Sex-specific differences in gait patterns of healthy older adults: results from the Baltimore Longitudinal Study of Aging

The effects of normal aging and orthopedic conditions on gait patterns during customary walking have been extensively investigated. Empirical evidence supports the notion that sex differences exist in the gait patterns of young adults but it is unclear as to whether sex differences exist in older adults. The aim of this study was to investigate sex-specific differences in gait among older adults. Study participants were 336 adults (50-96 years; 162 women) enrolled in the Baltimore Longitudinal Study of Aging (BLSA) who completed walking tasks at self-selected speed without assistance. After adjusting for significant covariates, women walked with higher cadence (p=0.01) and shorter stride length (p=0.006) compared to men, while gait speed was not significantly related to sex. Women also had less hip range of motion (ROM; p=0.004) and greater ankle ROM (p<0.001) in the sagittal-plane, and greater hip ROM (p=0.004) in the frontal-plane. Hip absorptive mechanical work expenditure (MWE) of the women was greater in the sagittal-plane (p<0.001) and lower in the frontal-plane (p<0.001), compared to men. In summary, women's gait is characterized by greater ankle ROM than men while men tend to have greater hip ROM than women. Characterizing unique gait patterns of women and men with aging may be beneficial for detecting the early stages of gait abnormalities that may lead to pathology.     

Dynamic accuracy of inertial measurement units during simple pendulum motion

A motion measurement system based on inertial measurement units (IMUs) has been suggested as an alternative to contemporary video motion capture. This paper reports an investigation into the accuracy of IMUs in estimating 3D orientation during simple pendulum motion. The IMU vendor's (XSens Technologies) accuracy claim of 3° root mean squared (RMS) error is tested. IMUs are integrated electronic devices that contain accelerometers, magnetometers and gyroscopes. The motion of a pendulum swing was measured using both IMUs and video motion capture as a reference. The IMU raw data were processed by the Kalman filter algorithm supplied by the vendor and a custom fusion algorithm developed by the authors. The IMU measurement of pendulum motion using the vendor's Kalman filter algorithm did not compare well with the video motion capture with a RMS error of between 8.5° and 11.7° depending on the length and type of pendulum swing. The maximum orientation error was greater than 30°, occurring approximately eight seconds into the motion. The custom fusion algorithm estimation of orientation compared well with the video motion capture with a RMS error of between 0.8° and 1.3°. Future research should concentrate on developing a general purpose fusion algorithm and vendors of IMUs should provide details about the errors to be expected in different measurement situations, not just those in a ‘best case’ scenario.     

Is knee biomechanics different in uphill walking on different slopes for older adults with total knee replacement?

The purpose of this study was to investigate knee biomechanics in uphill walking on slopes of 5°, 10° and 15° for total knee replacement (TKR) patients. Twenty-five post-TKR patients and ten healthy controls performed five walking trials on level ground and different slopes on an instrumented ramp system. A 2 × 2 × 4 (limb × group × incline slope) mixed model ANOVA was used to examine selected variables. The peak knee extension moment (KEM) was greater in 15° uphill walking compared to level, 5° and 10° uphill walking. TKR patients had lower peak KEM and smaller knee extension range of motion than healthy controls in all walking conditions. The Replaced Limb showed lower peak KEM in 10° and 15° uphill walking than the Non-replaced Limb and smaller knee extension range of motion (ROM) in 10° uphill walking. Knee extension and abduction ROM increased with increased incline angles. The greater peak loading-response vertical ground reaction force was found in level walking compared to three levels of uphill walking. The peak loading-response knee abduction moment was greater in level walking compared to 10° and 15° uphill walking. However, the medial knee contact force was greater in non-replaced limb compared to replaced limb in 10° and 15° uphill walking. The results suggest 5° uphill walking may have the potential to become a safe exercise for unilateral TKR patients.     

Hinged ankle braces do not alter knee mechanics during sidestep cutting

Lateral ankle sprains are common injuries in quick, dynamic movements and are caused by rapid ankle inversion. Ankle braces are used to reduce ankle inversion, while allowing normal plantar and dorsiflexion ranges of motion. Knee injuries, such as anterior cruciate ligament injuries, are also common in dynamic movements. It is important to understand how ankle braces affect injury risk at other proximal joints. There is limited and conflicting results on how ankle braces affect knee mechanics during these types of movements. Additionally, it is unknown if sex differences exist when using an ankle brace. Therefore, the purpose of this study was to determine the effects of a hinged ankle brace and sex during a 45° cutting movement. Three-dimensional kinematics and ground reaction forces were collected using a motion capture system and force plate on ten men and eight women during cutting trials. 2 × 2 repeated measures ANOVAs were used to detect differences in ground reaction forces, as well as knee and ankle kinematics between brace conditions and sex (p < 0.05). The brace condition exhibited greater initial contact ankle dorsiflexion (p = 0.011), decreased peak ankle inversion (p < 0.01), and increased vertical loading rate (p = 0.040). Females performed the cutting movement with less initial contact (p = 0.019) and peak knee flexion (p = 0.023) compared to males. Ankle bracing had no impact on the observed sex differences. Females exhibited decreased knee flexion compared to males, which has been well documented in the literature. The use of an ankle braces reduced ankle injury risk variables while not adversely impacting knee mechanics during a 45° sidecutting movement.     

Stance and swing phase knee flexion recover at different rates following total knee arthroplasty: An inertial measurement unit study

Total knee arthroplasty (TKA) is the most common joint replacement in the United States. Range of motion (ROM) monitoring includes idealized clinic measures (e.g. goniometry during passive ROM) that may not accurately represent knee function. Accordingly, a novel, portable, inertial measurement unit (IMU) based ROM measurement method was developed, validated, and implemented. Knee flexion was computed via relative motion between two IMUs and validated via optical motion capture (p > 0.05). Prospective analyses of 10 healthy individuals (5M, 50 ± 19 years) and 20 patients undergoing TKA (3 lost to follow up, 10M, 65 ± 6 years) were completed. Controls wore IMUs for 1-week. Patients wore IMUs for 1-week pre-TKA, 6-weeks immediately post-TKA, and 1-week at 1-year post-TKA. Flexion was computed continuously each day (8–12 h). Metrics included daily maximum flexion and flexion during stance/swing phases of gait. Maximum flexion was equal between cohorts at all time points. Contrastingly, patient stance and swing flexion were reduced pre-TKA, yet improved post-TKA. Specifically, patient stance and swing flexion were reduced below control/pre-TKA values during post-TKA week 1. Stance flexion exceeded pre-TKA and equaled control levels after week 2. However, swing flexion only exceeded pre-TKA and equaled control levels at 1-year post-TKA. This novel method improves upon the accuracy/portability of current methods (e.g. goniometry). Interestingly, surgery did not impact maximum ROM, yet improved the ability to flex during gait allowing more efficient and safe ambulation. This is the first study continuously monitoring long-term flexion before/after TKA. The results offer richer information than clinical measures about expected TKA rehabilitation.     

Differences in Lower Extremity and Trunk Kinematics between Single Leg Squat and Step Down Tasks

The single leg squat and single leg step down are two commonly used functional tasks to assess movement patterns. It is unknown how kinematics compare between these tasks. The purpose of this study was to identify kinematic differences in the lower extremity, pelvis and trunk between the single leg squat and the step down. Fourteen healthy individuals participated in this research and performed the functional tasks while kinematic data were collected for the trunk, pelvis, and lower extremities using a motion capture system. For the single leg squat task, the participant was instructed to squat as low as possible. For the step down task, the participant was instructed to stand on top of a box, slowly lower him/herself until the non-stance heel touched the ground, and return to standing. This was done from two different heights (16cm and 24cm). The kinematics were evaluated at peak knee flexion as well as at 60° of knee flexion. Pearson correlation coefficients (r) between the angles at those two time points were also calculated to better understand the relationship between each task. The tasks resulted in kinematics differences at the knee, hip, pelvis, and trunk at both time points. The single leg squat was performed with less hip adduction (p ≤ 0.003), but more hip external rotation and knee abduction (p ≤ 0.030), than the step down tasks at 60° of knee flexion. These differences were maintained at peak knee flexion except hip external rotation was only significant in the 24cm step down task (p ≤ 0.029). While there were multiple differences between the two step heights at peak knee flexion, the only difference at 60° of knee flexion was in trunk flexion (p < 0.001). Angles at the knee and hip had a moderate to excellent correlation (r = 0.51–0.98), but less consistently so at the pelvis and trunk (r = 0.21–0.96). The differences in movement patterns between the single leg squat and the step down should be considered when selecting a single leg task for evaluation or treatment. The high correlation of knee and hip angles between the three tasks indicates that similar information about knee and hip kinematics was gained from each of these tasks, while pelvis and trunk angles were less well predicted.     

Kinematics of the upper cervical spine during high velocity-low amplitude manipulation. Analysis of intra- and inter-operator reliability for pre-manipulation positioning and impulse displacements

To date, kinematics data analyzing continuous 3D motion of upper cervical spine (UCS) manipulation is lacking. This in vitro study aims at investigating inter- and intra-operator reliability of kinematics during high velocity low amplitude manipulation of the UCS.Three fresh specimens were used. Restricted dissection was realized to attach technical clusters to each bone (skull to C₂). Motion data was obtained using an optoelectronic system during manipulation. Kinematics data were integrated into specific-subject 3D models to provide anatomical motion representation during thrust manipulation. The reliability of manipulation kinematics was assessed for three practitioners performing two sessions of three repetitions on two separate days.For pre-manipulation positioning, average UCS ROM (SD) were 10° (5°), 22° (5°) and 14° (4°) for lateral bending, axial rotation and flexion–extension, respectively. For the impulse phase, average axial rotation magnitude ranged from 7° to 12°. Reliability analysis showed average RMS up to 8° for pre-manipulation positioning and up to 5° for the impulse phase.As compared to physiological ROM, this study supports the limited angular displacement during manipulation for UCS motion components, especially for axial rotation. Kinematics reliability confirms intra- and inter-operator consistency although pre-manipulation positioning reliability is slightly lower between practitioners and sessions.     

Upper limb joint angle measurement in occupational health

Usual human motion capture systems are designed to work in controlled laboratory conditions. For occupational health, instruments that can measure during normal daily life are essential, as the evaluation of the workers' movements is a key factor to reduce employee injury- and illness-related costs. In this paper, we present a method for joint angle measurement, combining inertial sensors (accelerometers and gyroscopes) and magnetic sensors. This method estimates wrist flexion, wrist lateral deviation, elbow flexion, elbow pronation, shoulder flexion, shoulder abduction and shoulder internal rotation. The algorithms avoid numerical integration of the signals, which allows for long-time estimations without angle estimation drift. The system has been tested both under laboratory and field conditions. Controlled laboratory tests show mean estimation errors between 0.06° and of 1.05°, and standard deviation between 2.18° and 9.20°. Field tests seem to confirm these results when no ferromagnetic materials are close to the measurement system.     

Trunk kinematic,kinetic, and neuro-muscular response to foot center of pressure translation along the medio-lateral foot axis during gait

Footwear devices that shift foot center of pressure (COP), thereby impacting lower-limb biomechanics to produce clinical benefit, have been studied regarding degenerative diseases of knee and hip joints, exhibiting evidence of clinical success. Ability to purposefully affect trunk biomechanics has not been investigated for this type of footwear. Fifteen healthy young male subjects underwent gait and electromyography analysis using a biomechanical device that shifts COP via moveable convex elements attached to the shoe sole. Analyses were performed in three COP configurations for pairwise comparison: (1) neutral (control) (2) laterally deviated, and (3) medially deviated. Sagittal and frontal-plane pelvis and spine kinematics, external oblique activity, and frontal and transverse-plane lumbar moments were affected by medio-lateral COP shift. Transverse-plane trunk kinematics, activity of the lumbar longissimus, latissimus dorsi, rectus abdominus, and quadratus lumborum, and sagittal-plane lumbar moment, were not significantly impacted. Two linear mixed effects models assessed predictive impact of (I) COP location, and (II) trunk kinematics and neuromuscular activity, on the significant lumbar moment parameters. The COP was a significant predictor of all modeled frontal and transverse-plane lumbar moment parameters, while pelvic and spine rotation, and lumbar longissimus activity were significant predictors of one frontal-plane lumbar moment parameter. Model results suggest that, although trunk biomechanics and muscle activity were altered by COP shift, COP offset influences lumbar kinetics directly, or via lower-limb changes not assessed in this study, but not by means of alteration of trunk kinematics or muscle activity. Further study may reveal implications in treatment of low back pain.     

Gender comparisons between unilateral and bilateral landings

The increased number of women participating in sports has led to a higher knee injury rate in women compared with men. Among these injuries, those occurring to the ACL are commonly observed during landing maneuvers. The purpose of this study was to determine gender differences in landing strategies during unilateral and bilateral landings. Sixteen male and 17 female recreational athletes were recruited to perform unilateral and bilateral landings from a raised platform, scaled to match their individual jumping abilities. Three-dimensional kinematics and kinetics of the dominant leg were calculated during the landing phase and reported as initial ground contact angle, ranges of motion (ROM) and peak moments. Lower extremity energy absorption was also calculated for the duration of the landing phase. Results showed that gender differences were only observed in sagittal plane hip and knee ROM, potentially due to the use of a relative drop height versus the commonly used absolute drop height. Unilateral landings were characterized by significant differences in hip and knee kinematics that have been linked to increased injury risk and would best be classified as "stiff" landings. The ankle musculature was used more for impact absorption during unilateral landing, which required increased joint extension at touchdown and may increase injury risk during an unbalanced landing. In addition, there was only an 11% increase in total energy absorption during unilateral landings, suggesting that there was a substantial amount of passive energy transfer during unilateral landings.     

The effects of core muscle activation on dynamic trunk position and knee abduction moments: Implications for ACL injury

Anterior cruciate ligament (ACL) injury is one of the most common serious lower-extremity injuries experienced by athletes participating in field and court sports and often occurs during a sudden change in direction or pivot. Both lateral trunk positioning during cutting and peak external knee abduction moments have been associated with ACL injury risk, though it is not known how core muscle activation influences these variables. In this study, the association between core muscle pre-activation and trunk position as well as the association between core muscle pre-activation and peak knee abduction moment during an unanticipated run-to-cut maneuver were investigated in 46 uninjured individuals. Average co-contraction indices and percent differences between muscle pairs were calculated prior to initial contact for internal obliques, external obliques, and L5 extensors using surface electromyography. Outside tilt of the trunk was defined as positive when the trunk was angled away from the cutting direction. No significant associations were found between pre-activations of core muscles and outside tilt of the trunk. Greater average co-contraction index of the L5 extensors was associated with greater peak knee abduction moment (p=0.0107). Increased co-contraction of the L5 extensors before foot contact could influence peak knee abduction moment by stiffening the spine, limiting sagittal plane trunk flexion (a motion pattern previously linked to ACL injury risk) and upper body kinetic energy absorption by the core during weight acceptance.     

Comparative assessment of knee joint models used in multi-body kinematics optimisation for soft tissue artefact compensation

Estimating joint kinematics from skin-marker trajectories recorded using stereophotogrammetry is complicated by soft tissue artefact (STA), an inexorable source of error. One solution is to use a bone pose estimator based on multi-body kinematics optimisation (MKO) embedding joint constraints to compensate for STA. However, there is some debate over the effectiveness of this method. The present study aimed to quantitatively assess the degree of agreement between reference (i.e., artefact-free) knee joint kinematics and the same kinematics estimated using MKO embedding six different knee joint models. The following motor tasks were assessed: level walking, hopping, cutting, running, sit-to-stand, and step-up. Reference knee kinematics was taken from pin-marker or biplane fluoroscopic data acquired concurrently with skin-marker data, made available by the respective authors. For each motor task, Bland-Altman analysis revealed that the performance of MKO varied according to the joint model used, with a wide discrepancy in results across degrees of freedom (DoFs), models and motor tasks (with a bias between −10.2° and 13.2° and between −10.2 mm and 7.2 mm, and with a confidence interval up to ±14.8° and ±11.1 mm, for rotation and displacement, respectively). It can be concluded that, while MKO might occasionally improve kinematics estimation, as implemented to date it does not represent a reliable solution to the STA issue.     

A wearable system to assess risk for anterior cruciate ligament injury during jump landing: measurements of temporal events, jump height, and sagittal plane kinematics

The incidence of anterior cruciate ligament (ACL) injury remains high, and there is a need for simple, cost effective methods to identify athletes at a higher risk for ACL injury. Wearable measurement systems offer potential methods to assess the risk of ACL injury during jumping tasks. The objective of this study was to assess the capacity of a wearable inertial-based system to evaluate ACL injury risk during jumping tasks. The system accuracy for measuring temporal events (initial contact, toe-off), jump height, and sagittal plane angles (knee, trunk) was assessed by comparing results obtained with the wearable system to simultaneous measurements obtained with a marker-based optoelectronic reference system. Thirty-eight healthy participants (20 male and 18 female) performed drop jumps with bilateral and unilateral support landing. The mean differences between the temporal events obtained with both systems were below 5 ms, and the precisions were below 24 ms. The mean jump heights measured with both systems differed by less than 1 mm, and the associations (Pearson correlation coefficients) were above 0.9. For the discrete angle parameters, there was an average association of 0.91 and precision of 3.5° for the knee flexion angle and an association of 0.77 and precision of 5.5° for the trunk lean. The results based on the receiver-operating characteristic (ROC) also demonstrated that the proposed wearable system could identify movements at higher risk for ACL injury. The area under the ROC plots was between 0.89 and 0.99 for the knee flexion angle and between 0.83 and 0.95 for the trunk lean. The wearable system demonstrated good concurrent validity with marker-based measurements and good discriminative performance in terms of the known risk factors for ACL injury. This study suggests that a wearable system could be a simple cost-effective tool for conducting risk screening or for providing focused feedback.     

Speed,age, sex,and body mass index provide a rigorous basis for comparing the kinematic and kinetic profiles of the lower extremity during walking

The increased use of gait analysis has raised the need for a better understanding of how walking speed and demographic variations influence asymptomatic gait. Previous analyses mainly reported relationships between subsets of gait features and demographic measures, rendering it difficult to assess whether gait features are affected by walking speed or other demographic measures. The purpose of this study was to conduct a comprehensive analysis of the kinematic and kinetic profiles during ambulation that tests for the effect of walking speed in parallel to the effects of age, sex, and body mass index. This was accomplished by recruiting a population of 121 asymptomatic subjects and analyzing characteristic 3-dimensional kinematic and kinetic features at the ankle, knee, hip, and pelvis during walking trials at slow, normal, and fast speeds. Mixed effects linear regression models were used to identify how each of 78 discrete gait features is affected by variations in walking speed, age, sex, and body mass index. As expected, nearly every feature was associated with variations in walking speed. Several features were also affected by variations in demographic measures, including age affecting sagittal-plane knee kinematics, body mass index affecting sagittal-plane pelvis and hip kinematics, body mass index affecting frontal-plane knee kinematics and kinetics, and sex affecting frontal-plane kinematics at the pelvis, hip, and knee. These results could aid in the design of future studies, as well as clarify how walking speed, age, sex, and body mass index may act as potential confounders in studies with small populations or in populations with insufficient demographic variations for thorough statistical analyses.     

A normative database of hip and knee joint biomechanics during dynamic tasks using anatomical regression prediction methods

Many methodologies exist to predict the hip joint center (HJC), of which regression based on anatomical landmarks appear most common. Despite the fact that predicted HJC locations vary depending upon chosen method, inter-study comparisons and inferences about populations are commonly made. The purpose of this study was to create a normative database of hip and knee biomechanics during walking, running, and single leg landings based on five commonly utilized HJC methods to serve as a reference for inter-study comparisons. Secondarily, we devised to provide comparisons of peak knee angles and hip angles, moments, and powers from the five HJC methods. Thirty healthy young adults performed walking, running, and single leg landing tasks at self-selected speeds (walking/running) and at 90% of their maximum jump height (landing). Three-dimensional motion capture and ground reaction forces were collected during all tasks. Five different HJC prediction methods: Bell, Davis, Hara, Harrington, and Greater Trochanter were implemented separately in a 6 degree of freedom model. Predicted HJC locations, direct kinematics, and inverse dynamics were computed for all tasks. Predicted HJC mediolateral, anteroposterior, and superior-inferior locations differed between methods by an average of 1.3, 2.9, and 1.4 cm, respectively. A database was created using the mean of all subjects for all five methods. In addition, one-way ANOVAs were used to compare triplanar peak angles, moments, and powers between the methods. The database of hip and knee biomechanics illustrates (1) variability between methods increases with more dynamic tasks (running/landing vs. walking) and (2) frontal and transverse plane hip and knee biomechanics are more variable between methods. Comparisons between methods found 38 and 16 main effect differences in hip and knee biomechanics, respectively. The Greater Trochanter method provided the most differences compared with other methods, while the Davis method provided the least differences. The database constructed provides an important reference for inter-study comparisons and details the impact of anatomical regression methods for predicting the HJC.     

Antagonist muscle co-contraction during a double-leg landing maneuver at two heights

Background: Knee injuries are common during landing activities. Greater landing height increases peak ground reaction forces (GRFs) and loading at the knee joint. As major muscles to stabilize the knee joint, Quadriceps and Hamstring muscles provide internal forces to attenuate the excessive GRF. Despite the number of investigations on the importance of muscle function during landing, the role of landing height on these muscles forces using modeling during landing is not fully investigated. Methods: Participant-specific musculoskeletal models were developed using experimental motion analysis data consisting of anatomic joint motions and GRF from eight male participants performing double-leg drop landing from 30 and 60 cm. Muscle forces were calculated in OpenSim and their differences were analyzed at the instances of high risk during landing i.e. peak GRF for both heights. Results: The maximum knee flexion angle and moments were found significantly higher from a double-leg landing at 60 cm compared to 30 cm. The results showed elevated GRF, and mean muscle forces during landing. At peak GRF, only quadriceps showed significantly greater forces at 60 cm. Hamstring muscle forces did not significantly change at 60 cm compared to 30 cm. Conclusions: Quadriceps and hamstring muscle forces changed at different heights. Since hamstring forces were similar in both landing heights, this could lead to an imbalance between the antagonist muscles, potentially placing the knee at risk of injury if combined with small flexion angles that was not observed at peak GRF in our study. Thus, enhanced neuromuscular training programs strengthening the hamstrings may be required to address this imbalance. These findings may contribute to enhance neuromuscular training programs to prevent knee injuries during landing.     

A musculoskeletal model customized for squatting task

Most musculoskeletal models (MSKM) are designed to evaluate gait and running, which have limited range of motion (ROM). The purpose of this study was to examine the effect of wrapping surfaces (WS) at the knee and hip joints in a MSKM, on the muscle moment arms (MA) and activations during squatting. The MSKM was then customized by changing parameters of the original WS and by implementing additional WS. The WS prevent muscles from crossing into the bones, providing realistic muscle MA for large ROM. The modified MSKM is suitable for analysis up to 138° hip and 145° knee flexions.     

Wide step width reduces knee abduction moment of obese adults during stair negotiation

PurposeAn increased likelihood of developing obesity-related knee osteoarthritis may be associated with increased peak internal knee abduction moments (KAbM). Increases in step width (SW) may act to reduce this moment. The purpose of this study was to determine the effects of increased SW on knee biomechanics during stair negotiation of healthy-weight and obese participants.MethodsParticipants (24: 10 obese and 14 healthy-weight) used stairs and walked over level ground while walking at their preferred speed in two different SW conditions – preferred and wide (200% preferred). A 2 × 2 (group × condition) mixed model analysis of variance was performed to analyze differences between groups and conditions (p < 0.05).ResultsIncreased SW increased the loading-response peak knee extension moment during descent and level gait, decreased loading-response KAbMs, knee extension and abduction range of motion (ROM) during ascent, and knee adduction ROM during descent. Increased SW increased loading-response peak mediolateral ground reaction force (GRF), increased peak knee abduction angle during ascent, and decreased peak knee adduction angle during descent and level gait. Obese participants experienced disproportionate changes in loading-response mediolateral GRF, KAbM and peak adduction angle during level walking, and peak knee abduction angle and ROM during ascent.ConclusionIncreased SW successfully decreased loading-response peak KAbM. Implications of this finding are that increased SW may decrease medial compartment knee joint loading, decreasing pain and reducing joint deterioration. Increased SW influenced obese and healthy-weight participants differently and should be investigated further.     

Assessment of the measurement accuracy of inertial sensors during different tasks of daily living

The low cost and ease of use of inertial measurement units (IMUs) make them an attractive option for motion analysis tasks that cannot be easily measured in a laboratory. To date, only a limited amount of research has been conducted comparing commercial IMU systems to optoelectronic systems, the gold standard, for everyday tasks like stair climbing and inclined walking. In this paper, the 3D joint angles of the lower limbs are determined using both an IMU system and an optoelectronic system for twelve participants during stair ascent and descent, and inclined, declined and level walking. Three different datasets based on different hardware and anatomical models were collected for the same movement in an effort to determine the cause and quantify the errors involved with the analysis. Firstly, to calculate software errors, two different anatomical models were compared for one hardware system. Secondly, to calculate hardware errors, results were compared between two different measurement systems using the same anatomical model. Finally, the overall error between both systems with their native anatomical models was calculated. Statistical analysis was performed using statistical parametric mapping. When both systems were evaluated based on the same anatomical model, the number of trials with significant differences decreased markedly. Thus, the differences in joint angle measurement can mainly be attributed to the variability in the anatomical models used for calculations and not to the IMU hardware.     

Influence of IMU position and orientation placement errors on ground reaction force estimation

Wearable inertial measurement units (IMU) have been proposed to estimate GRF outside of specialized laboratories, however the precise influence of sensor placement error on accuracy is unknown. We investigated the influence of IMU position and orientation placement errors on GRF estimation accuracy. Methods: Kinematic data from twelve healthy subjects based on marker trajectories were used to simulate 1848 combinations of sensor position placement errors (range ± 100 mm) and orientation placement errors (range ± 25°) across eight body segments (trunk, pelvis, left/right thighs, left/right shanks, and left/right feet) during normal walking trials for baseline cases when a single sensor was misplaced and for the extreme cases when all sensors were simultaneously misplaced. Three machine learning algorithms were used to estimate GRF for each placement error condition and compared with the no placement error condition to evaluate performance. Results: Position placement errors for a single misplaced IMU reduced vertical GRF (VGRF), medio-lateral GRF (MLGRF), and anterior-posterior GRF (APGRF) estimation accuracy by up to 1.1%, 2.0%, and 0.9%, respectively and for all eight simultaneously misplaced IMUs by up to 4.9%, 6.0%, and 4.3%, respectively. Orientation placement errors for a single misplaced IMU reduced VGRF, MLGRF, and APGRF estimation accuracy by up to 4.8%, 7.3%, and 1.5%, respectively and for all eight simultaneously misplaced IMUs by up to 20.8%, 23.4%, and 12.3%, respectively. Conclusion: IMU sensor misplacement, particularly orientation placement errors, can significantly reduce GRF estimation accuracy and thus measures should be taken to account for placement errors in implementations of GRF estimation via wearable IMUs.