Assessing Stability of the Knee In Vitro Using Envelope Normalized Kinematics and Principal Component Analysis

Abstract

This research focuses on creating a measure of dynamic stability that relates the passive envelope of constraint and kinematic motion during three gait simulations run on the Kansas Knee Simulator using PCA. Gait kinematics and envelope kinematics for fourteen natural knees were placed into one principal component model and the mode of variations were identified. This was done for anterior-posterior, internal-external, and varus-valgus kinematic directions. The coefficient matrix was used to score five implanted knees. Statistical analysis showed that there was no difference between natural and implanted knees when looking at the position of the kinematics relative to the envelope of constraint in any kinematic direction. This means that both the natural and implanted knees have kinematics that lie near or over the envelope of constraint. Maximum distance between the first four PCs was also calculated and statistical analysis showed that there was no difference between natural and implanted knees in the VV direction. However, there was a statistical difference in the AP and IE directions. There was a greater distance found in the implanted knees than in the natural knees, which was the opposite of what was hypothesized. This means that the natural knees were able to perform the three walk cycles with less kinematic variations. In addition, a suite of profiles for a variety of functional tasks were developed in order to simulate a broad range of activities of daily living, beyond standard walking. These tasks include crossover pivot, walks run at specific body weight, incline walking, and 120 degree squat. A PCM was created with kinematics collected from these profiles and knee specific envelopes. PCM showed that all of these profiles showed greater differences in the kinematics when compared to the original three walks.