A Wearable "Lab-in-Shoe" Gait Analysis System for Routine Clinical Assessment of People With Parkinson's Disease

Accurate and objective gait assessment is essential for managing Parkinson's disease (PD), yet conventional clinical evaluations rely heavily on clinical rating scales with bulky and costly laboratory-based gait analysis. This study presents a wearable Lab-in-Shoe system that integrates inertial measurement units (IMUs) and pressure sensors embedded within footwear for routine clinical gait analysis. The system segments gait cycles using plantar pressure data and applies Zero Velocity Update (ZUPT) and Principal Component Analysis (PCA) algorithms to mitigate IMU drift and reconstruct spatiotemporal gait trajectories. Validation experiments with healthy participants showed strong agreement with an optical motion capture system in stride length (ICC = 0.970, MAE = 0.04 m) and swing phase duration (ICC = 0.934, MAE = 0.02 s). Clinical assessment with PD patients revealed stage-dependent gait impairments and significant correlations between gait parameters and disease severity scores. A multiple linear regression model predicted MDS-UPDRS III scores with high accuracy (R ${}^{{2}} =0.87$ , RMSE = 6.75), indicating gait features quantitatively reflects motor symptom severity. Importantly, this study is the first to analyze plantar center of pressure (CoP) trajectories in PD using a wearable system, identifying progressive alterations in CoP patterns across Hoehn and Yahr stages, including signs of freezing of gait and a conservative gait balance strategy. These findings highlight the clinical potential of Lab-in-Shoe as a portable tool for continuous gait monitoring, enabling quantitative assessment of motor function, disease progression tracking, and therapeutic evaluation in PD patients.