Crosstalk is a leading source of error in motion analysis [1-2]. Due to incorrect flexion axis direction that develops from marker placement error, crosstalk results in a strong, anatomically incorrect correlation between flexion-extension (FE) and adduction-abduction (AA) motions [1-2]. Thus, crosstalk limits the ability of biomechanical models to reflect the “true” motion of the knee. Principal Component Analysis (PCA) has been proposed as a post-hoc correction for crosstalk in prior gait studies [1-2]; however, previous studies have not proposed a method to determine PCA corrected knee axes. Further, it is not clear how PCA should be implemented in motion analysis studies that involve several exercises, on the same subjects, involving a relatively high range of flexion angles.

The long-term goal of this study is to determine accurate knee kinematics in a variety of exercises performed by the same subjects. This study tests two hypotheses: (1) PCA corrects for crosstalk between FE and AA angles in gait and cycling and (2) PCA corrected knee axes are similar for gait and cycling. The aims are to (1) determine PCA corrected knee angles in gait and cycling for the same subjects and their corresponding FE-AA correlations, (2) develop and implement an algorithm for determining PCA corrected knee FE and AA axes, and (3) compare the PCA corrected FE and AA axes for the same subjects to determine if they are similar in gait and cycling.


Biomedical Engineering and Bioengineering

Number of Pages




URL: http://digitalcommons.calpoly.edu/bmed_fac/83