"The only certainty...," it is said, "is that nothing is certain."
And so it goes with computational forecasts of important events such as weather, finance, and climate. Among all of this uncertainty, however, there are patterns, likelihoods, and rarities that inform important decisions that may affect billions of dollars in resources and thousands, or even millions, of lives. In the hurricane season on the eastern U.S., computational forecasting plays a central role in critical decisions that can determine allocations of emergency resources and the movements of people. The uncertainty and accuracy of these forecasts is an important part in making effective use of these sophisticated tools.
Whether it’s coming up with the best design for a Formula 1 race car or understanding the effects of atrial fibrillation on the heart, developing the right simulation model for research sometimes involves equal parts applied math, engineering and computer science.
University of Utah School of Computing professor Mike Kirby sees himself as the person who connects these disciplines so he can take trailblazing ideas and help create better simulation software to aid researchers.
For many who suffer from debilitating neurological disorders such as Parkinson’s Disease, the constant muscular tremors are an unbearable symptom. Just drinking from a cup can be an overwhelming challenge.
When medication doesn’t work, brain surgery to destroy certain cells can be risky, and the results are irreversible. But there has been an emerging third option — deep brain stimulation (DBS), a therapy in which electrodes are implanted in the patient’s brain that deliver continuous electrical pulses to control motor function.
University of Utah bioengineering associate professor Christopher Butson has been researching ways to improve DBS systems to make them more effective and convenient for patients who wear them. He believes an answer lies in mobile tablets and smartphones.
Most farmers probably never thought they'd be in the market for a way to process huge digital images more quickly -- until, that is, inexpensive drones with high-resolution cameras gave them access to images they could use to micromanage irrigation and to detect the growth of crop-threatening diseases.
Didactic lecture sessions given by the three PIs (Rob MacLeod, Ross Whitaker and Jeff Weiss) as well as three invited instructors (Miriah Meyer, Steve Maas and Gerard Ateshian) experts in their fields
Laboratory exercises lead by a group of teaching assistants and developers,
Discussion session time for student-instructor interaction,
Visit to the experimental and computational laboratory facilities at the University of Utah, College of Engineering to give the participants an overview of the general academic background and research projects performed at the university,
Four Keynotes Lectures from leaders in the field,
Mentoring lectures on grant writing, responsible conduct of research, and simulation study design.
Over the course of four weeks, students from West High School and Skyline High School and a homeschooled junior year student were given an introduction to image-based modeling (IBM) as part of the SCI Institute Summer Internship. This course, which is given to undergraduate students during the academic year, was customized to fit the high school students' level. The goal of this internship was to help the students understand how computational simulation is used in the biomedical field to improve our knowledge of the body, allowing researchers to collaborate with medical doctors to provide patient-specific treatment solutions.
Image Analysis Tools for Understanding Connective Tissue Structure
Sponsored by the Burton Foundation
This summer, two Salt Lake area high school students from Copper Hills High School came to the University of Utah to participate in a hands-on research experience. The students learned how image analysis tools help biomechanics researchers understand the effects of structural features of musculoskeletal tissues (e.g. tendons, ligaments, and articular cartilage) on the functional behavior of these tissues.
Many musculoskeletal tissue injuries and diseases exhibit altered macroscopic and microscopic tissue structure. The Musculoskeletal Research Laboratories, a research center of the Scientific Computing and Imaging Institute, uses engineered tissue materials to study the effect of these structural changes on tissue behavior. Researchers use many image acquisition techniques to characterize the structure of native and engineered tissues, including optical microscopy, x-ray computed tomography (CT), and electron microscopy. Image analysis tools allow efficient detection and quantification of structural features from these images.
On September 27th, Hurricane Joaquin, a Category 4 storm developed over the Atlantic Ocean pounding the Bahamas. There were a number of predications as to which way the storm would travel, one of which was that the hurricane would head north along the east coast of the United States, but the path of the storm changed direction and dissipated on October 7th.
The results denoted possible predicted paths, based upon different models and/or conditions Joaquin might take as of Friday October 2, 2015. Using their Curve Boxplot analysis and visualization method, they show the median hurricane path and the 50 percent band (dark region) — denoting the spatial swath in which 50 percent of the predicted hurricane tracks lie. The light band denotes nearly 100 percent of the possible paths predicted. Red denotes outliers — those hurricane paths flagged as unlikely in reference to all other members of the ensemble.
The two-week course included the following activities:
Didactic lecture sessions given by the three PIs as well as four invited instructors and experts in their fields .
Laboratory exercises led by a group of 10 teaching assistants and developers.
Discussion session time for student-instructor interaction.
A visit to the experimental and computational laboratory facilities at the University of Utah, College of Engineering to give the participants an overview of the general academic background and research projects performed at the university.
Mentoring lectures on grant writing, responsible conduct of research, and simulation study design.
The two-week summer course hosted 39 participants this year: 31 graduate students, 1 MD/PhD student, 2 postdoctoral fellows, 3 junior faculty, and 2 developers from a research laboratory / industry. Participants came from 24 institutions, including 4 from universities in Belgium and England. After the first week of common classes, participants were divided into two tracks: Bioelectricity (10 participants) and biomechanics (29 participants).
IBBM is a dedicated two-week course in the area of image-based modeling and simulation applied to bioelectricity and biomechanics, providing participants with training in the numerical methods, image analysis, visualization, and computational tools necessary to carry out end-to-end, image-based, subject-specific simulations in either bioelectricity or orthopedic biomechanics. The course focuses on using freely available, open-source software developed under the research of the CIBC (P41 GM103545) and FEBio suite (RO1 GM083925). Students use this software to learn and apply the complete dataflow pipeline to particular sets of data with specific goals.
This summer, high school students from the Salt Lake area are coming to the University of Utah to participate in hands-on research in image analysis of the human brain. In conjunction with graduate students and faculty in the School of Computing and the Scientific Computing and Imaging Institute, the students are learning how computer science can help neuroscience researchers understand the brain and disorders that affect it, such as Alzheimer's disease and Autism.
Advances in medical imaging devices, such as magnetic resonance imaging (MRI), have led to our ability to acquire detailed information about the living human brain, including its anatomical structure, function, and connectivity. However, making sense of this complex data is a difficult task, especially in large imaging studies that may include hundreds or even thousands of participants. This is where computer science can play an important role. Image analysis algorithms can automatically quantify properties of the brain, such as the size of brain structures, or the functional activity in different brain regions. This provides neuroscience researchers with insights into how the brain functions and what abnormalities are present in diseased brains.
There has been a recent explosion of interest in the use of noninvasive transcranial brain stimulation (or "neurostimulation"), both in clinical settings and as a research tool. One of the two main ways to stimulate the brain transcranially is to run a current to the brain through the magnetic fields generated by a coil that is held near the scalp. This approach has been approved by the FDA for treating depression. The other main approach uses electrodes placed on the scalp to "inject" current into the head, or apply voltage on the scalp, which is known as transcranial direct current stimulation (tDCS) or transcranial alternating current stimulation (tACS), depending on the type of current or voltage source used. Various neurostimulation technologies have been tested in human experiments for a huge variety of applications including motor rehabilitation, speech therapy, enhancement of cognitive learning, and treating depression and other affective and behavioral disorders, chronic pain syndrome, post-traumatic stress disorder, and others.
Transcranial Magnetic Stimulation (TMS) of the human motor cortex.
SCIRun is the integrated programming environment that has been a core technology of the CIBC since its inception and each major version release is an enormous undertaking. The program now contains hundreds of thousands of lines of C++ code and a new release requires at least a review of all this code, with replacement or updating of larger portions of it. We are nearing the first release of such a major new version, SCIRun 5.
There must be considerable motivation for such a major release, motivation which comes from both our users, collaborators, and DBP partners but also from advances in software engineering and scientific computing, with which we must also keep pace. Our users continue to demand more efficiency, more flexibility in programming the workflows created with SCIRun, more support for big data, and more transparent access to large compute resources when simulations exceed the useful capacity of local resources. The evolution of software engineering has led to changes in computer languages, programming paradigms, visualization hardware and processing, user interface design (and tools to support this critical component), and the third party libraries that form the building blocks of complex scientific software. SCIRun 5 is a response to all these changing conditions and needs and also represents some long awaited refactoring that will provide greater flexibility and freedom as we move into the next generation of scientific computing.
Partial differential equations (PDEs) are ubiquitous in engineering applications. They mathematically model natural phenomena such as heat conduction, diffusion, and shock wave propagation. They also describe many bioelectrical and biomechanical functions and are a central element of the simulation research of the Center. Analytical solutions for most PDEs are known only for certain symmetric domains, such as a circle, square, or sphere. In order to obtain solutions to PDEs for more realistic domains, numerical approximations such as the finite element method (FEM) are used. In the FEM, both the domain and the PDE are discretized and a numerical solution is calculated using computational resources. The discretization of the geometric domain is called a mesh. Meshes play a vital role in the numerical solution of PDEs on a given geometric domain, the accuracy of which depends on parameters such as the shape and size of the mesh elements. The most commonly used meshes contain tetrahedral elements. While simple conceptually, mesh generation is one of the most computationally intensive tasks in solving a PDE numerically.
Introducing ViSOAR. As data acquisition advances, and data sizes increase, the need for tools to process and visualize the results in an effective and efficient manner is becoming increasingly important. The reliance on supercomputers for scientific visualization and analysis is already proving to be a hindrance for wide accessibility to researchers and scientists dealing with large data.
In collaboration with Dr. Don Tucker and his colleagues at Electrical Geodesics Inc (EGI) and the University of Oregon, this DBP is concerned with improving our ability to reconstruct and visualize neuroelectric sources (source localization) from EEG measurements and also our ability to stimulate specific brain regions using electrodes attached on to the scalp of the subject (transcranial direct current stimulation, tDCS).
For both research and clinical practice, EEG is a cost-effective tool to understand and excite brain activity. EEG advances have significantly improved the spatial resolution of source estimates and offer the promise of precise spatio-temporal monitoring and stimulation of cortical brain activity. By itself, high-resolution EEG would be affordable even for small hospitals in remote locations and could be easily managed by technicians in the field.
Cam Femoroacetabular Impingement Analysis using Statistical Shape Modeling
Femoroacetabular impingement (FAI) is caused by reduced clearance between the femoral head and acetabulum due to anatomic abnormalities of the femur (cam FAI), acetabulum (pincer FAI), or both (mixed FAI). Cam FAI is characterized by an aspherical femoral head or reduced femoral head-neck offset. During hip flexion, the abnormally shaped femur may cause shearing at the chondrolabral junction, thereby damaging articular cartilage and the acetabular labrum. Currently, diagnosis of cam FAI is largely accomplished using two-dimensional (2D) measurements of femur morphology acquired from radiographic projections or a series of radial planes from computed tomography (CT) or magnetic resonance (MR) images. Two- dimensional measures provide initial diagnosis of cam FAI, but their reliability has been debated. Also, there is no agreement on the range of measurements that should be considered normal. Furthermore, radiographic measures give only a limited description of femur anatomy or shape variation among cam FAI deformities. Together, these limitations of 2D measurements translate into a high misdiagnosis rate. In a series of FAI patients treated with surgery in our clinic, 40% had seen multiple previous musculoskeletal providers and 15% had undergone surgical procedures unrelated to the hip joint (hernia, etc.).
Mean control (left) and cam (right) shapes. Middle images show the mean control shape with color plots depicting how the mean cam shape differed across the femoral head, neck, and proximal shaft. Top and bottom rows show different rotations of the femoral head.
Volumetric CT images from a cam FAI patient. Validated threshold settings were applied to CT images to segment and reconstruct the bony morphology of each femur.
MRI Image Quantification Analysis for Atrial Fibrillation
Atrial fibrillation (AF) is a cardiac rhythm disturbance in which the atria, the upper chambers of the heart, undergo uncontrolled and uncoordinated electrical activation so that contraction of the atria contributes almost nothing to cardiac output. While not immediately atal (as is ventricular fibrillation) AF dramatically increases the risk of stroke, elevates mortality, and diminishes quality of life. Traditional diagnosis of AF as been limited to ECG-based determination of the time spent in this arrhythmia and there has previously been no other dependable biomarker capable of determining either the progression of the disease or of determining suitable treatment approaches. Therapy for AF consists of either antiarrhythmic drugs that may control the arrhythmia completely or at least reduce the resulting elevated heart rate combined with anticoagulation therapy or ablation. Ablation involves destroying targeted regions of the atria with the goal of either isolated triggers of spurious electrical activity or functionally separating the atrial wall into small enough segments that the putative mechanism of the arrhythmia cannot longer sustain. The latter approach is a form of substrate stabilization, and the management of this disease has suffered from a persistent lack of means to monitor or evaluate the stability of the tissue. It is precisely in this aspect that cardiologist at the University of Utah, with support from the CIBC have made their most significant contributions.
An interdisciplinary team at the Comprehensive Arrhythmia and MAnagement (CARMA) Center have made use of the segmentation, image analysis, and recently mesh generation and simulation capabilities of the CIBC to create a comprehensive program for AF management. The scope of the progress continues to expand each year and this application of CIBC technology has proven very fruitful even as it is very challenging.
Overview of the DBS system. The DBS electrode is implanted in the brain during stereotactic surgery. The electrode is attached via an extension wire to the IPG, which is implanted in the torso. The entire system is subcutaneous and is designed to deliver continuous stimulation for several years at a time.
In recent years, there has been significant growth in the use of patient-specific models to predict the effects of neuromodulation therapies, such as deep brain stimulation (DBS). However, translating these models from a research environment to the everyday clinical workflow is a challenge, primarily due to the complexity of the models and the expertise required in specialized visualization software. Recently, the CIBC has worked with Dr. Christopher Butson at the University of Wisconsin to deploy the interactive visualization system ImageVis3D Mobile for experimental use in the area of DBS planning. In addition to running on multi-node compute clusters and large desktop systems, ImageVis3D is also designed for mobile computing devices such as the iPhone or iPad. In this case, ImageVis3D was modified for an evaluation environment in order to visualize models of Parkinson's Disease (PD) patients who received DBS therapy1.
The selection of DBS settings is a significant clinical challenge that requires repeated revisions to achieve optimal therapeutic response, and is often performed without any visual representation of the stimulation system in the patient. We used ImageVis3D Mobile to provide models to movement disorders clinicians and asked them to use the software to determine: 1) which of the four DBS electrode contacts they would select for therapy and 2) what stimulation settings they would choose. We compared the stimulation protocol chosen from the software versus the stimulation protocol that was chosen via clinical practice (independent of the study). Lastly, we compared the amount of time required to reach these settings using the software versus the time required through standard practice. We found that the stimulation settings chosen using ImageVis3D Mobile were similar to those used in standard care, but were selected in drastically less time. We found that our visualization system, available directly at the point of care on a device familiar to the clinician, can be used to guide clinical decision-making for selecting DBS settings. The positive impact of the system could also translate to areas other than DBS.
A cross-section of a 3-dimensional, tetrahedral mesh of a torso. Each separate organ type is shown using a different color.
This year, an essential goal has been to enhance the generalized image-processing pipeline of software developed by CIBC and its partners. With the growing use of high quality medical imaging, practitioners around the globe are employing these acquired datasets for performing biomedical simulation. In its holistic approach to image-to-simulation pipelines, our software starts with image data and processing, constructs geometric models, performs simulation, and provides biophysical analysis of the data. A research highlight for the CIBC this year is the development of an improved scheme for mesh generation, a critical step within this pipeline. This research complements the software package BioMesh3D, but targets a completely different niche in the world of mesh generation algorithms.
High Quality Meshing
The problem of mesh generation has been widely studied, as a hybrid field of interest to the scientific, engineering, and computer science communities. In each of these fields, meshes are used to compute numerical approximations to solutions of partial differential equations. To do so, continuous mathematics are replaced with a discrete analogue, most commonly to facilitate the finite element method (FEM).
The FEM works by decomposing a domain of interest into discrete entities of various dimensions, such as points (0-dimensional), edges (1-dimensional), and cells of higher dimension (frequently triangles and quadrilaterals are used for 2-dimensionl elements, tetrahedra and hexahedra for 3-dimensional). Together, these elements form what is commonly called a mesh (see figure)Solutions to the complex system are solved piecewise on each element, and then aggregated together to form the final solution. The FEM has become an important tool in medical imaging as well. For example, CT scans of legs can be meshed so that orthopedic modeling can accurately simulate gait, MRI scans of the torso are frequently used in cardiac electrophysical modeling, and images of the skull can identify structures of the brain.
Because the FEM is a computational tool that processes individual elements to approximate a whole solution, it is deeply impacted by the mesh elements used to represent the space. Two principle concerns stand out in the meshing problem for medical images: