The Conventional Gait Model’s Sensitivity to Lower-limb Marker Placement

Clinical gait analysis supports treatment decisions for patients with motor disorders. Measurement reproducibility is affected by extrinsic errors such as marker misplacement—considered the main factor in gait analysis variability. However, how marker placement affects output kinematics is not completely understood. The present study aimed to evaluate the Conventional Gait Model’s sensitivity to marker placement. Using a dataset of kinematics for 20 children, eight lower-limb markers were virtually displaced by 10 mm in all four planes, and all the displacement combinations were recalculated. Root-mean-square deviation angles were calculated for each simulation with respect to the original kinematics. The marker movements with the greatest impact were for the femoral and tibial wands together with the lateral femoral epicondyle marker when displaced in the anterior–posterior axis. When displaced alone, the femoral wand was responsible for a deviation of 7.3° (± 1.8°) in hip rotation. Transversal plane measurements were affected most, with around 40% of simulations resulting in an effect greater than the acceptable limit of 5°. This study also provided insight into which markers need to be placed very carefully to obtain more reliable gait data.

Modelling and identification of characteristic kinematic features preceding freezing of gait with convolutional neural networks and layer-wise relevance propagation

Background Although deep neural networks (DNNs) are showing state of the art performance in clinical gait analysis, they are considered to be black-box algorithms. In other words, there is a lack of direct understanding of a DNN’s ability to identify relevant features, hindering clinical acceptance. Interpretability methods have been developed to ameliorate this concern by providing a way to explain DNN predictions. Methods This paper proposes the use of an interpretability method to explain DNN decisions for classifying the movement that precedes freezing of gait (FOG), one of the most debilitating symptoms of Parkinson’s disease (PD). The proposed two-stage pipeline consists of (1) a convolutional neural network (CNN) to model the reduction of movement present before a FOG episode, and (2) layer-wise relevance propagation (LRP) to visualize the underlying features that the CNN perceives as important to model the pathology. The CNN was trained with the sagittal plane kinematics from a motion capture dataset of fourteen PD patients with FOG. The robustness of the model predictions and learned features was further assessed on fourteen PD patients without FOG and fourteen age-matched healthy controls. Results The CNN proved highly accurate in modelling the movement that precedes FOG, with 86.8% of the strides being correctly identified. However, the CNN model was unable to model the movement for one of the seven patients that froze during the protocol. The LRP interpretability case study shows that (1) the kinematic features perceived as most relevant by the CNN are the reduced peak knee flexion and the fixed ankle dorsiflexion during the swing phase, (2) very little relevance for FOG is observed in the PD patients without FOG and the healthy control subjects, and (3) the poor predictive performance of one subject is attributed to the patient’s unique and severely flexed gait signature. Conclusions The proposed pipeline can aid clinicians in explaining DNN decisions in clinical gait analysis and aid machine learning practitioners in assessing the generalization of their models by ensuring that the predictions are based on meaningful kinematic features.

Impact of the Marker Set Configuration on the Accuracy of Gait Event Detection in Healthy and Pathological Subjects

For interpreting outcomes of clinical gait analysis, an accurate estimation of gait events, such as initial contact (IC) and toe-off (TO), is essential. Numerous algorithms to automatically identify timing of gait events have been developed based on various marker set configurations as input. However, a systematic overview of the effect of the marker selection on the accuracy of estimating gait event timing is lacking. Therefore, we aim to evaluate (1) if the marker selection influences the accuracy of kinematic algorithms for estimating gait event timings and (2) what the best marker location is to ensure the highest event timing accuracy across various gait patterns. 104 individuals with cerebral palsy (16.0 ± 8.6 years) and 31 typically developing controls (age 20.6 ± 7.8) performed clinical gait analysis, and were divided into two out of eight groups based on the orientation of their foot, in sagittal and frontal plane at mid-stance. 3D marker trajectories of 11 foot/ankle markers were used to estimate the gait event timings (IC, TO) using five commonly used kinematic algorithms. Heatmaps, for IC and TO timing per group were created showing the median detection error, compared to detection using vertical ground reaction forces, for each marker. Our findings indicate that median detection errors can be kept within 7 ms for IC and 13 ms for TO when optimizing the choice of marker and detection algorithm toward foot orientation in midstance. Our results highlight that the use of markers located on the midfoot is robust for detecting gait events across different gait patterns.

MovementRx: Versatile clinical movement analysis using Statistical Parametric Mapping

Background: Clinical gait analysis is an important field of biomechanics that Is influenced by subjectivity, which can lead to type I and II errors. Statistical Parametric Mapping (SPM) is a classical hypothesis testing method that can operate on all measured joint dynamics simultaneously, thereby overcoming errors associated with subjective reduction of these complex data and providing a quantitative and coherent assessment. Results: We present MovementRx, the first gait analysis modelling application that models joints in 3 degrees of freedom. It is a python-based versatile GUI-based movement analysis decision support system, that provides a holistic view of all lower limb joints fundamental to the kinematic/kinetic chain (i.e., related to functional gait). It utilizes the time varying statistical tool SPM1D combined with a visualizing software. The user can cascade the view from single 3D multivariate result down to specific single joint individual scalar component of movement in one dimension. It exports its API as a library for use by another python application or command line. We also presented a case study of a unilateral knee osteoarthritis (OA) patient with otherwise undetected contralateral OA predisposition. The intervention elevated the patient’s moments on the right (affected) limb, but it led to adverse compensation on the left (contralateral) limb, leaving the patient likely to develop OA in her left limb in the future, unless immediate preventive and / or corrective actions were taken. Conclusions: MovementRx is a clinical gait analysis tool for that provides objective representation of the magnitude of statistical effect of all relevant joints in a simple, coherent, objective, and visually intuitive manner. No other software correctly model joints in 3 degrees of freedom.

Podiatrists’ views and experiences of using real time clinical gait analysis in the assessment and treatment of posterior tibial tendon dysfunction

Background Real time clinical gait analysis (RTCGA) is often incorporated as part of a general or lower limb musculoskeletal (MSK) adult patient assessment. However, it is not known if RTCGA is clinically effective as a useful outcome measure or aids in decision making. Whether there is a clinical worth in conducting RTCGA in adult MSK consultations remains controversial. The aim of this study was to provide unique insights into MSK podiatrists use and opinions of RTCGA, using Posterior Tibial Tendon Dysfunction (PTTD) as an exemplar adult condition. Methods A qualitative approach was employed to explore MSK podiatrists’ views and experiences of RTCGA when assessing or treating patients with PTTD. Semi-structured interviews were conducted via Skype video calls which were transcribed using an orthographic transcription method. Thematic analysis was employed to identify key meanings and report patterns within the data. Results Twenty nine MSK podiatrists who used RTCGA in the assessment and treatment of PTTD participated in the study. Five themes were identified as 1) RTCGA Method; 2) Working with RTCGA; 3) RTCGA uses; 4) What could aid RTCGA; 5) How RTCGA skills are acquired. This is the first known study to explore this topic of relevance to clinicians and researchers alike. Clinical observations were not only kinematic, but also included patient perceived experiences such as pain and orthotic comfort with normative kinematic reference values not perceived as important to that management goal. The most common barefoot RTCGA observations performed were the rearfoot to leg angle, medial bulge, forefoot abduction and arch integrity. However, a high amount of variation in many gait observations was noted between participants. Documentation methods also varied with a four-point scale system to grade motion and position most often employed and RTCGA was most often learnt through experience. The main barriers to performing RTCGA were clinical time and space restrictions. Conclusion Findings from this study have provided a view of how podiatry MSK clinicians utilise RTCGA within their practice. MSK podiatrists use RTCGA as both an outcome measure and as an aid in decision making. This implies a perceived worth in conducting RTCGA, however further work is recommended that focusses on development of a national guideline to RTCGA to be adopted.