Quality-Controlled Pipeline for Multi-Segment Wearable Accelerometry in Physiological and Pathological Gait Assessment

Abstract

Wearable accelerometry brings gait assessment closer to clinical feasibility. Recording quality and unequal signal duration can still influence between-group comparisons. This study evaluated a quality-controlled, bilateral, multi-segment accelerometry pipeline in 31 participants (15 with pathological gait and 16 with physiological gait) who walked 30 m at a self-selected pace. Six triaxial accelerometers recorded bilateral acceleration at the hip, knee, and ankle levels, and identical analyses were performed using an up-to-30-s window and a fixed 15-s window. Eight of ten focused descriptors met the predefined robustness criteria in both configurations. The physiological group showed higher knee, ankle, and combined knee–ankle RMS, higher ankle and combined knee–ankle dominant frequency, and higher estimated accelerometric cadence. The pathological group, in turn, showed greater knee and ankle RMS asymmetry. High-frequency ratios and global DFA exponents were not statistically significant after false-discovery-rate correction. These group-level findings support explicit quality control and temporal sensitivity analysis in wearable gait assessment, but they cannot be attributed exclusively to pathology because the groups were not age-matched and walking speed was not independently measured. Individual clinical interpretation requires external validation.