Washington University Medical Center
Washington University St. Louis: Open ScholarshipNot a member yet
26344 research outputs found
Sort by
Biology Department Publications: 2007
Get PDFThis document is a noncomprehensive list of publications made by graduate students, post-doctoral researchers, and faculty members in the Biology Department in 2007. It indexes those publications available through major databases, and serves as an archival record of the Department’s robust scholarly contributions
Optimization-based Motion Planning with Homotopy and Homology Class Constraints
Get PDFMotion Planning is a fundamental problem in robotics that aims to find an optimal trajectory for a system to move on while avoiding obstacles in the environment. Often, a feasible trajectory connecting the start and target point with the shortest length is highly desirable. Additionally, in scenarios such as drone racing or surveillance, topology constraints may arise. At the low level, the LQR or PID controller is utilized to steer the agent to move along the designed trajectory. At a high level, optimization-based, search-based, or sample-based algorithms are utilized to synthesize the feasible trajectory. In this thesis, we deal with optimization-based trajectory synthesizing along with a special type of topology constraint named homotopy and homology class constraints. In the first part of the thesis, we just ignore topology constraints and emphasize how to transform motion planning tasks into optimization tasks while considering start-point end-point constraints and minimal energy or minimal time loss function. Although the loss function is a differential function, the first-order gradient optimization method, such as Adam, shows less ability to find the optimal. However, methods that utilize second-order information, such as the Gaussian-Newton method and the interior point optimizer, can solve the problem quickly and perfectly. The second part of the thesis emphasizes our proposed optimization method for motion planning with homotopy and homology class constraints. We first introduce the Auxiliary Energy Reduction Technique. The hallmark of our approach is that we first introduce virtual control terms to the original system dynamics that ensure that any preset state trajectory is dynamically feasible with respect to the new extended system. We then gradually shift the contribution of the artificial inputs to the actual original inputs, and in the end, the trajectory will be deformed to the one of the same homotopy class that is now also feasible with respect to the original system. However, the aforementioned method suffers from low efficiency when the required homotopy class is complex. Therefore, in the second method, we deal with two-dimensional obstacles by synthesizing auxiliary trajectories for obstacles then synthesizing optimal trajectories for the agent, and then gradually deforming obstacle trajectories to the original ones and keeping the agent’s trajectory optimal, which improves efficiency. To explore the homology class constraints, in the third method, we solve homology class constraints with respect to three-dimensional obstacles by embedding them in two- dimensional. In the fourth method, we combine our method of the third method to extend the second method to deal with special 3-dimensional obstacles with homotopy class constraints
Efficient Data-Driven Machine Vision: A Co-Design of Circuit, Algorithm, and Architecture for Edge Vision Sensors
No full textCMOS Image Sensors (CIS) are the most widely used visual devices in the modern sensing industry, playing a crucial role in converting light signals to digital images. In practical applications, CIS serve as the front end of a vision system, which also includes Image Signal Processors (ISPs) and vision algorithms before delivering the final output to end consumers. The quality of a vision system depends on the metrics used by the end consumer to evaluate the received images. In conventional vision systems, the end consumer is human, and the evaluation metric is image quality, which is based on visual fidelity and measured by factors such as peak signal-to-noise ratio and structural similarity index measure. To achieve higher visual-fidelity-based image quality, these systems allocate a significant portion of their hardware budget to high-resolution analog-to-digital conversions and advanced ISPs, resulting in large data transmission between CIS, ISPs, and end consumers, causing energy and latency bottlenecks in the vision system. Critically, these bottlenecks worsen as image resolution increases, which is a consequence of deploying large Computer Vision (CV) models in real-time scenarios as required by today’s edge-Artificial Intelligence (AI) applications. Therefore, the conventional human-centered vision system is unable to efficiently support edge-AI applications. This dissertation explores a new energy- and latency-efficient vision system based on the observation that in edge-AI applications, voluminous vision data are generated by intelligent edge CIS and consumed, not by humans, but by downstream CV algorithms to perform sophisticated tasks such as classification, recognition, and machine perception. The evaluation metric for the quality of the vision system thus becomes the accuracy of downstream tasks rather than visual-fidelity-based image quality. This observation motivates us to discard the concept of reconstructing high-fidelity images, but instead to compress raw images and preserve “task-specific” information to achieve energy and latency reduction without degrading downstream task accuracy. Such a vision system is termed a data-driven machine vision system and is constructed with high efficiency through a co-design paradigm of circuit, algorithm, and architecture. This dissertation analyzes the data-driven machine vision system by targeting different CV algorithms that extract the “task-specific” information and implementing these algorithms in CIS with circuit and architecture techniques to improve system efficiency. First, a vision system with a classic CV algorithm implemented in the sensor is presented, achieving 7.1× compression for the pedestrian detection task without accuracy loss. Next, two vision systems with deep learning-based CV algorithms implemented in the sensor are presented, achieving up to 8× compression for the image classification task and up to 20× compression for the eye-tracking task, respectively, with minimal task accuracy loss. Finally, frameworks for exploring computational CIS architectures and facilitating conventional CIS design space exploration are presented, forming the basis for systematically, rather than heuristically, choosing the optimal circuit and architecture for a given algorithm
Machine Learning Strategies for Enhancing Reconstruction and Classification in Diffuse Optical Tomography and Photoacoustic Tomography for Ovarian and Breast Cancer
Get PDFThis dissertation investigates the application of advanced deep learning techniques to enhance biomedical imaging technologies, specifically focusing on Diffuse Optical Tomography (DOT) and Photoacoustic Tomography (PAT) for the diagnosis and treatment of ovarian and breast cancers. By integrating convolutional neural networks (CNNs) and other machine learning architectures, the research addresses critical challenges in image reconstruction, 3D rendering, and classification tasks within medical imaging domains. For DOT, a novel machine learning model with physical constraints (ML-PC) was developed, significantly improving the quality and accuracy of image reconstructions from diffuse optical signals. This model effectively managed the inherently noisy and indirect measurements typical in DOT, providing clearer, higher-resolution images that enhance the detection and monitoring of medical conditions such as breast cancer. In the study of PAT, an ultrasound-enhanced Unet model was introduced, leveraging ultrasound features to enhance the reconstruction of optical absorption distributions in ovarian lesions. This model achieved high accuracy and diagnostic performance, effectively differentiating between malignant and benign ovarian lesions. The use of quantitative PAT, combined with the ultrasound-enhanced Unet model, provided a powerful tool for non-invasive imaging, offering significant improvements in detecting and characterizing ovarian cancer compared to traditional imaging methods. The PA-NeRF model addressed the limited-view problem in PAT, which is crucial for providing high-quality 3D reconstructions necessary for ovarian cancer diagnosis. By utilizing neural radiance fields (NeRF) for 3D PAT reconstruction from limited B-scan data, the model produced detailed and accurate 3D images from fewer scans. This advancement is particularly important for clinical applications where limited data acquisition is common, enabling better visualization and assessment of ovarian lesions and improving diagnostic capabilities. The USDOT-Transformer model was developed to predict pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) in breast cancer, integrating multimodal imaging data for a robust assessment of treatment response. This model combined ultrasound and diffuse optical tomography data with transformer-based deep learning, providing valuable insights into the effectiveness of NAC and demonstrating significant potential for personalized treatment planning. These advancements highlight the potential of combining machine learning with optical imaging modalities to improve diagnostic accuracy and treatment planning in cancer care. The integration of deep learning techniques has shown to be particularly beneficial in overcoming challenges associated with traditional imaging methods, providing more reliable and detailed insights into tissue properties and treatment responses. This dissertation illustrates the transformative impact of deep learning on biomedical imaging, setting the stage for future innovations that could further revolutionize this critical area of healthcare technology
Adaptation and Evaluation of Guideline-Based Family-Based Behavioral Treatment for Overweight and Obesity in Childhood Survivors of Acute Lymphoblastic Leukemia and Lymphoma
Get PDFChildren who survive acute lymphoblastic leukemia (ALL) and lymphoma are at considerable risk for chronic, treatment-related medical complications known as “late effects.” Not only are these late effects significant in and of themselves, but many are exacerbated by excess weight, a particularly concerning combination given the elevated rates of overweight and obesity within this population. Although the therapeutic regimens (e.g., steroid “pulses”) endured by children who survive ALL and lymphoma may contribute to excess weight, accumulating evidence also demonstrates the development of poor dietary habits and inadequate levels of physical activity during cancer treatment. Such weight-related health behaviors appear to contribute considerably to the high rates of overweight and obesity among these children, even long after cancer treatment has ended. The current study involved the adaptation and implementation of the first intensive (i.e. ≥ 26 contact hour), multicomponent, evidence-based, healthy lifestyle and weight management intervention for children who survive cancer and their families. Focus groups were conducted with 9 children who survived ALL and 11 of their parents to assess perceptions of weight, weight-related behaviors, and perceived barriers to family-based behavioral treatment (FBT) for overweight and obesity. The Framework for Reporting Adaptations and Modifications-Enhanced (FRAME), a theoretical framework to systematically guide the process of adapting of an evidence-based program, was then employed to inform adaptations to FBT. Finally, 12 children who survived ALL or lymphoma and their families completed a single-arm, non-randomized trial of the adapted intervention. Families completed height and weight measurements, 24-hour dietary recalls for the participating child, and additional measures of acceptability, feasibility, and weight-related health behaviors at baseline (BL), end of treatment (EoT), and follow up (FU; 6 months after baseline). Enrollment rates were only slightly lower (26%) than those demonstrated within previous trials of FBT and an analysis of possible predictors of intervention participation demonstrated that participants who were off cancer treatment for a shorter amount of time were more likely to enroll (Effect size: g = 0.706). Moreover, the most common reasons for non-participation were busy schedules (n = 11, 48%) and logistical considerations (n = 3, 13%), including the time commitment required by the study as well as its format (i.e., group format). High quantitative self-report ratings, qualitative feedback, and rates of attendance and retention indicated that the intervention was both acceptable and feasible for participating families. While the sample size was small, reductions in measures of relative weight for children across the intervention were particularly encouraging, with a significant linear reduction of PoM BMI across FBT (β = -1.03, 95% CI [-1.66, -0.40], t = 3.25; Effect size: g BL to EoT (0 - 4 months) = -0.617; g BL to FU (0 - 6 months) g = -0.630). Moreover, median change in child relative weight was 13.3 units, exceeding the previously established range of clinically meaningful reduction reliably achieved through past trials FBT. Reductions in guardian weight within the sample were also notable, demonstrating a reduction of 25 pounds, a level of change which exceeds the 10% and/or roughly 22-pound weight loss associated with maximal reduction of risk for most weight-related chronic illnesses. Curiously, changes in measures weight-related health behaviors, psychosocial dysfunction, and health-related quality of life across the intervention were otherwise non-significant and therefore warrant further investigation. The study highlights the potential of an intensive, multicomponent, evidence-based weight management intervention for children who survive cancer and their families as well as considerations for future intervention iterations involving a fully powered trial with a comparison group
El Trabajador Audible: Trabajo, Cultura Sonora y Modernidad en Argentina, Brasil y Uruguay (1920-1970)
No full textThe dissertation examines the politics of sound in the representation of labor and workers in twentieth-century South America, particularly Argentina, Brazil, and Uruguay. It delves into a sonic archive to study a period I termed the “Age of Labor,” a time between 1920 and 1970 when political organizations, governments, unions, intellectuals, and artists appealed to sound to portray workers as the key figures of twentieth-century modernity. I approach labor as a complex and multilayered category, central to State-sponsored governmentality policies that aimed to forge the citizen as a dominantly male, white, and industrial worker, but also as a category that allowed oppressed groups —workers, women, racialized minorities—to exert ways of resistance and agency. By analyzing an eclectic and trans-medial corpus of newspapers, songbooks, novels, radio broadcasts, and films, the dissertation underscores the pivotal role of sound culture in the dispute around labor in twentieth-century Argentina, Brazil, and Uruguay. It argues that listening to, capturing, and staging the voices of workers, as well as the soundscapes of labor, became central in the debates about capitalist modernization. The dissertation considers a vast array of sonic objects, listening practices, and ideas about the value/meaning of sound and the voice, and analyzes different print, sound, and audiovisual media
Optimal Beamforming Design for Stability and Performance of UAV Formations
No full textThis paper explores the application of multi-loop Wireless Networked Control Systems (WNCS) for managing Unmanned Aerial Vehicle (UAV) formations. A multi-antenna base station (BS) employs beamforming (BF) technology to communicate with and control multiple UAVs simultaneously. Given the inherent uncertainty and stochastic nature of wireless channels and control system dynamics, we propose a joint communication-control strategy to ensure an accurate, stable, and efficient system. Our objective is to minimize the weighted sum of state error and communication costs while maintaining system stability and adhering to transmission power constraints by adapting BF weight vectors. The benefits of the BF design are twofold: first, a good trade-off between communication costs and control stability is achieved due to the array gain from the multi-antenna controller. Second, the generation of multiple beams to control multiple UAVs simultaneously effectively resolves scheduling issues, enhancing the efficiency of limited radio spectrum and energy consumption. To address the resulting non-trivial probabilistic optimization problem, we adopt the Bernstein outage probability theorem to construct a semidefinite program relaxation. Simulation results confirm the effectiveness of the proposed joint strategy
Optimizing the Synthesis and Characterization of Garnet-Type Thin Film Ceramic Electrolytes for Solid-State Lithium Batteries
Get PDFThis study focuses on the synthesis and characterization of thin-film ceramic electrolytes, specifically Lithium Lanthanum Zirconium Oxide (LLZO) and Lithium Lanthanum Zirconium Tantalum Oxide (LLZTO), for application in solid-state lithium batteries (SSBs). These materials, known for their high ionic conductivity and chemical stability, are critical for enhancing the performance and safety of next-generation energy storage systems. However, stabilizing the high-conductivity cubic phase at room temperature remains a challenge.
To address this, innovative synthesis techniques, including electrospinning and tape casting, were employed to produce nanostructured fibers and films. Various experimental parameters, such as precursor composition, milling conditions, and environmental controls, were optimized to achieve desirable morphological and structural properties. Characterization of the synthesized materials using X-ray diffraction (XRD), scanning electron microscopy (SEM), and electrochemical impedance spectroscopy (EIS) revealed significant improvements in phase stability, mechanical flexibility, and ionic conductivity.
The results demonstrate the effectiveness of developing efficient method of producing solid electrolyte ceramic. This research not only advances the understanding of garnet-type electrolytes but also provides insights into scalable manufacturing techniques for high-performance solid-state lithium batteries
Investigating the Role of the I-II Linker in Nav1.5
Get PDFThe cardiac voltage-gated sodium channel, Nav1.5 is essential for initiating the cardiac action potential. Its dysfunction can lead to dangerous arrhythmias, sudden cardiac arrest, and death. Like many channels, its known properties depend heavily on its structure. The functional core of the alpha subunit consists of four homologous repeats (I, II, III, and IV), each formed from a voltage sensing domain and a pore domain. The channel also contains two cytoplasmic termini and three cytoplasmic linkers (I-II, II-III, and III-IV). While most of the channel has been resolved in multiple structures, the I-II and II-III linkers have remained conspicuously absent and are predicted to be disordered. Though some information regarding their roles in regulation and various interactions is available, the definitive role of these linkers is not well understood. We investigated the basic functional role of the I-II linker by starting with a detailed analysis of its sequence. We separated it into eight regions ranging in size from 32 to 52 residues, chosen based on their distinct properties. Since these regions had such distinct sequence properties, we hypothesized that they may have distinct effects on channel function. We tested this with experiments where we created individual Nav1.5 constructs with each region deleted, expressed them in Xenopus oocytes, and recorded the current with a COVG voltage-clamp. Interestingly, these deletions had very little effect on channel gating, though at least two (430 – 457del and 556 – 607del) resulted in a significant reduction in peak current. Disordered proteins and regions are typically less conserved overall than ordered proteins in terms of primary sequence conservation. In fact, there is some evidence that molecular features, such as residue content and net charge, may be more important for functional conservation than the exact sequence. We performed a multi-species sequence analysis and found that the different regions of the I-II linker had varying degrees of conservation. We also found that certain regions had noticeably more varied proline fractions that were higher on average for mammals than for non-mammals. Further, we identified charged residues and several interaction sites as exhibiting high levels of conservation. The pathogenicity of variants within the I-II linker is difficult to evaluate, so we attempted to correlate in-silico conformational changes for channel variants to in-vitro changes in channel function. We used a Metropolis Monte Carlo scheme to simulate portions of the I-II linker containing the variants. First, we investigated four clusters of phosphorylation sites in the I-II linker: S457-460, S483-486, S497-499, and S664-671. Comparing simulations of the wild-type sequence with simulations of an increasing number of phosphomimetic mutations, we found that only segments containing the S497-499 and S664-671 clusters showed conformational sensitivity to the mutations, while those containing S457-460 and S483-486 were unaffected. In Lorenzini et al. 2021, it was found that only mutations to S664-671 affected channel function when expressed in HEK293 cells. This experimental data is consistent with the simulations for S457-460, S483-486, and S664-671 but not for the S497-499. Additionally, we investigated five prolines (P627, P628, P637, P640, P648) that had been revealed in a multi-species sequence analysis to be conserved in mammals but notably absent from the Xenopus sequence. Simulations suggested that the presence of these prolines could have an expansionary effect on the conformational space occupied by the protein. We created mutant channels, where we replaced all or some of these prolines with their Xenopus counterparts. Ionic currents from mutant channels were recorded with the COVG voltage-clamp in Xenopus laevis oocytes. The only mutation that had a significant effect on channel gating was P627S, which depolarized channel activation (10.13 +/- 2.28 mV depolarizing activation V50 shift). Neither a phosphosilent (P627A) nor a phosphomimetic (P627E) mutation had a statistically significant effect, suggesting that either phosphorylation or another specific property of serine may be required to observe the effect. Since the deletion of large regions from the linker had little effect on channel gating while single interaction-associated mutations had conspicuous impacts, the I-II linker’s fundamental role may be to act as an interface for interactions with other proteins. If this is the case, without the specific interacting proteins present, removing the regions would likely have little effect on channel function, similar to what we observed in our deletion recordings in oocytes. Variants may have a larger impact if they create or disrupt these interactions. Determining whether they do so may be key in evaluating the likely pathogenicity of variants
Dual-Peptide Functionalized Hydrogel for Enhancing Biosynthesis and Immunomodulation in Primary and Stem Cells for Therapeutic Applications
No full textIntervertebral disc (IVD) degeneration significantly contributes to disability and the economic burden of musculoskeletal diseases. Early changes in IVD degeneration include reduced cellularity and altered phenotype in nucleus pulposus (NP) cells, leading to diminished capacity for extracellular matrix (ECM) and glycosaminoglycan production. Cell supplementation, particularly with bone marrow-derived human mesenchymal stem cells (MSCs), has been explored to promote ECM synthesis and repair. MSCs offer potential for either direct cell replacement or by secreting factors to support disc repair by restoring biosynthetic activity of remaining endogenous NP cells and ameliorating inflammation. Biomaterials have been extensively studied to preserve and restore the healthy phenotype of IVD cells and to serve as cell carriers to protect and localize transplanted cells from the harsh environment of degenerative IVDs. Previous research demonstrated that integrin or syndecan binding peptides from laminin can induce degenerative primary human NP cells to re-express a juvenile, healthy phenotype. Prior work has also suggested that hydrogels presenting cell adhesive peptides can enhance not just primary cell survival, but MSC survival and pro-reparative functions. Additionally, growth factors such as Insulin-like growth factor 1 (IGF-1), play a multifaceted role in stem cell biology and may promote proliferation, survival, migration, and immunomodulation for MSCs. Biomaterial cell carriers can be functionalized with not just cell adhesive peptides, but peptides mimicking growth factors such as IGF-1 to enhance MSC functions while promoting cell retention and minimizing off-target effects seen with direct administration of soluble growth factors. This dissertation investigates two alternative types of peptide mimetics in combination with integrin-binding peptides to modulate the fate of primary human IVD cells and MSCs. In the first part, bio-orthogonal chemistries were used to conjugate syndecan and integrin-binding peptides to alginate to regulate NP cell adhesion and fate in 2D and 3D. While alginate is the default carrier for in vitro NP cell culture, “naked” alginate lacks functional ECM-derived ligands that are necessary to support cell survival and promote biosynthesis. Results show that this combination of syndecan and integrin binding peptides are superior to integrin or syndecan binding peptides alone in inducing a healthy, juvenile phenotype and biosynthesis in degenerative primary human NP cells. In the second part, IGF-1 peptide mimetics and cRGD peptides were functionalized onto alginate hydrogels to enhance MSC therapeutic potential. One IGF-1 peptide mimetic, which lacks homology to full-length IGF-1, is shown to activate the PI3K/Akt signaling pathway, support MSC survival, and reduce inflammatory cytokine production in MSCs challenged with interleukin-1 beta (IL-1β). Co-culture studies further demonstrate that MSCs within alginate hydrogels grafted with IGF-1 and cRGD peptides significantly reduced inflammation in primary NP cells. The work presented here illustrates the value of incorporating cell-adhesive peptides to promote cell viability, biosynthetic activity, and juvenile NP- specific phenotypes. These findings build upon previous research that cell-adhesive laminin motifs presented to NP cells support expression of specific phenotypes. The results also highlight the potential of growth factor peptide mimetics tethered to hydrogels to enhance the survival, reparative functions, and anti-inflammatory capabilities of transplanted cells to the disc. These insights enable rational design of hydrogels used as a carrier for cell-transplantation based therapies for IVD degeneration and inflammatory diseases