Assessing full turning in Parkinson’s disease: development of a novel wearable-based pipeline

Parkinson’s disease (PD) is a neurodegenerative disorder causing progressive motor impairments affecting gait, balance, and turning. Wearable inertial measurement units (IMUs) provide an objective approach for quantitative movement assessment. The aim of this thesis was to develop an automated IMU-based pipeline for quantitative assessment of 360° turning in individuals with PD. The study investigated the effects of dopaminergic therapy and cognitive load, the relationship with clinical motor impairment, and the association between symptom-onset side and first-turn direction. The study included 103 individuals with PD undergoing preoperative evaluation for potential deep brain stimulation (DBS). Three synchronized IMUs were positioned on the lower back and both feet. Participants performed a standardized 360° turning task under Medication OFF and ON and Single Task and Dual Task conditions. The pipeline extracted temporal, velocity-related, and step-related features. Paired comparisons and Spearman correlations assessed medication, cognitive-load, and clinical associations, with Bonferroni correction. Results showed consistent differences between Medication OFF and ON. In the ON state, participants generally performed the task in less time, with higher mean and peak turn velocities and fewer steps. The cognitive task resulted in longer execution times, more steps, and lower turning velocities in both conditions. Higher MDS-UPDRS III scores were generally associated with longer durations, lower velocities, and more steps, whereas symmetry indices showed limited sensitivity. No systematic association was found between symptom-onset side and preferred rotation direction. Overall, the automated pipeline quantified 360° turning and identified changes associated with pharmacological condition and cognitive load. The results suggest that temporal, velocity-related, and step-related parameters may provide informative quantitative measures of motor performance in PD.