion of image acquisition.
We propose a Bayesian approach to incorporate anatomical information in the clustering of fiber trajectories. An expectation-maximization (EM) algorithm is used to cluster the trajectories, in which an atlas serves as the prior on the labels. The atlas guides the clustering algorithm and makes the resulting bundles anatomically meaningful. In addition, it provides the seed points for the tractography and initial settings of the EM algorithm. The proposed approach provides a robust and automated tool for tract-oriented analysis both in a single subject and over a population.
Both post-mortem and neuroimaging studies have contributed significantly to what we know about the brain and schizophrenia. MRI studies of volumetric reduction in several brain regions in schizophrenia have confirmed early speculations that the brain is disordered in schizophrenia. There is also a growing body of evidence suggesting that a disturbance in connectivity between different brain regions, rather than abnormalities within the separate regions themselves, are responsible for the clinical symptoms and cognitive dysfunctions observed in this disorder. Thus an interest in white matter fiber tracts, subserving anatomical connections between distant, as well as proximal, brain regions, is emerging. This interest coincides with the recent advent of diffusion tensor imaging (DTI), which makes it possible to evaluate the organization and coherence of white matter fiber tracts. This is an important advance as conventional MRI techniques are insensitive to fiber tract direction and organization, and have not consistently demonstrated white matter abnormalities. DTI may, therefore, provide important new information about neural circuitry, and it is increasingly being used in neuroimaging studies of psychopathological disorders. Of note, in the past five years 18 DTI studies in schizophrenia have been published, most describing white matter abnormalities. Questions still remain, however, regarding what we are measuring that is abnormal in this disease, and how measures obtained using one method correspond to those obtained using other methods? Below we review the basic principles involved in MR-DTI, followed by a review of the different methods used to evaluate diffusion. Finally, we review MR-DTI findings in schizophrenia.
This contribution presents finite element computation of the deformation field within the brain during craniotomy-induced brain shift. The results were used to illustrate the capabilities of non-linear (i.e. accounting for both geometric and material non-linearities) finite element analysis in non-rigid registration of pre- and intra-operative magnetic resonance images of the brain. We used patient-specific hexahedron-dominant finite element mesh, together with realistic material properties for the brain tissue and appropriate contact conditions at boundaries. The model was loaded by the enforced motion of nodes (i.e. through prescribed motion of a boundary) at the brain surface in the craniotomy area. We suggest using explicit time-integration scheme for discretised equations of motion, as the computational times are much shorter and accuracy, for practical purposes, the same as in the case of implicit integration schemes. Application of the computed deformation field to register (i.e. align) the pre-operative images with the intra-operative ones indicated that the model very accurately predicts the displacements of the tumour and the lateral ventricles even for limited information about the brain surface deformation. The prediction accuracy improves when information about deformation of not only exposed (during craniotomy) but also unexposed parts of the brain surface is used when prescribing loading. However, it appears that the accuracy achieved using information only about the deformation of the exposed surface, that can be determined without intra-operative imaging, is acceptable. The presented results show that non-linear biomechanical models can complement medical image processing techniques when conducting non-rigid registration. Important advantage of such models over the previously used linear ones is that they do not require unrealistic assumptions that brain deformations are infinitesimally small and brain stress-strain relationship is linear.
Many neuroanatomy studies rely on brain tissue segmentations of magnetic resonance images (MRI). We present a segmentation tool, which performs this task automatically by analyzing the MRIs as well as tissue specific spatial priors. The priors are aligned to the patient through a non-rigid registration method. The segmentation itself is parameterized by an XML file making the approach easily adjustable to various segmentation problems. The tool is hidden beneath a 'one-button' user interface, which is simple to install and is applicable to a wide variety of image acquisition protocols.
In this work, we extend a previously demonstrated entropy based groupwise registration method to include a free-form deformation model based on B-splines. We provide an efficient implementation using stochastic gradient descents in a multi-resolution setting. We demonstrate the method in application to a set of 50 MRI brain scans and compare the results to a pairwise approach using segmentation labels to evaluate the quality of alignment. Our results indicate that increasing the complexity of the deformation model improves registration accuracy significantly, especially at cortical regions.