Connectivity information derived from diffusion MRI can be used to parcellate the cerebral cortex into anatomically and functionally meaningful subdivisions. Acquisition and processing parameters can significantly affect parcellation results, and there is no consensus on best practice protocols. We propose a novel approach for evaluating parcellation based on measuring the degree to which parcellation conforms to known principles of brain organization, specifically cortical field homogeneity and interhemispheric homology. The proposed metrics are well behaved on morphologically generated whole-brain parcels, where they correctly identify contralateral homologies and give higher scores to anatomically versus arbitrarily generated parcellations. The measures show that individual cortical fields have characteristic connectivity profiles that are compact and separable, and that the topological arrangement of such fields is strongly conserved between hemispheres and individuals. The proposed metrics can be used to evaluate the quality of parcellations at the subject and group levels and to improve acquisition and data processing for connectivity-based cortical parcellation.

Large scale gene expression studies in the mammalian brain offer the promise of understanding the topology, networks and ultimately the function of its complex anatomy, opening previously unexplored avenues in neuroscience. High-throughput methods permit genome-wide searches to discover genes that are uniquely expressed in brain circuits and regions that control behavior. Previous gene expression mapping studies in model organisms have employed situ hybridization (ISH), a technique that uses labeled nucleic acid probes to bind to specific mRNA transcripts in tissue sections. A key requirement for this effort is the development of fast and robust algorithms for anatomically mapping and quantifying gene expression for ISH. We describe a neuroinformatics pipeline for automatically mapping expression profiles of ISH data and its use to produce the first genomic scale 3-D mapping of gene expression in a mammalian brain. The pipeline is fully automated and adaptable to other organisms and tissues. Our automated study of over 20,000 genes indicates that at least 78.8 percent are expressed at some level in the adult C56BL/6J mouse brain. In addition to providing a platform for genomic scale search, high-resolution images and visualization tools for expression analysis are available at the Allen Brain Atlas web site (http://www.brain-map.org).

BACKGROUND: Reverse-engineering regulatory networks is one of the central challenges for computational biology. Many techniques have been developed to accomplish this by utilizing transcription factor binding data in conjunction with expression data. Of these approaches, several have focused on the reconstruction of the cell cycle regulatory network of Saccharomyces cerevisiae. The emphasis of these studies has been to model the relationships between transcription factors and their target genes. In contrast, here we focus on reverse-engineering the network of relationships among transcription factors that regulate the cell cycle in S. cerevisiae. RESULTS: We have developed a technique to reverse-engineer networks of the time-dependent activities of transcription factors that regulate the cell cycle in S. cerevisiae. The model utilizes linear regression to first estimate the activities of transcription factors from expression time series and genome-wide transcription factor binding data. We then use least squares to construct a model of the time evolution of the activities. We validate our approach in two ways: by demonstrating that it accurately models expression data and by demonstrating that our reconstructed model is similar to previously-published models of transcriptional regulation of the cell cycle. CONCLUSION: Our regression-based approach allows us to build a general model of transcriptional regulation of the yeast cell cycle that includes additional factors and couplings not reported in previously-published models. Our model could serve as a starting point for targeted experiments that test the predicted interactions. In the future, we plan to apply our technique to reverse-engineer other systems where both genome-wide time series expression data and transcription factor binding data are available.

Automatic prostate segmentation in ultrasound images is a challenging task due to speckle noise, missing boundary segments, and complex prostate anatomy. One popular approach has been the use of deformable models. For such techniques, prior knowledge of the prostate shape plays an important role in automating model initialization and constraining model evolution. In this paper, we have modeled the prostate shape using deformable superellipses. This model was fitted to 594 manual prostate contours outlined by five experts. We found that the superellipse with simple parametric deformations can efficiently model the prostate shape with the Hausdorff distance error (model versus manual outline) of 1.32 +/- 0.62 mm and mean absolute distance error of 0.54 +/- 0.20 mm. The variability between the manual outlinings and their corresponding fitted deformable superellipses was significantly less than the variability between human experts with p-value being less than 0.0001. Based on this deformable superellipse model, we have developed an efficient and robust Bayesian segmentation algorithm. This algorithm was applied to 125 prostate ultrasound images collected from 16 patients. The mean error between the computer-generated boundaries and the manual outlinings was 1.36 +/- 0.58 mm, which is significantly less than the manual interobserver distances. The algorithm was also shown to be fairly insensitive to the choice of the initial curve.

Video object segmentation and tracking are essential for content-based video processing. This paper presents a framework for a semiautomatic approach to this problem. A semantic video object is initialized with human assistance in a key frame. The video object is then tracked and segmented automatically in the following frames. A new active contour model, VSnakes, is introduced as a segmentation method in this framework. The active contour energy is defined so as to reflect the energy difference between two contours instead of the energy of a single contour. Multiple-resolution wavelet decomposition is applied in generating the edge energy of the image frame. Contour relaxation is used to deal with the object deformation frame by frame, and the Viterbi algorithm is used to update the contour path during contour relaxation. Compared to the original snakes algorithm, semiautomatic video object segmentation with the VSnakes algorithm resulted in improved performance in terms of video object shape distortion (1.4% versus 2.9% in one experiment), which suggests that it could be a useful tool in many content-based video applications, e.g., MPEG-4 video object generation and medical imaging.

We have implemented and validated an algorithm for three-dimensional positron emission tomography transmission-to-computed tomography registration in the chest, using mutual information as a similarity criterion. Inherent differences in the two imaging protocols produce significant nonrigid motion between the two acquisitions. A rigid body deformation combined with localized cubic B-splines is used to capture this motion. The deformation is defined on a regular grid and is parameterized by potentially several thousand coefficients. Together with a spline-based continuous representation of images and Parzen histogram estimates, our deformation model allows closed-form expressions for the criterion and its gradient. A limited-memory quasi-Newton optimization algorithm is used in a hierarchical multiresolution framework to automatically align the images. To characterize the performance of the method, 27 scans from patients involved in routine lung cancer staging were used in a validation study. The registrations were assessed visually by two expert observers in specific anatomic locations using a split window validation technique. The visually reported errors are in the 0- to 6-mm range and the average computation time is 100 min on a moderate-performance workstation.

An algorithm is introduced for video object tracking and segmentation that automatically detects local deformations of a video object and adjusts the object contour to match the real object boundary without breaking the object into several pieces during tracking. An edge energy change ratio is defined for local deformation detection. The corresponding contour segment is then adjusted to the real object boundary based on a local affine correction. Incorporating this method into VSnakes, our general algorithm for video object tracking, resulted in improved performance.

Abstract Microarrays traditionally have been used to assay the transcript expression of coding regions of genes. Here, we use Escherichia coli oligonucleotide microarrays to assay transcript expression of both open reading frames (ORFs) and intergenic regions. We then use hidden Markov models to analyse this expression data and estimate transcription boundaries of genes. This approach allows us to identify 5′ untranslated regions (5′ UTRs) of transcripts as well as genes that are likely to be operon members. The operon elements we identify correspond to documented operons with 99% specificity and 63% sensitivity. Similarly we find that our 5′ UTR results accurately coincide with experimentally verified promoter regions for most genes. Contact: [email protected] Keywords: untranslated regions; operons; intergenic; microarrays; Escherichia coli. *To whom correspondence should be addressed.

Abstract Motivation: Many clustering algorithms have been proposed for the analysis of gene expression data, but little guidance is available to help choose among them. We provide a systematic framework for assessing the results of clustering algorithms. Clustering algorithms attempt to partition the genes into groups exhibiting similar patterns of variation in expression level. Our methodology is to apply a clustering algorithm to the data from all but one experimental condition. The remaining condition is used to assess the predictive power of the resulting clusters—meaningful clusters should exhibit less variation in the remaining condition than clusters formed by chance. Results: We successfully applied our methodology to compare six clustering algorithms on four gene expression data sets. We found our quantitative measures of cluster quality to be positively correlated with external standards of cluster quality. Availability: The software is under development. Contact: [email protected] Supplementary information: http://www.cs.washington.edu/homes/kayee/cluster or http://www.cs.washington.edu/homes/ruzzo/cluster * To whom correspondence should be addressed.

We have developed a new contour-based tracking algorithm that uses a sequence of template deformations to model and track generic video objects. We organize the deformations into a hierarchy: globally affine deformations, piecewise (locally) affine deformations, and arbitrary smooth deformations (snakes). This design enables the algorithm to track objects whose pose and shape change in time compared to the template. If the object is not a rigid body, we model the temporal evolution of its shape by updating the entire template after each video frame; otherwise, we only update the pose of the object. Experimental results demonstrate that our method is able to track a variety of video objects, including those undergoing rapid changes. We quantitatively compare our algorithm with its constituent pieces (e.g., the snake algorithm) and show that the complete algorithm can track objects with moving parts for a longer duration than partial versions of the hierarchy. It could be benefited from a higher level algorithm to dynamically adjust the parameters and template deformations to improve the segmentation accuracy further. The hierarchical nature of this algorithm provides a framework that offers a modular approach for the design and enhancement of future object-tracking algorithms.

Optical flow has been widely used in motion estimation. By initializing adaptive gradients, we have developed a technique to reduce the occlusion problem in computing optical flow with the Horn and Schunck (1981) method. The adaptive gradients are derived from the forward and backward image gradients based on the optical flow equation during iterations. Improved optical flow solutions with the elimination of 'double-image' artifacts caused by occlusions have been obtained.

Accurate detection of prostate boundaries is required in many diagnostic and treatment procedures for prostate disease. In this paper, a new paradigm for guided edge delineation is described, which involves presenting automatically detected prostate edges as a visual guide to the observer, followed by manual editing. This approach enables robust delineation of the prostate boundaries, making it suitable for routine clinical use. The edge-detection algorithm is comprised of three stages. An algorithm called sticks is used to enhance contrast and at the same time reduce speckle in the transrectal ultrasound prostate image. The resulting image is further smoothed using an anisotropic diffusion filter. In the third stage, some basic prior knowledge of the prostate, such as shape and echo pattern, is used to detect the most probable edges describing the prostate. Finally, patient-specific anatomic information is integrated during manual linking of the detected edges. The algorithm was tested on 125 images from 16 patients. The performance of the algorithm was statistically evaluated by employing five expert observers. Based on this study, we found that consistency in prostate delineation increases when automatically detected edges are used as visual guide during outlining, while the accuracy of the detected edges was found to be at least as good as those of the human observers. The use of edge guidance for boundary delineation can also be extended to other applications in medical imaging where poor contrast in the images and the complexity in the anatomy limit the clinical usability of fully automatic edge-detection techniques.

The use of stereotactic systems has been one of the main approaches for image-based guidance of the surgical tool within the brain. The main limitation of stereotactic systems is that they are based on preoperative images that might become outdated and invalid during the course of surgery. Ultrasound (US) is considered the most practical and cost-effective intraoperative imaging modality, but US images inherently have a low signal-to-noise ratio. Integrating intraoperative US with stereotactic systems has recently been attempted. In this paper, we present a new system for interactively registering two-dimensional US and three-dimensional magnetic resonance (MR) images. This registration is based on tracking the US probe with a dc magnetic position sensor. We have performed an extensive analysis of the errors of our system by using a custom-built phantom. The registration error between the MR and the position sensor space was found to have a mean value of 1.78 mm and a standard deviation of 0.18 mm. The registration error between US and MR space was dependent on the distance of the target point from the US probe face. For a 3.5-MHz phased one-dimensional array transducer and a depth of 6 cm, the mean value of the registration error was 2.00 mm and the standard deviation was 0.75 mm. The registered MR images were reconstructed using either zeroth-order or first-order interpolation. The ease of use and the interactive nature of our system (approximately 6.5 frames/s for 344 x 310 images and first-order interpolation on a Pentium II 450 MHz) demonstrates its potential to be used in the operating room.

Tracing of left-ventricular epicardial and endocardial borders on echocardiographic sequences is essential for quantification of cardiac function. The authors designed a method based on an extension of active contour models to detect both epicardial and endocardial borders on short-axis cardiac sequences spanning the entire cardiac cycle. They validated the results by comparing the computer-generated boundaries to the boundaries manually outlined by four expert observers on 44 clinical data sets. The mean boundary distance between the computer-generated boundaries and the manually outlined boundaries was 2.80 mm (sigma=1.28 mm) for the epicardium and 3.61 (sigma=1.68 mm) for the endocardium. These distances were comparable to interobserver distances, which had a mean of 3.79 mm (sigma=1.53 mm) for epicardial borders and 2.67 mm (sigma=0.88 mm) for endocardial borders. The correlation coefficient between the areas enclosed by the computer-generated boundaries and the average manually outlined boundaries was 0.95 for epicardium and 0.91 for endocardium. The algorithm is fairly insensitive to the choice of the initial curve. Thus, the authors have developed an effective and robust algorithm to extract left-ventricular boundaries from echocardiographic sequences.

Conventional diagnostic protocols in radiological viewing require the availability of a large display space, as exemplified by the film alternator (or lightbox). However, the recent trend in medical imaging workstation design is the use of relatively small display screens to present radiological images. In the study, the efficacy of an alternatorfilmstrip metaphor for navigating through and displaying the images of a patient study is evaluated. The metaphor relies on the availability of a ‘pictorial directory’ and accommodates the concurrent display of a variable number of images on the screen. In addition the study also evaluated the utility of two commonly available operations for manipulating the appearance of an image: contrast enhancement and image magnification. The study consisted of two experimental variables: availability of a pictorial directory (yes, no) and number of concurrently displayed images (1, 2, or 4). Twelve radiologists viewed two sets of X-ray computed tomography (CT) studies for each of the six treatment conditions and were asked to record an interpretation for each study. Results indicate that interpretations using the two-image display formats took the least amount of time per study, whereas the single-image format with access to the pictorial directory required the longest mean interpretation time. The image magnification capability was generally not found to be useful for image interpretation, whereas the contrast manipulation operations were judged extremely helpful. Implications of the results for medical imaging workstation design are discussed.

The problem of negative artifacts in emission tomography reconstructions computed by filtered backprojection (FBP) is of practical concern particularly in low count studies. Statistical reconstruction methods based on maximum likelihood (ML) are automatically constrained to be non-negative but their excessive computational overhead (orders of magnitude greater than FBP) has limited their use in operational settings. Motivated by the statistical character of the negativity artifact, the authors develop a simple post-processing technique that iteratively adjusts negative values by cancellation with positive values in a surrounding local neighborhood. The compute time of this approach is roughly equivalent to 2 applications of FBP. The approach was evaluated by numerical simulation in 1- and 2-dimensional (2D) settings. In 2D, the source distributions included the Hoffman, the Shepp and Vardi, and a digitized version of the Jaszczak cold spheres phantoms. The authors' studies compared smoothed versions of FBP, the post-processed FBP, and ML implemented by the expectation-maximization algorithm. The root mean square (RMS) error between the true and estimated source distribution was used to evaluate performance; in 2D, additional region-of-interest-based measures of reconstruction accuracy were also employed. In making comparisons between the different methods, the amount of smoothing applied to each reconstruction method was adapted to minimize the RMS error-this was found to be critical.