Ahmet Erdemir

Lerner Research Institute, Cleveland, Ohio, United States

Are you Ahmet Erdemir?

Claim your profile

Publications (64)127.72 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Understanding the mechanical environment of articular cartilage and chondrocytes is of the utmost importance in evaluating tissue damage which is often related to failure of the fibre architecture and mechanical injury to the cells. This knowledge also has significant implications for understanding the mechanobiological response in healthy and diseased cartilage and can drive the development of intervention strategies, ranging from the design of tissue-engineered constructs to the establishment of rehabilitation protocols. Spanning multiple spatial scales, a wide range of biomechanical factors dictate this mechanical environment. Computational modelling and simulation provide descriptive and predictive tools to identify multiscale interactions, and can lead towards a greater comprehension of healthy and diseased cartilage function, possibly in an individualized manner. Cartilage and chondrocyte mechanics can be examined in silico, through post-processing or feed-forward approaches. First, joint-tissue level simulations, typically using the finite-element method, solve boundary value problems representing the joint articulation and underlying tissue, which can differentiate the role of compartmental joint loading in cartilage contact mechanics and macroscale cartilage field mechanics. Subsequently, tissue-cell scale simulations, driven by the macroscale cartilage mechanical field information, can predict chondrocyte deformation metrics along with the mechanics of the surrounding pericellular and extracellular matrices. A high-throughput modelling and simulation framework is necessary to develop models representative of regional and population-wide variations in cartilage and chondrocyte anatomy and mechanical properties, and to conduct large-scale analysis accommodating a multitude of loading scenarios. However, realization of such a framework is a daunting task, with technical difficulties hindering the processes of model development, scale coupling, simulation and interpretation of the results. This study aims to summarize various strategies to address the technical challenges of post-processing-based simulations of cartilage and chondrocyte mechanics with the ultimate goal of establishing the foundations of a high-throughput multiscale analysis framework. At the joint-tissue scale, rapid development of regional models of articular contact is possible by automating the process of generating parametric representations of cartilage boundaries and depth-dependent zonal delineation with associated constitutive relationships. At the tissue-cell scale, models descriptive of multicellular and fibrillar architecture of cartilage zones can also be generated in an automated fashion. Through post-processing, scripts can extract biphasic mechanical metrics at a desired point in the cartilage to assign loading and boundary conditions to models at the lower spatial scale of cells. Cell deformation metrics can be extracted from simulation results to provide a simplified description of individual chondrocyte responses. Simulations at the tissue-cell scale can be parallelized owing to the loosely coupled nature of the feed-forward approach. Verification studies illustrated the necessity of a second-order data passing scheme between scales and evaluated the role that the microscale representative volume size plays in appropriately predicting the mechanical response of the chondrocytes. The tools summarized in this study collectively provide a framework for high-throughput exploration of cartilage biomechanics, which includes minimally supervised model generation, and prediction of multiscale biomechanical metrics across a range of spatial scales, from joint regions and cartilage zones, down to that of the chondrocytes.
    Interface focus: a theme supplement of Journal of the Royal Society interface 04/2015; 5(2):20140081. DOI:10.1098/rsfs.2014.0081 · 3.12 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: http://www.sciencedirect.com/science/article/pii/S175161611400335X#
    Journal of the Mechanical Behavior of Biomedical Materials 01/2015; 41:177-188. DOI:10.1016/j.jmbbm.2014.10.011 · 3.05 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Over the past two decades finite element (FE) analysis has become a popular tool for researchers seeking to simulate the biomechanics of the healthy and diabetic foot. The primary aims of these simulations have been to improve our understanding of the foot's complicated mechanical loading in health and disease and to inform interventions designed to prevent plantar ulceration, a major complication of diabetes. This article provides a systematic review and summary of the findings from FE analysis-based computational simulations of the diabetic foot.
    PLoS ONE 10/2014; 9(10):e109994. DOI:10.1371/journal.pone.0109994 · 3.53 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Therapeutic footwear is frequently prescribed in cases of rheumatoid arthritis and diabetes to relieve or redistribute high plantar pressures in the region of the metatarsal heads. Few guidelines exist as to how these interventions should be designed and what effect such interventions actually have on the plantar pressure distribution. Finite element analysis has the potential to assist in the design process by refining a given intervention or identifying an optimal intervention without having to actually build and test each condition. However, complete and detailed foot models based on medical image segmentation have proven time consuming to build and computationally expensive to solve, hindering their utility in practice. Therefore, the goal of the current work was to determine if a simplified patient-specific model could be used to assist in the design of foot orthoses to reduce the plantar pressure in the metatarsal head region. The approach is illustrated by a case study of a diabetic patient experiencing high pressures and pain over the fifth metatarsal head. The simple foot model was initially calibrated by adjusting the individual loads on the metatarsals to approximate measured peak plantar pressure distributions in the barefoot condition to within 3%. This loading was used in various shod conditions to identify an effective orthosis. Model results for metatarsal pads were considerably higher than measured values but predictions for uniform surfaces were generally within 16% of measured values. The approach enabled virtual prototyping of the orthoses, identifying the most favorable approach to redistribute the patient's plantar pressures.
    Journal of Biomechanics 07/2014; 47(12). DOI:10.1016/j.jbiomech.2014.07.020 · 2.50 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Recent advances in computational modeling and simulation of human movement makes it possible to isolate and predict the potential contributions of a prosthetic device to the overall system performance. The Mauch S-N-S knee is one of the most widely used prosthetic knees in the market. The goal of this study is to develop dynamic models of the Mauch S-N-S knee for predictive simulation of a transfemoral amputee׳s gait under idealized conditions. Based on the functional description of the Mauch S-N-S prosthetic knee from the literature, a combined bench test and data fitting approach employing modified slow, normal, and fast gait patterns and nine combinations of stance and swing damping settings were performed. Two types of dynamic models, 2-phase and 4-phase models, of the Mauch S-N-S prosthetic knee were developed. The range of the coefficient of determination of the two dynamic models, when compared to the test data, was from 39.9 to 95%. Both dynamic models of this study can be utilized in musculoskeletal modeling studies, to better understand amputee gait and the contributions and interactions of various prosthetic leg components to the ambulatory performance.
    Journal of Biomechanics 06/2014; 47(12). DOI:10.1016/j.jbiomech.2014.06.011 · 2.50 Impact Factor
  • Journal of Foot and Ankle Research 04/2014; 7(Suppl 1):A82. DOI:10.1186/1757-1146-7-S1-A82 · 1.83 Impact Factor
  • Craig J Bennetts, Scott Sibole, Ahmet Erdemir
    [Show abstract] [Hide abstract]
    ABSTRACT: Finite element analysis provides a means of describing cellular mechanics in tissue, which can be useful in understanding and predicting physiological and pathological changes. Many prior studies have been limited to simulations of models containing single cells, which may not accurately describe the influence of mechanical interactions between cells. It is desirable to generate models that more accurately reflect the cellular organisation in tissue in order to evaluate the mechanical function of cells. However, as the model geometry becomes more complicated, manual model generation can become laborious. This can be prohibitive if a large number of distinct cell-scale models are required, for example, in multiscale modelling or probabilistic analysis. Therefore, a method was developed to automatically generate tissue-specific cellular models of arbitrary complexity, with minimal user intervention. This was achieved through a set of scripts, which are capable of generating both sample-specific models, with explicitly defined geometry, and tissue-specific models, with geometry derived implicitly from normal statistical distributions. Models are meshed with tetrahedral (TET) elements of variable size to sufficiently discretise model geometries at different spatial scales while reducing model complexity. The ability of TET meshes to appropriately simulate the biphasic mechanical response of a single-cell model is established against that of a corresponding hexahedral mesh for an illustrative use case. To further demonstrate the flexibility of this tool, an explicit model was developed from three-dimensional confocal laser scanning image data, and a set of models were generated from a statistical cellular distribution of the articular femoral cartilage. The tools presented herein are free and openly accessible to the community at large.
    Computer Methods in Biomechanics and Biomedical Engineering 04/2014; 18(12). DOI:10.1080/10255842.2014.900545 · 1.79 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In cartilage tissue engineering studies, the stimulatory effect of a constant magnitude of mechanical perturbation declines after the first two weeks of culture. Similarly, it is known that chondrocyte-agarose constructs should not be loaded within the first days after seeding, to prevent considerable cell death, suggesting a mechanical threshold. This study aims to establish a relationship between chondrocyte deformation and death, and to evaluate the protective effect of the pericellular matrix (PCM) that is formed in 3D cultures. Chondrocyte viability was monitored every hour for 24 hours after applying a strain range of 0% to 25% to agarose constructs containing chondrocytes, cultured for 1, 3, 5, 7 or 10 days. At these culture time points, PCM thickness and chondrocyte deformation were assessed by means of histology and assayed for biochemical contents. Inverse finite element simulations were used to evaluate the change of mechanical properties of the chondrocyte and PCM over the 10 day culture duration. Chondrocyte death was demonstrated to be dependent on both the magnitude and duration of straining. The highest cell death was observed at day 1 (43%), reducing over culture duration (15% at day 3, and 2.5% at day 10). Cell deformation at 25% compression decreased significantly over culture duration (aspect ratio of 2.24 ± 0.67 at day 1 and 1.45 ± 0.24 at day 3) and with increased matrix production. Inverse finite element simulations showed an increasing PCM Young's modulus of 45 KPa at day 3 to 162 KPa at day 10. The current results provide evidence for a mechanical threshold for chondrocyte death and for the protective effect of the PCM. As such, these insights may help in establishing mechanical loading protocols for cartilage tissue engineering studies.
    Tissue Engineering Part A 01/2014; 20(13-14). DOI:10.1089/ten.TEA.2013.0436 · 4.64 Impact Factor
  • Ahmet Erdemir
    [Show abstract] [Hide abstract]
    ABSTRACT: Computational modeling and simulation (M&S) in biomechanics is extending its reach from research environments towards translational applications. In particular to the knee joint, finite element (FE) analysis has significant utility to understand joint and tissue function, explore pathological conditions and injury mechanisms, and investigate surgical interventions and implant performance [1]. Unfortunately, while FE models of joints, in particular of the knee, are ubiquitous in literature [1], they are not necessarily available publicly for the community to reuse and further develop. In response, Open Knee project, for FE analysis of the tibiofemoral joint, was launched [2].
    Journal of Medical Devices 12/2013; 7(4):0409101-409101. DOI:10.1115/1.4025767 · 0.62 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Anterior tears of the supraspinatus tendon are more likely to be clinically relevant than posterior tears of the supraspinatus. We hypothesized that anterior tears of the supraspinatus tendon involving the rotator cuff cable insertion are associated with greater tear gapping, decreased tendon stiffness, and increased regional tendon strain under physiologic loading conditions compared with equivalently sized tears of the rotator cuff crescent. Twelve human cadaveric shoulders were randomized to undergo simulation of equivalently sized supraspinatus tears of either the anterior rotator cuff cable (n = 6) or the adjacent rotator cuff crescent (n = 6). For each specimen, the supraspinatus tendon was cyclically loaded from 10 N to 180 N, and a custom three-dimensional optical system was used to track markers on the surface of the tendon. Tear gap distance, stiffness, and regional strains of the supraspinatus tendon were calculated. The tear gap distance of large cable tears (median gap distance, 5.2 mm) was significantly greater than that of large crescent tears (median gap distance, 1.3 mm) (p = 0.002), the stiffness of tendons with a small (p = 0.002) or large (p = 0.002) cable tear was significantly greater than that of tendons with equivalently sized crescent tears, and regional strains across the supraspinatus were significantly increased in magnitude and altered in distribution by tears involving the anterior insertion of the rotator cuff cable. These findings support our hypothesis that the rotator cuff cable, which is in the most anterior 8 to 12 mm of the supraspinatus tendon immediately posterior to the bicipital groove, is the primary load-bearing structure within the supraspinatus for force transmission to the proximal part of the humerus. Conversely, in the presence of an intact rotator cuff cable, the rotator cuff crescent insertion is relatively stress-shielded and plays a significantly lesser role in supraspinatus force transmission. Clinicians should consider early repair of rotator cuff cable tears, which may need surgical intervention to address their biomechanical pathology. In contrast, surgical treatment may be more safely delayed for rotator cuff crescent tears.
    The Journal of Bone and Joint Surgery 10/2013; 95(20):1817-24. DOI:10.2106/JBJS.L.00784 · 4.31 Impact Factor
  • Ahmet Erdemir
    [Show abstract] [Hide abstract]
    ABSTRACT: Prolonged mechanical loading of tissue in between a bony prominence and a support surface can lead to pressure ulcers. Despite recent initiatives to curb down incidence rates, the health care burden of pressure ulcer prevention remains significant [1]. Etiology of pressure ulcers are commonly attributed to interface pressures. As a result, interventions, e.g., support surfaces, routinely aim to reduce contact pressures. However, the clinical effectiveness of such an objective can be questionable [2]. Recent studies have shown that internal mechanics of the tissue can be associated with pressure ulcer development [3], potentially indicating the inefficacy of interventions targeted solely at contact pressure relief. Tissue characteristics at a bony prominence, e.g., tissue thickness and material properties, also influence load distribution within and on the surface of the tissue. Given the variability in patient populations and for a bony region of interest [4], it is possible that patient specific risk and load relief (with the use of support surface) may differ widely.
    ASME 2013 Conference on Frontiers in Medical Devices: Applications of Computer Modeling and Simulation; 09/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: Peripheral arterial disease (PAD), resulting from the accumulation of plaque, causes obstruction of blood flow in the large arteries in the arm and leg. In the United States, approximately 8.4 million people over the age of 40 have PAD [1]. If not treated, PAD can cause ischemic ulcerations and gangrene, which could eventually lead to amputation. Approximately, 25% of patients with PAD have worsening limb symptoms over 5 years, 7% requiring revascularization, and 4% requiring amputation [2].
    ASME 2013 Conference on Frontiers in Medical Devices: Applications of Computer Modeling and Simulation; 09/2013
  • Lealem Mulugeta, Ahmet Erdemir
    [Show abstract] [Hide abstract]
    ABSTRACT: Leading health institutions, government agencies, education and research institutions, as well as medical product developers around the world have recognized the substantial potential of computational modeling and simulation (M&S) to support clinical research and decision making in healthcare, e.g., [1]. Consequently, research activities in computational medicine are growing at a significant rate and notable discoveries are being made [2]. However, the mechanisms or processes necessary to appropriately translate these research activities and discoveries in computational methods to clinical practice are lacking. Moreover, there is substantial research diversity in the field such that subject matter experts within and across mathematical and biological disciplines tend to have their own interpretation of credible practice in M&S [3–5]. Additionally, tools and good practice guidelines established by individual disciplines or research areas do not readily transfer across other disciplines or are not adopted by different fields.
    ASME 2013 Conference on Frontiers in Medical Devices: Applications of Computer Modeling and Simulation; 09/2013
  • Snehal Chokhandre, Ahmet Erdemir
    [Show abstract] [Hide abstract]
    ABSTRACT: The tibiofemoral joint is a complex structure and its overall mechanical response is dictated by its numerous substructures at both macro and micro levels. An in-depth understanding of the mechanics of the joint is necessary to develop preventative measures and treatment options for pathological conditions and common injuries. Finite element (FE) analysis is a widely used tool in joint biomechanics studies focused on understanding the underlying mechanical behavior at joint, tissue and cell levels [1]. Studies, regardless of their purpose (descriptive or predictive), when employing FE analysis, require anatomical and mechanical data at single or multiple scales. It is also critical that FE representations are validated and closely represent specifics of the joint of interest, anatomically and mechanically. This is an utmost need if these models are intended to be used to support clinical decision making (in surgery or for rehabilitation) and for the development of implants.
    ASME 2013 Conference on Frontiers in Medical Devices: Applications of Computer Modeling and Simulation; 09/2013
  • Jason Halloran, Jack Andrish, Ahmet Erdemir
    [Show abstract] [Hide abstract]
    ABSTRACT: Patellofemoral complications remain the single largest reason for knee related clinical visits. Yet, robust clinical treatment remains a challenge [1]. To establish causal relationships and understand joint behavior, a complimentary approach utilizing simulation and experimentation may offer valuable insight. Simulation can be confirmed with experimental data and can also be exploited in a predictive capacity. For example, the medial patellofemoral ligament (MPFL) is a clinically relevant structure due to its role in patellofemoral stabilization [2]. MPFL reconstruction, which can be explored in a simulation framework, often utilizes a relatively stiff semitendinosus or gracilis tendon autograft [3]. The procedure is accepted to address patients with chronic patellar instability [4]. While joint stability may be achieved with such an approach, the underlying cartilage loading, and potential long term effects, are unknown. Previous simulation results found sensitivity in cartilage pressures during MPFL reconstruction [4], and these findings may be corroborated using a higher fidelity evaluation of clinically relevant factors. In the context of developing a general patellofemoral simulation framework, the goal of this study was to evaluate the effects of reconstructed MPFL zero force reference (“slack”) length on predicted joint mechanics across a range of potential values. To support the predictive simulation results, a preliminary model validation was also performed against specimen-specific in vitro joint mechanics.
    ASME 2013 Conference on Frontiers in Medical Devices: Applications of Computer Modeling and Simulation; 09/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: Understanding the mechanical behaviour of chondrocytes as a result of cartilage tissue mechanics has significant implications for both evaluation of mechanobiological function and to elaborate on damage mechanisms. A common procedure for prediction of chondrocyte mechanics (and of cell mechanics in general) relies on a computational post-processing approach where tissue-level deformations drive cell-level models. Potential loss of information in this numerical coupling approach may cause erroneous cellular-scale results, particularly during multiphysics analysis of cartilage. The goal of this study was to evaluate the capacity of first- and second-order data passing to predict chondrocyte mechanics by analysing cartilage deformations obtained for varying complexity of loading scenarios. A tissue-scale model with a sub-region incorporating representation of chondron size and distribution served as control. The post-processing approach first required solution of a homogeneous tissue-level model, results of which were used to drive a separate cell-level model (same characteristics as the sub-region of control model). The first-order data passing appeared to be adequate for simplified loading of the cartilage and for a subset of cell deformation metrics, for example, change in aspect ratio. The second-order data passing scheme was more accurate, particularly when asymmetric permeability of the tissue boundaries was considered. Yet, the method exhibited limitations for predictions of instantaneous metrics related to the fluid phase, for example, mass exchange rate. Nonetheless, employing higher order data exchange schemes may be necessary to understand the biphasic mechanics of cells under lifelike tissue loading states for the whole time history of the simulation.
    Computer Methods in Biomechanics and Biomedical Engineering 06/2013; DOI:10.1080/10255842.2013.809711 · 1.79 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Accurate prediction of plantar shear stress and internal stress in the soft tissue layers of the foot using finite element models would provide valuable insight into the mechanical etiology of neuropathic foot ulcers. Accurate prediction of the internal stress distribution using finite element models requires that realistic descriptions of the material properties of the soft tissues are incorporated into the model. Our investigation focused on the creation of a novel three-dimensional (3D) finite element model of the forefoot with multiple soft tissue layers (skin, fat pad, and muscle) and the development of an inverse finite element procedure that would allow for the optimization of the nonlinear elastic coefficients used to define the material properties of the skin muscle and fat pad tissue layers of the forefoot based on a Ogden hyperelastic constitutive model. Optimization was achieved by comparing deformations predicted by finite element models to those measured during an experiment in which magnetic resonance imaging (MRI) images were acquired while the plantar surface forefoot was compressed. The optimization procedure was performed for both a model incorporating all three soft tissue layers and one in which all soft tissue layers were modeled as a single layer. The results indicated that the inclusion of multiple tissue layers affected the deformation and stresses predicted by the model. Sensitivity analysis performed on the optimized coefficients indicated that small changes in the coefficient values (±10%) can have rather large impacts on the predicted nominal strain (differences up to 14%) in a given tissue layer.
    Journal of Biomechanical Engineering 06/2013; 135(6):61001-12. DOI:10.1115/1.4023695 · 1.75 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: High plantar pressures have been associated with foot ulceration in people with diabetes, who can experience loss of protective sensation due to peripheral neuropathy. Therefore, characterization of elevated plantar pressure distributions can provide a means of identifying diabetic patients at potential risk of foot ulceration. Plantar pressure distribution classification can also be used to determine suitable preventive interventions, such as the provision of an appropriately designed insole. In the past, emphasis has primarily been placed on the identification of individual focal areas of elevated pressure. The goal of this study was to utilize k-means clustering analysis to identify typical regional peak plantar pressure distributions in a group of 819 diabetic feet. The number of clusters was varied from 2 to 10 to examine the effect on the differentiation and classification of regional peak plantar pressure distributions. As the number of groups increased, so too did the specificity of their pressure distributions: starting with overall low or overall high peak pressure groups and extending to clusters exhibiting several focal peak pressures in different regions of the foot. However, as the number of clusters increased, the ability to accurately classify a given regional peak plantar pressure distribution decreased. The balance between these opposing constraints can be adjusted when assessing patients with feet that are potentially "at risk" or while prescribing footwear to reduce high regional pressures. This analysis provides an understanding of the variability of the regional peak plantar pressure distributions seen within the diabetic population and serves as a guide for the preemptive assessment and prevention of diabetic foot ulcers.
    Journal of Biomechanics 10/2012; 46(1). DOI:10.1016/j.jbiomech.2012.09.007 · 2.50 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In certain populations, open heart surgery to replace a diseased mitral valve is not an option, leaving percutaneous delivery a viable alternative. However, a surgical transcatheter based delivery of a metallic support frame incorporating a tissue derived valve puts considerable constraints on device specifications. Expansion to a large diameter from the catheter diameter without mechanical fracture involves advanced device design and appropriate material processing and selection. In this study, a new frame concept is presented with a desirable feature that incorporates wings that protrude during expansion to establish adequate fixation. Expansion characteristics of the design in relation to annulus fixation were quantified through finite element analysis predictions of the frame wing span and angles. Computational modeling and simulation was used to identify many favorable design features for the transcatheter mitral valve frame and obtain desired expansion diameters (35-45mm), acceptable radial stiffness (2.7N/mm), and ensure limited risk of failure based on predicted plastic deformations.
    Journal of Medical Devices 09/2012; 6(3):31005-31012. DOI:10.1115/1.4007182 · 0.62 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Patellofemoral complications are the single largest reason for knee related clinical visits. In spite of this, development of robust clinical treatments in this area remains a challenge [1]. Quantifying joint response across a wide range of conditions may lead to interventions specifically targeting desired or “normal” function. Previous patellofemoral studies often looked at joint mechanics as a function of specific quadriceps loaded flexion (e.g. deep knee bend) and/or during snapshots of loading representative of lifelike scenarios, e.g. gait, stair climb, etc. [2]. Sensitivity studies have been performed for these expected conditions [3,4] providing insight on the relationship between joint loading, geometry and potential contact mechanics. While patellofemoral biomechanics studies are prevalent, few, if any, have attempted to quantify joint response to systematic changes of two of the primary indicators of joint mechanics, namely quadriceps load and knee flexion. The overall joint response resulting from this type of approach could help quantify an envelope of natural function and also serves as an ideal data set for future computational model development. Once developed, probabilistic exploration of inherent uncertainties could be accomplished through a complimentary in vitro and in silico approach, offering quantification and classification of structure-function relationships. As a preliminary step, the goal of this study was to relate in vitro joint response, in terms of kinematics and contact mechanics, to systematic changes in knee flexion angle and quadriceps loading for a single specimen. Results from this study will offer insight into patellofemoral mechanics across a range of expected input and also serves as a starting point for future hypothesis driven studies.
    ASME 2012 Summer Bioengineering Conference; 06/2012

Publication Stats

846 Citations
127.72 Total Impact Points

Institutions

  • 2006–2015
    • Lerner Research Institute
      Cleveland, Ohio, United States
  • 2007–2014
    • Cleveland Clinic
      • Department of Biomedical Engineering
      Cleveland, Ohio, United States
  • 2002–2003
    • Pennsylvania State University
      • Center for Locomotion Studies
      University Park, Maryland, United States