R1, Subsurface Sensing and Modeling: · Neural Control of the Zebrafish Locomotive Repertoire Professor Donald O’Malley, Northeastern University This project will engage students in investigating the neural control of a sophisticated repertoire of locomotive behaviors exhibited by a model genetic organism, the larval zebrafish. Neurons in the brainstem and spinal cord are responsible for generating a range of behaviors including swimming, escaping, prey-tracking and prey capture. Because larval zebrafish are transparent, our group is able to employ a wide variety of optical techniques in our efforts to reveal the organization of the neural circuitry underlies these behaviors. In particular, we use calcium imaging to record neural activity patterns, and then use laser-ablation to optically dissect these control systems. Together with neuroanatomical, transgenic and computational modeling techniques, we are piecing together the neural components of the larva's Locomotor Control System. Our ultimate goal is to discover organizational principles that apply to all vertebrate animals. Such findings are likely to aid more clinically oriented researchers who are seeking to repair the damage suffered in human spinal cord injury. · Simulation of Electromagnetic Fields and Waves Professor Carey Rappaport, Northeastern University Student researchers will use computers to simulate EM fields and waves in realistic situations, including: detecting land mines and buried environmental hazards, examining human microwave exposure for cancer treatment and organ thermal therapy, and designing novel special-purpose antennas. Research projects involve using well-established modeling and visualization software, and writing new codes in MATLAB, Mathematica, FORTRAN, and C to pre-process, post-process and accelerate algorithms.

R2, Physics-Based Signal Processing and Image Understanding: · Hyperspectral Image Analysis at RPI Prof. Badri Roysam, Rensselaer Polytechnic Institute This project involves the development of algorithms for comparing spectra, and applying these algorithms to detect and understand changes in hyperspectral imagery. Applications for this type of work range from skin cancer detection, and microscopy to interpretation of satellite imagery. Students need to be able to use a spreadsheet like Excel, and program in MATLAB. Finally, the algorithms need to be validated. Students would construct test objects that undergo controlled changes that can be detected and verified by hyperspectral image analysis. · Development of Signal Processing and Imaging Algorithms for Biomedical and Nondestructive Testing Applications Professor Edwin Marengo, Northeastern University A major ongoing research area within the remote sensing disciplines is the search for non-iterative, and superresolution algorithms for inverse problems and imaging. This research area is expected to have immense impact in several applications, in the biomedical, geophysical, nondestructive testing and wireless communication fields, particularly in connection with algorithms that include multiple scattering. The inversion problem in question is nonlinear, and thus the algorithms in question also teach ways to cope with the nonlinear inverse problem. The non-iterative aspect enables real time, online application as required, for example, in ultrasound imaging of the breast, while the superresolution aspect corresponds to the extraction of more information or features about the subsurface objects being sensed than that associated to conventional technologies available in the market. This higher extraction of information carries with itself enhanced detection and estimation capabilities which can improve cancer diagnosis, among other applications. Our research group at the Gordon Center for Subsurface Sensing and Imaging Systems is working toward developing state-of-the-art non-iterative and superresolution imaging algorithms including multiple scattering effects, for remote sensing systems using acoustic or electromagnetic waves, and we are also interested in the development of an information theory of fields capable of rigorously characterizing the fundamental limitations in detection and estimation capability of a given remote sensing system. Research projects for undergraduates will focus on the development and testing of novel signal processing and imaging algorithms for biomedical and nondestructive testing applications, including the development of computer codes and the testing of the codes with synthetic or experimental data available to our group.

R3, Image and Data Information Management: · Semi-automatic Parallelization of Serial MATLAB Applications Prof. David Kaeli, Northeastern University In this project, students will learn how to develop parallel implementations of subsurface sensing and imaging applications utilizing semi-automatic parallelization tools. The programming environment is MATLAB. Students will learn how to utilize the Star-P parallel programming environment and will use this framework to parallelize serial MATLAB applications. Familiarity with MATLAB is recommended (though is not required) for this project. · Accelerating Imaging Codes Using GPUs Prof. David Kaeli, Northeastern University In this project, students will apply recent advances in graphics processing hardware (i.e., GPUs) to break computational barriers in medical imaging and environmental sensing problems. The platforms provide hundreds of processing cores that can be programmed in high level languages (e.g., C++). The student will learn how to develop parallel programs on NVIDIA, AMD, Intel and IBM GPU platforms. Programming will be done in C and C++, and will use threading and exploit data-level parallelism.

BioBED, validation testbed for solution-testing in biological microscopy: · Compact Confocal Microscope Scanner Professor Charles DiMarzio, Northeastern University Confocal microscopy, originally developed in the late 1950's, became practical with the advent of the laser, and it is now an established technique of microscopy for many applications. In particular, reflectance confocal microscopy has been developed for imaging skin cancers, and has the potential to guide surgery. Confocal microscopes remain large, difficult to use for imaging in vivo, and somewhat costly, and several efforts are under way to make them more suitable for clinical practice. The key components on which the success of confocal microscopy is based are those for scanning a focused light source to generate an image. These components are also the most significant contributors to the cost, size, and physical configuration of the instruments. This project has built a breadboard instrument based on rotating wedges. It generates a different scan pattern than other confocal scanners, but holds the promise of a very compact package. Ultimately, it could be reduced to a hand--held instrument, perhaps looking something like a dentist's drill. The REU student, directed by a graduate student, will be responsible for operating the breadboard, collecting and processing images, and attempting to answer the question of whether this type of scan pattern can produce images of sufficient quality to be useful in diagnosis of skin cancers in vivo. The student should bring to the project a strong interest in hands-on work with mechanical and electronic hardware, and computer interfaces. Experience in any of these areas is a plus. The student will have the opportunity to learn about medical optics, system integration, and scientific computer programming, and to work with engineers and scientists involved in many aspects of biomedical optics. · Correlation of Surface and Subsurface Features for Image Guided Therapy George T.Y. Chen, Ph.D. The primary goal of image guided radiotherapy is to accurately irradiate a target while sparing adjacent normal organs. Accuracy requires understanding issues related to patient setup, monitoring of patient position, dealing with internal physiologic motion, and identifying appropriate methods to track tumors. Technical approaches associated with this objective include image processing, 4D visualization, machine vision, and computer simulation. The students involved will assist in a focused aspect of data acquisition and analysis related to image guided radiotherapy, under supervision of Drs. Chen and Gierga. The Center's resources and expertise in video, image processing, and analysis will be utilized in conjunction with clinical physics expertise from MGH to provide the undergraduate with an experience that bridges clinical medicine with research on a real world problem. Student Skills: Strong PC skills, familiarity with basics of digital imaging, image processing, computer programming and physics would be helpful. Familiarity with Matlab and IDL would be helpful.

SoilBED, validation testbed for solution-testing in the underground: · 2D Sensing and Imaging of Underground Contamination Prof. Ingrid Padilla, University of Puerto Rico, Mayagüez Student researchers will work on the completion and testing of a 2D SoilBed designed to develop and validate computational tools for detection and imaging of underground contamination using cross well radar technology. The 2D SoilBed facility is at the Environmental Engineering Laboratory at the University of Puerto Rico, Mayagüez. The student will work with optimizing the antenna array in the SoilBed and with validation of results using image processing techniques.. Tasks would include placement and testing of antennas within flow and electromagnetic boundary conditions in a 2D SoilBed;acquiring and processing images and data collection and analysis.

Other Projects · Development of a Photonic Chip-Scale BioCalorimeter Professor Gregory J. Kowalski, Ph.D. In this project a student will be involved in developing a chip scale calorimeter based on extraordinary optical transmission (EOT) through an array of nanometric apertures. Students will gain experience in thermodynamic, thermal management, microfluidic, nanofabrication processes and bioinstrumentation design. Calorimetry is used in the study of binding interactions and is central to basic biology research and pharmaceutical R&D. Calorimetry provides detailed information on the nature of binding reactions by measuring the energy released/absorbed by a reaction over a range of reactant concentrations, and uses these data to determine the thermodynamic properties, stoichiometry of the reaction, and the binding constant for the reactants thereby providing valuable insight into reactions. Nanohole array devices have been used as affinity sensors where one of the binding partners is immobilized on the surface of the device. With these nanohole array sensors the signal is temperature dependent due to the temperature dependence of the dielectric function of the material in contact with the nanohole array surface. As the temperature changes, the plasmon excitation conditions change and the EOT signal is altered. For the nanohole array calorimeter the dielectric is an approximately 100nm thick layer of polycarbonate directly above the nanohole array surface that enables the use of EOT as a fast and sensitive temperature sensor to measure the heat of reaction (enthalpy, ?H) from a reaction. The technology can be easily multiplexed, placing many nanohole array sensors on a single chip, enabling the simultaneous measurement of controls to characterize confounding effects to determine the true heat of reaction and other thermodynamic binding properties. This multiplexing also indicates the possibility of using this for high throughput of molecular libraries. Early results indicate that a nanohole array calorimetry system has the potential to reduce the amount of protein required by 1000-fold and increase sensitivity by 100-fold. This will expand the use of calorimetry in biochemistry research and pharmaceutical R&D. Our research plan consists of three goals resulting in a prototype and performance assessment against five quantitative milestones. Goals 1 (thermal issues and EOT response) and 2 (sample delivery and mixing) explore the fundamental design options and tradeoffs involved in nanohole array device design and sample delivery. A prototype instrument is designed and evaluated in Goal 3. At the present we have developed a breadboard calorimeter and performed proof of concept experiments that we can make calorimeter measurements. During the next year we will be focusing on designing and implementing effective microfluidic sample delivery systems and test cell manufacturing techniques to perform quantitative measurements on biological materials. The test cell design needs to be compatible with a prototype device that has been developed on related project.