Bioengineering

Permanent URI for this collectionhttps://digital.lib.washington.edu/handle/1773/4895

Browse

Now showing 1 - 20 of 434

Feature Selection and Decoder Design for Closed-loop Neural Interface Learning
(2026-04-20) Li, Si Jia; Orsborn, Amy L.
Neural interfaces, ranging from non-invasive scalp electroencephalography (EEG) and pe-ripheral electromyography (EMG) to invasive intracortical arrays, hold transformative po- tential for restoring motor and communicative independence to individuals with neurological impairments. However, the translation of these systems from controlled laboratory settings to robust, everyday clinical use is hindered by a universal scalability bottleneck. Across all modalities, advances in recording hardware now permit the simultaneous monitoring of hundreds to thousands of channels, creating a “data deluge” that overwhelms current decoding frameworks. This expansion in feature space presents a critical challenge: clinical systems lack the computational efficiency to process high-dimensional feature sets within the strict power and latency constraints of portable or implantable hardware. Furthermore, the inherent non-stationarity of biological signals—whether cortical or peripheral—necessitates adaptive frameworks that can sustain performance over long durations without burdensome recalibration. This dissertation establishes principled computational frameworks to opti- mize feature selection and decoder design, effectively bridging the gap between expanding sensor capabilities and the requirements of real-time, robust control. First, to motivate the need for dimensionality reduction, we characterized the distribu-tion of task information across biophysical, spatial, and spectral scales. Using simultaneous recordings across multiple physiological scales, we demonstrated that neural feature vari- ance is highly redundant and spatially fractured. We identified “hub” electrodes that, while strongly correlated with broad population dynamics, often encode minimal task-relevant information. These findings challenge the standard practice of indiscriminate feature inclu- sion and provide the physiological justification for disentangling feature selection from task decoding. Second, we developed a novel framework for Adaptive Feature Selection to addressthe instability of high-dimensional inputs in closed-loop settings. We demonstrated that applying standard, static feature selection methods directly to online data frequently leads to performance degradation due to statistical volatility. To overcome this, we introduced a dynamic selection algorithm governed by temporal continuity constraints. This approach autonomously identifies and tracks the informative subspace in real-time, stabilizing control performance while significantly reducing the computational load required for decoding. Finally, we addressed the behavioral complexity of real-world use by developing a closed-loop decoder for Hybrid Multitasking. Moving beyond single-degree-of-freedom control, we designed a system that enables users to simultaneously perform continuous tracking and discrete classification. Our results indicate that participants can achieve efficient control over these complex, interfering task demands, and that the rate of user learning is strictly governed by specific decoder adaptation parameters. Collectively, this thesis advances the critical transition of neural interface technologytoward scalable, “plug-and-play” architectures. By solving the dual challenges of efficient online feature selection and robust multitasking control, this work contributes essential methodologies for the development of high-bandwidth clinical systems capable of restoring complex human agency.
Modulating electrospun fibrous scaffolds to mimic native meniscus properties and engineer an in vitro injury model
(2026-04-20) Meinhold, Katherine; Robinson, Jenny
Electrospun fibrous scaffolds can be produced and optimized for creating models of the fibrous connective tissues found in the body. Fibrous tissues, like the meniscus, are made up of highly complex and organized collagen fibers which aid their ultimate functionality. However, these tissues, and the meniscus in particular, are prone to injury and demonstrate poor regenerative outcomes. By optimizing the production of highly aligned electrospun polymer fibers with and without small molecules, like non-ionic surfactants, more effective and reproducible in vitro models of meniscal tissue may be built. These scaffolds can be used in conjunction with a tensile bioreactor to investigate how an altered microenvironment affects mechanically sensitive signaling pathways and meniscal cell phenotypes post-injury. In this work we present a detailed analysis of the primary driving factors of electrospun fiber alignment in rotating mandrel systems and how non-ionic surfactant can alter collection of random or aligned electrospun fibers at a macromolecular or macroscopic scale. To demonstrate the utility of surfactant for modifying properties, the concentration-based effects on scaffold physical properties and interactions with primary meniscal cells were assessed. Finally, a tensile bioreactor was used in conjunction with both unaligned and aligned polymer scaffolds to establish an injury model system. The work herein has shown that optimal alignment of micron scale fibers occurs with collection of a large diameter mandrel and surfactant has a destabilizing effect on collection of aligned fibers in systems with smaller diameter mandrels. These studies indicate that surfactant may be a powerful tool for modulating mechanical properties and primary meniscal cell response to fibrous scaffolds. Further, aligned scaffolds can be successfully used as an extracellular matrix-mimetic material in a meniscal injury tear model in a modified bioreactor system. This work sets the stage for reproducibly investigating the response of primary meniscal cells to fibrous scaffolds with altered force transmission.
Development of an LLM Framework for Clinical Hypothesis Testing using Multimodal Data
(2026-04-20) Gim, Nayoon; Wang, Ruikang K; Lee, Aaron Y
Electronic Health Records (EHRs) contain rapidly expanding volumes of structured clinical data with thepotential to accelerate evidence generation. However, translating clinical hypotheses into reproducible research remains a slow and resource-intensive process requiring manual cohort definition, data harmonization, and statistical coding. These labor-intensive steps limit scalability and transparency, contributing to challenges in reproducibility and auditability. This thesis investigates how large language model (LLM)-assisted workflows, combined with data standardization and privacy-preserving system design, can transform clinical research into a more scalable and transparent process. Chapter 1 introduces the motivation for LLM-assisted scientific workflows, reviews standards and interoperability challenges in health data, and defines the scope of the thesis. Chapter 2 establishes the data foundations that enable automated LLM-assisted clinical research by addressing two complementary requirements: data standardization and secure LLM interaction with health records. Clinical datasets are often fragmented and inconsistently structured, leading to dataset- specific analytic code that limits scalable automation. We address this by standardizing retinal imaging data in the AI-READI (Artificial Intelligence Ready and Exploratory Atlas for Diabetes Insights) cohort using structured DICOM (Digital Imaging and Communications in Medicine) representations. Building on this standardized structure, we develop a metadata preparation workflow that enables LLM-assisted analysis without exposing patient-level data. By aggregating schema information and natural language representations of data elements, this workflow provides the contextual information required for LLMs to generate executable analytical code without accessing any patient-level data. These approaches are demonstrated using two datasets: AI-READI and NHANES (National Health and Nutrition Examination Survey). To understand what aspects of clinical research can be effectively automated, it is first necessary to analyze existing manual workflows. Chapter 3 begins with a case study investigating the relationship between post-intraocular pressure elevation and the development of primary open-angle glaucoma using the IRIS Registry (Intelligent Research in Sight). This study was carried out using standard manual research workflows and serves as a representative example of real-world clinical research practice. Section 3.2 then builds on this work by shifting the focus from clinical outcomes to process analysis. Using the completed study from 3.1 as a reference, we examine the underlying research workflow to identify repetition inherent in manual pipelines, scalability bottlenecks, and recurring analytical steps. This secondary analysis highlights concrete opportunities for automation and directly motivates the development of an LLM-assisted framework to streamline hypothesis testing. Chapter 4 introduces LATCH (Large Language Model-Assisted Testing of Clinical Hypotheses), a framework that automates the translation of natural language research questions into executable statistical analyses. LATCH combines an LLM-driven semantic component that maps hypotheses to explicit cohort definitions and data extraction logic with a deterministic statistical engine that ensures reproducibility and auditability. We describe the system architecture and validate the framework by reproducing a set of published studies on diabetes using the NHANES data, demonstrating that LATCH can generate end-to- end analytical pipelines from natural language prompts without manual coding. We further characterize the system’s operational limits through targeted stress testing and behavior under edge-case conditions. Finally, Chapter 5 illustrates the application of LATCH in advancing biomedical knowledge. Beyond reproduction, LATCH enables extended analyses of existing studies, including cross-dataset generalizability testing between NHANES and AI-READI, temporal consistency evaluation, stratified analyses, and more granular exploration of prior findings. LATCH is also used to conduct exploratory, hypothesis-generating analyses of previously unexplored questions, including the identification of a nationwide vision-related trend in the diabetes population and associations between disease severity and retinal biomarkers using the AI-READI cohort. This thesis presents a framework that combines data standardization, privacy-aware infrastructure, and LLM-assisted analytics to improve the efficiency and reproducibility of clinical research. The work demonstrates that carefully designed AI-assisted systems can accelerate hypothesis testing, reduce repetitive manual effort, and support transparent real-world evidence generation while preserving human expert verification.
Genetically Encoded Stimuli-Responsive Biomaterials for Controlled Therapeutic Delivery
(2026-04-20) Ross, Murial L.; DeForest, Cole
Current therapeutics on the market are traditionally delivered through oral, parenteral, or pulmonary routes. While these methods allow for easy dosing, most of the therapeutic is not delivered to the targeted disease site, resulting in systemic side effects and the need for multiple doses. The utilization of biomaterials as therapeutic depots bypasses these issues by allowing for localized delivery of large doses of drugs over time. Although there are many published materials for drug delivery, composed of synthetic and/or naturally derived polymers, they struggle with long-term drug release, soft mechanical properties, batch-to-batch variability, biocompatibility, or lack of tunability. Recombinant protein-based polymers, expressed from host cells, address these issues through genetically designed polymers allowing for unprecedented tunability, monodisperse polymers and intrinsic biocompatibility/biodegradability. Genetically encoded polymers are a relatively new material type for controlled delivery, falling behind synthetic materials in terms of available material types and multi-stimuli responsiveness. This thesis focuses on expanding stimuli-responsive recombinant protein-based biomaterial types for controlled delivery of protein therapeutics. We first introduce a shear-thinning and self-healing interpenetrating polymer network (IPN) as an injectable biomaterial for controlled release of multiple protein therapeutics. We next present a method to control protein release following Boolean logic from a single protein chain that utilizes chemical biology tools to autonomously compiling molecular topologies that span 17 possible YES/OR/AND logic outputs in response to protease cues. This method was then extended to control the release of therapeutic proteins, such as growth factors, enzymes, therapeutic nanobodies, de novo-engineered cytokines and fluorescent proteins while maintaining bioactivity as well as incorporating light as a new cue. Further, this method was utilized to design material crosslinks to create multi-stimuli responsive drugamer hydrogels, where bulk degradation and drug release follows Boolean logic. These next-generation genetically encoded controlled release platforms will find a host of applications in precision delivery, regenerative medicine and tissue engineering.
Sensorimotor Perturbations to Study Neural Computations of Motor Learning
(2026-04-20) Rajeswaran, Pavithra; Orsborn, Amy L.
Motor learning depends on the brain’s ability to link actions to their outcomes and learn/refine sensorimotor maps through feedback. This thesis examines how sensorimotor perturbations using Brain Computer Interfaces (BCIs) and reaching behavioral tasks paradigms reveal the neural computations that support learning. We investigated BCI learning experiments with assistive manipulations that improve performance and found that assistive manipulations alter credit assignment learning and drive compaction of learned neural representations concentrating task information in fewer neurons. Second, we adapted visuomotor perturbations, commonly used in motor learning studies, to a new interface that preserves a portion of natural redundancy to study how task relevance and altering task relevance influenced learning. These perturbation paradigms revealed distinct signatures of error based and new controller learning when task relevance is manipulated in redundant spaces. Next, we introduce an electrocorticography based brain computer interface combined with optogenetic stimulation to investigate how rest and network connectivity contribute to learning that occurs outside active practice. Finally, we compared linear and nonlinear approaches for capturing the 'dimensionality'- complexity of activity patterns across neural populations.Together, these studies present a set of complementary tools that each uncover a different aspect of learning and offer findings that can shape future learning experiments and brain computer interface design. Future studies can integrate these tools, including redundancy-based perturbation paradigms, assistive interventions, measures of network connectivity, and analyses of population structure. Combining these approaches may enable frameworks that link neural computation with large scale network organization and deepen our understanding of the principles that support flexible learning in the brain.
Deciphering sequence determinants of alternative splicing and polyadenylation in health and disease with massively parallel reporter assays
(2026-04-20) Koplik, Samantha; Seelig, Georg
Splicing and polyadenylation are major co- and post-transcriptional processes that regulate gene expression. Alternative splicing (AS) and alternative polyadenylation (APA) are frequent drivers of human disease, yet systematic maps of how genetic variants affect these processes in different cell types remain limited. To address this gap, this thesis presents high-throughput perturbations of splicing and polyadenylation using massively parallel reporter assays (MPRAs) to measure the impact of genetic variation on both AS and APA in human cell lines of diverse tissue origin. First, I introduce Cell-type Oriented Massively Parallel Assay of Splicing Signatures (COMPASS), an MPRA that quantifies splicing outcomes for 87,546 variants across five human cell lines. COMPASS targets disease-relevant genes, including ACMG actionable and autism-associated genes, providing a resource to systematically dissect splicing impacts in health and disease. Benchmarking COMPASS data against predictive models highlights both strengths and weaknesses of current approaches. Biological relevance is further supported by prime editing experiments that validate selected variants in their native genomic context. Analyses of COMPASS data also reveal RNA-binding protein motifs whose disruption drives splicing changes and identify subsets of sequences that mediate cell type-specific splicing programs. Next, I applied a similar approach to dissect the cis-regulatory determinants of APA. APARENT2, a deep residual network for predicting APA, had previously identified variants enriched for gain-of-function for polyadenylation in autism GWAS cohorts. To validate these predictions, I developed an MPRA in multiple cell types, which confirmed these gain-of-function variants and uncovered additional cell type-specific effects. I further expanded this work to a larger APA MPRA that cataloged the effects of over 5,000 disease-associated variants, revealing both conserved and cell type-specific regulation. Together, these studies provide the most comprehensive cell type-resolved compendia of AS and APA to date. COMPASS delivers the largest atlas of splice-disrupting variants to date, and two APA MPRAs provide a complementary resource cataloging polyadenylation-disrupting variants. The resources developed in this thesis serve to support variant reclassification in clinical genomics, guide therapeutic target discovery, and aid in the refinement of predictive models.
Development of an Integrated Point-of-Care Diagnostic Framework for Bloodborne Pathogens
(2026-04-20) Gilligan-Steinberg, Shane David; Lutz, Barry
Diagnostic platforms are complex systems, integrating elements of assay chemistry, molecular biology, fluidics, and instrumentation for pathogen detection. Simplifying them is key to expanding their use to low-resource settings where infectious disease burden is the highest. In this thesis, I take a systems-approach to diagnostics development to leverage the strengths of innovations across technical domains into a platform for detection of bloodborne pathogens from a fingerstick sample. The first project demonstrates a highly multiplexed assay, designed to achieve coverage of HIV genetic sequence diversity by departing from empirical primer design strategies and employing a novel approach to multiplexing with cooperative neighboring assays. The second project developed an extraction-free simplification to complex sample preparation for bloodborne pathogens by directly diluting fingerstick plasma samples while scaling up reaction volumes to allow larger sample input and improved sensitivity. In the final project, I developed a prototype sample-to-result platform for detection of bloodborne pathogens from fingerstick samples. This integrated system was designed to leverage assay-level innovations in extraction-free amplification and the complementary strengths of centrifugal and paper-based microfluidics to create a platform for bloodborne pathogen detection.
Modeling Chromosomal Mosaicism During Early Human Embryogenesis on Microraft Platform
(2026-04-20) Jan, Ian; Allbritton, Nancy L
Aneuploidy and chromosomal mosaicism in human embryos complicate predicting pregnancy outcomes already plagued with frequent pregnancy losses. Given the rapid increase of assisted reproductive technology (ART) treatments, there is a critical need to better assess the developmental potential of embryo candidates, especially in patients with poor prognoses. Emerging evidence suggests that human mosaic embryos can selectively eliminate aneuploid cells for healthy development; however, the mechanisms mediating this embryonic self-correction have yet to be systematically studied in humans. Because both technical and ethical limitations restrict studying cellular processes during early human embryogenesis, I will develop a novel platform to quantitatively screen and sort gastruloids—an in vitro multicellular model recapitulating cell fate and signaling during gastrulation—comprised of euploid and aneuploid human pluripotent stem cells (hPSCs). Analyzing single gastruloids is challenging by most current technologies, which only allow for low-throughput sorting or bulk analyses. Thus, new tools are required to systematically study the heterogeneity among gastruloids undergoing dramatic changes during self-organization. The project was divided into three Aims. Firstly, I developed an automated system to perform image-based screens of single gastruloids by isolating individual colonies for downstream analyses. Secondly, the emergence of gastruloid heterogeneity derived from multiple aliquots of a single cell line was quantified by coupling both phenotypic and transcriptomic information. Low-dimensional latent representations were created from imaging time-series using deep learning (DL) of developing gastruloids derived from hPSCs expressing a SOX2-mCitrine reporter to track SOX2 dynamics. Among single gastruloids across multiple experimental batches, simple-to-complex variability in patterning behavior were assessed. Endpoint RNA-sequencing (RNA-seq) data from individually isolated gastruloids also revealed distinct clusters associated with batch and subclusters within batches. Thirdly, I assessed the phenotypic and transcriptomic differences between euploid and mosaic (derived from 50% euploid and 50% aneuploid cells) gastruloids to demonstrate the potential of this platform for studying chromosomal mosaicism during early human embryogenesis. The platform will enable future research to elucidate mechanisms for aneuploidy depletion critical for overcoming error-prone development and to improve reproductive treatments for an increasing number of patients.
MECHANISMS AND PROCESSES UNDERLYING THE REVERSIBILITY OF HYPOCONTRACTILITY-INDUCED DILATED CARDIOMYOPATHY
(2026-02-05) Reichardt, Isabella; Davis, Jennifer
Dilated cardiomyopathy (DCM) is a progressive heart disease driven by inherited mutations that impair cardiomyocyte contractility, ultimately leading to systolic dysfunction, dilation, and fibrosis. Current therapies, such as angiotensin inhibitors, β-blockers, and left ventricular assist devices (LVADs), are largely palliative and do not address the primary defect in myocyte force generation. Myosin activators have emerged as a promising new class of small molecules that can directly augment contractility at the myofilament level; however, despite encouraging preclinical data, clinical trials have yielded only modest functional improvements for patients. This gap highlights a longstanding assumption in the field that correcting the cardiomyocyte defect alone should be sufficient to reverse disease. Given that DCM often progresses over years to decades before diagnosis, it remains unclear whether targeted myocyte therapies can reverse established remodeling. This dissertation interrogates this assumption by defining how cardiomyocyte hypocontractility shapes fibroblasts behavior, ECM mechanics, and ultimately, the capacity of the heart to recover. Using transgenic mouse models and engineered heart tissues, we first demonstrate that fibroblasts act as early mechanosensitive rheostats that detect impaired myocyte tension and initiate changes in the ECM. Hypocontractile cardiomyocytes triggered the emergence of a hyperproliferative fibroblast state and increased myocardial stiffness well before the emergence of overt fibrosis. Importantly, conditional deletion of p38 in fibroblasts prevented fibroblast proliferation, ECM remodeling, and myocyte eccentric hypertrophy, demonstrating that interrupting fibroblast signaling halts DCM progression even when the myocyte defect persists. Next, we examined whether restoring myocyte contractility after disease onset is sufficient for recovery. While genetic suppression of the sarcomeric mutation fully restored cardiomyocyte structure, function, and chromatin accessibility, it did not reverse ECM stiffness or reduce fibroblast number. Persistent fibrosis and fibroblast survival limited whole-organ functional recovery. Full phenotypic reversal was achieved when myocyte correction was paired with an inhibitor of ECM crosslinking, revealing that non-myocyte adaptations impose a durable barrier to reverse remodeling. Finally, developmental studies demonstrated that transient myocyte correction during the postnatal period delayed DCM onset and attenuated remodeling, underscoring the importance of injury timing in determining recovery potential. Together, this body of work demonstrates that: 1) fibroblasts and fibrosis are central drivers of DCM progression, 2) adaptations within these compartments are far less amenable to reversal than cardiomyocyte dysfunction, and 3) early-life myocyte hypocontractility shapes the long-term severity of DCM, suggesting that pathological fibroblast and ECM states may be installed during the postnatal window. More broadly, this work can be used to help design new combinational therapeutic strategies and contextualize the performance of emerging myosin activators as they begin clinical trials for efficacy.
Probing Neurovascular Unit Dysfunction in Alzheimer’s Disease
(2026-02-05) Evitts, Kira; Young, Jessica E.; Zheng, Ying
Alzheimer’s disease (AD) is a progressive neurodegenerative disorder and the most common form of dementia. Nearly 80% of individuals with AD present with cerebrovascular pathologies such as microinfarcts, atherosclerosis, and cerebral microbleeds. The brain vasculature, which delivers oxygen and nutrients to surrounding neurons, is regulated by the neurovascular unit (NVU). The NVU is a multicellular system composed of neurons, astrocytes, microglia, pericytes, and brain endothelial cells (ECs). While NVU dysfunction is a hallmark of AD, the interactions between brain and vascular cell types remain incompletely understood. Current in vitro NVU models often lack perfusable 3D vasculature and omit critical cell types such as microglia, highlighting the need for more physiologically relevant systems. To address this gap, the body of work presented here reports a series of NVU models of increasing cellular complexity and applies them to investigate NVU dysfunction in AD. First, we examined the effects of AD neuronal secretomes on ECs using an engineered microvessel system. We perfused microvessels with conditioned medium (CM) generated from induced pluripotent stem cell (iPSC)-derived neurons containing an AD mutation that increased amyloid-beta (Aβ) production. We observed that increased Aβ production by AD neurons strongly correlated with features of EC activation. Further, when we depleted Aβ from our CM through several methods, we saw that the EC activation we observed was attenuated. This study provided a direct link between the EC activation observed in AD and Aβ. It also established a model of indirect EC-neuron interactions in the NVU. We next investigated EC-microglia interactions by developing a suite of models, including a 2D co-culture system, a 3D microvessel model containing iPSC-derived microglia-like cells (iMGL), and a multicellular 3D NVU model comprising neurons, astrocytes, iMGL, and ECs. Using these systems, we found that iMGL supported EC structure and barrier integrity under basal conditions. We then applied an AD-mimicking neuroinflammatory stimulus, TNFα, to our microglia-vessel model and determined that iMGL could help ameliorate the EC inflammatory response and prevent EC barrier breakdown. Incorporating neurons and astrocytes into the model revealed that iMGL also enhanced neuronal morphology, highlighting their broader supportive roles within the NVU. These findings deepen our understanding of how microglia interact with the brain vasculature and provide models to uncover the mechanisms driving these interactions. Together, these studies provide novel, physiologically relevant models for probing NVU dysfunction in AD. Beyond advancing mechanistic understanding of neurovascular interactions, these platforms offer a foundation for therapeutic discovery and for examining the impact of disease genetics on NVU cell types.
Structure-Based Prediction and Design of Adaptive Immune Receptors Targeting Peptide–MHC Complexes
(2026-02-05) Motmaen, Amir; Baker, David DB
Adaptive immune recognition depends on the specific interaction between peptides presented by major histocompatibility complexes (MHCs) and the receptors that survey them. Advances in protein structure prediction and design now allow us to computationally model these interactions with unprecedented fidelity and engineer them with remarkably higher success rates. In this dissertation, we develop two complementary structure-based deep learning approaches: one for predicting peptide–MHC specificity by fine-tuning structure prediction networks directly on binding data, and another for designing de novo T cell receptors and TCR-mimic antibodies that recognize peptide–MHC targets with high accuracy. Together, these methods illustrate how integrating structurally-informed deep learning protein design and structure prediction frameworks enable both robust generalization and precise molecular engineering. These tools lay the foundation for programmable, therapeutic recognition of diseased cells and expand our understanding of adaptive immune specificity.
Ultrasound mediated microbubble cavitation for treatment and monitoring of cancer
(2026-02-05) De Koninck, Lance Herman; Averkiou, Michalakis
Microbubbles, typically used as a diagnostic ultrasound contrast agent to improve blood flow visualization, also holds great promise as a therapeutic agent for treatment of solid tumors. When insonified by an acoustic pulse, microbubbles undergo volumetric oscillations, providing both an acoustic signal for image contrast enhancement and localized microscale forces during cavitation. These microscale forces can induce several biological effects that overcome the barriers to treating solid tumors, including increased cell membrane permeability, enhanced drug delivery, and tumor-specific vascular changes. In this thesis, we present several investigations of how ultrasound and microbubbles can be applied as both a therapeutic and monitoring strategy for treating cancer. We begin with an introduction on microbubble cavitation and its capacity to monitor and modulate the tumor microenvironment (Chapter 1). First, we implement acoustic conditions suitable for generating mechanical forces with cavitation therapy on a clinical scanner and determine which acoustic conditions maximize cavitation activity in vitro (Chapter 2). We then investigate the mechanical effects of cavitation therapy in vivo, using real-time monitoring to elucidate mechanisms of changes to the tumor microenvironment and enhancement of drug delivery to solid tumors (Chapter 3). Next, we evaluate the enhanced heat deposition during cavitation therapy in an ex vivo machine perfused liver model, evaluating acoustic pressure and microbubble delivery methods to increase temperature and heated area (Chapter 4). We then implement subharmonic imaging on a clinical scanner for noninvasive estimation of interstitial fluid pressure in solid tumors and evaluate the mode’s ambient pressure sensitivity in physiological ranges for a range of acoustic pressures and microbubble formulations (Chapters 5 and 6). We conclude with a summary of the accomplishments and future directions of this work (Chapter 7).
DETERMINING THE INTRAMOLECULAR MECHANISMS DRIVING ALTERED CONTRACTION IN THE MYOSIN MUTATIONS E525K AND V606M
(2026-02-05) Robeson, Kalen; Regnier, Michael; Davis, Jennifer
Force generation in the heart relies on the interaction of myosin and actin filaments, a tightly regulated process where subtle changes can lead to heart disease. Mutations in β-cardiac myosin impact the number of available myosin molecules, their binding to actin, and their ATP utilization rate. Understanding how this family of mutations alter heart contraction requires investigation of myosin at the single molecule, the sub-cellular, and the physiological level. This study investigates the mechanism of altered myosin function in two β-myosin mutations (E525K and V606M). The first project, presented in Chapter 2, uses molecular dynamics simulations of E525K and V606M myosin to highlight a regulatory role for the loop 2 structure in crossbridge binding. The second project, presented in chapter 3, involves a deep investigation of the E525K mutation using stem cell derived cardiomyocyte. Cells and tissues with the E525K mutation showed decreased force generation, consistent with dilated cardiomyopathy; however, single myofibril preparations demonstrated that myofibrils containing E525K myosin can generate more force than wild type under some conditions. These findings underscore the importance of multi-scale studies of myosin mutations. While single-molecule biochemical assays are informative, they may not always reflect the complete picture. As cardiac medicine moves towards personalized treatment, in-depth understanding of how specific myosin mutations alter chemomechanics is vital for designing tailored drugs for cardiomyopathy.
Electrical and Optogenetic Modulation of Cortical Activity to Promote Neuroprotection and Network Reorganization in Non-Human Primates
(2025-10-02) Zhou, Jasmine; Yazdan-Shahmorad, Azadeh AY
Many debilitating neurological conditions arise from abnormal network dynamics and connectivity. Novel neuromodulation techniques, including electrical and optogenetic stimulation, aim to harness the brain's innate plasticity to reorganize neural connections, reduce tissue damage, and improve patient outcomes. To test the feasibility of these emerging neuromodulation approaches while improving their translational potential, my dissertation work applies advanced electrophysiology, histology, and computational tools in the brains of non-human primates (NHPs), evaluating the cortical response to different stimulation paradigms. Chapter 2 addresses the critical need for acute interventions in ischemic stroke, which can prevent irreversible tissue injury and improve functional outcomes. By inducing focal lesions in the sensorimotor cortex of NHPs while recording electrocorticography signals, we evaluated the impact of acute, theta-burst electrical stimulation delivered adjacent to the ischemic infarct. Early stimulation significantly reduced peri-infarct neuronal depolarization and microglial activation, leading to smaller lesion volumes. This study demonstrated the therapeutic potential of acute electrical stimulation to mitigate excitotoxicity, inflammation, and neural damage following ischemic injury, offering a promising strategy to enhance patient outcomes after stroke. Chapter 3 explores the use of optogenetic tools to modulate cortical network dynamics. By delivering patterned laser illumination to regions expressing an inhibitory opsin, we successfully disrupted functional connectivity in the posterior parietal cortex, reducing gamma-band coherence between targeted locations and the broader network. This selective decoupling highlights the potential of optogenetics to weaken pathological synchrony, complementing prior work using excitatory strategies to enhance connectivity and facilitate recovery. Together, these findings advance the development of targeted neuromodulation therapies aimed at restoring healthy network function in neurological disorders.
Engaging Multiple Mechanisms of Plasticity to Promote Functional Recovery after Stroke
(2025-10-02) Khateeb, Karam; Yazdan-Shahmorad, Azadeh
The human brain is responsible for executing a vast range of functions such as movement, somatosensation, visual processing, and cognition. These and other abilities depend critically on a delicate balance between the stability and adaptability of neuronal connections. Neural injuries such as stroke disrupt these connections and often result in serious debilitating effects on the brain's ability to perform critical functions. A major challenge in developing effective rehabilitative treatments for stroke is the absence of a unifying framework for investigating multiple physiological processes in preclinical animal models. In this dissertation, I describe a framework by which we can study various aspects of cortical physiology in non-human primates (NHPs), a clinically relevant animal model, under healthy and stroke conditions (Chapter 2). Although neurons possess a diverse repertoire of plasticity mechanisms to modify and stabilize their connections, stimulation-based stroke therapies have largely focused on Hebbian forms of plasticity. Other mechanisms remain underexplored despite their potential relevance for recovery. I present two approaches to engage homeostatic and Hebbian plasticity mechanisms to induce targeted changes in functional connectivity in NHPs and rodents (Chapters 3 and 4). The combination of these tools and approaches can drive the development of effective rehabilitative stroke treatments to restore the loss of critical functions such as mobility, somatosensation, and visual processing.
Characterizing effects of puromycin selection on enrichment of astrocyte subtypes and extracellular matrix (ECM) generation with ECM incorporation in hyaluronic-acid (HA)-based hydrogels for axon growth of V2a interneurons
(2025-10-02) Vardhan, Sangamithra; Sakiyama-Elbert, Shelly
Astrocyte subtypes are key cellular players to study in vitro for identifying specific cues that increase axon growth of neuron populations in potential translation to develop a scalable astrocyte-derived solution for axon regeneration post SCI. I characterized the use of transgenic puromycin selectable cell lines to enrich protoplasmic and fibrous astrocytes differentiated from mouse embryonic stem cells (mESCs) through viability and glutamate uptake assays, immunocytochemistry, flow cytometry, quantitative polymerase chain reaction (qPCR), and calcium imaging. It was demonstrated that selected astrocyte subtypes maintain low levels of other cell types (mature neurons and oligodendrocytes and undifferentiated stem cells) and high levels of astrocyte markers with functionality by glutamate uptake, inflammatory response, and calcium transients. Furthermore, selected protoplasmic astrocyte components, specifically ECM, demonstrated significant increases in axon growth of mESC-derived V2a interneurons, a critical excitatory neuron population found in grey matter of spinal cord, compared to selected fibrous astrocytes in both 2D and 3D HA-based hydrogel setting. With bulk-RNA sequencing of astrocyte subtypes and proteomic analysis of selected astrocyte subtype ECM, selected protoplasmic astrocytes demonstrated upregulation of critical adhesion and ECM-related genes and proteins that could contribute to the positive impact on axon growth. Based on the criteria of protein availability and manufacturing, 5 selected protoplasmic ECM proteins at defined combinations and concentrations (based on design of experiments- DoE approach) in HA-based hydrogel were shown to significantly increased axon growth of V2a interneurons compared to selected protoplasmic ECM hydrogel. Overall, this thesis demonstrates the importance of studying in vitro properties of astrocyte subtypes for axon growth and the use of astrocyte-derived factors in developing a scalable, cell-free permissive biomaterial for supporting axon regeneration in future SCI applications.
The Center of Attention: the Locus Coeruleus' Role in Dopamine Dynamics
(2025-10-02) Matarasso, Avi; Bruchas, Michael
Arousal is essential for survival, and maladaptive arousal processing leads to inability to focus, anxiety-like behavior, and dysregulated affective states which implicate the locus coeruleus. The locus coeruleus (LC) is a major source of norepinephrine (NE) in the brain that projects to distinct brain regions which modularly influence arousal, anxiety, learning, and exploration, among other behavioral states. Recent timely studies have implicated dopamine (DA) as also co-released from the LC, yet definitive measures of release across regions, stimulus paradigms, and behaviors typically associated with the LC-NE system remain controversial. In this thesis, I utilize recently developed tools to establish catecholamine dynamics and more specifically the boundaries of LC-DA release, in appetitive and aversive behaviors. In Chapter 1, I review the current understanding of LC-evoked DA release and NE and DA release downstream of LC in a variety of behaviors. Next, I characterize the limitations of modern neuromodulator detection tools (Chapter 2). In Chapter 3, I establish baseline NE and DA dynamics in response to a variety of aversive, appetitive, and neutral stimuli in basolateral amygdala (BLA) and hippocampus (CA1). In Chapter 4, I address limitations in the biosensors used in detection of catecholamines. Then, I characterize evoked dopamine from LC terminals in CA1 and BLA. Lastly, I demonstrate that the LC releases DA during aversive and appetitive stimuli. In Chapter 5, I discuss the findings of this thesis in relationship with existing literature and outline future work that would meaningfully build on my dissertation work. My dissertation work expands our understanding of the LC's role in releasing DA and provides context for how maladaptive LC activity may affect DA signaling in aversive and appetitive conditions.
Optoretinography for the Functional Assessment of the Human Retina in Health and Disease
(2025-10-02) Liu, Teng; Sabesan, Ramkumar
Imaging retinal structure and function is crucial for detecting and diagnosing disease, guiding treatment strategies, and ultimately improving patient outcomes. Optoretinography (ORG) is an emerging paradigm capable of quantifying the light-evoked, nanometer-scale deformations of photoreceptors and offers a powerful, objective biomarker for retinal function. This dissertation introduces coarse-scale ORG (CoORG), an extended-field, line-scan OCT platform with high speed and phase sensitivity suitable for imaging retinal structure and function in a range of diseases. Normative data from healthy eyes reveal that cone ORG amplitude increases with retinal photon density and scales linearly with cone outer segment length. In healthy aging, ORG amplitudes decline, and the structure-function correction weakens compared to younger controls. Additionally, novel multi-layer ORG analyses extend the paradigm beyond photoreceptors and provide deeper physiological insight into outer retinal function. In patients with inherited retinal dystrophies (IRDs), ORG detects cone dysfunction even in areas with normal retinal structure and enables significantly greater sensitivity in monitoring longitudinal progression over conventional clinical imaging. Overall, this dissertation demonstrates the potential of ORG as a rapid, sensitive, patient-friendly biomarker for disease progression and outcome measure for clinical trials.
Mechanisms of contractile dysfunction of the G256E HCM-associated mutation
(2025-10-02) Kao, Kerry Yvonne; Regnier, Michael
Hypertrophic cardiomyopathy (HCM) affects approximately 1 in 500 individuals in the U.S. and is associated with adverse outcomes conditions such as atrial fibrillation, heart failure, and sudden cardiac death. About 60% of HCM cases result from inherited mutations with the MYH7 gene, encoding β-myosin, the molecular motor of the sarcomere, is the second most common mutational hotspot. Despite advancements in genetic sequencing, connecting specific mutations to clinical outcomes remains challenging, with many identified variants having unknown significance. Current therapeutic strategies for HCM are geared towards managing symptoms rather than targeting the underlying molecular causes of the disease. Our goal is to elucidate how HCM mutations in myosin lead to alterations at the protein and contractile organelle level, and how these alterations manifest at the cell and tissue level in order to find direct, druggable targets for more effective therapies.In the first body of work, we collaborated with academic groups from Stanford, UC Santa Barbara, Institut Curie, and others at UW to form a multi-institutional, multidisciplinary group to study how single missense mutations in myosin lead to HCM. We collaborated with the Allen Institute for Cell Science to engineer a CRISPR/Cas9-edited human induced pluripotent stem cell line (hiPSC) with the MYH7 G256E mutation as a HCM disease model. We (the Regnier Lab) focused on the contractile organelle, or myofibril, and molecular scale. In myofibrils isolated from hiPSC-cardiomyocytes (CMs), we saw greater and faster force development accompanied by impairment of the slow, early phase of relaxation. In molecular dynamics (MD) simulations of post-rigor human cardiac myosin (M.ATP), the G256E mutation caused instability in the transducer region, possibly altering signal transduction from the nucleotide pocket and the actin-binding cleft. Altogether, these data suggest that G256E leads to hypercontractility through impairment of relaxation resulting in a greater population of force generating myosin. The second body of work addresses how the G256E mutation alters ADP nucleotide handling. We challenged isolated hiPSC-CM myofibrils with elevated ADP to assess their response to product inhibition. G256E myofibrils demonstrated reduced sensitivity to ADP inhibition of the slow, early phase of relaxation, suggesting that ADP release is already impaired by the G256E mutation. In MD simulations of post-powerstroke myosin (A.M.ADP), we saw changes in ADP coordination in the nucleotide pocket due to the G256E mutation. Furthermore, we performed steered MD simulations to determine that G256E myosin requires on average 1.5x as much work to displace ADP compared to WT due to alterations in the molecular release pathway. This finding was validated through stopped-flow biochemistry with an observed 25% increase in ADP affinity with G256E sS1 compared to WT. In summary, by combining experimental approaches with molecular dynamics simulations, we uncovered the molecular structural alterations resulting from the G256E HCM mutation and related them to functional findings, pinpointing a specific step in the cross-bridge cycle that is impacted by this mutation–ADP release. This work serves as a framework for how identifying the structural basis of disease provides insights into genotype-specific therapeutic targets, paving the way for more precise and effective treatments for HCM.
Point-of-Care Diagnostic Device Development for Multiplexed Detection of Infectious Diseases
(2025-08-01) Jiang, Kevin Pengfei; Yager, Paul
The COVID-19 pandemic highlighted major weaknesses in the global healthcare system for patient accessibility to diagnostics. The pandemic led to a significant increase in demand for respiratory disease testing to facilitate treatment and limit transmission, demonstrating in the process that most existing test options in centralized facilities were too complex and expensive to perform in point-of-care settings. While laboratory-based nucleic acid amplification test (NAAT) assays remained the gold standard of clinical diagnosis, their reliance on complex equipment, cold-chain storage, and trained personnel initially limited their accessibility to testing in well-equipped labs and kept them out of low-resource or home environments. As transmission of SARS-CoV-2 led to national lockdown orders around the world, remote and home testing for infectious pathogens emerged as the new standard in patient care and clinical research. There was a major demand for multi-disease detection tools that could be implemented in point-of-care or home settings and could readily be adapted for new variants or different diseases entirely. In this thesis, I explored the design and development of multiple integrated point-of-care devices to perform simultaneous detection of multiple infectious diseases from a single patient sample. First, we demonstrated a point-of-care, paper-based rapid analysis device that could simultaneously detect multiple viral RNAs (for COVID-19 and influenza A) that had been spiked onto a commercial nasal swab. This compact and affordable device, enabled by novel valving innovations, not only enabled fast, sensitive detection of both SARS-CoV-2 and influenza A viruses from a single sample, but also opened the possibility for simple applications in other nucleic acid-based detection systems. Next, we adapted the multiplexing UbiNAAT platform for screening of two sexually transmitted illnesses (gonorrhea & chlamydia) in vaginal lysate samples from female clinical patients. We demonstrated the device’s compatibility with different assays and sample types, while incorporating additional processing steps for clinical sample treatment. The results highlighted the UbiNAAT’s potential for multiplexed screening of clinical samples across a multitude of infectious diseases. Lastly, we explored the development of an at-home testing platform for early-stage HIV using fingerstick blood samples. Toward this end, a novel point-of-care device with onboard plasma separation, chemical lysis, target concentration, isothermal amplification, and an innovative pull-tab valving system was integrated alongside extensive assay development for fluorescence and colorimetric detection. In future work, we anticipate that the findings and technologies developed in this thesis could be employed for a multitude of other targets, sample types, and detection systems that may enable large-scale, sensitive point-of-care testing.

Browse

Recent Submissions