Endothelial cells (ECs) form the innermost layer of all vasculature and constantly receive both biochemical and biomechanical signals, yielding a plethora of biomechanical responses. In response to various biochemical or biomechanical cues, ECs have been documented to generate biomechanical responses such as tractions and intercellular stresses between the cell and substrate and between adjacent cells in a confluent monolayer, respectively. Thus far, the ability of endothelial tight junctions and adherens junctions to transmit intercellular stresses has been actively investigated, but the role of gap junctions is currently unknown. In addition, there is no report of the independent influence of hyperglycemia on endothelial biomechanics present in the literature. To fill these gaps, we conducted a two-fold study where we investigated the influence of endothelial gap junction Cx43 and hyperglycemia in endothelial tractions and intercellular stress generation. In the first study, we selectively disrupted and enhanced EC gap junction Cx43 by using 2',5'-dihydroxychalcone and retinoic acid, respectively and in the second study, we cultured ECs in both normal glucose and hyperglycemic condition for 10 days. In both studies, tractions and intercellular stresses were calculated using traction force microscopy (TFM) and monolayer stress microscopy (MSM), respectively. Our results reveal that Cx43 downregulation increased as well as decreased endothelial avg. normal intercellular stresses in response to a low (0.83 µM) and a high dose (8.3 µM) chalcone treatment, respectively, while Cx43 upregulation decreases avg. normal intercellular stresses in both treatment conditions (2.5 µM and 25 µM) compared to control. In addition, we observed a decrease in intercellular stresses with hyperglycemic condition compared to control. The results we present here represent, for the first time, detailed and comprehensive biomechanical analysis of endothelial cells under the influence of glucose and the gap junction Cx43. We believe our results will provide valuable insights into endothelial permeability, barrier strength as well as leading to a greater understanding of overall endothelial mechanics.

Hypoplastic left heart syndrome (HLHS) is a type of heart defect where the left ventricle is underdeveloped or not developed, resulting in only a single functioning right ventricle. Approximately 7.5% of patients with congenital heart disease are born with a single ventricle (SV) which is accompanied by a spectrum of other malformations such as atrophied ascending aorta, atrial septal defects, and ventricular septal defects (VSD). The existing three-hybrid staged surgical approach serving as a palliative treatment for this anomaly entails multiple complications and achieves a survival rate of only 50%. To reduce the trauma associated with the second stage of the hybrid procedure the hybrid comprehensive stage 2 (HCSII) operation can be a novel palliation alternative for a select subset of SV patients with adequate antegrade aortic flow. The procedure reduces surgical trauma in newborns by introducing a stented intrapulmonary baffle to avoid dissection of the pulmonary arteries and reconstruction of the aortic arch while obviating the dissection of the ductal continuation and distal arch. It is the purpose of this dissertation to undertake a computational investigation to elucidate the complex hemodynamics of patients who have undergone HCS II. This was accomplished in a multiscale manner coupling a 0D lumped parameter model (LPM) of the peripheral circulation with 3D pulsatile Computational Fluid Dynamics (CFD) model providing the details and enabling investigation of the HCS II complex hemodynamics. The use of CFD allows modeling of blood flow, the study of the effect of different surgical procedures, suggestion of potential improvements from investigation of areas of concern which are: the pressure drop across the baffle, the loading of the baffle itself, shear stress and shear rates that might lead to thrombus formation, as well as oxygen transport and particle residence time. A 3D anatomical model representative of a patient having undergone the HCSII was rendered utilizing the solid modeling software Solidworks based on anatomical landmarks from CT scans, and a 0D LPM was tuned to produce flowrates and waveforms that matched catheter data. The pulsatile CFD computations were carried out using the commercial STARCCM+ solver. Several cases of baffle strictures relevant to surgical implementations were considered and results showed that the largest pressure drop across the baffle reported was about 3 mmHg while for the same narrowing size and accounting for the distal arch kink, a four-fold increase is observed yielding a 12.15 mmHg drop. Moreover, the analysis showed that for averaged blood flow velocity of 0.5 m/s, no vortex shedding from the baffle was observed in the computational model due to the short distance from the baffle to the aortic arch apex. The velocity and pressure-flow fields were examined at different points throughout the cardia cycle: late diastole, early systole, peak systole, and early diastole. Reverse flow was observed towards late diastolic phase due to the presence of an adverse pressure gradient, and a stagnant flow in the aortic arch apex was also noticed. For the pulmonary circulation and due to the low flow velocity and low pulsatility, the T-junction shape of the SVC presented no risk of recirculation or swirling that may promote thrombogenesis. The wall shear stress on the baffle surface was also reported in pulsatile flow. It was observed that the flow detaches in systole and subsequently reattaches to the baffle surface. Moreover, the baffle surface experiences high wall shear stress magnitudes during systole and uneven distribution of WSS during diastole. The variation in the baffle related narrowing had a little impact on the flow hemodynamics, as shown by the nearly constant oxygen transport across the models. The geometrical modification applied to the models had little effect on the oxygen delivery for up to a 15% change between a 4 mm increment of MPA minimum diameter. The results showed consistency with the published data of Glenn patients. Particle residence time was also reported to identify any blood recirculation or flow stagnation that may lead to platelet activation leading to clot formation rate. On average particles take about 0.5(s) to exit the fluid domain. This time span is equal to the time of one cardiac cycle. Finally, the energy loss and energy efficiency were calculated as a function of split ratio and baffle related narrowing. Across all models, the efficiency was shown to be high.

Airborne diseases transmitted through tiny respiratory droplets such as the Coronavirus disease can not only rapidly infect others but also worsen the symptoms in already affected individuals. People with COVID-19 are more likely to develop severe acute respiratory syndrome or SARS through aspiration pneumonia, which refers to when some virus-laden droplets are inhaled into the airway and lungs. This project aims to study droplet behavior in the upper airway in order to investigate methods to reduce the risk of infected droplets entering the upper airway in a patient. The project involves the design and development of a cough emulator that can simulate a human cough accurately, build a physical model of the upper airway using a material similar in texture to the human windpipe, and measure and track the generated particles as they transverse through the upper airway and exit the mouth. The criteria needed to be met to design, manufacture, and evaluate a cough emulator reproducing a human-like cough include the volume, pressure, and flow rate of a cough. To evaluate the validity and accuracy of the device, the number, size, and spread of cough droplets are compared to that of a real cough. The upper airway is fabricated using Elastic 50A resin due to its flexible and durable properties, and texture similarity to the tissue of the human trachea. In addition, particles are tracked in the upper airway using a Charged Couple Device (CCD) Camera.

Surgical robots are increasingly becoming an integral part of an operating room to assist surgeons in performing dexterous surgical procedures and improve surgical precision. Research and development of novel and innovative surgical devices have become essential to further the frontiers of knowledge in this field and enable enhanced surgical outcomes. Precise and accurate positioning of tools is a key requirement during the design of a surgical robot, and this often demands design of new mechanisms. This dissertation explores development of two robotic surgical devices to improve the current surgical procedure of an osteochondral autograft transplantation and transperineal prostate biopsy needle insertion. We design and develop a robotically assisted novel graft removal mechanism to harvest a personalized autologous graft of any shape and size for osteochondral autograft transplantation. To provide robotic precision, greater access, and compact design, we design and develop a robotic mechanism that can provide four Degrees of Freedom manipulation in a compact form comparable to size of manual templates for transperineal prostate biopsy.

Prostate cancer is one of the most commonly diagnosed cancers among men and a major area of concern for healthcare system. Magnetic Resonance Imaging (MRI) that has superior soft tissue imaging capability, enables higher cancer detection rate. MRI guided intervention devices that require device-to-image registration, essential for accurate and precise targeting, complicates the overall process while increasing total time for intervention. We design and develop a simplified, MRI guided, registration-free transperineal prostate biopsy needle guide and perform experiments. A preliminary simulation is performed to understand whether the interaction between guide hole and biopsy needle during the firing step affects the needle motion characteristics. The proposed device could be incorporated for rapid adoption into the clinical environment.

Medical Robots are transforming the healthcare landscape by applying robotic technologies across several aspects of patient care. Though medical robots can be classified based on several aspects such as field of application, target anatomical region or surgical and non-surgical robots, an important classification is based on the compliant nature of the robotic systems i.e., soft robotics and robots with rigid structures. Through this dissertation we present the development of two medical robotic devices, each falling under two separate categories i.e., a soft robotic device applied to tele-palpation and training and a rigid robotic device for orthognathic surgery. We design and develop a novel tactile display apparatus to facilitate medical palpation and enable early diagnosis of a possibly cancerous tumor or thyroid lump. We also design and develop a prototype of a surgical robotic device for orthognathic surgery, that can eliminate the intra-operative device registration, thereby simplifying the robotic procedure with a smaller footprint.

The complexity of the human musculoskeletal system presents challenges in accurately identifying its characteristics, particularly due to the presence of redundant actuators on a single joint. Non-invasive measures are necessary to overcome these challenges. Optimization algorithms have emerged as a crucial tool to advance subject-specific musculoskeletal modeling allows a more realistic representation of biomechanical behaviors, enhancing our understanding of human movement and enabling better clinical decision-making. Furthermore, optimization algorithms play a vital role in customizing rehabilitation and assistive devices, such as orthoses and prostheses. The current ankle-foot orthosis (AFO) stiffness measurement methods require bulky, complex designs, and often permanent modification of the AFO. To address this, we proposed the Ankle Assistive Device Stiffness (AADS) test method, which utilizes a simple design jig and motion capture system. In our method we employed a static optimization algorithm to estimate external forces and AFO torque, providing reliable stiffness quantification. The AADS test demonstrated high precision among different operators and trials, with an overall percent error within ±6%. In the pursuit of accurately measuring muscle-tendon parameters, various techniques, including shear waves, have been utilized. However, these techniques often are invasive or lack the ability to provide quantitative measurements. In our second study, we introduced a noninvasive method for estimating passive muscle-tendon parameters (PMPs) in knee flexors/extensors and the Achilles tendon. We employed a direct collocated optimal control algorithm and evaluated the precision of the proposed method through simulation, replica leg experiments, and in-vivo experiments involving 10 subjects. The estimated range for tendon slack length was reported between 0.59 and 1.13, while the median of tendon stiffness was 421 KN/m. Muscle stiffness ranged between 473 N/m and 1200 N/m. The average root mean square error (RMSE) between experimentally collected joint kinematics and kinetics and forward dynamic verification was less than 0.56° and 12 mN.m/Kg, respectively.

<p>Increased muscle stiffness helps prevent excessive movement that can lead to ligament and soft-tissue damage. There is empirical evidence suggesting that muscles are important in preventing injuries caused by excessive translational movements. Very little is known, however, on how our muscles provide translational stiffness. This thesis uses complementary theoretical (Chapters 2 and 3) and experimental (Chapter 4) techniques to address how muscles provide translational joint stiffness.</p> <p>In Chapters 2 and 3, we used an elastic energy approach to successfully derive equations that quantify muscular contributions to joint stiffness. From the equations, we were able to determine how the geometric orientation and mechanical properties of an individual muscle allows it to provide translational stiffness. In Chapter 4, using the techniques developed in the previous chapters, we test the notion that the nervous system is responsive to translational loading.</p> <p>From these works, several important discoveries were found. We are the first to find that muscles with large squared projections (alignment) over a degree-of-freedom are well suited to provide translational stiffness. Further, by explicitly describing the interactions between the translational and rotational stiffnesses we found that ignoring these interactions resulted in an overestimation of principal stiffnesses. This has large implication for stability analyses, where such overestimations could suggest that an unsafe task is actually safe. Experimentally, we found that the nervous system is responsive to translational loading. This was accomplished through increased activity of muscle well suited to provide translational stiffness.</p> <p>Collectively, the works presented provide much needed knowledge on the role muscle play in stabilizing and protecting our joints. This thesis provides a strong foundation for continued joint stiffness, stability, and impedance research.</p> / Doctor of Science (PhD)

A combination of surface electromyography (EMG) and pattern recognition algorithms have led to improvements in the functionality of upper limb prosthetics. This method of control relies on user's ability to repeatedly generate consistent muscle contractions. Research in EMG based control of prosthesis has mainly utilized adult subjects who have fully developed neuromuscular control. Little is known about children's ability to generate consistent EMG signals necessary to control artificial limbs with multiple degrees of freedom. To address this gap, two experiments were designed to validate and benchmark an experimental protocol that quantifies the ability to coordinate forearm muscle contractions in able-bodied children across adolescent ages. Able-bodied, healthy adults (n = 8) and children (n = 9) participated in the first experiment that aimed to measure the subject's ability to produce distinguishable EMG signals. Each subject performed 8 repetitions of 16 different hand/wrist movements. We quantify the number of movement types that can be classified by Support Vector Machine with > 90% accuracy. Additional adults (n=8) and children (n=12) were recruited for the second experiment which measured the subjects' ability to control the position of a virtual cursor on a 1-DoF slide using proportional EMG control under three different gain levels. We demonstrated that children had a smaller number of highly independent movements than adults did, due to higher variability. Furthermore, we found that children had higher failure rates and slower average target acquisitions due to increased time-to-target and follow-up correction time. We also found significant correlation between forearm circumference/age and performance. The results of this study provide novel insights into the technical and empirical basis to better understand neuromuscular development in pediatric upper-limb amputees.

Pharmaceutical aerosols provide a number of advantages for treating respiratory diseases that include targeting high doses directly to the lungs and reducing exposure of other organs to the medication, which improve effectiveness and minimize side effects. However, difficulties associated with aerosolized drug delivery to the lungs include drug losses in delivery devices and in the extrathoracic region of human upper airways. Intersubject variability of extrathoracic and thoracic drug deposition is a key issue as well and should be minimized. Improvements to respiratory drug delivery efficiency have been recently proposed by Dr. P. Worth Longest and Dr. Michael Hindle through the use controlled condensational growth methods, which include enhanced condensational growth (ECG) and excipient enhanced growth (EEG). These methods reduce inhaled drug loss through the introduction of an aerosol with an initial submicrometer aerodynamic diameter, which then experiences condensational growth to increase droplet size and enhance thoracic deposition. Tracheobronchial and nasal human airway computational models were developed for this study to assess drug delivery using conventional and EEG methods. Computational versions of these models are used to assess drug delivery and variability with computational fluid dynamics (CFD) simulations, which are validated with experimental data where possible. Using CFD, steady state delivery of albuterol sulfate (AS) during high flow therapy (HFT) through a nasal cannula was characterized with four nasal models developed for this study, with results indicating an increase in average delivered dose from 24.0% with a conventional method to 82.2% with the EEG technique and an initially sized 0.9 µm aerosol, with a corresponding decrease in the coefficient of variation from 15% to 3%. Transient CFD simulations of nebulized AS administration through a mask during noninvasive positive pressure ventilation (NPPV) were performed and validated with experimental data, which resulted in 40.5% delivered dose with the EEG method as compared with 19.5% for a conventional method and a common inhalation profile. Using two newly created face-nose-mouth-throat models, dry powder delivery of ciprofloxacin during NPPV was assessed for the first time with steady state CFD predictions, which showed an increase in average delivered lung dose through a new mask design of 78.2% for the EEG method as compared with 36.2% for conventional delivery, while corresponding differences in delivered dose between the two models were reduced from 45.4% to 12.8% with EEG. In conclusion, results of this study demonstrate (i) the use of highly realistic in silico and in vitro models to predict the lung delivery of inhaled pharmaceutical aerosols, (ii) indicate that the EEG approach can reduce variability in nose-to-lung aerosol delivery through a nasal cannula by a factor of five, and (iii) introduce new high efficiency methods for administering aerosols during NPPV, which represents an area of current clinical need.