Reducing knee joint crosstalk using PCA correction

摘要

Multiple marker protocols exist for analysing gait kinematics, however most models show crosstalk, particularly in the knee joint due to difficulties in aligning the flexion/extension (F/E) axis. These difficulties are enlarged in patient groups, such as children with cerebral palsy (CP) due to difficulties in palpation and deviations from model assumptions. Functional calibration might be effective in improving knee axes alignment [1], but is time consuming and difficult to perform. Therefore post hoc optimisation methods are preferable. The aim of this study is to outline the effects of a newly proposed post hoc method based on principal component analysis (PCA) to reduce crosstalk [2] in commonly used models for gait analysis.