Assessment and validation of a simple automated method for the detection of gait events and intervals

Pelvic Drop Changes due to Proximal Muscle Strengthening Depend on Foot-Ankle Varus Alignment

Background

Strengthening of hip and trunk muscles can modify pelvis and hip movements. However, the varus alignment of the foot-ankle complex (FAC) may influence the effects of muscle strengthening, due to the relationship of FAC alignment with pelvic and hip kinematics. This study evaluated the effects of hip and trunk muscle strengthening on pelvis and hip kinematics during walking, in subgroups with larger and smaller values of FAC varus alignment. In addition, this study evaluated the effects of hip and trunk muscle strengthening on hip passive and active properties, in the same subgroups.

Methods

Fifty-three women, who were divided into intervention and control groups, participated in this nonrandomized controlled trial. Each group was split into two subgroups with larger and smaller values of FAC varus alignment. Hip and trunk muscle strengthening was performed three times a week for two months, with a load of 70% to 80% of one repetition maximum. Before and after strengthening, we evaluated (1) pelvis and hip excursions in the frontal and transverse planes during walking, (2) isokinetic hip passive external rotator torque, and (3) isokinetic concentric and eccentric peak torques of the hip external rotator muscles. Mixed analyses of variance (ANOVAs) were carried out for each dependent variable related to walking kinematics and isokinetic measurements (α = 0.05).

Results

The subgroup with smaller varus alignment, of the intervention group, presented a reduction in pelvic drop after strengthening (P = 0.03). The subgroup with larger varus alignment increased pelvic drop after strengthening, with a marginal significance (P = 0.06). The other kinematic excursions did not change (pelvic anterior rotation P = 0.30, hip internal rotation P = 0.54, and hip adduction P = 0.43). The intervention group showed increases in passive torque (P = 0.002), peak concentric torque (P < 0.001), and peak eccentric torque (P < 0.001), independently of FAC alignment. These results suggest that FAC varus alignment influences the effects of strengthening and should be considered when hip and trunk muscle strengthening is used to reduce pelvic drop during walking.

Identification of gait events in children with spastic cerebral palsy: comparison between the force plate and algorithms

OBJECTIVE

To compare the gait event identification of five algorithms recommended in the literature with those provided by force plate (gold standard) in children with unilateral or bilateral spastic cerebral palsy (SCP).

METHODS

This was a cross-sectional study of the gait of three girls and four boys with a mean age of 8.6±4.7 years. Four children had unilateral SCP with an equinus gait pattern, and the remaining three children exhibited bilateral SCP with a slide/drag gait pattern. Kinematic and kinetic gait data were collected during barefoot walking at a comfortable speed. From a total of 202 steps, the detection of 202 foot-strike (FS) and 194 toe-off (TO) events by each algorithm was compared with the detection of these same events by the force plate. The error between the events detected by the algorithms and those detected by the force plate was determined in milliseconds. Repeated measures ANOVA was used to compare the errors among the algorithms.

RESULTS

The algorithm reported by Ghoussayni et al. showed the best performance in all situations, except for the identification of FS events on the unaffected side in children with unilateral SCP. For these events, the algorithms reported by Desailly et al. and Zeni et al. showed the best performance.

CONCLUSION

Ghoussayni et al.'s algorithm can be used to detect gait events in children with SCP when a force plate is not available.

An improved technique for increasing the accuracy of joint-to-ground distance tracking in kinect v2 for foot-off and foot contact detection

The Kinect sensor has been widely used in different applications such as rehabilitation and gait analysis. Whilst Kinect v2 was released with several improvements over its predecessor, it still incorporates depth-map intrinsic inaccuracies. This results in inconsistencies in skeletal-data acquisition, especially in joint localisation and distance-to-ground tracking with respect to the Kinect's 3-D Cartesian coordinate reference point. This research proposes a correction technique based on the two-point linear equation derived from the information gathered from different subjects' skeletal data and data regression analysis to compensate the inaccuracies in joint-to-ground data collection. The research also proposes a new footsteps detection method based on skeletal data and plane detection techniques that calculates a footstep by using the ankle's Euclidean distance from the floor, regardless of the subject's distance from the camera. The results show that after the correction technique was applied, data acquisition proved to be consistent and more accurate within a distance range of 1.6-2.9 m from the Kinect camera, regardless of the subject's location to the camera's reference point. Moreover, the inconsistency of joint data read by the Kinect was reduced from 25.69% to 5.25% and the footsteps detection accuracy increased from 42.85% to 79.76% on average for both legs.

Foot-ground clearance characteristics in women: A comparison across different ages

BACKGROUND

Tripping is a common event leading to falls amongst elderly. Minimum foot clearance (MFC) is a critical swing phase control factor associated with tripping and falls.

RESEARCH QUESTION

Are there differences in MFC characteristics among three age groups of women and are there association between MFC and lower limb kinematics?

METHODS

Cross-sectional observational study. Three-dimensional gait analysis of 55 healthy women. ANOVA was used to compare (p<0.05) MFC characteristics among young, middle-aged and elderly groups. Multiple Linear Regression Analysis was used to test prediction over MFC.

RESULTS

Elderly women walked slower, with lower MFC and lower maximum foot velocity during swing (MFV) than young and middle-aged women. There were more hip flexion and less ankle dorsiflexion during MFC among elderly. There is a strong positive relationship between dorsiflexion and MFC. And ankle dorsiflexion was the most predictive variable over MFC.

SIGNIFICANCE

Elderly women walk slower with lower MFC value and less ankle dorsiflexion than gender-matched young controls. Increased hip flexion may represent a gait adaptation to avoid tripping. Gait speed had no effect on those findings.

Postural adjustments in adolescent idiopathic thoracic scoliosis during walking

INTRODUCTION

Adolescent idiopathic scoliosis (AIS) is the most common type of three-dimensional spinal deformity. Identifying the postural adjustments or changes for different phases and events is needed for developing programs to improve the AIS gait, but such information has been limited. The current study aimed to fill the gap via three-dimensional motion analysis of quiet standing and level walking in patients with severe thoracic AIS.

MATERIALS AND METHODS

Sixteen female adolescents with AIS (Lenke 1 or 2, age: 14.9 ± 1.7 years, height: 154.7 ± 5.0 cm, mass: 41.7 ± 7.2 kg) and sixteen sex-, age- and BMI-matched healthy controls (age: 14.8 ± 2.7 years, height: 154.9 ± 5.6 cm, mass: 44.7 ± 6.3 kg) participated in the current study with informed written consent. The kinematic and kinetic changes between the trunk, pelvis, and lower limb segments, and at the lumbosacral level at different gait events were measured during quiet standing and level walking.

RESULTS

The homogeneity of the current patient group helped reduce the effects of the level and severity of spinal deformity on inter-subject variability that has been associated with controversies over reported gait variables in AIS. The current results support the hypothesis that postural adjustments involving the trunk, pelvis and lower limb segments were needed in severe thoracic AIS during both quiet standing and level walking, and differed between concave and convex sides at different key gait events during level walking.

CONCLUSIONS

Although scoliotic spinal deformity occurred mainly in the frontal plane, postural adjustments in all three planes were present at key events during level walking with associated joint loading changes in patients with severe thoracic AIS. Monitoring of such adjustments and the associated joint kinetic changes will be helpful for assessing the disease and treatment outcomes.

Automatic real-time gait event detection in children using deep neural networks

Annotation of foot-contact and foot-off events is the initial step in post-processing for most quantitative gait analysis workflows. If clean force plate strikes are present, the events can be automatically detected. Otherwise, annotation of gait events is performed manually, since reliable automatic tools are not available. Automatic annotation methods have been proposed for normal gait, but are usually based on heuristics of the coordinates and velocities of motion capture markers placed on the feet. These heuristics do not generalize to pathological gait due to greater variability in kinematics and anatomy of patients, as well as the presence of assistive devices. In this paper, we use a data-driven approach to predict foot-contact and foot-off events from kinematic and marker time series in children with normal and pathological gait. Through analysis of 9092 gait cycle measurements we build a predictive model using Long Short-Term Memory (LSTM) artificial neural networks. The best-performing model identifies foot-contact and foot-off events with an average error of 10 and 13 milliseconds respectively, outperforming popular heuristic-based approaches. We conclude that the accuracy of our approach is sufficient for most clinical and research applications in the pediatric population. Moreover, the LSTM architecture enables real-time predictions, enabling applications for real-time control of active assistive devices, orthoses, or prostheses. We provide the model, usage examples, and the training code in an open-source package.

Impact of Load Variation on the Accuracy of Gait Recognition from Surface EMG Signals

As lower-limb exoskeleton and prostheses are developed to become smarter and to deploy man–machine collaboration, accurate gait recognition is crucial, as it contributes to the realization of real-time control. Many researchers choose surface electromyogram (sEMG) signals to recognize the gait and control the lower-limb exoskeleton (or prostheses). However, several factors still affect its applicability, of which variation in the loads is an essential one. This study aims to (1) investigate the effect of load variation on gait recognition; and to (2) discuss whether a lower-limb exoskeleton control system trained by sEMG from different loads works well in multi-load applications. In our experiment, 10 male college students were selected to walk on a treadmill at three different speeds (V3 = 3 km/h, V5 = 5 km/h, and V7 = 7 km/h) with four different loads (L0 = 0, L20 = 20%, L30 = 30%, L40 = 40% of body weight, respectively), and 50 gait cycles were performed. Back propagation neural networks (BPNNs) were used for gait recognition, and a support vector machine (SVM) and k-nearest neighbor (k-NN) were used for comparison. The result showed that (1) load variation has significant effects on the accuracy of gait recognition (p < 0.05) under the three speeds when the loads range in L0, L20, L30, or L40, but no significant impact is found when the loads range in L0, L20, or L30. The least significant difference (LSD) post hoc, which can explore all possible pair-wise comparisons of means that comprise a factor using the equivalent of multiple t-tests, reveals that there is a significant difference between the L40 load and the other three loads (L0, L20, L30), but no significant difference was found among the L0, L20, and L30 loads. The total mean accuracy of gait recognition of the intra-loads and inter-loads was 91.81%, and 69.42%, respectively. (2) When the training data was taken from more types of loads, a higher accuracy in gait recognition was obtained at each speed, and the statistical analysis shows that there was a substantial influence for the kinds of loads in the training set on the gait recognition accuracy (p < 0.001). It can be concluded that an exoskeleton (or prosthesis) control system that is trained in a single load or the parts of loads is insufficient in the face of multi-load applications.

Differential effects of speed on two-dimensional foot strike pattern during barefoot and shod running in recreationally active men

The majority of barefoot running studies have not considered speed as an influential factor on foot strike pattern. The aim of this study was to investigate differences in foot strike pattern and spatiotemporal characteristics between barefoot and shod overground running at varying speeds. We first determined maximal running speed (Vm) over 50 m in 15 recreationally active men who self-reported as habitual rearfoot strikers. Participants then completed shod and barefoot running trials at different speeds equivalent to approximately 90%, 80%, 70% and 60% of Vm. Sagittal plane two-dimensional (2D) foot-ground contact angle, ankle plantar-dorsi flexion angle, contact time, flight time, step length and step rate variables for each trial were recorded. A significant interaction effect of running speed and footwear condition (p < 0.05) on foot-ground contact angle, ankle plantar-dorsi flexion angle and contact time was observed. There was a main effect of running speed (p < 0.01) on flight time, step length and step rate. There was a main effect of footwear condition on step length (p < 0.01). Participants were more inclined to plantarflex the ankle and contact the ground with the forefoot at higher percentages of Vm, especially when running barefoot.

Validity of the Microsoft Kinect™ in assessing spatiotemporal and lower extremity kinematics during stair ascent and descent in healthy young individuals

Stair negotiation is one of the most challenging, yet frequently encountered, locomotor tasks in daily life. This study is the first attempt to investigate the capacity of the Kinect^™ sensor to assess stair negotiation spatiotemporal and sagittal plane kinematic variables. The goal of this study was to examine the validity of the Kinect^™ v2 sensor in assessing lower extremity kinematics and spatiotemporal parameters in healthy young individuals; and to demonstrate its potential as a low-cost stair gait analysis tool. Twelve healthy participants ascended and descended a 3-step custom-built staircase at their preferred speed, as spatiotemporal parameters and kinematics were extracted simultaneously using the Kinect^™ and a three-dimensional motion analysis. Spatiotemporal measures included gait speed, swing phase time, and double stance time. Kinematic outcomes included hip, knee, and ankle joint angles in the sagittal plane. Consistency (ICC_2,1) and absolute agreement (ICC_3,1) between the two systems were assessed using separate interclass correlations coefficients. In addition, ensemble curves and associated 90% confidence intervals (CI90) were generated for the hip, knee, and ankle kinematics to enable between system comparisons throughout the gait cycle. Results showed that the Kinect^™ has the potential to be an effective clinical assessment device for sagittal plane hip and knee joint kinematics and for some spatiotemporal parameters during the stair gait negotiation.

System for automatic gait analysis based on a single RGB-D camera

Human gait analysis provides valuable information regarding the way of walking of a given subject. Low-cost RGB-D cameras, such as the Microsoft Kinect, are able to estimate the 3-D position of several body joints without requiring the use of markers. This 3-D information can be used to perform objective gait analysis in an affordable, portable, and non-intrusive way. In this contribution, we present a system for fully automatic gait analysis using a single RGB-D camera, namely the second version of the Kinect. Our system does not require any manual intervention (except for starting/stopping the data acquisition), since it firstly recognizes whether the subject is walking or not, and identifies the different gait cycles only when walking is detected. For each gait cycle, it then computes several gait parameters, which can provide useful information in various contexts, such as sports, healthcare, and biometric identification. The activity recognition is performed by a predictive model that distinguishes between three activities (walking, standing and marching), and between two postures of the subject (facing the sensor, and facing away from it). The model was built using a multilayer perceptron algorithm and several measures extracted from 3-D joint data, achieving an overall accuracy and F1 score of 98%. For gait cycle detection, we implemented an algorithm that estimates the instants corresponding to left and right heel strikes, relying on the distance between ankles, and the velocity of left and right ankles. The algorithm achieved errors for heel strike instant and stride duration estimation of 15 ± 25 ms and 1 ± 29 ms (walking towards the sensor), and 12 ± 23 ms and 2 ± 24 ms (walking away from the sensor). Our gait cycle detection solution can be used with any other RGB-D camera that provides the 3-D position of the main body joints.

What is the Best Configuration of Wearable Sensors to Measure Spatiotemporal Gait Parameters in Children with Cerebral Palsy?

Wearable inertial devices have recently been used to evaluate spatiotemporal parameters of gait in daily life situations. Given the heterogeneity of gait patterns in children with cerebral palsy (CP), the sensor placement and analysis algorithm may influence the validity of the results. This study aimed at comparing the spatiotemporal measurement performances of three wearable configurations defined by different sensor positioning on the lower limbs: (1) shanks and thighs, (2) shanks, and (3) feet. The three configurations were selected based on their potential to be used in daily life for children with CP and typically developing (TD) controls. For each configuration, dedicated gait analysis algorithms were used to detect gait events and compute spatiotemporal parameters. Fifteen children with CP and 11 TD controls were included. Accuracy, precision, and agreement of the three configurations were determined in comparison with an optoelectronic system as a reference. The three configurations were comparable for the evaluation of TD children and children with a low level of disability (CP-GMFCS I) whereas the shank-and-thigh-based configuration was more robust regarding children with a higher level of disability (CP-GMFCS II–III).

Identification of heel strike under a slippery condition

Kinematics at heel strike instant (HSI) has been used to quantify slip severity. However, methods to identify HSI remain ambiguous and have not been evaluated under slippery conditions. A glass force plate was used to observe the contact interface between shoe and floor under slippery conditions. HSIs identified from the video captured beneath the force plate and from the force plate and kinematics were compared. The results showed that HSIs identified with the video were closer to those identified with the normal force threshold (NFT) (9.0 ms ± 5.5 ms) than were most of those identified with kinematics. Slips with a longer distance travelled between NFT HSI and video HSI had a larger heel horizontal velocity (>0.8 m/s) and a smaller foot angular velocity (<100deg/s) at the NFT instant, and were still part of the forward swing. The results show that improved methods are needed over NFT to identify HSI, especially under slippery conditions.

The application of inertial measurements unit for the clinical evaluation and assessment of gait events

Foot drop is one of the common gait abnormalities which are difficult to detect, diagnose and evaluate. While various gait monitoring systems are available, many are computationally expensive and difficult to implement outside laboratory environments. This study introduces an in-house designed system based on inertial measurement units to capture the gait symptoms, specifically in the case of foot drop symptoms. The system specification and communication results, as well as filtering methods are discussed. Also, the pitch angle of thigh, shank and foot from a subject with no reported foot problem have been compared (gathered from identical equipment under similar conditions) to the same angle from a foot drop subject.

Validation of simplified centre of mass models during gait in individuals with chronic stroke

BACKGROUND

The feasibility of using a multiple segment (full-body) kinematic model in clinical gait assessment is difficult when considering obstacles such as time and cost constraints. While simplified gait models have been explored in healthy individuals, no such work to date has been conducted in a stroke population. The aim of this study was to quantify the errors of simplified kinematic models for chronic stroke gait assessment.

METHODS

Sixteen individuals with chronic stroke (>6months), outfitted with full body kinematic markers, performed a series of gait trials. Three centre of mass models were computed: (i) 13-segment whole-body model, (ii) 3 segment head-trunk-pelvis model, and (iii) 1 segment pelvis model. Root mean squared error differences were compared between models, along with correlations to measures of stroke severity.

FINDINGS

Error differences revealed that, while both models were similar in the mediolateral direction, the head-trunk-pelvis model had less error in the anteroposterior direction and the pelvis model had less error in the vertical direction. There was some evidence that the head-trunk-pelvis model error is influenced in the mediolateral direction for individuals with more severe strokes, as a few significant correlations were observed between the head-trunk-pelvis model and measures of stroke severity.

INTERPRETATION

These findings demonstrate the utility and robustness of the pelvis model for clinical gait assessment in individuals with chronic stroke. Low error in the mediolateral and vertical directions is especially important when considering potential stability analyses during gait for this population, as lateral stability has been previously linked to fall risk.

A New Proxy Measurement Algorithm with Application to the Estimation of Vertical Ground Reaction Forces Using Wearable Sensors

Measurement of the ground reaction forces (GRF) during walking is typically limited to laboratory settings, and only short observations using wearable pressure insoles have been reported so far. In this study, a new proxy measurement method is proposed to estimate the vertical component of the GRF (vGRF) from wearable accelerometer signals. The accelerations are used as the proxy variable. An orthogonal forward regression algorithm (OFR) is employed to identify the dynamic relationships between the proxy variables and the measured vGRF using pressure-sensing insoles. The obtained model, which represents the connection between the proxy variable and the vGRF, is then used to predict the latter. The results have been validated using pressure insoles data collected from nine healthy individuals under two outdoor walking tasks in non-laboratory settings. The results show that the vGRFs can be reconstructed with high accuracy (with an average prediction error of less than 5.0%) using only one wearable sensor mounted at the waist (L5, fifth lumbar vertebra). Proxy measures with different sensor positions are also discussed. Results show that the waist acceleration-based proxy measurement is more stable with less inter-task and inter-subject variability than the proxy measures based on forehead level accelerations. The proposed proxy measure provides a promising low-cost method for monitoring ground reaction forces in real-life settings and introduces a novel generic approach for replacing the direct determination of difficult to measure variables in many applications.

Mechanisms contributing to gait speed and metabolic cost in children with unilateral cerebral palsy

BACKGROUND

Gait speed and metabolic cost are indicators of functional capacity in children with cerebral palsy. Uncovering their mechanisms helps guide therapeutic actions.

OBJECTIVES

To investigate the contributions of energy-generating and energy-conserving mechanisms to gait speed and metabolic cost of children with unilateral cerebral palsy.

METHODS

Data on eccentric and concentric muscle work, co-contraction, elastic torque and vertical stiffness of the affected-limb, forcing torque of the non-affected limb, gait speed and metabolic cost were collected from 14 children with unilateral cerebral palsy, aged 6-12 years. Analyses included two groups of multiple regression models. The first group of models tested the association between each dependent variable (i.e., speed and metabolic cost) and the independent variables that met the input criteria. The second group verified the contribution of the non-selected biomechanical variables on the predictors of the first model.

RESULTS

Gait speed (R²=0.80) was predicted by elastic torque (β=0.62; 95%CI: 0.60, 0.63), vertical stiffness (β=-0.477; 95%CI: -0.479, -0.474) and knee co-contraction (β=0.27; 95%CI: -1.96, 2.49). The production of eccentric work by the affected limb proved relevant in adjusting the vertical stiffness (R²=0.42; β=-0.64; 95%CI: 0.86, -0.42); elastic torque of the affected-leg was associated with impulsive torque of the non-affected leg (R²=0.31; β=0.55; 95%CI: 0.46, 0.64). Metabolic cost of gait (R²=0.48) was partially predicted by knee co-contraction (β=0.69; 95%CI: 0.685, 0.694).

CONCLUSIONS

The chain of associations revealed by the two steps models helped uncover the mechanisms involved in the locomotion of children with unilateral cerebral palsy. Intervention that changes specific energy conserving and generating mechanisms may improve gait of these children.

Real-Time Gait Event Detection Based on Kinematic Data Coupled to a Biomechanical Model

Real-time detection of multiple stance events, more specifically initial contact (IC), foot flat (FF), heel off (HO), and toe off (TO), could greatly benefit neurorobotic (NR) and neuroprosthetic (NP) control. Three real-time threshold-based algorithms have been developed, detecting the aforementioned events based on kinematic data in combination with a biomechanical model. Data from seven subjects walking at three speeds on an instrumented treadmill were used to validate the presented algorithms, accumulating to a total of 558 steps. The reference for the gait events was obtained using marker and force plate data. All algorithms had excellent precision and no false positives were observed. Timing delays of the presented algorithms were similar to current state-of-the-art algorithms for the detection of IC and TO, whereas smaller delays were achieved for the detection of FF. Our results indicate that, based on their high precision and low delays, these algorithms can be used for the control of an NR/NP, with the exception of the HO event. Kinematic data is used in most NR/NP control schemes and is thus available at no additional cost, resulting in a minimal computational burden. The presented methods can also be applied for screening pathological gait or gait analysis in general in/outside of the laboratory.

MULTI-SEGMENTED TRUNK MOTION OF HEALTHY NON-ELDERLY ADULTS IN DIFFERENT DECADES OF LIFE

Traditionally, gait analysis models the trunk as one rigid body segment. This approach has limitations; it does not capture all the movements of this area of the body throughout locomotion. Lower-extremity-gait kinematics do not routinely change in healthy non-elderly adults in different decades of life; however, it is unknown if trunk kinematics will be altered during different activities of daily living as a function of age. The purpose of this study was to determine if a previously validated multi-segmented trunk model would detect trunk movement variations in non-elderly healthy adults in different decades of life. Thirty-four non-elderly healthy adults in various decades of life (20-29 years, 30-39 years, 40-49 years, and 50-59 years) completed two tasks of ambulatory daily living (level walking and stair descent). Trunk maximum angle during the gait cycle, timing of the trunk maximum angle during the gait cycle and trunk range of motion were examined using analysis of variance procedures. Findings are that age group did not affect the trunk kinematics of individuals in different decades of life, but that may not represent the experiences of elderly individuals.

Validation of a commercial inertial sensor system for spatiotemporal gait measurements in children

Although inertial sensor systems are becoming a popular tool for gait analysis in both healthy and pathological adult populations, there are currently no data on the validity of these systems for use with children. The purpose of this study was to validate spatiotemporal data from a commercial inertial sensor system (MobilityLab) in typically-developing children. Data from 10 children (5 males; 3.0-8.3 years, mean=5.1) were collected simultaneously from MobilityLab and 3D motion capture during gait at self-selected and fast walking speeds. Spatiotemporal parameters were compared between the two methods using a Bland-Altman method. The results indicate that, while the temporal gait measurements were similar between the two systems, MobilityLab demonstrated a consistent bias with respect to measurement of the spatial data (stride length). This error is likely due to differences in relative leg length and gait characteristics in children compared to the MobilityLab adult reference population used to develop the stride length algorithm. A regression-based equation was developed based on the current data to correct the MobilityLab stride length output. The correction was based on leg length, stride time, and shank range-of-motion, each of which were independently associated with stride length. Once the correction was applied, all of the spatiotemporal parameters evaluated showed good agreement. The results of this study indicate that MobilityLab is a valid tool for gait analysis in typically-developing children. Further research is needed to determine the efficacy of this system for use in children suffering from pathologies that impact gait mechanics.

THREE-DIMENSIONAL MULTI-SEGMENTED SPINE JOINT REACTION FORCES DURING COMMON WORKPLACE PHYSICAL DEMANDS/ACTIVITIES OF DAILY LIVING

Objective

The quantification of inter-segmental spine joint reaction forces during common workplace physical demands.

Background

Many spine reaction force models have focused on the L5/S1 or L4/L5 joints to quantify the vertebral joint reaction forces. However, the L5/S1 or L4/L5 approach neglects most of the intervertebral joints.

Methods

The current study presents a clinically applicable and noninvasive model which calculates the spinal joint reaction forces at six different regions of the spine. Subjects completed four ambulatory activities of daily living: level walking, obstacle crossing, stair ascent, and stair descent.

Results

Peak joint spinal reaction forces were compared between tasks and spine regions. Differences existed in the bodyweight normalized vertical joint reaction forces where the walking (8.05±3.19N/kg) task had significantly smaller peak reaction forces than the stair descent (12.12±1.32N/kg) agreeing with lower extremity data comparing walking and stair descent tasks.

Conclusion

This method appears to be effective in estimating the joint reaction forces using a segmental spine model. The results suggesting the main effect of peak reactions forces in the segmental spine can be influenced by task.

A comparison of kinematic-based gait event detection methods in a self-paced treadmill application

Kinematic-based algorithms for detecting gait events are efficient and useful in the absence of (reliable) kinetic data. However, the validity of these kinematic-based algorithms for self-paced treadmill walking is unknown, particularly given the influence of walking speed on such data. We quantified offsets in event detection of four foot kinematics-based algorithms (horizontal position, horizontal velocity, vertical velocity, and sagittal resultant velocity) relative to events determined by a threshold in vertical ground reaction force among seven uninjured individuals - and nine with unilateral transtibial amputation - walking on a self-paced treadmill. Across walking speeds from 0.48-1.64m/s (0.5-31.7% CV), offsets ranged from -7 to +3 frames (≈83.3ms) in heel strike, and -3 to +5 frames (≈66.6 ms) in toe off. Regardless of method, offsets in heel strike were not influenced (-0.01<r<0.01, all P>0.61) by variability in walking speed. However, offsets in toe-off were positively correlated with variability in walking speed for the horizontal position (r=0.539; P<0.001) and velocity (r=0.463; P<0.001) algorithms, and negatively correlated (r=-0.317; P<0.001) for the vertical velocity algorithm; offsets from the sagittal resultant velocity algorithm, with thresholds adjusted for walking speed, were not strongly associated (r=0.126; P=0.27). Although relatively minimal offsets support the applicability of these algorithms to self-paced walking, for individuals with asymptomatic and pathological gait patterns, sagittal resultant velocity of the foot produces the most consistent event detection over the widest range of (and variability in) walking speeds.

External rotation elastic bands at the lower limb decrease rearfoot eversion during walking: a preliminary proof of concept

Background

Reducing rearfoot eversion is a commonly desired effect in clinical practice to prevent or treat musculoskeletal dysfunction. Interventions that pull the lower limb into external rotation may reduce rearfoot eversion.

Objective

This study investigated whether the use of external rotation elastic bands, of different levels of stiffness, will decrease rearfoot eversion during walking. We hypothesized that the use of elastic bands would decrease rearfoot eversion and that the greater the band stiffness, the greater the eversion reduction.

Method

Seventeen healthy participants underwent three-dimensional kinematic analysis of the rearfoot and shank. The participants walked on a treadmill with and without high- and low-stiffness bands. Frontal-plane kinematics of the rearfoot-shank joint complex was obtained during the stance phase of walking. Repeated-measures ANOVAs were used to compare discrete variables that described rearfoot eversion-inversion: mean eversion-inversion; eversion peak; and eversion-inversion range of motion.

Results

The low-stiffness and high-stiffness bands significantly decreased eversion and increased mean eversion-inversion (p≤0.037) and eversion peak (p≤0.006) compared with the control condition. Both bands also decreased eversion-inversion range of motion (p≤0.047) compared with control by reducing eversion. The high-stiffness band condition was not significantly different from the low-stiffness band condition for any variables (p≥0.479).

Conclusion

The results indicated that the external rotation bands decreased rearfoot eversion during walking. This constitutes preliminary experimental evidence suggesting that increasing external rotation moments at the lower limb may reduce rearfoot eversion, which needs further testing.

The Effect of Walking Speed on Foot Kinematics is Modified When Increased Pronation is Induced

BACKGROUND

The relation between walking speed and foot kinematics during gait is not well established, and neither is it clear whether this relation is modified in the presence of factors expected to increase pronation (eg, abnormal foot alignment). Understanding how foot kinematics is affected by walking speed under varying conditions could contribute to our understanding of stresses to the musculoskeletal system during walking. We evaluated the effect of walking speed on foot kinematics in the frontal plane during gait and determined whether this effect is modified by using medially inclined insoles that force the foot into increased pronation.

METHODS

Twenty-six healthy young adults were assessed while walking on a treadmill wearing flat insoles and wearing medially inclined insoles. Foot kinematics in the frontal plane was measured with a three-dimensional motion analysis system. Data were analyzed during the stance phase of gait.

RESULTS

There was no main effect of speed on average calcaneal position. However, there was a significant insole type × walking speed interaction effect. In the flat insole condition, increased walking speed was associated with a less inverted average calcaneal position (or greater magnitudes of eversion), whereas in the inclined insole condition, higher speeds were associated with a less everted average calcaneal position (or increased magnitudes of inversion).

CONCLUSIONS

The magnitude of foot eversion increases at faster gait speeds under typical conditions. In the presence of factors that induce excessive pronation, however, this effect is reversed. Results suggest that individuals use greater active control of foot motion at faster speeds in the presence of excessive pronation to improve push-off efficiency. Potential clinical consequences of this functional strategy are discussed.

FreeWalker: a smart insole for longitudinal gait analysis

Gait analysis is an important diagnostic measure to investigate the pattern of walking. Traditional gait analysis is generally carried out in a gait lab, with equipped force and body tracking sensors, which needs a trained medical professional to interpret the results. This procedure is tedious, expensive, and unreliable and makes it difficult to track the progress across multiple visits. In this paper, we present a smart insole called FreeWalker, which provides quantitative gait analysis outside the confinement of traditional lab, at low- cost. The insole consists of eight pressure sensors and two motion tracking sensors, i.e. 3-axis accelerometer and 3-axis gyroscope. This enables measurement of under-foot pressure distribution and motion sequences in real-time. The insole is enabled with onboard SD card as well as wireless data transmission, which help in continuous gait-cycle analysis. The data is then sent to a gateway, for analysis and interpretation of data, using a user interface where gait features are graphically displayed. We also present validation result of a subject's left foot, who was asked to perform a specific task. Experiment results show that we could achieve a data-sampling rate of over 1 KHz, transmitting data up to a distance of 20 meter and maintain a battery life of around 24 hours. Taking advantage of these features, FreeWalker can be used in various applications, like medical diagnosis, rehabilitation, sports and entertainment.

Estimation of Temporal Gait Events from a Single Accelerometer Through the Scale-Space Filtering Idea

The purpose of this paper is to develop an accelerometry system capable of performing gait event demarcation and calculation of temporal parameters using a single waist-mounted device. Particularly, a mobile phone positioned over the L2 vertebra is used to acquire trunk accelerations during walking. Signals from the acceleration magnitude and the vertical acceleration are smoothed through different filters. Cut-off points between filtered signals as a result of convolving with varying levels of Gaussian filters and other robust features against temporal variation and noise are used to identify peaks that correspond to gait events. Five pre-frail older adults and five young healthy adults were recruited in an experiment. Cadence, step/stride time, step/stride CV, step asymmetry and percentages of the stance/swing and single/double support phases, among the two groups of different mobility were quantified by the system.

Towards Real-Time Detection of Gait Events on Different Terrains Using Time-Frequency Analysis and Peak Heuristics Algorithm

Real-time detection of gait events can be applied as a reliable input to control drop foot correction devices and lower-limb prostheses. Among the different sensors used to acquire the signals associated with walking for gait event detection, the accelerometer is considered as a preferable sensor due to its convenience of use, small size, low cost, reliability, and low power consumption. Based on the acceleration signals, different algorithms have been proposed to detect toe off (TO) and heel strike (HS) gait events in previous studies. While these algorithms could achieve a relatively reasonable performance in gait event detection, they suffer from limitations such as poor real-time performance and are less reliable in the cases of up stair and down stair terrains. In this study, a new algorithm is proposed to detect the gait events on three walking terrains in real-time based on the analysis of acceleration jerk signals with a time-frequency method to obtain gait parameters, and then the determination of the peaks of jerk signals using peak heuristics. The performance of the newly proposed algorithm was evaluated with eight healthy subjects when they were walking on level ground, up stairs, and down stairs. Our experimental results showed that the mean F1 scores of the proposed algorithm were above 0.98 for HS event detection and 0.95 for TO event detection on the three terrains. This indicates that the current algorithm would be robust and accurate for gait event detection on different terrains. Findings from the current study suggest that the proposed method may be a preferable option in some applications such as drop foot correction devices and leg prostheses.

Concurrent validity and reliability of using ground reaction force and center of pressure parameters in the determination of leg movement initiation during single leg lift

Postural adjustment evaluations during single leg lift requires the initiation of heel lift (T1) identification. T1 measured by means of motion analyses system is the most reliable approach. However, this method involves considerable workspace, expensive cameras, and time processing data and setting up laboratory. The use of ground reaction forces (GRF) and centre of pressure (COP) data is an alternative method as its data processing and setting up is less time consuming. Further, kinetic data is normally collected using frequency samples higher than 1000Hz whereas kinematic data are commonly captured using 50-200Hz. This study describes the concurrent-validity and reliability of GRF and COP measurements in determining T1, using a motion analysis system as reference standard. Kinematic and kinetic data during single leg lift were collected from ten participants. GRF and COP data were collected using one and two force plates. Displacement of a single heel marker was captured by means of ten Vicon(©) cameras. Kinetic and kinematic data were collected using a sample frequency of 1000Hz. Data were analysed in two stages: identification of key events in the kinetic data, and assessing concurrent validity of T1 based on the chosen key events with T1 provided by the kinematic data. The key event presenting the least systematic bias, along with a narrow 95% CI and limits of agreement against the reference standard T1, was the Baseline COPy event. Baseline COPy event was obtained using one force plate and presented excellent between-tester reliability.

Comparison between passive vision-based system and a wearable inertial-based system for estimating temporal gait parameters related to the GAITRite electronic walkway

Quantitative gait analysis allows clinicians to assess the inherent gait variability over time which is a functional marker to aid in the diagnosis of disabilities or diseases such as frailty, the onset of cognitive decline and neurodegenerative diseases, among others. However, despite the accuracy achieved by the current specialized systems there are constraints that limit quantitative gait analysis, for instance, the cost of the equipment, the limited access for many people and the lack of solutions to consistently monitor gait on a continuous basis. In this paper, two low-cost systems for quantitative gait analysis are presented, a wearable inertial system that relies on two wireless acceleration sensors mounted on the ankles; and a passive vision-based system that externally estimates the measurements through a structured light sensor and 3D point-cloud processing. Both systems are compared with a reference clinical instrument using an experimental protocol focused on the feasibility of estimating temporal gait parameters over two groups of healthy adults (five elders and five young subjects) under controlled conditions. The error of each system regarding the ground truth is computed. Inter-group and intra-group analyses are also conducted to transversely compare the performance between both technologies, and of these technologies with respect to the reference system. The comparison under controlled conditions is required as a previous stage towards the adaptation of both solutions to be incorporated into Ambient Assisted Living environments and to provide continuous in-home gait monitoring as part of the future work.

Reliability of ground reaction forces in the aquatic environment

The aim of this study was to verify the reliability of the kinetic parameters of gait using an underwater force platform. A total of 49 healthy participants with a median age of 21years were included. The kinetic gait data were collected using a 0.6×0.6×0.1m aquatic force plate (Bertec®), set in a pool (15×13×1.30m) with a water depth of 1.20m and water temperature of 32.5°C. Participants walked 10m before reaching the platform, which was fixed to the ground. Participants were instructed to step onto the platform with their preferred limb and data from three valid attempts were used to calculate the average values. A 48-h interval between tests was used for the test-retest reliability. Data were analyzed using interclass correlation coefficients (ICC) and results demonstrated that reliability ranged from poor to excellent, with ICC scores of between 0.24 and 0.87 and mean differences between (d¯)=-0.01 and 0.002. The highest reliability values were found for the vertical (Fz) and the lowest for the mediolateral components (Fy). In conclusion, the force platform is reliable for assessing the vertical and anteroposterior components of power production rates in water, however, caution should be applied when using this instrument to evaluate the mediolateral component in this environment.

Predicting ground contact events for a continuum of gait types: An application of targeted machine learning using principal component analysis

An ongoing challenge in the application of gait analysis to clinical settings is the standardized detection of temporal events, with unobtrusive and cost-effective equipment, for a wide range of gait types. The purpose of the current study was to investigate a targeted machine learning approach for the prediction of timing for foot strike (or initial contact) and toe-off, using only kinematics for walking, forefoot running, and heel-toe running. Data were categorized by gait type and split into a training set (∼30%) and a validation set (∼70%). A principal component analysis was performed, and separate linear models were trained and validated for foot strike and toe-off, using ground reaction force data as a gold-standard for event timing. Results indicate the model predicted both foot strike and toe-off timing to within 20ms of the gold-standard for more than 95% of cases in walking and running gaits. The machine learning approach continues to provide robust timing predictions for clinical use, and may offer a flexible methodology to handle new events and gait types.

INFLUENCE OF VARIOUS DAILY TASKS ON SEGMENTED TRUNK KINEMATICS

Back pain can affect up to 65% of the American population and cost the health care system approximately fifty billion dollars each year. Due to the difficulty with recording spine/trunk movement, several methods and models exist. The myriad of methods and the need for understanding of spine/trunk motion has led to a lack in a ‘gold-standard’ of treatment for individuals with back pain. Therefore, the purpose of this study was to examine the effect of different activities of daily living on the kinematics of individual trunk segments in young adults to determine how common ambulatory tasks will alter trunk motion compared to level walking.

Young healthy adults completed, in a random order, four activities of daily living: level walking, obstacle crossing, stair ascent and descent using a previously validated model. Subjects were outfitted with a full body marker set which included a segmented trunk. Multi-segmented trunk angles between the three inferior segments, sacrum to lower lumbar [SLL], lower lumbar to upper lumbar [LLUL] and upper lumbar to lower thorax [ULLT], were calculated and compared between tasks. Peak flexion angles, instance of peak angle and range of motion were analyzed.

The overall hypothesis that different spine levels will have altered kinematics during various activities of daily living was supported. Stair descent had smaller peak flexion angles than obstacle crossing and stair ascent. The instance of peak angle were different depending on trunk angle and daily task. The most inferior trunk angle — Sacrum-to-Lower Lumbar — had the largest range of motion during all four tasks in all three (sagittal, frontal and transverse) planes of motion.

This study was able to show how various activities of daily living produce different motions in the three inferior segments of a multi-segmented trunk method. The results of this study are the first steps in understanding how the trunk responds on a daily basis and how those responses could lead to back pain.

Gait Partitioning Methods: A Systematic Review

In the last years, gait phase partitioning has come to be a challenging research topic due to its impact on several applications related to gait technologies. A variety of sensors can be used to feed algorithms for gait phase partitioning, mainly classifiable as wearable or non-wearable. Among wearable sensors, footswitches or foot pressure insoles are generally considered as the gold standard; however, to overcome some inherent limitations of the former, inertial measurement units have become popular in recent decades. Valuable results have been achieved also though electromyography, electroneurography, and ultrasonic sensors. Non-wearable sensors, such as opto-electronic systems along with force platforms, remain the most accurate system to perform gait analysis in an indoor environment. In the present paper we identify, select, and categorize the available methodologies for gait phase detection, analyzing advantages and disadvantages of each solution. Finally, we comparatively examine the obtainable gait phase granularities, the usable computational methodologies and the optimal sensor placements on the targeted body segments.

Analysis of several methods and inertial sensors locations to assess gait parameters in able-bodied subjects

PURPOSE

The purpose of this paper was to determine which types of inertial sensors and which advocated locations should be used for reliable and accurate gait event detection and temporal parameter assessment in normal adults. In addition, we aimed to remove the ambiguity found in the literature of the definition of the initial contact (IC) from the lumbar accelerometer. Acceleration and angular velocity data was gathered from the lumbar region and the distal edge of each shank. This data was evaluated in comparison to an instrumented treadmill and an optoelectronic system during five treadmill speed sessions.

RESULTS

The lumbar accelerometer showed that the peak of the anteroposterior component was the most accurate for IC detection. Similarly, the valley that followed the peak of the vertical component was the most precise for terminal contact (TC) detection. Results based on ANOVA and Tukey tests showed that the set of inertial methods was suitable for temporal gait assessment and gait event detection in able-bodied subjects. For gait event detection, an exception was found with the shank accelerometer. The tool was suitable for temporal parameters assessment, despite the high root mean square error on the detection of IC (RMSEIC) and TC (RMSETC). The shank gyroscope was found to be as accurate as the kinematic method since the statistical tests revealed no significant difference between the two techniques for the RMSE off all gait events and temporal parameters.

CONCLUSION

The lumbar and shank accelerometers were the most accurate alternative to the shank gyroscope for gait event detection and temporal parameters assessment, respectively.

Effect of walking speed on gait sub phase durations

Gait phase durations are important spatiotemporal parameters in different contexts such as discrimination between healthy and pathological gait and monitoring of treatment outcomes after interventions. Although gait phases strongly depend on walking speed, the influence of different speeds has rarely been investigated in literature. In this work, we examined the durations of the stance sub phases and the swing phase for 12 different walking speeds ranging from 0.6 to 1.7 m/s in 21 healthy subjects using infrared cinematography and an instrumented treadmill. We separated the stance phase into loading response, mid stance, terminal stance and pre-swing phase and we performed regression modeling of all phase durations with speed to determine general trends. With an increasing speed of 0.1m/s, stance duration decreased while swing duration increased by 0.3%. All distinct stance sub phases changed significantly with speed. These findings suggest the importance of including all distinct gait sub phases in spatiotemporal analyses, especially when different walking speeds are involved.

Using horizontal heel displacement to identify heel strike instants in normal gait

Heel strike instants are an important component of gait analyses, yet accurate detection can be difficult without a force plate. This paper presents two novel techniques for kinematic heel strike instant (kHSI) detection which examined maximal resultant horizontal heel displacement (HHD). Each of these HHD techniques calculates HHD from a selected reference location of either the stance ankle or stance heel to the swing heel. The proposed techniques, along with other previously established techniques, were validated against a 10N force plate threshold. Fifty-four healthy adults walked overground at both normal and fast speeds while wearing athletic shoes. The reported true and absolute errors were as low as 3.2 (4.4) and 5.7 (3.4)ms, respectively, across 8678kHSI when using the stance ankle as a reference, which significantly outperformed (p<0.0001) the established techniques. Gait speed was shown to have a significant effect (p<0.0001) on HHD-determined kHSI, as well as the three other techniques evaluated, highlighting the need for condition-specific identification of kHSI.

A universal approach to determine footfall timings from kinematics of a single foot marker in hoofed animals

The study of animal movement commonly requires the segmentation of continuous data streams into individual strides. The use of forceplates and foot-mounted accelerometers readily allows the detection of the foot-on and foot-off events that define a stride. However, when relying on optical methods such as motion capture, there is lack of validated robust, universally applicable stride event detection methods. To date, no method has been validated for movement on a circle, while algorithms are commonly specific to front/hind limbs or gait. In this study, we aimed to develop and validate kinematic stride segmentation methods applicable to movement on straight line and circle at walk and trot, which exclusively rely on a single, dorsal hoof marker. The advantage of such marker placement is the robustness to marker loss and occlusion. Eight horses walked and trotted on a straight line and in a circle over an array of multiple forceplates. Kinetic events were detected based on the vertical force profile and used as the reference values. Kinematic events were detected based on displacement, velocity or acceleration signals of the dorsal hoof marker depending on the algorithm using (i) defined thresholds associated with derived movement signals and (ii) specific events in the derived movement signals. Method comparison was performed by calculating limits of agreement, accuracy, between-horse precision and within-horse precision based on differences between kinetic and kinematic event. In addition, we examined the effect of force thresholds ranging from 50 to 150 N on the timings of kinetic events. The two approaches resulted in very good and comparable performance: of the 3,074 processed footfall events, 95% of individual foot on and foot off events differed by no more than 26 ms from the kinetic event, with average accuracy between −11 and 10 ms and average within- and between horse precision ≤8 ms. While the event-based method may be less likely to suffer from scaling effects, on soft ground the threshold-based method may prove more valuable. While we found that use of velocity thresholds for foot on detection results in biased event estimates for the foot on the inside of the circle at trot, adjusting thresholds for this condition negated the effect. For the final four algorithms, we found no noteworthy bias between conditions or between front- and hind-foot timings. Different force thresholds in the range of 50 to 150 N had the greatest systematic effect on foot-off estimates in the hind limbs (up to on average 16 ms per condition), being greater than the effect on foot-on estimates or foot-off estimates in the forelimbs (up to on average ±7 ms per condition).

Automated gait temporal-spatial assessment from non-motorized treadmill belt speed data

Non-motorized treadmills (NMT) provide belt speed data that can be used to estimate work output, and potentially, gait temporal-spatial parameters that provide an improved understanding of gait performance. The purpose of this study was to determine the validity of an automated technique that uses belt speed data from an NMT to estimate temporal-spatial gait parameters. Seventeen injury-free adult participants performed a series of 20-s, metronome-guided walking and running trials for each of eight predetermined cadence conditions (72-200 steps/min). Two NMT-based cadence algorithms [PSD estimated cadence (PEC) and threshold estimated cadence (TEC)], and one NMT-based step length algorithm (NMT_SL) were evaluated for their ability to predict traditional motion analysis-based measures of cadence and step length (MAC and MA_SL, respectively). The results of this study demonstrate that both the PEC and TEC algorithms were capable of predicting MAC with a standard error of the estimate (SEE) less than four steps/min (R(2) = 0.997 and R(2) = 0.993, respectively). Predictions of MA_SL from NMT_SL were separated by gait type (walking vs. running) to account for an obvious separation in the step length data with a qualitative gait change. When applied to walking data, NMT_SL was capable of predicting MA_SL with an SEE of 23 mm (R(2) = 0.96). When applied to running data, NMT_SL was capable of predicting MA_SL with an SEE of 44 mm (R(2) = 0.80). The assessment of the novel technique suggests that it is feasible to use non-motorized treadmill belt speed data to predict gait events and analyze simple gait metrics. Future research should evaluate the applicability of these algorithms for use with participants/patients presenting with pathological gait.

A Novel Approach for Toe Off Estimation During Locomotion and Transitions on Ramps and Level Ground

Identification of the toe off event is critical in many gait applications. Accelerometer threshold-based algorithms lack adaptability and have not been tested for transitions between locomotion states. We describe a new approach for toe off identification using one accelerometer in over ground and ramp walking, including transitions. The method uses invariant foot acceleration features in the segment of gait, where toe off is probable. Wavelet analysis of foot acceleration is used to derive a unique feature in a particular frequency band, yielding estimated toe off occurrence. We tested the new method for five conditions: over ground walking (W), ramp ascending (RA), ramp descending (RD); transitions between states (W-RA, W-RD). Mean absolute estimation error was 17.4 ± 12.5, 13.8 ± 8.5, and 22.0 ± 16.4 ms for steady states W, RA, and RD, 20.1 ± 15.5, and 17.1 ± 13.7 ms for transitions W-RA and W-RD, respectively. Algorithm performance was equivalent across all pairs of transition and locomotion state except between RA and RD ( p = 0.03), demonstrating adaptability. The db1 wavelet outperformed db2 across states and transitions (p < 0.01). The presented algorithm is a simple, robust approach for toe off detection.

Can sacral marker approximate center of mass during gait and slip-fall recovery among community-dwelling older adults?

Falls are prevalent in older adults. Dynamic stability of body center of mass (COM) is critical for maintaining balance. A simple yet accurate tool to evaluate COM kinematics is essential to examine the COM stability. The purpose of this study was to determine the extent to which the COM position derived from body segmental analysis can be approximated by a single (sacral) marker during unperturbed (regular walking) and perturbed (gait-slip) gait. One hundred eighty seven older adults experienced an unexpected slip after approximately 10 regular walking trials. Two trials, the slip trial and the preceding regular walking trial, monitored with a motion capture system and force plates, were included in the present study. The COM positions were calculated by using the segmental analysis method wherein, the COM of all body segments was calculated to further estimate the body COM position. These body COM positions were then compared with those of the sacral marker placed at the second sacral vertebra for both trials. Results revealed that the COM positions were highly correlated with those of the sacrum׳s over the time intervals investigated for both walking (coefficient of correlation R>0.97) and slip (R>0.90) trials. There were detectable kinematic difference between the COM and the sacral for both trials. Our results indicated that the sacral marker can be used as a simple approximation of body COM for regular walking, and to somewhat a lesser extent, upon a slip. The benefits from the simplicity appear to overweigh the limitations in accuracy.

Method to classify elderly subjects as fallers and non-fallers based on gait energy image

Falls are one of the leading causes of injuries among the elderly. Therefore, distinguishing fallers and performing preventive actions is vitally important. A new variation of the gait energy image (GEI) called coloured gait energy image (CGEI) is proposed for classifying subjects as fallers and non-fallers and for visualising their gait patterns. Eight elderly fallers, eight elderly non-fallers and eight young subjects performed timed up and go (TUG) test, which is one of the well-known clinical tools for fall risk assessment and contains two gait sequences. Subjects were also asked to perform two other variations of the TUG test, namely TUG with manual load and TUG with cognitive load. Gait sequences were extracted from the TUG test based on the opinion of three human observers. Then the gait cycles were automatically extracted from the walking sequence and divided into three phases, corresponding to double support and first and second half of single support. Next, the GEI of each phase was generated and formed one of the colour components of CGEI. Histogram-based features obtained from CGEI were then used to classify the video collected from walking sequences of elderly fallers and non-fallers. Correct classification rate was improved by approximately 27% compared with the standard TUG test.

Automatic initial contact detection during overground walking for clinical use

The division of gait into cycles is crucial for identifying deficits in locomotion, particularly to monitor disease progression or rehabilitative recovery. Initial contact (IC) events are often used to separate movement into repetitive cycles yet automatic methods for IC identification in pathological gait are limited in both number and capacity. The aim of this work was to develop a more precise algorithm in IC detection. A projected heel markers distance (PHMD) algorithm is presented here and compared for accuracy to the high pass algorithm (HPA) in IC identification. Kinematic gait data from two clinical cohorts were analyzed and processed automatically for IC detection: (1) unilateral total hip arthroplasty (THA) patients (n=27) and (2) cerebral palsy pediatric (CPP) patients (n=20). IC events determined by the two algorithms were benchmarked against the IC events detected manually and from force plates. The PHMD method detected 96.6% IC events in THA patients and 99.1% in CPP patients with an average error of 5.3 ms and 18.4 ms. The HPA method detected 99.1% IC events in THA patients and 97.3% IC events in CPP patients, with an average error of 57.5 ms and 10.2 ms. PHMD identified no superfluous IC events, whereas 51.5% of all THA IC and 47.6% of CPP IC were superfluous events requiring manual deletion with HPA. With the superior comparison against the current gold standard, the PHMD algorithm appears valid for a wide spectrum of clinical data sets and allows for precise, fully automatic processing of kinematic gait data without additional sensors, triggers, or force plates.

Adaptive control of center of mass (global) motion and its joint (local) origin in gait

Dynamic gait stability can be quantified by the relationship of the motion state (i.e. the position and velocity) between the body center of mass (COM) and its base of support (BOS). Humans learn how to adaptively control stability by regulating the absolute COM motion state (i.e. its position and velocity) and/or by controlling the BOS (through stepping) in a predictable manner, or by doing both simultaneously following an external perturbation that disrupts their regular relationship. Post repeated-slip perturbation training, for instance, older adults learned to forward shift their COM position while walking with a reduced step length, hence reduced their likelihood of slip-induced falls. How and to what extent each individual joint influences such adaptive alterations is mostly unknown. A three-dimensional individualized human kinematic model was established. Based on the human model, sensitivity analysis was used to systematically quantify the influence of each lower limb joint on the COM position relative to the BOS and the step length during gait. It was found that the leading foot had the greatest effect on regulating the COM position relative to the BOS; and both hips bear the most influence on the step length. These findings could guide cost-effective but efficient fall-reduction training paradigm among older population.