Inter-segmental coordination patterns in Parkinson’s disease are particularly disturbed during preferred walking speed: a data-driven network approach

Background: Human gait involves complex coordination between musculoskeletal segments. This coordination is disturbed in Parkinson's disease (PD) and likely influenced by different walking speeds. Objectives: To investigate inter-segmental coordination during different walking speeds in people with PD (pwPD) using an unconstrained and data-driven network theory approach. Methods: Twenty-nine pwPD and 29 controls walked at preferred, fast and slow speeds. Data was collected using optical motion capture. Body segment accelerations were correlated pairwise to build kinectomes for each speed and movement direction. Anatomical body segments were defined as nodes and their co-accelerations as edges to build network graphs. The kinectomes and maximum-weighted graph patterns were compared between groups. Results: Permutation testing revealed no significant kinectome differences between groups across speeds or directions. Coordination deficits in the PD group were observed predominantly at preferred walking speed (162 significantly different graph patterns) in anteroposterior and mediolateral directions. At fast walking speed, 4 significantly different graph patterns were found in anteroposterior and vertical directions. Slow walking speed showed 1 significantly different pattern in mediolateral direction. Conclusions: PD affects inter-segmental coordination, becoming most apparent at preferred walking speed. This is surprising and highly relevant, as it is the most common gait condition in real life. 'Non-preferred' walking speeds in PD exhibit more control-like patterns, which could inform future treatment studies. The direction-specific coordination deficits could provide novel insights into patho- and compensatory mechanisms in PD gait. Trial registration: The study is registered in the German Clinical Trials Register (DRKS00022998, registered on 04 Sep 2020).