Brachial plexus birth injury (BPBI) is a traumatic perinatal neuromuscular injury causing muscle paralysis and lifelong arm impairment. Muscle paralysis in these children also leads to bone and joint consequences, including deformed growth of the scapula and humerus. BPBI occurs during a critical period of rapid musculoskeletal growth, but the parallel postnatal interactions of muscle and bone that drive these persistent deformities are not understood. Clinical reports and preliminary work suggest that short, contracted muscles after injury can alter mechanical loading of the shoulder consistent with observed bone deformity at macro- and microstructural levels. Altered active limb function with reduced range of motion and load bearing is also present; disuse is known to alter tissue growth and maturation. Finally, nerve injury in other conditions also affects bone growth directly, and direct effects in the postnatal period are not clear. Identifying appropriate targets for future treatment requires understanding which factors associated with altered bone and muscle development are most critical for driving altered growth. Almost nothing is known about the timing and progression of changes in underlying bone and muscle structure or metabolism following nerve injury occurring at birth to provide a foundation for clinical decision-making. Our primary hypothesis is that the bone deformity following BPBI is driven primarily by the mechanical environment, derived from impaired longitudinal growth of paralyzed muscle and altered active functional loading beginning shortly after injury. We will apply our unique rodent and computational models of BPBI that probe the separate contributions of nerve injury and muscle contracture to perform complementary assessments of the relative contributions of nerve injury, passive muscle loading, and active functional loading following BPBI to bone deformity. We will do so using 1) previously validated rat neurectomy models of brachial plexus injury and our unique disarticulation model of altered loading and 2) an integrated computational model to determine which specific features of bone deformity following BPBI are driven primarily by each potential driver. This R01 project, conducted by a multidisciplinary team with expertise in orthopedic surgery and biomechanical engineering, has high potential to elucidate the role of denervation in the parallel development of bone and muscle that occurs postnatally. Our innovative study design permits us to isolate both direct neural effects and indirect effects from altered passive and active mechanical loading on bone development in a way that has not previously been possible. Ultimately, this work has the potential to shift current research and treatment paradigms from an isolated focus on muscle as a treatment target to an integrated muscle and bone approach based on driving factors of deformity and loss of function. We anticipate our results will provide new candidates for improved treatment of BPBI and other neuromuscular injuries.

One of the defining characteristics of the human species is the unprecedented level of manual dexterity. The exquisite sensorimotor control of the hand enables manipulation of objects and tools to perform a variety of fundamental tasks, all with little conscious effort. Following a stroke, however, a majority of stroke survivors will experience permanent loss of manual dexterity despite months of therapy. In an effort to restore hand dexterity, advanced assistive devices (e.g., exoskeletons) have been developed. Unfortunately, few of these state-of-the-art devices have been translated to end users. One key factor limiting user acceptance is the lack of robust and personalized human-machine interfaces enabling intuitive control of the device. Surface electromyographic (EMG) signals hold promise for providing intuitive control of exoskeletons, but EMG robustness is far from ideal. Intrinsic interference, such as changes in EMG amplitudes at different arm postures, complicate signal interpretation. Moreover, current myoelectrically based control strategies typically rely on a triggering signal to initiate assistance or on pattern recognition to select a specific movement. These approaches are unreliable and can lead to user disengagement, thereby limiting functional improvement. In addition, the biomechanical properties of the musculoskeletal system are not captured in the majority of controllers. Without accommodating the neuromechanical constraints of stroke survivors, the control of assistive devices can be inaccurate and nonintuitive. As a result, hand exoskeletons have not fulfilled their assistive promise, and their clinical impact remains limited. Accordingly, we propose a new, transformative path, in which the neuromechanical principles underlying finger motor control are applied to drive a novel hand exoskeleton. We will develop a personalized hybrid (neural data- & model-based) interface that combines the decoded neural command with a musculoskeletal model. The developed interface will be used to control a soft-hard hybrid exoskeleton to enable dexterous finger movements in stroke survivors.

Our group has been working collaboratively with the Department of Defense to conduct case reviews of Warfighters injured in under-body blast loading events. These cases elucidated the mechanism of injury for a high-rate vertical loads and provided a valuable comparison point for experimental work. One limitation of that study was the paucity of follow-up information on the Warfighter injury outcomes. We propose to assess these injuries within the framework of immediate, short-term, and long-term Warfighter outcomes and functionality. Additionally, we would like to expand the injury set from high rate vertical loading to other commonly experienced in-theater injury types. Finally, we will develop a framework to assess functional deficit using a computational musculoskeletal model. To complete this work we have an existing collaboration between the Department of Defense, engineers with a background in injury biomechanics, and orthopaedic surgeons. Their clinical expertise will be critical to the success of this study.

Many colleges of engineering across the country have begun to use the Entrepreneurial Mindset (EM) in their teaching. The vocabulary of EM and entrepreneurially-minded learning (EML) has been developed and promoted by the Keen Foundation. We propose to bring EM and EML to NC State. Many faculty already approve of the concepts but are not yet well-versed in the specific educational methods and language of EM/EML. This project will host the first EM training on NC State's campus. The PI and co-PIs on this project will serve as local experts. An outside consultant with many years experience of teaching EM will come to campus to help us offer this first year of training. Participating faculty will be broken into two cohorts of seven who will meet with coaches over the year to develop new EM content in their own classes. Each participating faculty will publish their work to Engineering Unleashed to conclude their year with the program.

Rotator cuff injuries affect 4.5 million Americans and the increase in prevalence of rotator cuff injuries correlates with increasing age. Around 250,000 rotator cuff repairs are performed each year, equating to over three billion dollars in economic cost. Trauma to the rotator cuff results in a variety of injuries. The rotator cuff is critically important in activities of daily living (ADLs), and therefore damage to the rotator cuff can limit independence and quality of life. Rotator cuff tears are commonly treated with surgical reattachment of the tendon; however, surgery is not always advisable due to the high probability of retear. High repair tension during reattachment surgery is not only technically challenging during repair but also is often followed by post-surgical retear, limited functional capacity, and/or shoulder instability due to altered joint contact forces. Thus, a method to predict repair tension could play a key role in surgical selection. Repair tension is intraoperatively measured; therefore, it cannot inform the decision to operate. Surgeons currently must select patients for surgical repair based on other MR imaging parameters associated with poor outcomes and more severe injury. Following injury of the rotator cuff, the muscle-tendon unit undergoes physiological changes including retraction, muscle atrophy, and increased intramuscular fat and fibrous tissue. However, no research directly relating all three main clinical parameters typically used for determining surgical candidacy (intramuscular fat, muscle atrophy, and retraction) have been related to repair tension. Given the financial costs associated with surgical intervention and the physical burden experienced by patients, an improved understanding of the factors that influence rotator cuff injury and repair tension are needed to provide a higher level of care for surgery patients by positively impacting the quality of life for millions of Americans. Our overall objective is to understand the mechanical underpinnings of rotator cuff repair to identify relationships between physiological changes and post-surgical outcomes. This research will provide crucial information needed to inform likelihood of surgical success and better predict surgical outcomes, thereby guiding development of a standard of practice for rotator cuff repair surgery. The proposed research combines an experimental and computational framework to (1) determine the relationship between rotator cuff injury-induced physiological changes and repair tension, and (2) determine how physiological changes due to rotator cuff injury and elevated repair tension influences post-surgical outcomes.

Both the interpretation and clinical value of magnetic resonance (MR) imaging for muscle anatomy and structure are limited by the bottleneck that arises from muscle segmentation. In particular, because of the lack of contrast between connective tissue and muscles, discrimination of individual muscles is more burdensome and challenging than segmentation when adjacent tissues exhibit high contrast. For example, in rotator cuff injury, clinicians assess individual muscle size and fatty infiltration because they are indicators of both injury and likely response to treatment. However, manual segmentation of 3D muscle anatomy is too time-consuming for clinical demands. Thus, clinical assessments of muscle structure in rotator cuff injuries use a single, 2D image, at a location selected primarily for consistency of anatomic landmarks. Even limited to a single image, the level of segmentation accuracy and repeatability required for robust quantitative assessments has impelled the application of deep learning methods. While encouraging, the demonstrated success of deep learning for this specific application does not address the more fundamental problem that clinical measures derived from a single, 2D view do not accurately reflect either total volume or fatty infiltration for the rotator cuff muscles. Despite our capacity to acquire high quality images, there remains a clear lack of appropriate resources and methods for the automatic segmentation of upper limb muscles. As such, there is a critical need for effective, accurate, and fast methods for this purpose. Without such an advance, the obstacles associated with segmentation will continue to prevent valuable information, important for both diagnosis and treatment, from being integrated into clinical decision-making. Our long-term goal is to develop a shareable framework that enables accurate, automated segmentation of MR images of upper limb muscles, on a timescale that makes image analysis tractable for the clinic. The overall objective of this application, the next step toward attainment of this goal, is to leverage our existing, fully annotated, upper limb MR images in 38 healthy individuals and 10 persons with rotator cuff tears, to develop, assess, and share successful machine learning approaches. Our central hypothesis is that supervised methods trained on our unique data sets will outperform unsupervised approaches and that, when applied to clinical images, both accuracy and speed of quantitative analysis will be improved.

The goal of this KEEN Faculty fellowship is to integrate physics and other prerequisite concepts for just-in-time delivery to students in the introductory engineering mechanics sequence via an adaptive learning platform. This work will be extended through delivery at 3 North Carolina institutions (UNCC, NCAT, NCSU). We will perform rigorous evaluation of student success and faculty perceptions at 3 institutions, and disseminate learnings, and develop documentation needed for wider dissemination of course materials. This work will support student success through personalized learning support, and will accelerate adoption of new learning technology through collaborative innovation across the three universities.

Stroke is a leading cause of chronic disability in the United States, with nearly 610,000 new cases annually and over 7 million stroke survivors. Approximately 80% of stroke survivors are affected by hemiparesis, or muscle weakness on one side of the body. Hemiparetic walking is slow and asymmetric when compared to unimpaired gait. Preferred walking speeds following stroke range between < 0.2 m/s and ~0.8 m/s compared to ~1.4 m/s in healthy adults, and high interlimb asymmetry has been documented in ankle joint power output. These altered hemiparetic gait characteristics are associated with walking performance limitations including increased metabolic cost and altered joint loading. Increases in metabolic cost lead to rapid exhaustion, less activity and limited mobility. Additionally, changes in joint loading are associated with secondary diseases including osteoarthritis, lower-back pain, and due to reduced activity, cardiovascular disease. Interventions designed to reduce mechanical asymmetries may reduce metabolic cost and normalize joint loading for improved walking performance post stroke. Ankle exoskeletons have been used to reduce metabolic cost in healthy controls, and preliminary studies have demonstrated their potential for restoring ankle function by applying torque at the paretic ankle during the propulsive phase of gait. The long term goal of this research is to develop a robotic ankle exoskeleton that can improve walking performance post stroke in order to pave the way for a portable permanent walking aid. This goal is rooted in part on recent success reducing metabolic cost of intact human walking using a clutch-based elastic ankle exoskeleton. Exoskeleton design for stroke is limited, however, by a knowledge gap regarding the influence of exoskeleton assistance timing and magnitude on mechanical gait symmetry, metabolic cost, and joint contact loading. Understanding this relationship is critical because exoskeleton assistance provided at the wrong time, or of insufficient magnitude could be either useless, or even detrimental to walking performance. Although we cannot measure joint contact force experimentally in vivo, we will use a combined experimental and computational approach to investigate the following aims: (1) Determine how the timing and magnitude of exoskeleton assistance influence the mechanical symmetry and metabolic energetics of post-stroke walking. (2) Determine how timing and magnitude of exoskeleton assistance influences joint contact forces. Taken together, these aims provide new and essential information to enable the development of exoskeletons capable of minimizing comorbidities and restoring mobility for stroke survivors.

Brachial plexus birth palsy (BPBP) is a traumatic perinatal neuromuscular injury causing muscle paralysis that results in lifelong impairment of arm function. Muscle paralysis in these children leads to negative consequences for bones and joints including shoulder stiffness, deformed growth of the scapula and humerus, joint dislocation, and muscle weakness. These impairments greatly limit critical activities of daily living, such as eating and bathing. Notably, these musculoskeletal impairments, and their negative impact on function, persist even if the nerve recovers. Almost nothing is known about changes in underlying bone microstructure or metabolism following nerve injury occurring at birth. Perinatal paralysis occurs during a critical period of rapid musculoskeletal development and the profound bony deformities profoundly interfere with use of the arm, but the parallel postnatal development of muscle and bone that drives these persistent deformities and impaired function is not understood. Because the biomechanical and metabolic environment is so important to proper development of muscle and bone, detailed characterization of both muscle and bone structure and physiology in the same animals is essential for understanding the interactions that lead to deformity and loss of function following perinatal nerve injury. Our hypothesis is that nerve injury directly affects bone formation and metabolism, while restricted muscle growth induces additional deformity through mechanical and cell-signaling pathways. The proposed work will elucidate mechanisms by which deformity develops and persists following BPBP, which will ultimately enable us to develop more effective treatments for BPBP and other perinatal neuromuscular injuries that address both muscle and bone.

During the period of musculoskeletal development, the mechanical environment experienced by the developing bone and soft tissues plays a critical role in guiding proper cellular, mechanical, and morphological characteristics. Injury to nerve or musculoskeletal structures during this critical period is known to be associated with substantial changes that in many cases lead to permanent deformity or impaired function. Previous work in the field by us and other researchers have relied on animal models to characterize progression of injury during development, or have used computational models of either soft tissue or bone to understand the mechanical environment following injury in developing tissue. However, the reciprocal interactions between soft tissue and bone development are understudied, especially with regard to the influence on functional behavior. A computational framework that represents developmental changes to both bone and soft tissue in normal development and following injury and links these changes to joint function would provide a powerful platform for investigating the effects of altered mechanical environment on musculoskeletal structure and function. The aims of this pilot project are to Aim 1. Develop a novel platform for modeling musculoskeletal growth and loading (Integrated Iterative Musculoskeletal (I2M) modeling) using an integrated OpenSim and finite element analysis framework to describe and predict growth of individual tissues over time and derive functional outcome measures. Aim 2. Validate this growth modeling platform using two unique existing comprehensive datasets describing musculoskeletal growth: 1) Model of scapular development following neonatal nerve injury and 2) Model of knee joint growth following ligament injury This project would represent an innovative approach for exploring the effects of a variety of neuromusculoskeletal injury mechanisms developed using a strong foundation of unique data describing growth changes following both nerve and soft tissue injury. Such a platform will be broadly useful to researchers and clinical scientists interested in improving understanding of injury mechanisms during development and predicting response to rehabilitative interventions. Our team is a new collaboration with expertise in OpenSim model development and simulation, finite element analysis, and bone and soft tissue growth.