Experimental testing with cadaveric tissue allows the application of controlled loads and/or motions while still maintaining the inherent variability in the anatomy and soft tissue of the specimens. Multi-axial dynamic loading of tissue allows for experiments to be conducted that simulate conditions approaching physiological. Knee simulators have been used to generate physiological loading on the human knee to study kinematics, soft tissue loading, and joint contact pressure. These machines have been used to investigate injury, surgical outcomes, and prosthetic design. While there are a number of different geometries for knee loading devices, most are based on the Oxford rig design [1] with a vertical orientation of the leg where the hip is able to translate up and down while allowing flexion at the hip, knee, and ankle. The foot or ankle can have a variety of constraints and degrees of freedom. One of the recent areas of interest in knee biomechanics is the role different structures of the knee play during deep knee flexion activities. This is of particular interest to the orthopedic industry because of the common complaint regarding a feeling of a loss of stability during high flexion activities for post-TKR patients and the prevalence of high-flexion activities in emerging worldwide markets. The objective of this abstract is to describe two knee loading devices that have been used to study knee biomechanics, and most recently high flexion motion, and present some representative data from these tools.

This is the presentation of a prototype wrist, which has powered rotation and flexion. Powered flexion is an important addition to an externally powered prosthesis. Flexion and extension, along with rotation give the prosthesis two additional degrees of freedom, which when added to degrees of freedom for the shoulder and elbow yield a six-degree of freedom system. Six-degrees of freedom are important because they allow placement of the hand throughout the entire workspace. Without flexion, amputees cannot reach some positions, and must use compensatory motions for many other movements.

Verified computational models of total knee replacement serve as the primary design-phase tool for parametric analysis of implant geometry. Previously, dynamic finite element models of the Kansas Knee Simulator (KKS) were developed and tibiofemoral (TF) and patellofemoral (PF) kinematic predictions were verified by comparison with experimental measurements [1,2]. In this prior work, the implants were mounted in metallic fixtures to assess the ability of the model to accurately predict the TF and PF kinematics without the additional complexity of variable cadaver specimens and soft-tissue constraint. The next step in the systematic model verification procedure was to verify kinematic predictions with multiple specimen-specific models. Specifically, the objectives of the present study were: 1) to develop an explicit finite element (FE) model of the KKS capable of recreating experimental loading protocols for a deep knee bend activity and 2) to verify predicted six degree-of-freedom (DOF) TF and PF kinematics of two cruciate retaining (CR) and two posterior stabilized (PS) implanted specimen-specific models with deformable, wrapping soft tissue constraint.

Numerical simulation of blood flow in arteries has important applications in disciplines such as surgical planning, medical device design, and disease research. However, there are many challenges involved in accurately characterizing the motion of blood within the arterial system. First, computational models can require finite element meshes with several million degrees of freedom. Furthermore, surgical planning applications require solving these problems as quickly as possible. Second, the boundary conditions must faithfully represent the flow and pressure characteristics of the vascular trees external to the computational model. These conditions must also be consistent with the intrinsic wave propagation phenomena within the model. Third, the fluid-solid interactions between blood, vessel wall and surrounding tissues and organs need to be modeled. In order to represent the pulsation of blood within arteries, one must consider the variations in arterial compliance (especially significant when dealing with large models), points of attachment of the vessels to the surrounding organs, etc. Prior vascular fluid-structure interaction models have considered uniform vessel wall properties and a zero pressure reference state.

For some applications of musculoskeletal modeling it may be important to simulate the passive responses of joint range-of-motion limits. For example, Lemay and Crago [1] enforced dynamic limits of flexion and extension in an elbow model. In cases such as the elbow where the joint can be modeled as a simple hinge, the range of motion can be easily specified in terms of a minimum and maximum joint angle, and the motion limits can be enforced using simple visco-elastic restraining torques against any limit violations. For joints such as the shoulder girdle, however, which involve multiple articulating bones and multiple degrees of freedom, dynamic enforcement of joint range-of-motion limits requires a more complex approach.

Robotic technology has been adopted for studying the biomechanics of the knee joint by our research center and others since 1993 to gain a fundamental understanding of the knee, as well as to provide orthopedic surgeons with scientific data on the efficacy of various reconstructive techniques [1–4]. A robotic testing system generally consists of a six degree-of-freedom (DOF) robotic manipulator with a universal force-moment sensor (UFS) attached to the end effector [5]. This testing system offers a wide range of motion with precision path and position repeatability.

Musculoskeletal computational modeling can be a powerful and useful tool to study joint behavior, examine muscle and ligament function, measure joint contact pressures, simulate injury, and analyze the biomechanical results of reconstructive procedures. Commonly, biomechanical models are based on either finite element analysis (FEA) or three-dimensional rigid body dynamics. While each approach has advantages for specific applications, rigid body dynamics algorithms are highly efficient [1], thus significantly reducing solution time. Many musculoskeletal models of the elbow have been developed [2, 3], but all have constrained the articulations to have particular degrees of freedom and ignored the effects of ligaments. An accurate and robust model without these limitations has potential as a clinical tool to predict the outcome of injuries and/or surgical procedures. This work develops and validates an accurate computational model of the elbow joint whereby joint kinematics are dictated by three-dimensional bony geometry contact, ligamentous constraints, and muscle loading.

Knowledge of patient-specific muscle and joint contact forces during activities of daily living could improve the treatment of movement-related disorders (e.g., osteoarthritis, stroke, cerebral palsy, Parkinson’s disease). Unfortunately, it is currently impossible to measure these quantities directly under common clinical conditions, and calculation of these quantities using computer models is limited by the redundant nature of human neural control (i.e., more muscles than theoretically necessary to actuate the available degrees of freedom in the skeleton). Walking is a particularly important task to understand, since loss of mobility is associated with increased morbidity and decreased quality of life [1]. Though numerous musculoskeletal computer modeling studies have used optimization methods to resolve the neural control redundancy problem, these estimates remain unvalidated due to the lack of internal force measurements that can be used for validation purposes.

Whole body vibration (WBV) can induce a host of pathologies, including muscle fatigue and neck and low back pain [1,2]. A new model of WBV in the rat has been developed to define relationships between WBV exposures, kinematics, and behavioral sensitivity (i.e. pain) [3]. Although in vivo studies provide valuable associations between biomechanics and physiology, they are not able to fully define the mechanical loading of specific spinal regions and/or the tissues that may undergo injurious loading or deformation. Mathematical models of seated humans and primates have been used to estimate spinal loads and design measures that mitigate them during WBV [4–6]. Although such models provide estimates of relative spinal motions, they have limited utility for relating potentially pathological effects of vibration-induced kinematics and kinetics since those models do not enable simultaneous evaluation of relevant spinal tissues with the potential for injury and pain generation. As such, the goal of this work was to develop and validate a three degree of freedom (3DOF) lumped-parameter model of the prone rat undergoing WBV directed along the long-axis of the spine. The model was constructed with dimensions of a generalized rat and model parameters optimized using kinematics over a range of frequencies. It was validated by comparing predicted and measured transmissibility and further used to predict spinal extension and compression, as well as acceleration, during WBV for frequencies known to produce resonance in the seated human and pain in the rat [3,7].

Segmental spine instability — especially that of torsional nature — is difficult to diagnose in a degenerative lumbar spine in vivo . The motion of the lumbar segment is of a coupled nature and is described by a total of six degrees of freedom. Relative motion between two bodies in space is described by translations and rotations, and this includes the concept of the instantaneous axis of rotation (IAR). References 1–3 are representative reports of the many available in the literature about calculations of the IAR in cadaveric tissue, usually single motion segments [1–3]. In contrast, results of the IAR for lumbar motion measured in vivo (and by extension, for the whole lumbar spine ) are actually scarce. Based on the preliminary report presented by group [4], in the present study we describe the characteristics of the IAR for lumbar spine torsion measured in vivo in a much larger cohort of normal subjects through the application of the method originally described by Kinzel et al. [5].

When fibroblasts are added to a fibrin gel, the cells rapidly compact the gel and produce a fiber alignment pattern that depends in part on the cell traction forces, gel geometry, and gel mechanical constraints [1]. Over time the fibrin is digested and replaced with cell synthesized collagen and other extracellular matrix (ECM) proteins that follow the initial alignment pattern of the gel [2]. This remodeling process proceeds in a complex and integrated manner that is influenced by the mechanical environment [3]. In order to better understand fibroblast-fibrin interactions and the remodeling process, we obtained time-lapse images of the development of fiber alignment between clusters of dermal fibroblasts (i.e., explants) in a fibrin gel. The experimental results were then compared to a model that incorporated the effects of traction forces on ECM reorganization.

In-vitro simulation of active shoulder joint motion is critical to gaining an understanding of the effects of surgical procedures and implant designs. However, development of systems for the accurate simulation of active shoulder motion has lagged well behind those implemented for the lower limb and elbow, which have used principles of closed-loop joint angle control 1,4 . In contrast, active shoulder motion has been confined to simulators that can hold static joint angles through the application of loads based on computer model outputs 2 , or that use constant velocity of the middle deltoid while using open-loop control to apportion other muscle loads as a function of a-priori physiologic loading ratios 3 . Neither of these schemes utilizes real-time feedback of kinematic data in order to follow smooth, predefined profiles. The lack of more refined shoulder simulators, based on control theory, can primarily be attributed to the complexity of shoulder motion and the number of degrees of freedom (DOFs) ( i.e. plane of abduction, abduction angle, and axial rotation) which must be controlled.

In this study, we present an innovative methodology (Figure 1) to calculate patient specific tibio-femoral (TF) contact forces by integrating medical image data, 3D skin-mounted marker trajectories, ground reaction forces, electromyography (EMG) data and finite element analysis (FEA). The muscle redundancy problem is solved through an EMG-constrained optimization approach. Calculated muscle forces are input to a FEA to calculate TF contact forces. Kinematics of the degrees of freedom (DOFs) of the knee that cannot be accurately assessed from the trajectories of skin-mounted markers, are estimated using a novel iterative procedure which combines muscle force calculation with dynamic FEA. The presented methodology is applied to analyze TF contact forces of a walking trial performed on an instrumented treadmill of which the speed was sequentially ramped up and down. The results presented in this abstract will be validated against the in-vivo measured TF contact forces.

Computational models are needed to estimate soft tissue loads during movement. It would be ideal to perform such estimates on a subject-specific basis, where the information could be used clinically for assessing injury risk, planning treatment and monitoring rehabilitation outcomes. Musculoskeletal simulation software has evolved to the point that it is now relatively straight forward to estimate muscle forces needed to emulate subject-specific joint kinematics and kinetics [1]. These muscle forces can subsequently be used as boundary conditions in a knee mechanics model to estimate the associated ligament and cartilage loads [2]. However, this serial simulation approach may ignore inherent interactions between musculoskeletal dynamics and internal joint mechanics. That is, cartilage contact forces and ligament tension can potentially contribute to joint moment equilibrium [3]. Further, ligament stretch may allow joint kinematics to vary in a way that affects muscle moment arms and lines of action about the joint. Thus, it would be preferable if muscle, ligament and cartilage contact loads were estimated simultaneously so that these interactions are accounted for. The objective of this study was to incorporate a six degree of freedom tibiofemoral model into an existing subject-specific gait simulation framework [4]. In this study, we introduce the computational model, and then use it to track measured gait dynamics of a subject with an instrumented knee joint replacement. A comparison of the model-predicted and measured tibiofemoral contact forces provides a basis for assessing the validity of this novel co-simulation framework.

The two kinematic axes of the tibiofemoral joint, the flexion-extension (F-E) and longitudinal rotation (LR) axes [1], are unrelated to the anatomic landmarks often used to align prostheses during total knee arthroplasty (TKA) [1, 2]. As a result, conventional TKA changes the position and orientation of the joint line, thus changing the position and orientation of the F-E and LR axes and consequently the kinematics of the knee. However, the extent to which TKA changes these axes is unknown. An instrument that can measure the locations of and any changes to these axes is an instrumented spatial linkage (ISL), a series of six instrumented revolute joints that can measure the six degrees of freedom of motion (DOF) between two rigid bodies without constraining motion. Previously, we computationally determined how best to design and use an ISL such that rotational and translational errors in locating the F-E and LR axes were minimized [3]. However, this ISL was not constructed and therefore its ability to measure changes in the axes has not been validated. Therefore the objective was to construct the ISL and quantify the errors in measuring changes in position and orientation of the F-E axis.

The orientation of collagen fibers is essential for tissue mechanical functioning. Tissues are able to adapt this network to changes in the mechanical environment. Two mechanisms for this adaptation that have been proposed are cell-orientation dependent cell-traction (eg (6)), and strain-protected enzymatic collagen degradation (eg (11)). The premise is that these two mechanisms together are able to predict a transient and equilibrium responses of tissue adaptation to mechanical constraints. To evaluate this, they are captured in a numerical model and predictions are corroborated against distinct experimental observations. This abstract overviews the versatility of the model, using already presented (5) and new data.

A multi degree of freedom skeletal muscle system stimulated via optical control is presented. These millimeter-scale, optically excitable 3D skeletal muscle bio-actuators are created by culturing genetically modified precursory muscle cells that are activated with light: optogenetics. These muscle bio-actuators are networked together to create a distributed muscle system. Muscle systems can manipulate loads having no fixed joint. These types of loads include shoulders, the mouth, and the jaw.

A wheelchair-mounted robotic arm was designed and developed to enhance the capabilities of mobility-impaired persons with limited upper extremities limitations exceeding previous models specifications and performance [1]. The major enhancements of the wheelchair’s mechanical design are the incorporation of DC servo drive with encoders at each individual joint. The arm has seven degrees of freedom (DoF) and is side-mounted on a power wheelchair (fig 1). The control system allows coordinated Cartesian control, and offers expandability for research in combined mobility and manipulation. This paper discusses the control scheme and a virtual simulation of the existing WMRA prototype performing several activities of daily living.

Models to represent the mechanical behavior of the annulus fibrosus are important tools to understand the biomechanics of the spine. Many hyperelastic constitutive equations have been proposed to simulate the mechanical behavior of the annulus that incorporate the anisotropic nature of the tissue. Recent approaches [1,2] have included terms into the energy function which take into account fiber-fiber and fiber-matrix interactions, leading to complex functions that cannot be readily implemented into commercial finite element codes for an efficient simulation of nonlinear realistic models of the spine (which are generally composed of 100,000+ degrees of freedom). An effort is undertaken here to test the capability of a relatively simple strain energy function [3] for the description of the annulus fibrosus. This function has already been shown to successfully represent the mechanical behavior of the arterial tissue and can be readily implemented into existing finite element codes.

Musculoskeletal computer models capable of predicting muscle and joint contact forces accurately during human movement could facilitate the design of improved joint replacements and new clinical treatments for articular cartilage defects or movement-related disorders [1]. A primary challenge to developing such predictions is the non-uniqueness of the calculated muscle forces, often referred to as the “muscle redundancy problem” [2]. Since more muscles act on the skeleton than the number of degrees of freedom in the skeleton, an infinite number of possible muscle force solutions exist.