Darshan Pai

A Reference Architecture for Scientific Workflow Management Systems and the VIEW SOA Solution

Scientific workflows have recently emerged as a new paradigm for scientists to formalize and structure complex and distributed scientific processes to enable and accelerate many scientific discoveries. In contrast to business workflows, which are typically control flow oriented, scientific workflows tend to be dataflow oriented, introducing a new set of requirements for system development. These requirements demand a new architectural design for scientific workflow management systems (SWFMSs). Although several SWFMSs have been developed that provide much experience for future research and development, a study from an architectural perspective is still missing. The main contributions of this paper are: 1) based on a comprehensive survey of the literature and identification of key requirements for SWFMSs, we propose the first reference architecture for SWFMSs; 2) according to the reference architecture, we further propose a service-oriented architecture for View (a VIsual sciEntific Workflow management system); 3) we implemented View to validate the feasibility of the proposed architectures; and 4) we present a View-based scientific workflow application system (SWFAS), called FiberFlow, to showcase the application of our View system.

Abnormal Fronto-striatal Connectivity in Children with Histories of Early Deprivation: A Diffusion Tensor Imaging Study

An Inattentive/Overactive (I/O) behavioral phenotype has been reported in a significant percentage of children raised from birth in orphanages. While a number of studies have identified both functional and structural brain abnormalities in children raised from birth in orphanages, no published studies have evaluated potential neural correlates of the I/O phenotype. We applied diffusion tensor imaging (DTI) to 15 pre-teen children raised in orphanages in Eastern Europe or Asia and later adopted to the US. Fiber tracts were constructed from DTI data using probabilistic fiber tracking and the cortical fiber distribution of fibers originating from the head of the caudate was compared between the early deprivation (ED) group and 12 age-matched controls. The ED group showed a more diffuse connectivity pattern, especially in the right hemisphere, potentially related to incomplete neuronal pruning during development. These structural abnormalities may be associated with inattention and overactivity encountered in children with ED.

A Task Abstraction and Mapping Approach to the Shimming Problem in Scientific Workflows

Recently, there has been an increasing need in scientific workflows to solve the shimming problem, the use of a special kind of adaptors, called shims, to link related but incompatible workflow tasks. However, existing techniques produce scientific workflows that are cluttered with many visible shims, which distract a scientist’s focus on functional components. Moreover, these techniques do not address a new type of shimming problem that occurs due to the incompatibility between the ports of a task and the inputs/outputs of its internal task component. To address these issues, 1) we propose a task template model which encapsulates the composition and mapping of shims and functional task component within a task interface; 2) we design an XML based task specification language, called TSL, to realize the proposed task template model; 3) we propose a service oriented architecture for task management to enable the distributed execution of shims and functional components; and 4) we implement the proposed model, language and architecture and present a case study to validate them. Our technique uniquely addresses both types of shimming problems. To our best knowledge, this is the first shimming technique that makes shims invisible at the workflow level, resulting in scientific workflows that are more elegant and readable.

Intracranial EEG analysis in tumor-related epilepsy: Evidence of distant epileptic abnormalities.

OBJECTIVE In patients with tumor-related epilepsy (TRE), surgery traditionally focuses on tumor resection; but identification and removal of associated epileptogenic zone may improve seizure outcome. Here, we study spatial relationship of tumor and seizure onset and early spread zone (SOSz). We also perform quantitative analysis of interictal epileptiform activities in patients with both TRE and non-lesional epilepsy in order to better understand the electrophysiological basis of epileptogenesis. METHODS Twenty-five patients (11 with TRE and 14 with non-lesional epilepsy) underwent staged surgery using intracranial electrodes. Tumors were outlined on MRI and images were coregistered with post-implantation CT images. For each electrode, distance to the nearest tumor margin was measured. Electrodes were categorized based on distance from tumor and involvement in seizure. Quantitative EEG analysis studying frequency, amplitude, power, duration and slope of interictal spikes was performed. RESULTS At least part of the SOSz was located beyond 1.5 cm from the tumor margin in 10/11 patients. Interictally, spike frequency and power were higher in the SOSz and spikes near tumor were smaller and less sharp. Interestingly, peritumoral electrodes had the highest spike frequencies and sharpest spikes, indicating greatest degree of epileptic synchrony. A complete resection of the SOSz resulted in excellent seizure outcome. CONCLUSIONS Seizure onset and early spread often involves brain areas distant from the tumor. SIGNIFICANCE Utilization of epilepsy surgery approach for TRE may provide better seizure outcome and study of the intracranial EEG may provide insight into pathophysiology of TRE.

Intra-patient supine-prone colon registration in CT colonography using shape spectrum

CT colonography (CTC) is a minimally invasive screening technique for colorectal polyps and colon cancer. Since electronic colon cleansing (ECC) cannot completely remove the presence of pseudo-polyps, most CTC protocols acquire both prone and supine images to improve the visualization of the lumen wall and to reduce false positives. Comparisons between the prone and supine images can be facilitated by computerized registration between the scans. In this paper, we develop a fully automatic method for registering colon surfaces extracted from prone and supine images. The algorithm uses shape spectrum to extract the shape characteristics which are employed as the surface signature to find the correspondent regions between the prone and supine lumen surfaces. Our experimental results demonstrate an accuracy of 12.6 +/- 4.20 mm over 20 datasets. It also shows excellent potential in reducing the false positive when it is used to determine polyps through correspondences between prone and supine images.

Colon Segmentation for Prepless Virtual Colonoscopy

A novel segmentation framework for a prepless virtual colonoscopy (VC) is presented, which reduces the necessity for colon cleansing before the CT scan. The patient is injected rectally with a water-soluble iodinated contrast medium using manual insufflators and a small rectal catheter. Compared to the air-based contrast medium, this technique can better preserve the color lumen and reduce the partial volume effect. However, the contrast medium, together with the fecal materials and air, makes colon wall segmentation challenging. Our solution makes no assumptions about the shape, size, and location of the fecal material in the colon. This generality allows us to label the fecal material accurately and extract the colon wall reliably. The accuracy of our technique has been verified on 60 human subjects. Compared with current VC technologies, our method is shown to be better in terms of both sensitivity and specificity. Further, in our experiments, the accuracy of the technique was comparable to that of optical colonoscopy results.

Evaluation of fiber bundles across subjects through brain mapping and registration of diffusion tensor data.

This paper presents a visualization and analysis framework for evaluating changes in structural organization of fiber bundles in human brain white matter. Statistical analysis of fiber bundle organization is conducted using an anisotropy measure, volume ratio (VR), which is ratio of anisotropic and isotropic components. Initially fiber bundles are tracked using a probabilistic algorithm starting from seed voxels. To ensure accurate selection of seed voxels and to prevent operator bias, a reference brain (MNI_152) is used when marking ROIs. Individual structural MRI brain scans are mapped to the reference using volumetric conformal parameterization. This mapping preserves topology and aligns features perfectly making it a robust and accurate registration technique. One-to-one mapping to the template allows ROI selection and subsequent transfer of ROI to structural MRI of subject. Affine registration coregisters structural MRI and DTI. Seed voxels are mapped to DTI using the resulting transformation parameters. To evaluate the proposed approach, MRI and DTI of 12 normal volunteers and 15 medial temporal lobe epilepsy patients are used. First, a statistical hypothesis testing is conducted to test for anisotropy changes in cingulum and fornix fiber bundles of epileptic patients. Experimental results reveal a 40% decrease in anisotropy levels of cingulum in patients compared to volunteers. They also show a 25% overall decrease in anisotropy of fornix. Secondly, shapes of the bundles are visualized in 3D illustrating that the bundles of epileptic patients are bumpy while those of normal volunteers are smooth.

Coclustering for Cross-Subject Fiber Tract Analysis Through Diffusion Tensor Imaging

One of the fundamental goals of computational neuroscience is the study of anatomical features that reflect the functional organization of the brain. The study of physical associations between neuronal structures and the examination of brain activity in vivo have given rise to the concept of anatomical and functional connectivity, which has been invaluable for our understanding of brain mechanisms and their plasticity during development. However, at present, there is no robust and accurate computational framework for the quantitative assessment of cortical connectivity patterns. In this paper, we present a quantitative analysis and modeling tool that is able to characterize anatomical connectivity patterns based on a newly developed coclustering algorithm, termed the business model-based coclustering algorithm (BCA). We apply BCA to diffusion tensor imaging (DTI) data in order to provide an automated and reproducible assessment of the connectivity patterns between different cortical areas in human brains. BCA not only partitions the cortical mantel into well-defined clusters, but also maximizes the connectivity strength between these clusters. Moreover, BCA is computationally robust and allows both outlier detection as well as parameter-independent determination of the number of clusters. Our coclustering results have showed good performance of BCA in identifying major white matter fiber bundles in human brains and facilitate the detection of abnormal connectivity patterns in patients suffering from various neurological diseases.

Quantitative analysis of diffusion tensor images across subjects using probabilistic tractography

White matter tractography can generate a 3D macroscopic view of the white matter connections in the brain by measuring the diffusion of cellular fluid in tissue and visualize the connections with line strips, tubes, etc. In the past, tractography techniques based on deterministic methodologies have been extensively used. Deterministic tracking follows the primary eigen direction of diffusion that limits its clinical applicability to areas where anisotropy is not clear. More recently, probabilistic tractography methods have been emerging as new tools for extracting fiber tracts. This paper presents a new probability-based framework for quantitative, inter-subject analysis of diffusion tensor imaging data. Advantages of the probabilistic tractography are in its inherent ability to model noise and uncertainty in its global estimation model. This facilitates more accurate tracking in complex neighborhoods, such as branching and crossing fibers. With a combination of statistical measures, our approach can pinpoint abnormalities through the comparison between a set of normal population and patient data. The experiments demonstrate its excellent performance in identifying imaging symptoms of epilepsy and Tourette Syndrome.

Segmentation of brain structures using PDE-driven surface growing

This paper describes an innovative technique for segmentation of the thalamus (a brain structure) from high resolution MRI brain images. Based on a powerful PDE-driven surface growing, this method of segmentation works really well with highly noisy and low-contrast datasets. The presented technique is essentially rooted in the PDE-based surface flow technique. It starts with an initial seed - a real 3D surface model. The growing velocity is computed automatically by the simulation of the PDE flow on the to-be-segmented dataset. Our experiments have demonstrated that it is a very efficient and general segmentation algorithm for many different brain structures.