Projects
My research focuses on human motion biomechanics; primarily using motion capture, force plates, EMG, and VO2 consumption to investigate the integration of clinical interventions (assistive devices, surgery, rehabilitation) with models of human movement; in order to understand and optimize functional outcomes for individuals with musculoskeletal movement pathologies.
Dissertation Work
Many individuals with walking disorders, due to neuromuscular disabilities, such as cerebral palsy (CP), use assistive devices, such as walkers or crutches, to aid their mobility. These devices are prescribed and utilized because they prevent the user from falling over. However, the stability gained by the user comes at the additional metabolic cost of manipulating the walker during walking. This increases the workload of walking on the user. There is a critical gap in the understanding of the interface between the walker and the user. Individuals with CP or other disabilities can have compromised motor control, impaired balance, and weakness, all of which can make ambulating difficult – even while using a walker. Currently, a major obstacle to designing devices that optimally aid users is the lack of understanding of how the user interacts with the walker.
The goal of this proposed work is to first understand the effect of using a walker on both the stability and efficiency of the gait of individuals with walking disorders, then devise and implement a control scheme that can add assistive forces to the user to improve the efficiency of gait, while maintaining the stability benefits of the device. First, we will quantify the kinematics and kinetics, as well as changes in stability and efficiency of walking with and without a passive posterior walker to gain a complete understanding of how gait changes while walking with a posterior walker. With this understanding, we will then develop a predictive simulation framework to characterize how assistive forces change gait. This framework will require developing a foot-floor contact model as well as defining the motion tracking optimization method and cost function. Finally, we will develop assistance methods inside of this predictive framework and apply a subset of those methods to a physical powered posterior walker to validate trends found in the predictive simulations.
Ultimately, this work will result in a validated predictive simulation framework to assess how possible assistance methods may alter the efficiency of a person’s gait. In the use case presented here, this framework is used to develop a posterior walker with motorized rear wheels driven by a control scheme optimized to increase efficiency of ambulation without losing stability benefits of the device. It is important to keep individuals with walking disabilities ambulatory for as long as possible. The significance of the proposed effort is to provide a posterior walker that reduces the workload of walking, keeping this population walking longer, providing critical exercise and continued muscle development. This will help the individual stay active longer, and perhaps prevent the use of a wheelchair. This will greatly improve their quality of life, providing key physiological, mental, and social benefits.
Beyond My Dissertation
Capturing the Foot
A Foot in a Cleat
In order to investigate the internal foot motion of athletes as close to in-game scenarios as possible, we needed a way to capture the foot as multiple segments, while the athlete was wearing cleats.
I developed a marker set that could capture the foot as four separate segments: hindfoot, medial midfoot, lateral midfoot, and toes. This marker set was applied to the cleat, as well as through the cleat to determine the amount of motion between the foot and the cleat.
Ultimately, this marker set was implemented in a project exploring how internal foot motion compares as athletes repeat the same drill on different turfs.
Ankle Fusion vs. Replacement
Goal: Identify changes in foot range of motion with ankle fusion (arthrodesis) and replacement (arthroplasty).
Total ankle arthroplasty (TAA), or ankle replacement is an increasingly popular surgical treatment of end-stage ankle arthritis, as this procedure aims to return patients to full function in the ankle joint. Ankle arthrodesis (AA), or ankle fusion, is the traditional surgical treatment for end-stage ankle arthritis and it eliminates mobility of internal ankle joints. TAA and AA are effective 2 years after implementation, but TAA has more improved patient reported outcomes.
During the fusion procedure, the mobility in the ankle joint is removed. This loss of motion in the ankle increases motion in the rest of the foot. In order to capture and study the internal foot motion of patients we need to implement a multi-segment foot model. Specifically, we are implementing the Oxford foot model in this case. The Oxford Foot Model is a validated 4-segment model consisting of the tibia (TB), hindfoot (HF), forefoot (FF), and hallux (HX, toes) connected by 3 spherical joints (Carson 2001).
In this study we will be examining the altered foot structure and resulting internal foot motion of TAA and AA candidates as they complete three ambulation tasks: walking on level ground, climbing stairs, and walking on a tilted surface. We do this through placing reflective markers on the bony landmarks of their feet and capturing this motion with our Vicon Motion Capture system. Tracked foot motion can then be used to drive our inverse kinematics model.
Moving forward we are investigating how foot kinematics change as patients recover from TAA and AA. We also are interested in comparing how the range of motion at the ankle changes with TAA versus AA. Ultimately, we will use the insight gained in this study to inform total ankle implant designs.
Initial findings from this work were presented at the 45th Meeting of the American Society of Biomechanics.
Foot-Ground Contact Modeling
How can we predict the joint kinetics through an athlete, given only the motion of the athlete?
How could we test different assistance methods before bringing subjects into the lab, so as to only have humans try the assistance most likely to actually help them?
Both of these questions hinge on the ability to predict the interactions between the human and the surface on which they are trying to move. In order to predict the reaction forces of these interactions, I am developing a simulation method that includes modeling foot-ground contact.
This work is developing a method of predicting ground reaction forces (GRF) in 3D, such that GRF are dependent only on the motion and mass of the model, and free to change when additional external forces are applied to the model in the simulation.
We have developed a full-body model that includes a ground contact model at each foot.
We then implement this model into the OpenSim Moco framework, which allows us to set up and run a torque driven simulation where the cost function of the optimization scheme is set to minimize:
(1) The deviation from the reference motion
(2) The control effort (the torque given to each joint)
Additionally, the controls between the pelvis and ground (trying to follow the external kinematic constraints (aka “Hand-of-God”)) are further penalized by weighting these controllers 10x the weight of all other controllers in the system.
This method results in reasonable predictions of the ground reaction force in 3D during walking.
Progress on this work will be presented at the 5th Meeting of the North American Society on Biomechanics in Ottawa, ON, Canada in August 2022.
Knee Support for Baseball Catchers
This work quantifies the effects of knee support on the lower-body joint kinematics and kinetics in the deep squat position.
This work aims to quantify the efficacy of knee support for baseball catchers in the deep squat position. Motion capture and analog force collection were used to develop a model for human squatting. This new model allowed us to determine the extent to which the application of the support of the knee, between the upper and lower leg while squatted, changed the joint moments of the lower body.
Little work has been done to examine the deep squat position (>130○ sagittal knee flexion). In baseball and softball, catchers perform this squat an average of 146 times per nine-inning game. To alleviate some of the stress on their knees caused by this repetitive loading, some catchers wear foam knee supports.
Subjects in this study performed the deep squat with no support, foam support, and instrumented support. In order to measure the force through the knee support, instrumented knee supports were designed and fabricated. We then developed an inverse dynamic model to incorporate the support loads. From the model, joint angles and moments were calculated for the three conditions.
There is a significant reduction in the sagittal moment at the knee of 43% on the dominant side and 63% on the non-dominant side when comparing the kinetics of the instrumented support condition with the support for active and absent. These reductions are a result of the foam supports carrying approximately 20% of body weight on each side.
Knee support reduces the moment necessary to generate the deep squat position common to baseball catchers. Given the short moment arm of the patella femoral tendon, even small changes in moment can have a large effect in the tibial-femoral contact forces, particularly at deep squat angles. Reducing knee forces may be effective in decreasing incidence of osteochondritis dissecans.
This work has been completed and published in the Journal of Biomechanics. The article was selected as the highlighted article in the publishing issue.
Work with Physicians in Orthopedics
>>> Note: Some photographs in this section contain images of cadaver specimens. <<<
Motion Analysis of 2-Tine Staple and K-Wire Fixation for Scapholunate Ligament Stabilization
Numeruous strategies for repair or reconstruction of the scapholunate interosseous ligament (SLIL) after an acute injury have been described in the literature, but no consensus currently exists regarding optimal treatment. Many of these treatment methods include the need for prolonged immobilization postoperatively, which may lead to loss of wrist range of motion. Previous studies have demonstrated benefits of two and four-tine staple fixation in SLIL reconstruction, including improved rotational control and avoidance of the articular surface. The purpose of this study was to compare scaphoid and lunate kinematics in a cadaver model after SLIL fixation with traditional Kirschner wire (K-wire) fixation or two-tine staple fixation.
Eight fresh cadaver arms with normal scapholunate (SL) intervals were included and infrared motion capture was used to assess kinematics between the scaphoid and lunate as the wrists were moved through a simulated dart throwers motion. Kinematic data was recorded for each wrist in four states: SLIL intact, SLIL sectioned, K-wire fixation across SL interval and scaphocapitate joint, and with two-tine staple fixation across SL interval. Strength of the SL staple fixation was evaluated using an axial load machine to assess load to failure of the staple construct across the SL interval.
Staple fixation across the SL interval more closely maintains physiologic motion of the scaphoid and lunate compared to K-wire fixation. Additionally, staple fixation provides adequate strength following injury to the SLIL as no primary failures of the staple were demonstrated at forces simulating a ground level fall. Therefore, staple fixation has the potential to allow incorporation of earlier wrist range of motion into rehabilitation following treatment of SLIL injury with staple fixation based on the results in a cadaver model.
This work has been completed and published in the Journal of Hand Surgery. My contribution to this work was designing and building the motion capture arena, collection of all motion capture data, and all analysis of wrist kinematics.
Antegrade vs. Retrograde Femoral Nailing Following Traumatic Femur Fracture
Femoral nailing is a surgical procedure conducted to repair a fractured femur. There are two main approaches within femoral nailing: antegrade (nail enters through hip) and retrograde (nail enters through knee). Choosing the correct procedure allows for easier nail insertion, prevents complications and affects fracture reductions. The antegrade procedure is more popular and used to fix proximal femur fractures. However, hip pain and hip abductor atrophy has been reported following surgery. The retrograde procedure has been shown to be better suited for obese patients and is often used for distal femur fractures, ipsilateral fractures of the femur and tibia, and patients with multiple fractures. Following retrograde procedures, quadriceps atrophy, as well as knee pain and stiffness have been reported.
In this study, we aim to quantify the difference in how well patients recover following antegrade and femoral nailing. We hypothesize the antegrade procedure will produce a greater disruption of the hip and knee kinematics and kinetics compared to a control group, meanwhile the retrograde procedure will cause a greater disruption of the knee kinematics and kinetics.
To accomplish this goal we will be collecting kinematics and kinetics of femoral nailing procedures at multiple time points through 6 months of their recovery. We will collect kinematics using our 18-camera Vicon Motion Capture system. We will collect kinetics using the 5 Bertec force plates embedded in the MAMP Lab floor.
The results of this study will provide insight to surgeons regarding the functional recovery of patients. This will give them another tool to utilize in deciding the best procedure to perform on future femur fracture patients.
Moving forward with this project we hope to expand this from a pilot project to a clinical study with a larger data set and evaluate patients through 18 months of recovery. We are also interested in investigating whether femoral nailing patients are at a greater risk of suffering additional intra-articular knee injuries following surgery.
The results from the pilot portion of this project were presented at the 44th Annual Meeting of the American Society of Biomechanics.