Effect of walking speed on gait sub phase durations

Gait signature changes with walking speed are similar among able-bodied young adults despite persistent individual-specific differences

Understanding individuals' distinct movement patterns is crucial for health, rehabilitation, and sports. Recently, we developed a machine learning-based framework to show that "gait signatures" describing the neuromechanical dynamics governing able-bodied and post-stroke gait kinematics remain individual-specific across speeds. However, we only evaluated gait signatures within a limited speed range and number of participants, using only sagittal plane (i.e., 2D) joint angles. Here we characterized changes in gait signatures across a wide range of speeds, from very slow (0.3 m/s) to exceptionally fast (above the walk-to-run transition speed) in 17 able-bodied young adults. We further assessed whether 3D kinematic and/or kinetic (ground reaction forces, joint moments, and powers) data would improve the discrimination of gait signatures. Our study showed that gait signatures remained individual-specific across walking speeds: Notably, 3D kinematic signatures achieved exceptional accuracy (99.8%, confidence interval (CI) 99.1-100%) in classifying individuals, surpassing both 2D kinematics and 3D kinetics. Moreover, participants exhibited consistent, predictable linear changes in their gait signatures across the entire speed range. These changes were associated with participants' preferred walking speeds, balance ability, cadence, and step length. These findings support gait signatures as a tool to characterize individual differences in gait and predict speed-induced changes in gait dynamics.

Gait phase normalization resolves the problem of different phases being compared in gait cycle normalization

For time-continuous analysis of gait, the problem of variations in cycle durations is resolved by normalizing to the gait cycle, but results depend on the definition of the cycle start. Gait cycle normalization ignores variations in gait phase durations, which results in averaging and comparing data across different phases. We propose gait phase normalization as part of a comprehensive method for independently analyzing magnitude and timing differences. First, gait phases are identified and differences in absolute and/or relative timing of phase durations or any point of interest between conditions or groups are analyzed using standard statistics. Next, time-continuous gait data is normalized to gait phases, and statistical parametric mapping (SPM) is used to assess magnitude differences in gait data. This approach is demonstrated on data recorded from ten young healthy adults walking on a treadmill at five different speeds. Sagittal knee angle was normalized to gait cycle or gait phase using five different gait cycle start events. Walking at different speeds resulted in significant changes in gait phase durations, highlighting a problem ignored by gait cycle normalization. SPM results for knee angle normalized to gait cycle varied from normalization to gait phases. Gait phase normalized SPM results were robust to the definition of the cycle start, in contrast to gait cycle normalized data. The approach of analyzing phase durations and normalizing data to gait phases overcomes previous limitations and enables a comprehensive analysis of magnitude and timing differences in time-continuous gait data and could be readily adapted to other tasks.

Instantaneous phase of rhythmic behaviour under volitional control

The phase of signals representing cyclic behavioural patterns provides valuable information for understanding the mechanisms driving the observed behaviours. Methods usually adopted to estimate the phase, which are based on projecting the signal onto the complex plane, have strict requirements on its frequency content, which limits their application. To overcome these limitations, input signals can be processed using band-pass filters or decomposition techniques. In this paper, we briefly review these approaches and propose a new one. Our approach is based on the principles of Empirical Mode Decomposition (EMD), but unlike EMD, it does not aim to decompose the input signal. This avoids the many problems that can occur when extracting a signal's components one by one. The proposed approach estimates the phase of experimental signals that have one main oscillatory component modulated by slower activity and perturbed by weak, sparse, or random activity at faster time scales. We illustrate how our approach works by estimating the phase dynamics of synthetic signals and real-world signals representing knee angles during flexion/extension activity, heel height during gait, and the activity of different organs involved in speech production.

Gait signature changes with walking speed are similar among able-bodied young adults despite persistent individual-specific differences

Understanding individuals' distinct movement patterns is crucial for health, rehabilitation, and sports. Recently, we developed a machine learning-based framework to show that "gait signatures" describing the neuromechanical dynamics governing able-bodied and post-stroke gait kinematics remain individual-specific across speeds. However, we only evaluated gait signatures within a limited speed range and number of participants, using only sagittal plane (i.e., 2D) joint angles. Here we characterized changes in gait signatures across a wide range of speeds, from very slow (0.3 m/s) to exceptionally fast (above the walk-to-run transition speed) in 17 able-bodied young adults. We further assessed whether 3D kinematic and/or kinetic (ground reaction forces, joint moments, and powers) data would improve the discrimination of gait signatures. Our study showed that gait signatures remained individual-specific across walking speeds: Notably, 3D kinematic signatures achieved exceptional accuracy (99.8%, confidence interval (CI): 99.1-100%) in classifying individuals, surpassing both 2D kinematics and 3D kinetics. Moreover, participants exhibited consistent, predictable linear changes in their gait signatures across the entire speed range. These changes were associated with participants' preferred walking speeds, balance ability, cadence, and step length. These findings support gait signatures as a tool to characterize individual differences in gait and predict speed-induced changes in gait dynamics.

Medial Temporal Lobe Atrophy in Older Adults With Subjective Cognitive Impairments Affects Gait Parameters in the Spatial Navigation Task

Both navigation abilities and gait can be affected by the atrophy in the medial temporal cortex. This study aimed to determine whether navigation abilities could differentiate seniors with and without medial temporal lobe atrophy who complained about their cognitive status. The participants, classified to either the medial temporal atrophy group (n = 23) or the control group (n = 22) underwent neuropsychological assessment and performed a spatial navigation task while their gait parameters were recorded. The study showed no significant differences between the two groups in memory, fluency, and semantic knowledge or typical measures of navigating abilities. However, gait parameters, particularly the propulsion index during certain phases of the navigation task, distinguished between seniors with and without medial temporal lobe lesions. These findings suggest that the gait parameters in the navigation task may be a valuable tool for identifying seniors with cognitive complaints and subtle medial temporal atrophy.

Bi-Directional Long Short-Term Memory-Based Gait Phase Recognition Method Robust to Directional Variations in Subject's Gait Progression Using Wearable Inertial Sensor

Inertial Measurement Unit (IMU) sensor-based gait phase recognition is widely used in medical and biomechanics fields requiring gait data analysis. However, there are several limitations due to the low reproducibility of IMU sensor attachment and the sensor outputs relative to a fixed reference frame. The prediction algorithm may malfunction when the user changes their walking direction. In this paper, we propose a gait phase recognition method robust to user body movements based on a floating body-fixed frame (FBF) and bi-directional long short-term memory (bi-LSTM). Data from four IMU sensors attached to the shanks and feet on both legs of three subjects, collected via the FBF method, are processed through preprocessing and the sliding window label overlapping method before inputting into the bi-LSTM for training. To improve the model's recognition accuracy, we selected parameters that influence both training and test accuracy. We conducted a sensitivity analysis using a level average analysis of the Taguchi method to identify the optimal combination of parameters. The model, trained with optimal parameters, was validated on a new subject, achieving a high test accuracy of 86.43%.

Immediate and Short-Term Effects of In-Shoe Heel-Lift Orthoses on Clinical and Biomechanical Outcomes in Patients With Insertional Achilles Tendinopathy

Background

Physical therapists frequently employ heel lifts as an intervention to reduce Achilles tendon pain and restore function.

Purpose

To determine the short-term effect of heel lifts on clinical and gait outcomes in participants with insertional Achilles tendinopathy (IAT).

Study Design

Case series; Level of evidence, 4.

Methods

Participants with IAT underwent eligibility screening and completed assessments at baseline and 2 weeks later. Primary outcomes included symptom severity (Victoria Institute of Sports Assessment-Achilles [VISA-A]), gait analysis with the 10-m walk-test at 2 speeds (normal and fast), and pain during walking. Pain and gait analysis were assessed under 3 conditions: before fitting 20-mm heel lifts, immediately after heel-lift fitting, and after 2 weeks of wearing heel lifts. Ultrasound images and measurements at the Achilles insertion were obtained from prone and standing positions (with and without heel lifts). Spatiotemporal gait parameters and tibial tilt angles were evaluated at normal speed using inertia measurement units during the 3 study conditions. Differences between the conditions were analyzed using paired t test or analysis of variance.

Results

Overall, 20 participants (12 female, 13 with bilateral IAT; mean age, 51 ± 9.3 years; mean body mass index 31.6 ± 6.8 kg/m2) completed all assessments. Symptom severity (VISA-A) of the more symptomatic side significantly improved at 2 weeks (60 ± 20.6) compared with baseline (52.2 ± 20.4; P < .01). Pain during gait (Numeric Pain Rating Scale) was significantly reduced immediately after heel-lift fitting (0.7 ± 2.0) when compared with baseline (2.2 ± 2.7, P = .043). Spatiotemporal gait parameters and tibial tilt angle before and after using heel lifts at normal walking speed were not significantly different; however, gait speed, stride length, and tibial tilt angle on both sides increased significantly immediately after using heel lifts and were maintained after 2 weeks of wear.

Conclusion

Using heel lifts not only improved symptom severity after 2 weeks but also immediately reduced pain during gait and had a positive impact on gait pattern and speed.

Centre of pressure changes during stance but not during gait in young women after alcohol intoxication

Background

Women are underrepresented in research focused on alcohol (e.g., Brighton, Moxham & Traynor, 2016; DOI 10.1097/JAN.0000000000000136) despite the changing patterns of alcohol consumption, which has been increasing in women in recent decades. The purpose of this study was to analyse the relationship between habitual alcohol consumption and centre of pressure (CoP) parameters during stance and gait while intoxicated by alcohol.

Methods

Thirty women (24.39 ± 2.93 years) participated in this study. All participants were asked to answer the AUDIT questionnaire. Stance and gait analysis were repeated under two conditions on a Zebris platform (FDM GmbH; Munich, Germany): when the participants were sober (0.00% breath alcohol concentration, BrAC) and when they were in an intoxicated state (0.11% BrAC). Participants were divided by their AUDIT score into a low-risk alcohol consumption group (n = 15; AUDIT score: 3 to 6) and a hazardous alcohol consumption group (n = 15; AUDIT score: 7 to 13).

Results

No statistical difference was observed in stance and gait parameters when comparing the low-risk and hazardous groups under 0.00% BrAC and 0.11% BrAC conditions. A statistically significant difference was observed when comparing 0.00% BrAC and 0.11% BrAC conditions within each group. This significant difference was found in CoP path length and CoP average velocity during quiet stance. However, no statistically significant differences were observed in CoP parameters during gait. An alcohol intoxication of 0.11% BrAC was not sufficient to cause statistically significant impairments in butterfly parameters of gait.

State-Based Versus Time-Based Estimation of the Gait Phase for Hip Exoskeletons in Steady and Transient Walking

The growing demand for online gait phase (GP) estimation, driven by advancements in exoskeletons and prostheses, has prompted numerous approaches in the literature. Some approaches explicitly use time, while others rely on state variables to estimate the GP. In this article, we study two novel GP estimation methods: a State-based Method (SM) which employs the phase portrait of the hip angle (similar to previous methods), but uses a stretching transformation to reduce the nonlinearity of the estimated GP; and a Time-based Method (TM) that utilizes feature recognition on the hip angle signal to update the estimated cadence twice per gait cycle. The methods were tested across various speeds and slopes, encompassing steady and transient walking conditions. The results demonstrated the ability of both methods to estimate the GP in a range of conditions. The TM outperformed the SM, exhibiting a root-mean-squared error below 3% compared to 8.5% for the SM. However, the TM exhibited diminished performance during speed transitions, whereas the SM performed consistently in steady and transient conditions. The SM displayed a better performance in inclined walking and demonstrated higher linearity at faster speeds. Through the assessment of these methods in diverse conditions, this study lays the groundwork for further advancements in GP estimation methods and their application in assistive controllers.

Influence of walking speed on gait spatiotemporal parameters and the functional rockers of the foot in healthy adults

BACKGROUND

The gait cycle is generally divided into stance phase and swing phase. The stance phase can also be divided into three functional rockers, each with a distinct fulcrum. It has been shown that walking speed (WS) influences both stance and swing phase but its influence on the functional foot rockers duration is unknown. The aim of the study was to analyze the WS influence on functional foot rockers duration.

METHODS

a cross-sectional study is completed with 99 healthy volunteers to assess the effect of WS on kinematics and foot rockers duration in treadmill walking at 4, 5, and 6 km·h^-1 RESULTS: Friedman test exhibited that all spatiotemporal variables and the length of the foot rockers changed significantly with WS (p < 0.05) except rocker 1 at 4 and 6 km·h^-1.

CONCLUSION

Every spatiotemporal parameter and the duration of the three functional rockers are affected by walking speed, although not all rockers are affected equally. The findings of this study reveal that Rocker 2 is the primary rocker whose duration is influenced by changes in gait speed.

Assessing real-world gait with digital technology? Validation, insights and recommendations from the Mobilise-D consortium

BACKGROUND

Although digital mobility outcomes (DMOs) can be readily calculated from real-world data collected with wearable devices and ad-hoc algorithms, technical validation is still required. The aim of this paper is to comparatively assess and validate DMOs estimated using real-world gait data from six different cohorts, focusing on gait sequence detection, foot initial contact detection (ICD), cadence (CAD) and stride length (SL) estimates.

METHODS

Twenty healthy older adults, 20 people with Parkinson's disease, 20 with multiple sclerosis, 19 with proximal femoral fracture, 17 with chronic obstructive pulmonary disease and 12 with congestive heart failure were monitored for 2.5 h in the real-world, using a single wearable device worn on the lower back. A reference system combining inertial modules with distance sensors and pressure insoles was used for comparison of DMOs from the single wearable device. We assessed and validated three algorithms for gait sequence detection, four for ICD, three for CAD and four for SL by concurrently comparing their performances (e.g., accuracy, specificity, sensitivity, absolute and relative errors). Additionally, the effects of walking bout (WB) speed and duration on algorithm performance were investigated.

RESULTS

We identified two cohort-specific top performing algorithms for gait sequence detection and CAD, and a single best for ICD and SL. Best gait sequence detection algorithms showed good performances (sensitivity > 0.73, positive predictive values > 0.75, specificity > 0.95, accuracy > 0.94). ICD and CAD algorithms presented excellent results, with sensitivity > 0.79, positive predictive values > 0.89 and relative errors < 11% for ICD and < 8.5% for CAD. The best identified SL algorithm showed lower performances than other DMOs (absolute error < 0.21 m). Lower performances across all DMOs were found for the cohort with most severe gait impairments (proximal femoral fracture). Algorithms' performances were lower for short walking bouts; slower gait speeds (< 0.5 m/s) resulted in reduced performance of the CAD and SL algorithms.

CONCLUSIONS

Overall, the identified algorithms enabled a robust estimation of key DMOs. Our findings showed that the choice of algorithm for estimation of gait sequence detection and CAD should be cohort-specific (e.g., slow walkers and with gait impairments). Short walking bout length and slow walking speed worsened algorithms' performances. Trial registration ISRCTN - 12246987.

Exploring variations in gait patterns and joint motion characteristics in school-aged children across different walking speeds: a comprehensive motion analysis study

This study aimed to investigate differences in gait patterns among individuals with different walking speeds and identify the range of motion (ROM) and angular velocity for various joints during gait. Forty-five schoolchildren were randomly selected for this study. To capture their walking patterns, two FDR-AX700 4K HDR camcorders were positioned to observe the predetermined walkway. Each participant completed a 5-meter walk at various speeds, including slow, normal, and fast, while maintaining a straight stride. There were significantly higher ROM and angular velocity (p<0.05) at the hip, knee, and ankle joints across most stages of walking at a faster speed compared to slow and normal speeds. At the same time, the angular velocity was significantly higher at the hip joint during hip extension terminal stance at normal speed compared to slow and fast speeds (p<0.05, ƞ2 =0.74). Similarly, the ROM of knee flexion swing, ankle plantar flexion loading response, and ankle dorsiflexion midswing angular velocity were significantly higher during normal walking speed (p<0.05). Conversely, slow-speed walking showed significantly higher ROM at knee extension terminal swing (ƞ2=0.52) and ankle dorsiflexion terminal stance (ƞ2=0.78) (p<0.05). The results indicate that individuals with different walking speeds exhibit significant differences in gait patterns. Slower walking speeds resulted in lower gait velocity and different joint motions compared to faster walking speeds.

A diverse and multi-modal gait dataset of indoor and outdoor walks acquired using multiple cameras and sensors

Gait datasets are often limited by a lack of diversity in terms of the participants, appearance, viewing angle, environments, annotations, and availability. We present a primary gait dataset comprising 1,560 annotated casual walks from 64 participants, in both indoor and outdoor real-world environments. We used two digital cameras and a wearable digital goniometer to capture visual as well as motion signal gait-data respectively. Traditional methods of gait identification are often affected by the viewing angle and appearance of the participant therefore, this dataset mainly considers the diversity in various aspects (e.g., participants' attributes, background variations, and view angles). The dataset is captured from 8 viewing angles in 45° increments along-with alternative appearances for each participant, for example, via a change of clothing. The dataset provides 3,120 videos, containing approximately 748,800 image frames with detailed annotations including approximately 56,160,000 bodily keypoint annotations, identifying 75 keypoints per video frame, and approximately 1,026,480 motion data points captured from a digital goniometer for three limb segments (thigh, upper arm, and head).

In-shoe plantar pressure depends on walking speed and type of weight-bearing activity in people with diabetes at high risk of foot ulceration

BACKGROUND

In evaluating therapeutic footwear, in-shoe plantar pressure is usually obtained during mid-gait steps at self-selected walking speed in a laboratory setting. However, this may not accurately represent plantar pressures or indicate the cumulative stress experienced in daily life. We investigated the effects of walking speed and different weight-bearing activities on in-shoe plantar pressure in people with diabetes at high risk of ulceration.

METHODS

In a cross-sectional study including 30 participants we compared in-shoe plantar pressures between three standardized walking speeds (0.8, 0.6 and 0.4 m/s) and between walking at self-selected speed and eight other weight-bearing activities (3 components of the Timed Up and Go test, accelerating, decelerating, stair ascending and descending, and standing). Mean forefoot regional peak plantar pressure and pressure-time integral were statistically assessed per foot using linear mixed models (α < 0.05) with Holm-Bonferroni correction.

FINDINGS

With increasing walking speed, peak pressures increased and pressure-time integrals decreased (P ≤ 0.014). Peak pressures during standing, decelerating, stair ascending and Timed Up and Go test were lower (P ≤ 0.001), and with other activities not different to walking at self-selected speed. Pressure-time integrals during stair ascending and descending were higher (P ≤ 0.001), during standing lower (P ≤ 0.009), and with other activities not different to walking at self-selected speed.

INTERPRETATION

In-shoe plantar pressure depends on walking speed and type of weight-bearing activity. Only measuring pressures to evaluate footwear at self-selected walking speed in a laboratory setting may not accurately represent the stress on the foot in daily life of the high-risk patient; a more comprehensive assessment is suggested.

Analysis of Diabetic Foot Deformation and Plantar Pressure Distribution of Women at Different Walking Speeds

Official guidelines state that suitable physical activity is recommended for patients with diabetes mellitus. However, since walking at a rapid pace could be associated with increased plantar pressure and potential foot pain, the footwear condition is particularly important for optimal foot protection in order to reduce the risk of tissue injury and ulceration of diabetic patients. This study aims to analyze foot deformation and plantar pressure distribution at three different walking speeds (slow, normal, and fast walking) in dynamic situations. The dynamic foot shape of 19 female diabetic patients at three walking speeds is obtained by using a novel 4D foot scanning system. Their plantar pressure distributions at the three walking speeds are also measured by using the Pedar in-shoe system. The pressure changes in the toes, metatarsal heads, medial and lateral midfoot, and heel areas are systematically investigated. Although a faster walking speed shows slightly larger foot measurements than the two other walking speeds, the difference is insignificant. The foot measurement changes at the forefoot and heel areas, such as the toe angles and heel width, are found to increase more readily than the measurements at the midfoot. The mean peak plantar pressure shows a significant increase at a faster walking speed with the exception of the midfoot, especially at the forefoot and heel areas. However, the pressure time integral decreases for all of the foot regions with an increase in walking speed. Suitable offloading devices are essential for diabetic patients, particularly during brisk walking. Design features such as medial arch support, wide toe box, and suitable insole material for specific area of the foot (such as polyurethane for forefoot area and ethylene-vinyl acetate for heel area) are essential for diabetic insole/footwear to provide optimal fit and offloading. The findings contribute to enhancing the understanding of foot shape deformation and plantar pressure changes during dynamic situations, thus facilitating the design of footwear/insoles with optimal fit, wear comfort, and foot protection for diabetic patients.

Human Gait Activity Recognition Machine Learning Methods

Human gait activity recognition is an emerging field of motion analysis that can be applied in various application domains. One of the most attractive applications includes monitoring of gait disorder patients, tracking their disease progression and the modification/evaluation of drugs. This paper proposes a robust, wearable gait motion data acquisition system that allows either the classification of recorded gait data into desirable activities or the identification of common risk factors, thus enhancing the subject's quality of life. Gait motion information was acquired using accelerometers and gyroscopes mounted on the lower limbs, where the sensors were exposed to inertial forces during gait. Additionally, leg muscle activity was measured using strain gauge sensors. As a matter of fact, we wanted to identify different gait activities within each gait recording by utilizing Machine Learning algorithms. In line with this, various Machine Learning methods were tested and compared to establish the best-performing algorithm for the classification of the recorded gait information. The combination of attention-based convolutional and recurrent neural networks algorithms outperformed the other tested algorithms and was individually tested further on the datasets of five subjects and delivered the following averaged results of classification: 98.9% accuracy, 96.8% precision, 97.8% sensitivity, 99.1% specificity and 97.3% F1-score. Moreover, the algorithm's robustness was also verified with the successful detection of freezing gait episodes in a Parkinson's disease patient. The results of this study indicate a feasible gait event classification method capable of complete algorithm personalization.

Comparison of IMU set-ups for the estimation of gait spatio-temporal parameters in an elderly population

The increasing average age emphasizes the importance of gait analysis in elderly populations. Inertial Measurement Units (IMUs) represent a suitable wearable technology for the characterization of gait by estimating spatio-temporal parameters (STPs). However, the location of inertial sensors on the human body and the associated algorithms for the estimation of gait STPs play a fundamental role and are still open challenges. Accordingly, the aim of this work was to compare three IMUs set-ups (trunk, shanks, and ankles) and correspondent algorithms to a gold standard optoelectronic system for the estimation of gait STPs in a healthy elderly population. In total, 14 healthy elderly subjects walked barefoot at three different speeds. Gait parameters were assessed for each IMUs set-up and compared to those estimated with the gold standard. A statistical analysis based on Pearson correlation, Root Mean Square Error and Bland Altman plots was conducted to evaluate the accuracy of IMUs. Even though all tested set-ups produced accurate results, the IMU on the trunk performed better in terms of correlation (R ≥ 0.8), RMSE (0.01-0.06 s for temporal parameters, 0.03-0.04 for the limp index), and level of agreement (-0.01 s ≤ mean error ≤ 0.01 s, -0.02 s ≤ standard deviation error ≤ 0.02 s), also allowing simpler preparation of subjects and minor encumbrance during gait. From the promising results, a similar experiment might be conducted in pathological populations in the attempt to verify the accuracy of IMUs set-ups and algorithms also in non-physiological patterns.

Children with autism display altered ankle strategies when changing speed during over-ground gait

BACKGROUND

Examining gait mechanics when altering speed has been used in various clinical populations to understand the pervasiveness of neurological impairments. Few studies have examined whether different gait mechanics exist when altering speed in children with Autism Spectrum Disorder, although autism may present as a movement disorder due to abnormalities in the central nervous system. Most autism gait-related research has used preferred walking speed, while different speeds may yield discernible patterns that can be used for future interventions. The purpose of this study was to examine kinematic strategies used by children with autism in preferred, fast, and slow walking speeds.

METHODS

Three-dimensional kinematic data were obtained on 14 children (aged 8-17 years) during preferred, fast, and slow walking. Hip, knee, and ankle angular joint positions were examined at loading response, pre-swing, and terminal swing sub-phases due to their importance on forward propulsion and weight transfer. Repeated measures analyses of variance (α = 0.05) were used to test for statistical differences and effect sizes were interpreted with Cohen's d.

FINDINGS

Although significant differences were observed for each joint and sub-phase, the left and right ankle joints during pre-swing displayed the most consistent differences among conditions (p < 0.001, and p < 0.001), respectively. Additionally, the left ankle displayed a moderate effect size (η² = 0.71) and the right ankle displayed a large effect size (η² = 0.80).

INTERPRETATIONS

These findings reveal that the ankle joint, during pre-swing, is the primary kinematic strategy used by children with autism when altering gait speed, whereas previous evidence suggests that the hip joint was the primary strategy.

Gender differences in the effect of a 0.11% breath alcohol concentration on forward and backward gait

Alcohol contributes to a large number of diseases and health conditions related to injuries. The aim of our study was to evaluate gender differences in forward and backward gait when sober and at a breath alcohol concentration (BrAC) of 0.11%. Fifty females and fifty males participated in our study. The gait analysis was performed twice, when sober and after drinking a given amount of vodka mixed with orange juice. Under both conditions, participants were asked to walk forward and then backward on a Zebris platform. Multivariate analysis and the Mann-Whitney U test were used to compare the differences between genders when walking forward and backward. The Wilcoxon Signed Ranks test was used to compare the differences between 0.00% BrAC and 0.11% BrAC. Spearman's Rho was used to analyze the relationship between the AUDIT score, anthropometrical characteristics and the subjective score of drunkenness and gait parameters. The results show different strategies to improve stability during gait in women and men when intoxicated with alcohol. When intoxicated, males in forward gait increase their stability by increasing their foot rotation, while females increase their step width. A decrease in balance-related variables was observed in females when walking backward with a BrAC of 0.11%. Additionally, females tended to perform an increase in balance-related gait variables when subjectively feeling more drunk in both forward and backward gait. Different strategies to maintain stability during gait were observed in women and men. The results of our study show that alcohol intoxication has a greater impact on gait in females who tended to perform an increase in balance-related variables with an increase in their subjective score of drunkenness.

Physical factors that differentiate body kinematics between treadmill and overground walking

Treadmills are widely used in rehabilitation and gait analysis. However, previous studies have reported differences in terms of kinematics and kinetics between treadmill and overground walking due to physical and psychological factors. The aim of this study was to analyze gait differences due to only the physical factors of treadmill walking. Walking motions of a male participant were captured at 0.63, 1.05, 1.33, and 3.91 m/s. A gait controller of a virtual subject (63 kg) was trained for ground walking at each walking speed via a reinforcement learning method. Additionally, the gait controllers of virtual subjects with different body masses of 47, 79, and 94 kg were trained for ground walking at 1.05 m/s. The gait controllers and virtual subjects were tested for treadmill walking, and their lower-limb joint kinematics were compared with those for ground walking. Treadmill conditions of maximum allowable belt force and feedback control frequency of belt speed were set between 100 and 500 N and between 10 and 50 Hz, respectively. The lower-limb kinematics were identical between the two conditions regardless of the body mass and walking speed when the belt could provide a constant speed regardless of external perturbation in the ideal treadmill. However, kinematic differences were observed when simulation was performed on a non-ideal treadmill with a relatively low belt force and control frequency of belt speed. The root-mean-square differences of the hip, knee, and ankle flexion angles between treadmill and overground running at 3.91 m/s increased by 3.76°, 3.73°, and 4.91°, respectively, when the maximum belt force and control frequency decreased from infinity to 100 N and 10 Hz, respectively. At a maximum belt force exceeding 400 N or a control frequency exceeding 25 Hz, the root-mean-square difference of the joint kinematics was less than 3° for all body masses and walking speeds. Virtual subjects walking on non-ideal treadmills showed different joint kinematics from ground walking. The study identified physical factors that differentiate treadmill walking from overground walking, and suggested the belt forces and control frequencies of a treadmill to achieve the desired limit of kinematic differences.

Continuous Gait Phase Estimation Using LSTM for Robotic Transfemoral Prosthesis Across Walking Speeds

User gait phase estimation plays a key role for the seamless control of the lower-limb robotic assistive devices (e.g., exoskeletons or prostheses) during ambulation. To achieve this, several studies have attempted to estimate the gait phase using a thigh or shank angle. However, their estimation resulted in some deviation from the actual walking and varied across the walking speeds. In this study, we investigated the different setups using for the machine learning approach to obtain more accurate and consistent gait phase estimation for the robotic transfemoral prosthesis over different walking speeds. Considering the transfemoral prosthetic application, we proposed two different sensor setups: i) the angular positions and velocities of both thigh and torso (S1) and ii) the angular positions and velocities of both thigh and torso, and heel force data (S2). The proposed setups and method are experimentally evaluated with three healthy young subjects at four different walking speeds: 0.5, 1.0, 1.5, and 2.0 m/s. Both results showed robust and accurate gait phase estimation with respect to the ground truth (loss value of S1: 4.54e-03 Vs. S2: 4.70e-03). S1 had the advantage of a simple equipment setup using only two IMUs, while S2 had the advantage of estimating more accurate heel-strikes than S1 by using additional heel force data. The choice between the two sensor setups can depend on the researchers' preference in consideration of the device setup or the focus of the interest.

Research on Kinematic Parameters of Multiple Gait Pattern Transitions

Gait recognition technology is the key technology in the field of exoskeletons. In the current research of gait recognition technology, there is less focus on the recognition of the transition between gait patterns. This study aims to determine which kinematic parameters have significant differences in the transitions (between level and stair walking and between level and ramp walking) of different gait patterns, to determine whether these parameters change differently in different gait pattern transitions, and the order the significant differences occur through a comparative analysis of various kinematic parameters between the transition stride and the before stride in the former pattern. We analyzed 18 parameters concerning both lower limbs and trunk. We compared each time point of the transition strides to the corresponding time points of the before stride using a series of two-sample t-tests, and we then evaluated the difference between the transition stride and the before stride based upon the number of time points within the gait cycle that were statistically different. We found that the sagittal plane angular velocity and the angular acceleration of all joints and the resultant velocity of the thigh and shank of the leading limb had significant differences in the process of transition; the sagittal plane angular velocity of all joints of the trailing limb and the velocity of the trunk in the coronary axis direction also showed a significant difference. The angular acceleration of all joints, the sagittal plane angular velocity of the ankle joint of the leading limb, and the acceleration of the trunk in the coronal axis direction showed a difference in the early stage of the transition. In general, the leading limb had a significant difference earlier than the trailing limb, and the acceleration parameters changed earlier than the velocity parameters. These parameters showed different combinations of changes in the transition of different gait patterns, and the changes in these parameters reflected different gait pattern transitions. Therefore, we believe that the results of this study can provide a reference for the gait pattern transition recognition of wearable exoskeletons.

Optimal Swing Support During Walking Using Wireless Pneumatic Artificial Muscle Driver

This study evaluates the effect of swing support during walking using a wireless pneumatic artificial muscle (PAM) driver on hip and knee flexion angles. This driver can control two contraction parameters of the PAM: delay of contraction from the trigger and duration of contraction through a smartphone. Eleven healthy young individuals participated in this study. We asked the participants to walk with two PAMs attached to the left hip joint and a pressure sensor placed under the right heel to trigger the contraction. During the experiment, the contraction parameters were randomly changed: 0, 100, or 200 ms for the delay and 0, 100, 200, or 300 ms for the duration. The experimental results revealed significant differences in the hip and knee flexion angles, hip joint angular excursion, and stride length among the conditions. In addition, the optimal parameter differed among the subjects. It was confirmed that this individual variation was related to the walking speed of the subject, without PAM assistance.

An artificial neural network approach to detect presence and severity of Parkinson's disease via gait parameters

Introduction

Gait deficits are debilitating in people with Parkinson’s disease (PwPD), which inevitably deteriorate over time. Gait analysis is a valuable method to assess disease-specific gait patterns and their relationship with the clinical features and progression of the disease.

Objectives

Our study aimed to i) develop an automated diagnostic algorithm based on machine-learning techniques (artificial neural networks [ANNs]) to classify the gait deficits of PwPD according to disease progression in the Hoehn and Yahr (H-Y) staging system, and ii) identify a minimum set of gait classifiers.

Methods

We evaluated 76 PwPD (H-Y stage 1–4) and 67 healthy controls (HCs) by computerized gait analysis. We computed the time-distance parameters and the ranges of angular motion (RoMs) of the hip, knee, ankle, trunk, and pelvis. Principal component analysis was used to define a subset of features including all gait variables. An ANN approach was used to identify gait deficits according to the H-Y stage.

Results

We identified a combination of a small number of features that distinguished PwPDs from HCs (one combination of two features: knee and trunk rotation RoMs) and identified the gait patterns between different H-Y stages (two combinations of four features: walking speed and hip, knee, and ankle RoMs; walking speed and hip, knee, and trunk rotation RoMs).

Conclusion

The ANN approach enabled automated diagnosis of gait deficits in several symptomatic stages of Parkinson’s disease. These results will inspire future studies to test the utility of gait classifiers for the evaluation of treatments that could modify disease progression.

Gait Phase Recognition Using Deep Convolutional Neural Network with Inertial Measurement Units

Gait phase recognition is of great importance in the development of assistance-as-needed robotic devices, such as exoskeletons. In order for a powered exoskeleton with phase-based control to determine and provide proper assistance to the wearer during gait, the user’s current gait phase must first be identified accurately. Gait phase recognition can potentially be achieved through input from wearable sensors. Deep convolutional neural networks (DCNN) is a machine learning approach that is widely used in image recognition. User kinematics, measured from inertial measurement unit (IMU) output, can be considered as an ‘image’ since it exhibits some local ‘spatial’ pattern when the sensor data is arranged in sequence. We propose a specialized DCNN to distinguish five phases in a gait cycle, based on IMU data and classified with foot switch information. The DCNN showed approximately 97% accuracy during an offline evaluation of gait phase recognition. Accuracy was highest in the swing phase and lowest in terminal stance.

A Pilot Study of Musculoskeletal Abnormalities in Patients in Recovery from a Unilateral Rupture-Repaired Achilles Tendon

The purpose of this study was to compare the inter-limb joint kinematics, joint moments, muscle forces, and joint reaction forces in patients after an Achilles tendon rupture (ATR) via subject-specific musculoskeletal modeling. Six patients recovering from a surgically repaired unilateral ATR were included in this study. The bilateral Achilles tendon (AT) lengths were evaluated using ultrasound imaging. The three-dimensional marker trajectories, ground reaction forces, and surface electromyography (sEMG) were collected on both sides during self-selected speed during walking, jogging and running. Subject-specific musculoskeletal models were developed to compute joint kinematics, joint moments, muscle forces and joint reaction forces. AT lengths were significantly longer in the involved side. The side-to-side triceps surae muscle strength deficits were combined with decreased plantarflexion angles and moments in the injured leg during walking, jogging and running. However, the increased knee extensor femur muscle forces were associated with greater knee extension degrees and moments in the involved limb during all tasks. Greater knee joint moments and joint reaction forces versus decreased ankle joint moments and joint reaction forces in the involved side indicate elevated knee joint loads compared with reduced ankle joint loads that are present during normal activities after an ATR. In the frontal plane, increased subtalar eversion angles and eversion moments in the involved side were demonstrated only during jogging and running, which were regarded as an indicator for greater medial knee joint loading. It seems after an ATR, the elongated AT accompanied by decreased plantarflexion degrees and calf muscle strength deficits indicates ankle joint function impairment in the injured leg. In addition, increased knee extensor muscle strength and knee joint loads may be a possible compensatory mechanism for decreased ankle function. These data suggest patients after an ATR may suffer from increased knee overuse injury risk.

Parametric generation of three-dimensional gait for robot-assisted rehabilitation

For robot-assisted rehabilitation and assessment of patients with motor dysfunction, the parametric generation of their normal gait as the input for the robot is essential to match with the features of the patient to a greater extent. In addition, the gait needs to be in three-dimensional space, which meets the physiological structure of the human better, rather than only on a sagittal plane. Thus, a method for the parametric generation of three-dimensional gait based on the influence of the motion parameters and structure parameters is presented. First, the three-dimensional gait kinematic of participants is collected, and trajectories of ankle joint angle and ankle center position are calculated. Second, for the trajectories, gait features are extracted including gait events indicating the physiological features of walking gait, in addition to extremes indicating the geometrical features of the trajectories. Third, regression models are derived after using leave-one-out cross-validation for model optimization. Finally, cubic splines are fitted between the predicted gait features to generate the trajectories for a full gait cycle. It is inferred that the generated curves match the measured curves well. The method presented herein gives an important reference for research into lower limb rehabilitation robots.

A PID Controller Approach to Explain Human Ankle Biomechanics across Walking Speeds

Lower-limb robotic prostheses and exoskeletons depend on controllers to function in synchrony with their users. Recent advancements in control technology permit embodiment and more intuitive control for the user. In this study, we utilize a control engineering perspective to propose a phase-dependent muscle-driven proportional, integral, and derivative (PID) controller to regulate human ankle joint trajectories across walking speeds. We calculated the correlation coefficients that relate the tibialis and gastrocnemius muscle activation to the ankle joint angle error, integral of the error, and rate of change of the error between an average ankle joint trajectory and the ankle angle at two walking speeds: 1.5 m/s and 2.0 m/s. We noted that preswing (PSW) was the only gait period that had high absolute values for the correlation coefficients (> 0.7) across all three relationships. Other gait periods had varying high and low correlation coefficients across the different relationships. These results present a promising justification to utilize the classic control technique in a non-conventional manner. A phase-dependent and muscle-driven PID controller influenced by the PSW phase may be used to modulate the ankle joint trajectory with muscle activation across walking speeds in lower-limb robotic prostheses and exoskeletons.

Data-driven spectral analysis for coordinative structures in periodic human locomotion

Living organisms dynamically and flexibly operate a great number of components. As one of such redundant control mechanisms, low-dimensional coordinative structures among multiple components have been investigated. However, structures extracted from the conventional statistical dimensionality reduction methods do not reflect dynamical properties in principle. Here we regard coordinative structures in biological periodic systems with unknown and redundant dynamics as a nonlinear limit-cycle oscillation, and apply a data-driven operator-theoretic spectral analysis, which obtains dynamical properties of coordinative structures such as frequency and phase from the estimated eigenvalues and eigenfunctions of a composition operator. Using segmental angle series during human walking as an example, we first extracted the coordinative structures based on dynamics; e.g. the speed-independent coordinative structures in the harmonics of gait frequency. Second, we discovered the speed-dependent time-evolving behaviours of the phase by estimating the eigenfunctions via our approach on the conventional low-dimensional structures. We also verified our approach using the double pendulum and walking model simulation data. Our results of locomotion analysis suggest that our approach can be useful to analyse biological periodic phenomena from the perspective of nonlinear dynamical systems.

Estimation of Vertical Ground Reaction Force Using Low-Cost Insole With Force Plate-Free Learning From Single Leg Stance and Walking

For the evaluation of pathological gait, a machine learning-based estimation of the vertical ground reaction force (vGRF) using a low-cost insole is proposed as an alternative to costly force plates. However, learning a model for estimation still relies on the use of force plates, which is not accessible in small clinics and individuals. Therefore, this paper presents a force plate-free learning from a single leg stance (SLS) and natural walking measured only by the insoles. This method used a linear least squares regression that fits insole measurements during SLS to body weight in order to learn a model to estimate vGRF during walking. Constraints were added to the regression so that vGRF estimates during walking were of proper magnitude, and the constraint bounds were newly defined as a linear function of stance duration. Moreover, a lower bound for the estimated vGRF in mid-stance was added to the constraints to enhance estimation accuracy. The vGRF estimated by the proposed method was compared with force platforms for 4 healthy young adults and 13 elderly adults including patients with mild osteoarthritis, knee pain, and valgus hallux. Through the experiments, the proposed learning method had a normalized root mean squared error under 10% for healthy young and elderly adults with stance durations within a certain range (600-800 ms). From these results, the validity of the proposed learning method was verified for various users requiring assessment in the field of medicine and healthcare.

Systematic differences of gluteal muscle activation during overground and treadmill walking in healthy older adults

Guteal muscle activation during walkway and treadmill walking was compared by means of Surface EMG (SEMG). Healthy older adults (50-75 years, n = 54; 29 females, 25 males) walked on a walkway (WW) at their self-selected slow, normal, and fast walking speeds and on a treadmill (TM) at 2, 3, 4, 5, and 6 km/h. Subject-individual, best-matched speed pairs were constituted and named SLOW, NORMAL, and FAST. Hip muscle activation was measured on both sides at mid-distance between the greater trochanter and the iliac crest by applying eight equally-spaced bipolar SEMG channels from ventral to dorsal (P1-P8). Grand averaged amplitude curves and mean amplitudes over the complete stride were analyzed to compare WW and TM walking. TM walking evoked significantly elevated mean amplitude levels, particularly at the ventral positions P1 to P4, which were disproportionately increased at SLOW. In grand averaged curves, corresponding significant amplitude differences between WW and TM were observed during load acceptance (SLOW; NORMAL), mid-stance (all speeds), and late swing phase (SLOW), with the number of significant differences decreasing for all electrode positions from SLOW to FAST. Compared to WW walking, TM walking may thus require systematically elevated effort of gluteal muscles, in particular at slow walking speed.

Abnormal gait pattern emerges during curved trajectories in high-functioning Parkinsonian patients walking in line at normal speed

Background

Several patients with Parkinson´s disease (PD) can walk normally along straight trajectories, and impairment in their stride length and cadence may not be easily discernible. Do obvious abnormalities occur in these high-functioning patients when more challenging trajectories are travelled, such as circular paths, which normally implicate a graded modulation in the duration of the interlimb gait cycle phases?

Methods

We compared a cohort of well-treated mildly to moderately affected PD patients to a group of age-matched healthy subjects (HS), by deliberately including HS spontaneously walking at the same speed of the patients with PD. All participants performed, in random order: linear and circular walking (clockwise and counter-clockwise) at self-selected speed. By means of pressure-sensitive insoles, we recorded walking speed, cadence, duration of single support, double support, swing phase, and stride time. Stride length-cadence relationships were built for linear and curved walking. Stride-to-stride variability of temporal gait parameters was also estimated.

Results

Walking speed, cadence or stride length were not different between PD and HS during linear walking. Speed, cadence and stride length diminished during curved walking in both groups, stride length more in PD than HS. In PD compared to HS, the stride length-cadence relationship was altered during curved walking. Duration of the double-support phase was also increased during curved walking, as was variability of the single support, swing phase and double support phase.

Conclusion

The spatio-temporal gait pattern and variability are significantly modified in well-treated, high-functioning patients with PD walking along circular trajectories, even when they exhibit no changes in speed in straight-line walking. The increased variability of the gait phases during curved walking is an identifying characteristic of PD. We discuss our findings in term of interplay between control of balance and of locomotor progression: the former is challenged by curved trajectories even in high-functioning patients, while the latter may not be critically affected.

Foot Placement Modulation Diminishes for Perturbations Near Foot Contact

Whenever a perturbation occurs during walking we have to maintain our balance using the recovery strategies that are available to us. Foot placement adjustment is often considered an important recovery strategy. However, because this strategy takes time it is likely a poor option if the foot is close to contact at the instant a perturbation occurs. The main goal of this study is to gain a better understanding of how humans deal with balance perturbations during walking if foot placement adjustments are constrained by time. Ten healthy subjects walked on an instrumented treadmill and received mediolateral and anteroposterior pelvis perturbations at various instances during the single support phase. The results show that foot placement modulation in the first recovery step following anteroposterior perturbations is fairly invariant of the perturbation magnitude and direction, regardless of the onset instance. For mediolateral perturbations, foot placement adjustments strongly modulate with the perturbation magnitude and direction, but these effects diminish when the perturbation onset is closer to the instant of foot contact. For most perturbations the first recovery step was consistent across subjects for all onset instances. However, in the second step various strategies arose that were not consistent across subjects, nor within subjects, especially for perturbations applied close to foot contact. Despite these different strategies, the COP location following foot contact strongly related to the COM velocity throughout these strategies. The results show that humans have various ways to compensate for limited availability of a foot placement strategy, with strategy selection highly dependent on the instant during the gait phase at which the perturbation is applied.

A public dataset of overground and treadmill walking kinematics and kinetics in healthy individuals

In a typical clinical gait analysis, the gait patterns of pathological individuals are commonly compared with the typically faster, comfortable pace of healthy subjects. However, due to potential bias related to gait speed, this comparison may not be valid. Publicly available gait datasets have failed to address this issue. Therefore, the goal of this study was to present a publicly available dataset of 42 healthy volunteers (24 young adults and 18 older adults) who walked both overground and on a treadmill at a range of gait speeds. Their lower-extremity and pelvis kinematics were measured using a three-dimensional (3D) motion-capture system. The external forces during both overground and treadmill walking were collected using force plates and an instrumented treadmill, respectively. The results include both raw and processed kinematic and kinetic data in different file formats: c3d and ASCII files. In addition, a metadata file is provided that contain demographic and anthropometric data and data related to each file in the dataset. All data are available at Figshare (DOI: 10.6084/m9.figshare.5722711). We foresee several applications of this public dataset, including to examine the influences of speed, age, and environment (overground vs. treadmill) on gait biomechanics, to meet educational needs, and, with the inclusion of additional participants, to use as a normative dataset.

Psychometric properties of outcome measures evaluating decline in gait in cerebellar ataxia: A systematic review

Cerebellar ataxia often results in impairment in ambulation secondary to gait pattern dysfunction and compensatory gait adjustments. Pharmaceutical and therapy-based interventions with potential benefit for gait in ataxia are starting to emerge, however evaluation of such interventions is hampered by the lack of outcome measures that are responsive, valid and reliable for measurement of gait decline in cerebellar ataxia. This systematic review aimed for the first time to evaluate the psychometric properties of gait and walking outcomes applicable to individuals with cerebellar ataxia. Only studies evaluating straight walking were included. A comprehensive search of three databases (MEDLINE, CINAHL and EMBASE) identified 53 studies meeting inclusion criteria. Forty-nine were rated as 'poor' as assessed by the COnsensus-based Standards for the selection of health Measurement INstruments checklist. The primary objective of most studies was to explore changes in gait related to ataxia, rather than to examine psychometric properties of outcomes. This resulted in methodologies not specific for psychometric assessment. Thirty-nine studies examined validity, 11 examined responsiveness and 12 measured reliability. Review of the data identified double and single support and swing percentage of the gait cycle, velocity, step length and the Scale for Assessment and Rating of Ataxia (SARA) gait item as the most valid and responsive measures of gait in cerebellar ataxia. However, further evaluation to establish their reliability and applicability for use in clinical trials is clearly warranted. We recommend that inter-session reliability of gait outcomes should be evaluated to ensure changes are reflective of intervention effectiveness in cerebellar ataxia.

A systematic review of the gait characteristics associated with Cerebellar Ataxia

BACKGROUND

Cerebellar Ataxias are a group of gait disorders resulting from dysfunction of the cerebellum, commonly characterised by slowly progressing incoordination that manifests as problems with balance and walking leading to considerable disability. There is increasing acceptance of gait analysis techniques to quantify subtle gait characteristics that are unmeasurable by current clinical methods This systematic review aims to identify the gait characteristics able to differentiate between Cerebellar Ataxia and healthy controls.

METHODS

Following systematic search and critical appraisal of the literature, gait data relating to preferred paced walking in Cerebellar Ataxia was extracted from 21 studies. A random-effect model meta-analysis was performed for 14 spatiotemporal parameters. Quality assessment was completed to detect risk of bias.

RESULTS

There is strong evidence that compared with healthy controls, Cerebellar Ataxia patients walk with a reduced walking speed and cadence, reduced step length, stride length, and swing phase, increased walking base width, stride time, step time, stance phase and double limb support phase with increased variability of step length, stride length, and stride time.

CONCLUSION

The consensus description provided here, clarifies the gait pattern associated with ataxic gait disturbance in a large cohort of participants. High quality research and reporting is needed to explore specific genetic diagnoses and identify biomarkers for disease progression in order to develop well-evidenced clinical guidelines and interventions for Cerebellar Ataxia.

Golden Gait: An Optimization Theory Perspective on Human and Humanoid Walking

Human walking is a complex task which includes hundreds of muscles, bones and joints working together to deliver harmonic movements with the need of finding equilibrium between moving forward and maintaining stability. Many different computational approaches have been used to explain human walking mechanisms, from pendular model to fractal approaches. A new perspective can be gained from using the principles developed in the field of Optimization theory and in particularly the branch of Game Theory. In particular we provide a new insight into human walking showing as the trade-off between advancement and equilibrium managed during walking has the same solution of the Ultimatum game, one of the most famous paradigms of game theory, and this solution is the golden ratio. The golden ratio is an irrational number that was found in many biological and natural systems self-organized in a harmonic, asymmetric, and fractal structure. Recently, the golden ratio has also been found as the equilibrium point between two players involved into the Ultimatum Game. It has been suggested that this result can be due to the fact that the golden ratio is perceived as the fairest asymmetric solution by the two players. The golden ratio is also the most common proportion between stance and swing phase of human walking. This approach may explain the importance of harmony in human walking, and provide new perspectives for developing quantitative assessment of human walking, efficient humanoid robotic walkers, and effective neurorobots for rehabilitation.

Smart Annotation of Cyclic Data Using Hierarchical Hidden Markov Models

Cyclic signals are an intrinsic part of daily life, such as human motion and heart activity. The detailed analysis of them is important for clinical applications such as pathological gait analysis and for sports applications such as performance analysis. Labeled training data for algorithms that analyze these cyclic data come at a high annotation cost due to only limited annotations available under laboratory conditions or requiring manual segmentation of the data under less restricted conditions. This paper presents a smart annotation method that reduces this cost of labeling for sensor-based data, which is applicable to data collected outside of strict laboratory conditions. The method uses semi-supervised learning of sections of cyclic data with a known cycle number. A hierarchical hidden Markov model (hHMM) is used, achieving a mean absolute error of 0.041 ± 0.020 s relative to a manually-annotated reference. The resulting model was also used to simultaneously segment and classify continuous, ‘in the wild’ data, demonstrating the applicability of using hHMM, trained on limited data sections, to label a complete dataset. This technique achieved comparable results to its fully-supervised equivalent. Our semi-supervised method has the significant advantage of reduced annotation cost. Furthermore, it reduces the opportunity for human error in the labeling process normally required for training of segmentation algorithms. It also lowers the annotation cost of training a model capable of continuous monitoring of cycle characteristics such as those employed to analyze the progress of movement disorders or analysis of running technique.

Real-Time Continuous Gait Phase and Speed Estimation from a Single Sensor

Human gait involves a repetitive cycle of movements, and the phase of gait represents the location in this cycle. Gait phase is measured across many areas of study (e.g., for analyzing gait and controlling powered lower-limb prosthetic and orthotic devices). Current gait phase detection methods measure discrete gait events (e.g., heel strike, flat foot, toe off, etc.) by placing multiple sensors on the subject's lower-limbs. Using multiple sensors can create difficulty in experimental setup and real-time data processing. In addition, detecting only discrete events during the gait cycle limits the amount of information available during locomotion. In this paper we propose a real-time and continuous measurement of gait phase parameterized by a mechanical variable (i.e., phase variable) from a single sensor measuring the human thigh motion. Human subject experiments demonstrate the ability of the phase variable to accurately parameterize gait progression for different walking/running speeds (1 to 9 miles/hour). Our results show that this real-time method can also estimate gait speed from the same sensor.

Sensor-Based Gait Parameter Extraction With Deep Convolutional Neural Networks

Measurement of stride-related, biomechanical parameters is the common rationale for objective gait impairment scoring. State-of-the-art double-integration approaches to extract these parameters from inertial sensor data are, however, limited in their clinical applicability due to the underlying assumptions. To overcome this, we present a method to translate the abstract information provided by wearable sensors to context-related expert features based on deep convolutional neural networks. Regarding mobile gait analysis, this enables integration-free and data-driven extraction of a set of eight spatio-temporal stride parameters. To this end, two modeling approaches are compared: a combined network estimating all parameters of interest and an ensemble approach that spawns less complex networks for each parameter individually. The ensemble approach is outperforming the combined modeling in the current application. On a clinically relevant and publicly available benchmark dataset, we estimate stride length, width and medio-lateral change in foot angle up to -0.15 ± 6.09 cm, -0.09 ± 4.22 cm and 0.13 ± 3.78° respectively. Stride, swing and stance time as well as heel and toe contact times are estimated up to ±0.07, ±0.05, ±0.07, ±0.07 and ±0.12 s respectively. This is comparable to and in parts outperforming or defining state of the art. Our results further indicate that the proposed change in the methodology could substitute assumption-driven double-integration methods and enable mobile assessment of spatio-temporal stride parameters in clinically critical situations as, e.g., in the case of spastic gait impairments.

Evaluation of measurements of propulsion used to reflect changes in walking speed in individuals poststroke

Recent rehabilitation approaches for individuals poststroke have focused on improving walking speed because it is a reliable measurement that is associated with quality of life. Previous studies have demonstrated that propulsion, the force used to propel the body forward, determines walking speed. However, there are several different ways of measuring propulsion and no studies have identified which measurement best reflects differences in walking speed. The primary purposes of this study were to determine for individuals poststroke, which measurement of propulsion (1) is most closely related to their self-selected walking speeds and (2) best reflects changes in walking speed within a session. Participants (N=43) with chronic poststroke hemiparesis walked at their self-selected and maximal walking speeds on a treadmill. Propulsive impulse, peak propulsive force, and mean propulsive value (propulsive impulse divided by duration) were analyzed. In addition, each participant׳s cadence was calculated. Pearson correlation coefficients were used to determine the relationships between different measurements of propulsion versus walking speed as well as changes in propulsion versus changes in walking speed. Stepwise linear regression was used to determine which measurement of propulsion best predicted walking speed and changes in walking speed. The results showed that all 3 measurements of propulsion were correlated to walking speed, with peak propulsive force showed the strongest correlation. Similarly, when participants increased their walking speeds, changes in peak propulsive forces showed the strongest correlation to changes in walking speed. In addition, multiplying each measurement by cadence improved the correlations. The present study suggests that measuring peak propulsive force and cadence may be most appropriate of the variables studied to characterize propulsion in individuals poststroke.