Identification and interpretation of gait analysis features and foot conditions by explainable AI.

Clinical gait analysis is a crucial step for identifying foot disorders and planning surgery. Automating this process is essential for efficiently assessing the substantial amount of gait data. In this study, we explored the potential of state-of-the-art machine learning (ML) and explainable artificial intelligence (XAI) algorithms to automate all various steps involved in gait analysis for six specific foot conditions. To address the complexity of gait data, we manually created new features, followed by recursive feature elimination using Support Vector Machines (SVM) and Random Forests (RF) to eliminate low-variance features. SVM, RF, K-nearest Neighbor (KNN), and Logistic Regression (LREGR) were compared for classification, with a Majority Voting (MV) model combining trained models. KNN and MV achieved mean balanced accuracy, recall, precision, and F1 score of 0.87. All models were interpreted using Local Interpretable Model-agnostic Explanation (LIME) method and the five most relevant features were identified for each foot condition. High success scores indicate a strong relationship between selected features and foot conditions, potentially indicating clinical relevance. The proposed ML pipeline, adaptable for other foot conditions, showcases its potential in aiding experts in foot condition identification and planning surgeries.

Machine learning-based prediction of joint moments based on kinematics in patients with cerebral palsy.

Joint moments during gait provide valuable information for clinical decision-making in patients with cerebral palsy (CP). Joint moments are calculated based on ground reaction forces (GRF) using inverse dynamics models. Obtaining GRF from patients with CP is challenging. Typically developed (TD) individuals' joint moments were predicted from joint angles using machine learning, but no such study has been conducted on patients with CP. Accordingly, we aimed to predict the dorsi-plantar flexion, knee flexion-extension, hip flexion-extension, and hip adduction-abduction moments based on the trunk, pelvis, hip, knee, and ankle kinematics during gait in patients with CP and TD individuals using one-dimensional convolutional neural networks (CNN). The anonymized retrospective gait data of 329 TD (26 years ± 14, mass: 70 kg ± 15, height: 167 cm ± 89) and 917 CP (17 years ± 9, mass:47 kg ± 19, height:153 cm ± 36) individuals were evaluated and after applying inclusion-exclusion criteria, 132 TD and 622 CP patients with spastic diplegia were selected. We trained specific CNN models and evaluated their performance using isolated test subject groups based on normalized root mean square error (nRMSE) and Pearson correlation coefficient (PCC). Joint moments were predicted with nRMSE between 18.02% and 13.58% for the CP and between 12.55% and 8.58% for the TD groups, whereas with PCC between 0.85 and 0.93 for the CP and between 0.94 and 0.98 for the TD groups. Machine learning-based joint moment prediction from kinematics could replace conventional moment calculation in CP patients in the future, but the current level of prediction errors restricts its use for clinical decision-making today.