Objective Identification of Pharmacological Interference in Parkinsonian Gait Using Continuous Digital Phenotyping
Abstract
Zhenghua Li and Kenichi Yamamura
Building upon our previous demonstration that Apple HealthKit enables longitudinal monitoring of Parkinsonian gait [1], this study evaluates its utility for short-term pharmacological assessment. Using a self-controlled longitudinal design, gait performance was quantified across four pharmacological phases: pre-medication, on-medication, withdrawal day, and post-withdrawal recovery. Objective gait parameters—including coefficient of variation (CV), walking speed, asymmetry, and step length—were derived from real-world walking data. Rotigotine administration resulted in marked deterioration in gait stability, characterized by a substantial increase in CV and worsening asymmetry, accompanied by reduced step length. These abnormalities were further exacerbated on the withdrawal day and were rapidly reversible following discontinuation, with CV decreasing to below baseline levels. These findings demonstrate that continuous digital phenotyping enables objective identification of pharmacological interference and supports data-driven, individualized treatment strategies in Parkinson’s disease.