In a world of war and turmoil in developing nations, land mines are becoming a concern, as millions of them are scattered in countries all over the world. Moreover, land mine prevention programs cannot clear land mine fields as fast as they are detonated each day. As a result, there are thousands that fall victim each year. There is immense demand for newer technologies to replace the aging prostheses used in these war torn nations. The active prosthetic knee (APK) design project is a novel design that utilizes simple, robust one degree of freedom (DOF) mechanics, while providing fully active knee torque control. The APK utilizes a high-speed brushed servomotor, providing the necessary torque and dynamics to provide the necessary gait characteristics of human locomotion. The main research contributions of this thesis are: 1) the mechanics and 2) the intelligence of the APK. This thesis investigates and highlights the prosthetic’s design process. The human biological system is studied and used as the design criteria when designing the device. Anthropometric data was used to determine the sizing and other physical properties. Moreover, Adaptive-Network-based Fuzzy-Interference Systems (ANFIS) and Fuzzy Interference Systems (FIS) are used to provide control to the APK. Finite element analysis (FEA) verifies the structural integrity of the APK. Four simulations are analyzed: equivalent stress, equivalent strain, shear stress and total deformation. These four simulations provide a mathematical interpretation of the physical system. We found that the first prototype, although a sound design, can be further improved to allow greater loading capabilities. Moreover, cyclical loading and total life cycles would also be significantly improved. A modular test stand is also designed and prototyped to allow APK testing. Since the APK prototype cannot be immediately placed on a human test subject, the test stand allows for experimentation in replicating human gait cycles.

Purpose: Biomechanical differences between people with low back pain (LBP) and healthy controls have been shown previously. LBP has been associated with standing postures in occupational settings. A transient pain-generating model allows for comparisons between pain developers (PD) and non-pain developers (NPD). The first objective was to utilize a multifactorial approach to characterize differences between PD and NPD individuals. The second objective was to investigate the impact of exercise on LBP during standing. Methods: Forty-three participants without any history of LBP volunteered for this study. Participants performed pre- and post-standing functional movements and 2-hours of standing. Continuous electromyography (EMG) data were collected from 16 trunk and hip muscles, kinematic and kinetic data were used to construct an 8-segment rigid link model. Vertebral joint rotation stiffness (VJRS) measures were calculated. Participants completed visual analog scales (VAS) rating LBP every 15 minutes during the 2-hr standing. Participants were classified as PD or NPD based on greater than 10 mm increase in VAS. Participants were assigned to exercise (EX) or control (CON) groups. All participants returned for a second data collection following 4-weeks. Results: Forty percent of participants developed LBP during standing. The PD group had elevated muscle co-activation prior to reports of pain (p < 0.05). Following standing, there was a decrease in VJRS about the lateral bend axis during unilateral stance. PDEX had decreased VAS scores during the second data collection (p = 0.007) compared with PDCON. Male PDEX had decreased gluteus medius co-activation during standing (p < 0.05). Between-day repeatability for the CON groups was excellent with intraclass correlation coefficients > 0.80 for the majority of the outcome measures. Conclusions: There were clear differences between PD/NPD groups in muscle activation patterns, prior to subjective reports of LBP, supporting the hypothesis that some of the differences observed between these groups may be predisposing rather than adaptive. An exercise intervention resulted in positive changes in the PD group, both in subjective pain scores as well as muscle activation profiles. Elevated muscle co-activation in the first 15-30 minutes of standing may indicate that an individual is at increased risk for LBP during standing.

Low back pain is highly prevalent in the developed world, with 80% of the population being affected at some point in their lives. Herniation, a common injury to the intervertebral disc, is characterized as the posterior migration of the nucleus pulposus through the layers of the annulus fibrosus. Various risk factors have been associated with the development of disc herniation, but the mechanisms are largely not understood. For example, exposure to vibration has been linked to the occurrence of herniation, yet our understanding of this association is not clear. It is hypothesized that vibration cyclically loads the tissues of the intervertebral disc until failure occurs as a result of fatigue. Tissues at risk of fatigue failure may include the intra-lamellar matrix, the connective tissue found between collagen fibres within a single lamella, and the inter-lamellar matrix, the connective tissue found between adjacent lamellae. In order to determine the mechanistic link between vibration and herniation, a firm understanding of the properties of the intervertebral disc as well as the intra and inter-lamellar matrices are of utmost importance. Further, it is important to determine these properties under physiological loading scenarios. This thesis consists of five studies, which have each provided a unique piece to the intervertebral disc herniation puzzle in order to better understand this mechanistic link. First, it was discovered that annular tissue is subject to significantly higher stresses and is stiffer under biaxial strain as compared to uniaxial strain. Biaxial strain is more representative of the in vivo loading scenario and provides more accurate information regarding scenarios that the annulus can tolerate and those that can result in injury. It was also revealed that, when strained at physiological strain rates (up to 4%/sec), these mechanical properties do not change such that they are independent of strain rate. Therefore, when strained at varying rates akin to voluntary movement, the annulus is not subject to higher stresses or altered stiffness. Second, the effect of vibration, an acknowledged risk factor for herniation, was examined on the mechanical properties of the intra and inter-lamellar matrices. It was discovered that vibration altered these matrices such that they were more extensible and strained to greater magnitudes, yet did not reach higher stresses before failing. It was hypothesized that this increased extensibility was due to damage to elastin, as elastin assists in minimizing tissue deformation and helps tissues recover from tensile strain. The final study assessed the effect of exposure to vibration on the development of disc herniation. The initiation of herniation was observed in a significantly greater number of intervertebral discs exposed to vibration as compared to a control condition. Although epidemiological studies had documented a correlation between exposure to vibration and herniation, this was the first study to conclude that exposure to vibration is in fact a mechanical risk factor for the development of herniation and increases the incidence of herniation. Further, based on the findings of the mechanical properties of the intra and inter-lamellar matrices, and in particular the observed 15-20 times greater failure strength of the intra as compared to inter-lamellar matrix, it would appear that the inter-lamellar matrix, and thus delamination, may be the weakest link in the herniation pathway. This thesis has uncovered new information regarding physiological mechanical properties of the annulus. Further, new information regarding the intra and inter-lamellar matrices was obtained, improving our understanding of the healthy disc. Last, by subjecting the disc to a known risk factor for herniation, hypotheses were generated regarding the initiation and progression of disc herniation, specifically related to the roles of the intra and inter-lamellar matrices.

General body discomfort increases over time during prolonged sitting and it is typically accepted that no single posture can be comfortably maintained for long periods. Despite this knowledge, workplace exposure to prolonged sitting is very common. Sedentary occupations that expose workers to prolonged sitting are associated with an increased risk of developing low back pain (LBP), disc degeneration and lumbar disc herniation. Given the prevalence of occupations with a large amount of seated work and the propensity for a dose-response relationship between sitting and LBP, refining our understanding of the biomechanics of the lumbar spine during sitting is important. Sitting imposes a flexed posture that, when held for a prolonged period of time, may cause detrimental effects on the tissues of the spine. While sitting is typically viewed as a sedentary and constrained task, several researchers have identified the importance of investigating movement during prolonged sitting. The studies in this thesis were designed to address the following two global questions: (1) How do the lumbar spine and pelvis move during sitting? (2) Can lumbar spine movement and postures explain LBP and injury associated with prolonged sitting? The first study (Study 1) examined static X-ray images of the lower lumbo-sacral spine in a range of standing and seated postures to measure the intervertebral joint angles that contribute to spine flexion. The main finding was that the lower lumbo-sacral joints approach their total range of motion in seated postures. This suggests that there could be increased loading of the passive tissues surrounding the lower lumbo-sacral intervertebral joints, contributing to low back pain and/or injury from prolonged sitting. Study 2 compared external spine angles measured using accelerometers from L3 to the sacrum with corresponding angles measured from X-ray images. While the external and internal angles did not match, the accelerometers were sensitive to changes in seated lumbar posture and were consistent with measurements made using similar technology in other studies. This study also provided an in-depth analysis of the current methods for data treatment and how these methods affect the outcomes. A further study (Study 3) employed videofluoroscopy to investigate the dynamic rotational kinematics of the intervertebral joints of the lumbo-sacral spine in a seated slouching motion in order to determine a sequence of vertebral motion. The pelvis did not initiate the slouching motion and a disordered sequence of vertebral rotation was observed at the initiation of the movement. Individuals performed the slouching movement using a number of different motion strategies that influenced the IVJ angles attained during the slouching motion. From the results of Study 1, it would appear as though the lowest lumbar intervertebral joint (L5/S1) contribute the most to lumbo-sacral flexion in upright sitting, as it is at approximately 60% of its end range in this posture. However, the results from Study 3 suggest that there is no consistent sequence of intervertebral joint rotation when flexing the spine from upright to slouched sitting. When moving from standing to sitting, lumbar spine flexion primarily occurs at the lowest joint (i.e. L5/S1); however, a disordered sequence of vertebral motion the different motion patterns observed may indicate that different joints approach their end range before the completion of the slouching movement. In order to understand the biomechanical factors associated with sitting induced low back pain, Study 4 examined the postural responses and pain scores of low back pain sufferers compared with asymptomatic individuals during prolonged seated work. The distinguishing factor between these two groups was their respective time-varying seated lumbar spine movement patterns. Low back pain sufferers moved more than asymptomatic individuals did during 90 minutes of seated work and they reported increased low back pain over time. Frequent shifts in lumbar spine posture could be a mechanism for redistributing the load to different tissues of the spine, particularly if some tissues are more vulnerable than others. However, increased movement did not completely eliminate pain in individuals with pre-existing LBP. The LBP sufferers’ seated spine movements increased in frequency and amplitude as time passed. It is likely that these movements became more difficult to properly control because LBP patients may lack proper lumbar spine postural control. The results of this study highlight the fact that short duration investigations of seated postures do not accurately represent the biological responses to prolonged exposure. Individuals with sitting-induced low back pain and those without pain differ in how they move during seated work and this will have different impacts on the tissues of the lumbar spine. A tissue-based rational for the detrimental effects on the spinal joint of prolonged sitting was examined in Study 5 using an in vitro spine model and simulated spine motion patterns documented in vivo from Study 4. The static protocol simulated 2 hours of sitting in one posture. The shift protocol simulated infrequent but large changes in posture, similar to the seated movements observed in a group of LBP sufferers. The fidget protocol replicated small, frequent movements about one posture, demonstrated by a group of asymptomatic individuals. Regardless of the amount of spine movement around one posture, all specimens lost a substantial amount of disc height. Furthermore, the passive range of motion of a joint changed substantially after 2 hours of simulated sitting. Specifically, there were step-like regions of reduced stiffness throughout the passive range of motion particularly around the adopted “seated flexion” angle. However, small movements around a posture (i.e. fidgeting) may mitigate the changes in the passive stiffness in around the seated flexion angle. The load transferred through the joint during the 2-hour test was varied either by changing postures (i.e. shifting) or by a potential creep mechanism (i.e. maintaining one static posture). Fidgeting appeared to reduce the variation of load carriage through the joint and may lead to a more uniform increase in stiffness across the entire passive range of motion. These changes in passive joint mechanics could have greater consequences for a low back pain population who may be more susceptible to abnormal muscular control and clinical instability. Nevertheless, the observed disc height loss and changes in joint mechanics may help explain the increased risk of developing disc herniation and degeneration if exposure to sitting is cumulative over many days, months and years. In summary, this work has highlighted that seated postures place the joints of the lumbar spine towards their end range of motion, which is considered to be risky for pain/injury in a number of tissue sources. In-depth analyses of both internal and external measurements of spine postures identified different seated motion patterns and self-selected seated postures that may increase the risk for developing LBP. The model of seated LBP/discomfort development used in this thesis provided evidence that large lumbar spine movements do not reduce pain in individuals with pre-existing LBP. Tissue-based evidence demonstrated that 2 hours of sitting substantially affects IVJ mechanics and may help explain the increased risk of developing disc herniation and degeneration if exposure to sitting is cumulative over many days, months and years. The information obtained from this thesis will help develop and refine interventions in the workplace to help reduce low back pain during seated work.

Abstract: Previous lower limb weighting studies have placed a load on the legs bilaterally and tested different placement locations. It was previously determined that kinematic changes occur with greater masses and at joints proximal to weight placement [1]. Other studies have determined that these changes exist for a short adaptation period before parameters revert to a steady state [2]. Tasks that require voluntary gait modifications such as obstacle clearance have also been performed with lower leg bilateral weight addition [4]. In cases of normal obstacle clearance increased flexion at all three joints in the lower limb is needed to safely traverse the obstacle [3]. The goal of this study was to investigate joint kinematics and kinetics of unilaterally weighted participants using level ground force platform collection techniques, rather than a treadmill. It was hoped that this would allow for new insight into the adaptation periods and strategic motor pattern changes seen at the ankle, knee and hip. Kinematic and force platform data were collected on two groups of 10 healthy male subjects. Group 1 (mean age = 23years, mean weight = 82.181kg, mean height = 1.798m) was a normal walking group and group 2 (mean age = 24.8years, mean weight = 79.901kg, mean height = 1.773m) was an obstacle clearance group. Both groups participated in 20 trials each of three different conditions; normal, weighted and weight off using a 2.27kg limb mass attached just proximal to the right maleoli markers. A repeated-measures two-way ANOVA was carried out on relevant variables in order to determine statistical significance. Weight addition and removal affected the kinematics and kinetics of the normal walking and obstacle clearance groups. This effect was more prominent in the normal walking group. If changes were seen, trials 1 through 3 were the locations showing a quick adaptation followed by a leveling off back to a new steady state in later trials. Participants in the normal walking group chose to utilize the hip joint in order to control for weight addition and removal. Kinematically, changes in the hip joint angle occurred at all instances analyzed throughout the gait cycle with this effect being more prominent in the weight off condition. In conjunction with this, the hip joint energy generation increased during all phases of the gait cycle while the ankle and knee joints either decreased energy generation or increased energy absorption. In the obstacle group, participants also chose to increase flexion at the hip joint. However, the ankle joint also had either decreased plantarflexion or increased dorsiflexion at all the instances analyzed during the gait cycle. However, joint energy generation increases at these joints were only found during stance and at heel contact. The toe obstacle clearance values also showed a marked increase in trial 1 for the weighted condition which demonstrates a voluntary gait modification made by participants to safely traverse the obstacle that was quickly adapted for. Overall, the results found by previous studies using treadmill collection techniques were still seen in overground force platform data but they were not as robust. References: 1.Martin PE et al. J Biomech. 1990; 23(6):529-536. 2.Noble et al. Exp Brain Res. 2006; 169: 482-495. 3.Patla AE et al. Exp Brain Res. 1995; 196: 499-504. 4.Reid MJ et al. Neurosci Res Comm. 2001; 29(2): 79-87.

In this Thesis we develop mathematical models to analyze two proposed causative mechanisms for the ventricular expansion observed in hydrocephalus: cerebrospinal fluid pulsations and small transmantle pressure gradients. To begin, we describe a single compartment model and show that such simple one-dimensional models cannot represent the complex dynamics of the brain. Hence, all subsequent models of this Thesis are spatio-temporal. Next, we develop a poroelastic model to analyze the fluid-solid interactions caused by the pulsations. Periodic boundary conditions are applied and the system is solved analytically for the tissue displacement, pore pressure, and fluid filtration. The model demonstrates that fluid oscillates across the brain boundaries. We develop a pore flow model to determine the shear induced on a cell by this fluid flow, and a comparison with data indicates that these shear forces are negligible. Thus, only the material stresses remain as a possible mechanism for tissue damage and ventricular expansion. In order to analyze the material stresses caused by the pulsations, we develop a fractional order viscoelastic model based on the linear Zener model. Boundary conditions appropriate for infants and adults are applied and the tissue displacement and stresses are solved analytically. A comparison of the tissue stresses to tension data indicates that these stresses are insufficient to cause tissue damage and thus ventricular expansion. Using age-dependent data, we then determine the fractional Zener model parameter values for infant and adult cerebra. The predictions for displacement and stresses are recomputed and the infant displacement is found to be unphysical. We propose a new infant boundary condition which reduces the tissue displacement to a physically reasonable value. The model stresses, however, are unchanged and thus the pulsation-induced stresses remain insufficient to cause tissue damage and ventricular expansion. Lastly, we develop a fractional hyper-viscoelastic model, based on the Kelvin- Voigt model, to obtain large deformation predictions. Using boundary conditions and parameter values for infants, we determine the finite deformation caused by a small pressure gradient by summing the small strain deformation resulting from pressure gradient increments. This iterative technique predicts that pediatric hydrocephalus may be caused by the long-term existence of small transmantle pressure gradients. We conclude the Thesis with a discussion of the results and their implications for hydrocephalus research as well as a discussion of future endeavors.

A great deal of research has been done on the adaptive strategies of individuals who have been affected by a gait altering ailment, but there is little research on the adaptive strategies to imposed restrictions in the healthy population. The role of the ankle in healthy gait is to generate a “push-off” force to create forward propulsion of the body (Winter, 2004). The purpose of this thesis was to identify adaptation patterns and compensation strategies in individuals while wearing and not wearing a device to reduce ankle motion(Ankle Motion Minimizer – AMM). Motion capture and force plate data were collected to determine the lower body kinematics and joint powers during both level ground walking and obstacle avoidance tasks. Repeated Measure ANOVAs with an alpha level of 0.05 determined that differences in the ankle angles and the ankle, knee, and hip powers existed between the various conditions. Results showed that participants had a decreased range of motion and power production at the ankle joint while wearing the AMM. Meanwhile, an increase in the power bursts from the ipsilateral knee were observed during the AMM conditions as well as small increases at the contralateral ankle and ipsilateral hip during the unilateral AMM condition. EMG analysis showed a distinct muscle activation pattern for each individual muscle during the different conditions. From this investigation, individuals who are unable to produce power through the ankle joint, were able to increase power propulsion predominately at the knee to compensate for the lack of propulsion provided by the ankle, therefore allowing ambulation to continue.

Maximum voluntary forces and psychophysically acceptable forces are often used to set force guidelines for exertions as a means to protect against overexertion injuries in the workplace. The focus of this dissertation was the exploration of the roles of whole body balance, shoe-floor friction and joint strength in limiting the capacity of a person to produce maximum voluntary hand forces and psychophysically acceptable hand forces. The underlying goal was to advance knowledge regarding how physical exertion capacity is biomechanically governed, then to use this information to develop models to predict capability based on these governing principles. The hypothesis underscoring this work was that maximum voluntary hand force capability is governed by whole body balance, shoe-floor friction and joint strength; and consequently, psychophysically acceptable forces would be chosen proportionally to this maximum voluntary force capability, where the magnitude of the proportionality was dependent on the limiting factor, or ‘weakest link’. To investigate this hypothesis, both experimental and mathematical modeling paradigms were used. Initially, an experimental study was used to investigate how biomechanical factors governed maximum hand force capability across a range of exertions. It revealed that each governing factor differentially limited maximum force capability. Moreover, this study identified how foot placement, handle height, distance from the handle, friction, and body posture all influence the underlying biomechanical weakest link, and ultimately force producing capability. Data gathered in the experimental study was next used to evaluate a mathematical model that was developed to predict maximum force capability, given information on posture and direction of force application. In addition, the model also predicted population variability in maximum capacity based on the inclusion of a novel approach to probabilistically represent population variability. The evaluation demonstrated that the model underestimated maximum hand force capability compared to measured hand forces by approximately 18, 26, and 41% during medial, pulling and downward exertions respectively. However, it appeared that the ‘weakest link’ principle for predicting maximum force capacity was plausible, as evidenced by significant rank ordered correlations between the measured and predicted hand forces. Further research investigated if psychophysically acceptable forces were selected as a proportion of task specific maximum voluntary force capability, where the proportionality was related to the biomechanical weakest link. Using an experimental design, psychophysically acceptable forces and corresponding maximum forces were measured. Participants chose psychophysically acceptable forces that were 4/5ths of their task specific maximum voluntary force capability when capability was limited by balance. Additionally, they choose psychophysically acceptable forces that were 2/3rds of their maximum voluntary force capability when capability was limited by joint strength. The identification and confirmation of a weakest link proportionality principle represents an important contribution to the field of occupational biomechanics. The weakest link proportionality principle was integrated into the model to allow prediction of: maximum voluntary hand force capability, the limiting factor, and psychophysically acceptable hand force capability. The updated model underestimated empirically measured psychophysically acceptable forces by 24% and 43% during downward and pulling exertions respectively. However, the original model underestimated the maximum hand force capacity by 23% and 34% during the same exertions, without the proportional relationships. This underestimation may be a result of the underlying assumption that joint strength is independent, resulting in an underestimation of maximum joint strength capacity and a corresponding underestimation of maximum hand force capacity. The underestimation may also be due to differences in strength capacities between the participants tested during this thesis compared to those tested in past research used to determine the maximum strength indices reported in the literature. This body of work supported the hypothesis that psychophysically acceptable forces are selected as a proportion of the maximum voluntary hand force, where the proportionality depends on the underlying biomechanical weakest link. The model is a promising first step towards predicting maximum and psychophysically acceptable occupational force threshold limits.

One of the most poorly understood co-morbidities associated with obesity is the pathway to osteoarthritis of the knee. To implement appropriate preventative strategies, it is important to explore how obesity is a causal factor for osteoarthritis. The present research compared the kinematics and kinetics of a group of young obese, but otherwise healthy, adults to a group of young, healthy-weight adults, in an attempt to identify mechanical abnormalities at the knee during walking that may predispose the obese to osteoarthritis of the knee. Optotrak motion capture (Northern Digital Inc. Waterloo, Ontario) and a forceplate (AMTI OR6-7, Advanced Mechanical Technology Inc, Watertown, MA) were used to measure ground reaction forces and moments of 16 participants – 8 obese and 8 sex-, age- and height-matched healthy-weight – to analyze knee joint kinematics and kinetics at three walking speeds. Participants wore an accelerometer (ActiGraph GT3X, Fort Walton Beach, USA) for seven days to measure daily steps counts. Dependent t-tests were performed to determine group differences in ground reaction forces, knee angles and knee moments, as well as knee adduction moment impulse and cumulative knee adductor load (CKAL). The obese group walked at a significantly slower self-selected speed (p=0.013). While not statistically significant, the obese group did present with a more valgus mean dynamic knee alignment than the health-weight group. A significantly greater maximum abduction angle (p=0.009) and smaller minimum knee flexion angle at heel contact (p=0.001) was found in the obese group. A significant difference was found in the peak medial rotation moment in the transverse plane (p=0.003). A greater stance duration lead to a significantly greater knee adduction moment impulse (p=0.049) in the obese group. While significant group differences were not found in the steps per day, the obese group had a significantly greater CKAL (p=0.025). Obese young adults with healthy knees demonstrated a gait pattern of reduced medial knee joint compartment loading through greater knee abduction, medial knee rotation and a slower walking speed compared to matched controls. The ramifications of gait modifications on long-term musculoskeletal health remain unknown, but compensations may lead to increased risk of osteoarthritis of the knee.

The present research quantified the shoulder rhythm for arm postures that represent the right-handed reachable workspace and compared 3 methods of scapular tracking: acromion marker cluster (AMC), stylus and scapular locator. The shoulder rhythm models can be incorporated into existing and future shoulder biomechanical models to determine shoulder geometry when simulating postures experienced in workplaces and thus have ergonomic implications for correctly identifying risk factors. The results of this research also provide guidance for future studies involving scapular tracking. Fourteen male and 14 female participants performed static arm postures spread over 5 elevation angles: 0, 45, 90, 135, 180 degrees, three elevation planes: 0, 45, 90 degrees to the frontal plane and, three axial rotations: maximum internal, neutral, and maximum external rotation. Kinematic data was recorded using a Vicon MX20+ motion-tracking system. Bone rotations were calculated using Euler angles and continuous prediction models were generated to estimate scapular and clavicular orientations based primarily on thoracohumeral relative orientations. Methods of scapular tracking were compared using repeated measures analysis of variance. Participant characteristics did not influence any of the scapular or the clavicular angles (p>.05). Axial rotation did not influence scapular retraction/protraction and elevation plane did not influence clavicular elevation (p>.05). Elevation angle was the largest contributor to lateral rotation and posterior tilt of the scapula and all clavicular angles. Plane of elevation was the largest contributor to scapular protraction. Using the stylus as the gold standard, the locator and the AMC underestimated lateral rotation, with a maximum difference of 11 degrees and 9 degrees between the locator and the stylus and AMC and the stylus measurements, respectively. The AMC and the locator overestimated posterior tilt at overhead postures and underestimated it at low elevation angles. The maximum difference between the AMC- and the locator- and the stylus-measured tilt was 10 degrees. The scapular locator consistently overestimated protraction by approximately 5 degrees. The AMC underestimated protraction in the frontal plane at low elevation angle but overestimated it at all other postures and the overestimation increased with plane of elevation, internal rotation and elevation angle. Overall, it is recommended to use AMC rather than the scapular locator to measure scapular position.