Markerless Video-Based Gait Analysis for Motor Phenotyping in Individuals with Schizophrenia Using Interpretable Machine Learning

Background/Objectives: Motor abnormalities are frequently observed in schizophrenia, but accessible methods for objective gait quantification remain limited. This exploratory controlled-setting study examined whether markerless smartphone-based video analysis combined with interpretable machine learning could quantify gait-related motor phenotypes in individuals with schizophrenia. Methods: Gait videos were collected from 100 individuals with schizophrenia and 35 healthy controls using a standardized recording setup. MediaPipe Pose was used to extract skeletal landmarks and derive 12 image-plane spatiotemporal and estimated two-dimensional knee-kinematic gait features. After temporal segmentation and quality control, 404 usable gait segments were analyzed. Decision Tree and Support Vector Machine models were applied for exploratory segment-level group-separation analysis using 15-fold cross-validation after pre-cross-validation resampling. Results: Several extracted gait features differed between groups, particularly image-plane ankle displacement, mean step displacement, displacement velocity, step characteristics, and knee-joint motion. In the Decision Tree model, image-plane ankle displacement served as the primary root node, indicating its central role in internal segment-level group separation. However, because the healthy control group was substantially younger and not age-matched, multiple gait segments from the same participant could appear across cross-validation folds, and resampling was performed before fold partitioning, model performance should be interpreted as exploratory internal segment-level behavior rather than participant-level classification evidence. Conclusions: Markerless video-based gait analysis with interpretable machine learning may provide a feasible research-support approach for quantifying gait-related motor phenotypes in individuals with schizophrenia. These findings should not be interpreted as evidence for clinical classification or screening. Future studies require matched controls, psychiatric comparison groups, subject-wise validation, external datasets, calibrated gait measures, and privacy-preserving data governance.