Image registration is typically formulated as an optimization problem with multiple tunable, manually set parameters. We present a principled framework for learning thousands of parameters of registration cost functions, such as a spatially-varying tradeoff between the image dissimilarity and regularization terms. Our approach belongs to the classic machine learning framework of model selection by optimization of cross-validation error. This second layer of optimization of cross-validation error over and above registration selects parameters in the registration cost function that result in good registration as measured by the performance of the specific application in a training data set. Much research effort has been devoted to developing generic registration algorithms, which are then specialized to particular imaging modalities, particular imaging targets and particular postregistration analyses. Our framework allows for a systematic adaptation of generic registration cost functions to specific applications by learning the "free" parameters in the cost functions. Here, we consider the application of localizing underlying cytoarchitecture and functional regions in the cerebral cortex by alignment of cortical folding. Most previous work assumes that perfectly registering the macro-anatomy also perfectly aligns the underlying cortical function even though macro-anatomy does not completely predict brain function. In contrast, we learn 1) optimal weights on different cortical folds or 2) optimal cortical folding template in the generic weighted sum of squared differences dissimilarity measure for the localization task. We demonstrate state-of-the-art localization results in both histological and functional magnetic resonance imaging data sets.

We develop a method for unsupervised analysis of functional brain images that learns group-level patterns of functional response. Our algorithm is based on a generative model that comprises two main layers. At the lower level, we express the functional brain response to each stimulus as a binary activation variable. At the next level, we define a prior over the sets of activation variables in all subjects. We use a Hierarchical Dirichlet Process as the prior in order to simultaneously learn the patterns of response that are shared across the group, and to estimate the number of these patterns supported by data. Inference based on this model enables automatic discovery and characterization of salient and consistent patterns in functional signals. We apply our method to data from a study that explores the response of the visual cortex to a collection of images. The discovered profiles of activation correspond to selectivity to a number of image categories such as faces, bodies, and scenes. More generally, our results appear superior to the results of alternative data-driven methods in capturing the category structure in the space of stimuli.

In this paper, we analyze Markov Random Field (MRF) as a spatial regularizer in fMRI detection. The low signal-to-noise ratio (SNR) in fMRI images presents a serious challenge for detection algorithms, making regularization necessary to achieve good detection accuracy. Gaussian smoothing, traditionally employed to boost SNR, often produces over-smoothed activation maps. Recently, the use of MRF priors has been suggested as an alternative regularization approach. However, solving for an optimal configuration of the MRF is NP-hard in general. In this work, we investigate fast inference algorithms based on the Mean Field approximation in application to MRF priors for fMRI detection. Furthermore, we propose a novel way to incorporate anatomical information into the MRF-based detection framework and into the traditional smoothing methods. Intuitively speaking, the anatomical evidence increases the likelihood of activation in the gray matter and improves spatial coherency of the resulting activation maps within each tissue type. Validation using the receiver operating characteristic (ROC) analysis and the confusion matrix analysis on simulated data illustrates substantial improvement in detection accuracy using the anatomically guided MRF spatial regularizer. We further demonstrate the potential benefits of the proposed method in real fMRI signals of reduced length. The anatomically guided MRF regularizer enables significant reduction of the scan length while maintaining the quality of the resulting activation maps.

Extracting the prostate from magnetic resonance (MR) imagery is a challenging and important task for medical image analysis and surgical planning. We present in this work a unified shape-based framework to extract the prostate from MR prostate imagery. In many cases, shape-based segmentation is a two-part problem. First, one must properly align a set of training shapes such that any variation in shape is not due to pose. Then segmentation can be performed under the constraint of the learnt shape. However, the general registration task of prostate shapes becomes increasingly difficult due to the large variations in pose and shape in the training sets, and is not readily handled through existing techniques. Thus, the contributions of this paper are twofold. We first explicitly address the registration problem by representing the shapes of a training set as point clouds. In doing so, we are able to exploit the more global aspects of registration via a certain particle filtering based scheme. In addition, once the shapes have been registered, a cost functional is designed to incorporate both the local image statistics as well as the learnt shape prior. We provide experimental results, which include several challenging clinical data sets, to highlight the algorithm’s capability of robustly handling supine/prone prostate registration and the overall segmentation task.

Two key aspects of coupled multi-object shape analysis and atlas generation are the choice of representation and subsequent registration methods used to align the sample set. For example, a typical brain image can be labeled into three structures: grey matter, white matter and cerebrospinal fluid. Many manipulations such as interpolation, transformation, smoothing, or registration need to be performed on these images before they can be used in further analysis. Current techniques for such analysis tend to trade off performance between the two tasks, performing well for one task but developing problems when used for the other. This article proposes to use a representation that is both flexible and well suited for both tasks. We propose to map object labels 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 representation, which is routinely used in fuzzy classification, is ideally suited for representing and registering multiple shapes. On closer examination, this representation reveals several desirable properties: algebraic operations may be done directly, label uncertainty is expressed as a weighted mixture of labels (probabilistic interpretation), interpolation is unbiased toward any label or the background, and registration may be performed directly. We demonstrate these properties by using label space in a gradient descent based registration scheme to obtain a probabilistic atlas. While straightforward, this iterative method is very slow, could get stuck in local minima, and depends heavily on the initial conditions. To address these issues, two fast methods are proposed which serve as coarse registration schemes following which the iterative descent method can be used to refine the results. Further, we derive an analytical formulation for direct computation of the "group mean" from the parameters of pairwise registration of all the images in the sample set. We show results on richly labeled 2D and 3D data sets.

We demonstrate an automated, multi-level method to segment white matter brain lesions and apply it to lupus. The method makes use of local morphometric features based on multiple MR sequences, including T1-weighted, T2-weighted, and fluid attenuated inversion recovery. After preprocessing, including co-registration, brain extraction, bias correction, and intensity standardization, 49 features are calculated for each brain voxel based on local morphometry. At each level of segmentation a supervised classifier takes advantage of a different subset of the features to conservatively segment lesion voxels, passing on more difficult voxels to the next classifier. This multi-level approach allows for a fast lesion classification method with tunable trade-offs between sensitivity and specificity producing accuracy comparable to a human rater.

Automatic segmentation of the heart’s left atrium offers great benefits for planning and outcome evaluation of atrial ablation procedures. However, the high anatomical variability of the left atrium presents significant challenges for atlas-guided segmentation. In this paper, we demonstrate an automatic method for left atrium segmentation using weighted voting label fusion and a variant of the demons registration algorithm adapted to handle images with different intensity distributions. We achieve accurate automatic segmentation that is robust to the high anatomical variations in the shape of the left atrium in a clinical dataset of MRA images.

We propose an alternative to univariate statistics for identifying population differences in functional connectivity. Our feature selection method is based on a procedure that searches across subsets of the data to isolate a set of robust, predictive functional connections. The metric, known as the Gini Importance, also summarizes multivariate patterns of interaction, which cannot be captured by univariate techniques. We compare the Gini Importance with univariate statistical tests to evaluate functional connectivity changes induced by schizophrenia. Our empirical results indicate that univariate features vary dramatically across subsets of the data and have little classification power. In contrast, relevant features based on Gini Importance are considerably more stable and allow us to accurately predict the diagnosis of a test subject.

We propose a novel method, fMRI-Informed Regional Estimation (FIRE), which utilizes information from fMRI in E/MEG source reconstruction. FIRE takes advantage of the spatial alignment between the neural and the vascular activities, while allowing for substantial differences in their dynamics. Furthermore, with a region-based approach, FIRE estimates the model parameters for each region independently. Hence, it can be efficiently applied on a dense grid of source locations. The optimization procedure at the core of FIRE is related to the re-weighted minimum-norm algorithms. The weights in the proposed approach are computed from both the current source estimates and fMRI data, leading to robust estimates in the presence of silent sources in either fMRI or E/MEG measurements. We employ a Monte Carlo evaluation procedure to compare the proposed method to several other joint E/MEG-fMRI algorithms. Our results show that FIRE provides the best trade-off in estimation accuracy between the spatial and the temporal accuracy. Analysis using human E/MEG-fMRI data reveals that FIRE significantly reduces the ambiguities in source localization present in the minimum-norm estimates, and that it accurately captures activation timing in adjacent functional regions.

Pain assessment in patients who are unable to verbally communicate is a challenging problem. The fundamental limitations in pain assessment in neonates stem from subjective assessment criteria, rather than quantifiable and measurable data. This often results in poor quality and inconsistent treatment of patient pain management. Recent advancements in pattern recognition techniques using relevance vector machine (RVM) learning techniques can assist medical staff in assessing pain by constantly monitoring the patient and providing the clinician with quantifiable data for pain management. The RVM classification technique is a Bayesian extension of the support vector machine (SVM) algorithm, which achieves comparable performance to SVM while providing posterior probabilities for class memberships and a sparser model. If classes represent "pure" facial expressions (i.e., extreme expressions that an observer can identify with a high degree of confidence), then the posterior probability of the membership of some intermediate facial expression to a class can provide an estimate of the intensity of such an expression. In this paper, we use the RVM classification technique to distinguish pain from nonpain in neonates as well as assess their pain intensity levels. We also correlate our results with the pain intensity assessed by expert and nonexpert human examiners.