Patient-specific optimized kinematic models of the leg

INTRODUCTION Patient-specific musculoskeletal models have the potential to assist in many clinical problems within rehabilitation and orthopedics. The predictive power of musculoskeletal models depends upon how well the morphology of the underlying kinematic model matches that of the patient. This paper presents a method to create a patient-specific kinematic model of the leg based on motion capture data in the AnyBody Modeling System (AMS). The method uses optimisation of joint positions and orientations, while minimising the error between model markers and the recorded markers from a motion capture experiment in a least-squares sense [1]. A commonly used method to create patient-specific kinematic models is to use anatomical landmarks to identify joint positions and orientations. The Plug-in Gait method by Vicon is one such method. The disadvantage is that the anatomical methods does not account for anatomical differences between subjects i.e. landmarks may have different positions relative to the actual functional axis. Similar optimisation methods to find patient-specific kinematic models have previously been proposed [2-4]. However, they are not readily available for practitioners of gait analysis, and have high computational costs.