<p>This thesis provides new regression methods for the removal of structured noise in datasets. With multivariable data, the variables and the noise can be both temporally correlated (i.e. auto correlated in time) and contemporaneously correlated (i.e. cross-correlated at the same time). In many occasions it is possible to acquire measurements of the noise, or some function of it, during the data collection. Several new constrained latent variable methods (LVM) that are built upon previous LVM regression frameworks are introduced. These methods make use of the additional information available about the noise to decompose a dataset into basis for the noise and signal. The properties of these methods are investigated mathematically, and through both simulation and application to actual biomedical data.</p> <p>In Chapter Two, linear, constrained LVM methods are introduced. The performance of these methods are compared to the other similar LVM methods as well as ordinary PLS throughout several simulation studies. In Chapter Three, a NIPALS type algorithm is developed for the soft constrained PLS method which is also able to account for missing data as well as datasets with large covariance matrices. Chapter Four introduces the nonlinear-kernelized constrained LVM methods. These methods are capable of handling severe nonlinearities in the datasets. The performance of these methods are compared to nonlinear kernel PLS method. In Chapter Five the constrained methods are used to remove ballistocardiographic and muscle artifacts from EEG datasets in combined EEG-fMRI as well as single EEG experiments on patients. The results are shown and compared to the standard noise removal methods used in the field. Finally in Chapter Six, the overall conclusion and scope of the future work is laid out.</p> / Doctor of Philosophy (PhD)

<p>A microfluidic device capable of drug delivery to multiple wells in a concentration gradient was designed for automated high content and high throughput screening. The design was proposed to utilize a nanoporous polycarbonate membrane to spatially and temporally control drug dosage from the microchannels below to the wells above. Microchannels were to hold to the drugs or reagents, while wells were to culture cells. An array of 16 wells was to fit in the equivalent area of a single well of a 96 well plate. Two simpler devices were created to validate electrokinetic drug delivery to a single well and to characterize cell proliferation and viability in micro-wells. The first device tested drug delivery to a single well with methylene blue dye at applied voltages of 100V, 125V, and 150V. It was validated that the dosage of dye could be controlled by increasing the voltage and by increasing the duration the voltage was applied. The second devices were a series of 9-well arrays, each testing a different diameter (1.2 mm – 0.35 mm). These devices were cultured with MCF-7 breast cancer cells over 5 days. At the end of the 5 day study, all diameters except for 0.5 mm and 0.35 mm measured a cell viability of 99% and exhibited cell growth patterns similar to coverslip glass controls. The proposed integrated cell culture and drug delivery device could have application towards early stage drug discovery and could have compatibility with lab equipment originally designed for well plates.</p> / Master of Applied Science (MASc)

Total ankle replacement (TAR) implants are an effective option to restore the range of motion of the ankle joint for arthritic patients. An effective tool for analyzing these implants’ mechanical performance and longevity in-silico is finite element analysis (FEA). ABAQUS FEA was used to statically analyze the von Mises stress and contact pressure on the articulating surface of the bearing component in two newly installed fixed-bearing total ankle replacement implants (the Wright Medical INBONE II and the Exactech Vantage). This bearing component rotates on the talar component to induce primary ankle joint motion of plantarflexion and dorsiflexion. The stress response was analyzed on this bearing component since it is made of the least strong material in the implant assembly (ultra-high molecular weight polyethylene (UHMWPE). This bearing component commonly fails and is the cause for surgeon revisions. Six different FEA models for various gait percentages during stance (10%, 20%, 30%, 40%, 50%, and 60%) were created. They varied in magnitude of the compressive load and the ankle dorsiflexion/plantarflexion angle. This study captured the variation in stress magnitudes based on the portion of the stance phase. The results indicated that the stress distribution on the articulating surface increased as compressive load increased, and the largest magnitudes occurred at high dorsiflexion angles (15-30°). The von Mises stress and contact pressure tended to occur in regions where the thickness of the bearing was the least. Additionally, high contact pressures were examined in areas near the talar component's edge or at the bearing's edges. To the author’s knowledge, this is the first study available to the research community that analyzes the Vantage implant with FEA. This study lays an essential foundation for future researchers in presenting a thorough literature review of TAR and for a simple model setup to capture the stress distributions of two TAR implants. This study provides valuable information that can be beneficial to medical company designers and orthopedic surgeons in understanding the stress response of TAR patients.

Chronic low back pain (cLBP) affects the quality of life of over 23% of American adults and places an immense burden on the economy.1–3 To more effectively diagnose and treat cLBP, quantitative biomarkers are needed to objectively measure pain and better identify the underlying mechanisms of cLBP. The overarching goal of this work is to create an analytical pipeline to investigate brain activity patterns measured with electroencephalogram (EEG) data of cLBP patients treated with transcutaneous spinal cord stimulation (tSCS). The specific aims of this project are to 1) investigate whether there is a difference between baseline EEG measurements and EEG measurements after transcutaneous spinal cord stimulation, 2) investigate potential brain activity patterns that may be associated with patient-reported pain scores, and 3) discuss the potential use of these brain activity patterns in wearable devices. Resting state EEG data was collected from four patients before and after completion of all tSCS therapy sessions, as well as before and after acute stimulation, for eyes open and eyes closed conditions. Patients underwent 12 or 21 tSCS sessions, with daily collection of patient-reported outcomes using a visual analog scale. EEG analysis has been done in MATLAB using the EEGLAB toolbox. A statistically significant (p

This thesis proposes an algorithm that determine the geometry of 3D-printed, custom-designed spring element bands made of thermoplastic elastomer (TPE) for use in a wearable orthotic device to aid in the physical therapy of a human hand exhibiting spasticity after stroke. Each finger of the hand is modeled as a mechanical system consisting of a triple-rod pendulum with nonlinear stiffness at each joint and forces applied at the attachment point of each flexor muscle. The system is assumed quasi-static, which leads to a torque balance between the flexor tendons in the hand, joint stiffness and the design force applied to the fingertip by the 3D-printed spring element. To better understand material properties of the spring element’s material, several tests are performed on TPE specimens printed with different infill geometries, including tensile tests and cyclic loading tests. The data and stress-strain curves for each geometry type are presented, which yield a nonlinear relationship between stress and strain as well as apparent hysteresis. Polynomial curves are used to fit the data, which allows for the band geometry to be designed. A hypothetical hand is presented along with how input measurements might be taken for the algorithm. The inputs are entered into the algorithm, and the geometry of the bands for each finger are generated. Results are discussed, and future work is noted, providing a means for the design of a customized orthotic device.

Chemo-radiation therapy is a principal modality for the treatment of head and neck cancers, and its efficacy depends on the interaction of tumor oxygen with free radicals. In this study, we adopted a novel hybrid diffuse optical instrument combining a commercial frequency-domain tissue oximeter (Imagent) and a custom-made diffuse correlation spectroscopy (DCS) flowmeter, which allowed for simultaneous measurements of tumor blood flow and blood oxygenation. Using this hybrid instrument we continually measured tumor hemodynamic responses to chemo-radiation therapy over the treatment period of 7 weeks. We also explored monitoring dynamic tumor hemodynamic changes during radiation delivery. Blood flow data analysis was improved by simultaneously extracting multiple parameters from one single autocorrelation function curve measured by DCS. Patients were classified into two groups based on clinical outcomes: a complete response (CR) group and an incomplete response (IR) group with remote metastasis and/or local recurrence within one year. Interestingly, we found human papilloma virus (HPV-16) status largely affected tumor homodynamic responses to therapy. Significant differences in tumor blood flow index (BFI) and reduced scattering coefficient (μs’) between the IR and CR groups were observed in HPV-16 negative patients at Week 3. Significant differences in oxygenated hemoglobin concentration ([HbO2]) and blood oxygen saturation (StO2) between the two groups were found in HPV-16 positive patients at Week 1 and Week 3, respectively. Receiver operating characteristic curves were constructed and results indicated high sensitivities and specificities of these hemodynamic parameters for early (within the first three weeks of the treatment) prediction of one-year treatment outcomes. Measurement of tumor hemodynamics may serve as a predictive tool allowing treatment selection based on biologic tumor characteristics. Ultimately, reduction of side effects in patients not benefiting from radiation treatment may be feasible.

In immersive audio environments with distributed microphones, Independent Component Analysis (ICA) can be applied to uncover signals from a mixture of other signals and noise, such as in a cocktail party recording. ICA algorithms have been developed for instantaneous source mixtures and convolutional source mixtures. While ICA for instantaneous mixtures works when no delays exist between the signals in each mixture, distributed microphone recordings typically result various delays of the signals over the recorded channels. The convolutive ICA algorithm should account for delays; however, it requires many parameters to be set and often has stability issues. This thesis introduces the Channel Aligned FastICA (CAICA), which requires knowledge of the source distance to each microphone, but does not require knowledge of noise sources. Furthermore, the CAICA is combined with Time Frequency Masking (TFM), yielding even better SOI extraction even in low SNR environments. Simulations were conducted for ranking experiments tested the performance of three algorithms: Weighted Beamforming (WB), CAICA, CAICA with TFM. The Closest Microphone (CM) recording is used as a reference for all three. Statistical analyses on the results demonstrated superior performance for the CAICA with TFM. The algorithms were applied to experimental recordings to support the conclusions of the simulations. These techniques can be deployed in mobile platforms, used in surveillance for capturing human speech and potentially adapted to biomedical fields.

Made available in DSpace on 2014-12-17T14:55:52Z (GMT). No. of bitstreams: 1 JoaoMTL_DISSERT.pdf: 5377292 bytes, checksum: 39e8ffd078c1ad7b51b4db49d8d4f654 (MD5) Previous issue date: 2011-06-30 / The great diversity in the architecture of biomedical devices, coupled with their different communication protocols, has hindered the implementation of systems that need to make access to these devices. Given these differences, the need arises to provide access to such a transparent manner. In this sense, this paper proposes an embedded architecture, service-oriented, for access to biomedical devices, as a way to abstract the mechanism for writing and reading data on these devices, thereby contributing to the increase in quality and productivity of biomedical systems so as to enable that, the focus of the development team of biomedical software, is almost exclusively directed to its functional requirements / A grande diversidade na arquitetura de dispositivos biom?dicos, aliada aos seus diferentes protocolos de comunica??o, tem dificultado a implementa??o de sistemas que necessitam realizar o acesso a esses dispositivos. Diante dessas diferen?as, surge a necessidade de prover o acesso a esses de forma transparente. Neste sentido, o presente trabalho prop?e uma arquitetura embarcada, orientada a servi?o, para acesso a dispositivos biom?dicos, como forma de abstrair o mecanismo de escrita e leitura de dados nesses dispositivos, contribuindo desta maneira, para o aumento na qualidade e produtividade dos sistemas biom?dicos, de forma a possibilitar com que, o foco da equipe de desenvolvimento de softwares biom?dicos, seja quase que exclusivamente voltado aos seus requisitos funcionais

Increasing useof nanomaterials in consumer products and biomedical applications creates the possibilities of intentional/unintentional exposure to humans and the environment. Beyond the physiological limit, the nanomaterialexposure to humans can induce toxicity. It is difficult to define toxicity of nanoparticles on humans as it varies by nanomaterialcomposition, size, surface properties and the target organ/cell line. Traditional tests for nanomaterialtoxicity assessment are mostly based on bulk-colorimetric assays. In many studies, nanomaterials have found to interfere with assay-dye to produce false results and usually require several hours or days to collect results. Therefore, there is a clear need for alternative tools that can provide accurate, rapid, and sensitive measure of initial nanomaterialscreening. Recent advancement in single cell studies has suggested discovering cell properties not found earlier in traditional bulk assays. A complex phenomenon, like nanotoxicity, may become clearer when studied at the single cell level, including with small colonies of cells. Advances in lab-on-a-chip techniques have played a significant role in drug discoveries and biosensor applications, however, rarely explored for nanomaterialtoxicity assessment. We presented such cell-integrated chip-based approach that provided quantitative and rapid response of cellhealth, through electrochemical measurements. Moreover, the novel design of the device presented in this study was capable of capturing and analyzing the cells at a single cell and small cell-population level. We examined the change in exocytosis (i.e. neurotransmitterrelease) properties of a single PC12 cell, when exposed to CuOand TiO2 nanoparticles. We found both nanomaterials to interfere with the cell exocytosis function. We also studied the whole-cell response of a single-cell and a small cell-population simultaneously in real-time for the first time. The presented study can be a reference to the future research in the direction of nanotoxicity assessment to develop miniature, simple, and cost-effective tool for fast, quantitative measurements at high throughput level. The designed lab-on-a-chip device and measurement techniques utilized in the present work can be applied for the assessment of othernanoparticles' toxicity, as well.

This research used a focused ion beam in order to fabricate record small nano-magnetic structures, investigate the properties of magnetic materials in the rarely studied range of nanometer size, and exploit their extraordinary characteristics in medicine and nano-electronics. This study consists of two parts: (i) Fabrication and study of record small magnetic tunnel junctions (ii) Introduction of a novel method for detection of magnetoelectric nanoparticles (MENs) in the tissue. A key challenge in further scaling of CMOS devices is being able to perform non-volatile logic with near zero power consumption. Sub-10-nm nanomagnetic spin transfer torque (STT) magnetic tunneling junctions (MTJs) have the potential for a universal memory that can address this key challenge. The main problem is to decrease the switching current density. This research studied these structures in sub-10-nm size range. In this range, spin related excitations consume considerably smaller amounts of energy as compared to the larger scale. This research concluded that as predicted a decrease in switching current superior to that of the linear scaling will happen in this size range. Magneto-electric nanoparticles (MENs) can be used to directly couple intrinsic electric-field-driven processes with external magnetic fields for controlling neural activity deep in the brain. These particles have been proven to be capable of inducing deep brain stimulation non-invasively. Furthermore, these magneto-electric nano-particles can be used for targeted drug delivery and are contenders to replace conventional chemotherapy. The circulatory system can deliver a drug to almost every cell in the body; however, delivering the drug specifically into the tumor cell and then releasing it on demand remains a formidable task. Nanomedicine can accomplish this, but ensuring that the drug is released at an appropriate rate once at the target site is an important task. In order to have a complete understanding of the behavior of these MENs when injected into the body, a comprehensive bio-distribution study was performed. This study introduced a novel spectroscopy method for tracing the nanoparticles in the bloodstream. This study investigated the post injection distribution of the MENs in vital organs throughout a period of two months.