Detecting the presence of anterior cruciate ligament deficiency based on a double pendulum model, intrinsic time-scale decomposition (ITD) and neural networks

摘要

The anterior cruciate ligament (ACL) possesses the function of stabilizing the knee joint through limiting anterior tibial translation and controlling tibial rotation. Patients with unilateral ACL deficiency often demonstrate alterations of knee kinematics, kinetics and gait patterns in the deficient side in comparison to the unaffected contralateral side. This also leads to the early onset of osteoarthritis. In order to detect and monitor the progression of ACL deficiency over time, various classification approaches using spatiotemporal gait variables have been presented. In this study we propose a novel method for classifying gait patterns between ACL-deficient (ACLD) knee and unaffected contralateral ACL-intact (ACLI) knee based upon gait system dynamics, intrinsic time-scale decomposition (ITD) and neural networks. First, human leg is modeled as a double-pendulum to imitate and simplify the human walking. Since the lower extremities act as a kinetic chain during dynamic tasks, control of the hip joint will interact with knee motion. Related gait kinematic parameters including knee and hip joint angle and angular velocity are decomposed into a series of proper rotation components (PRCs) and a baseline signal by using the ITD method. The first PRCs of knee and hip joint angle and angular velocity are extracted, which contain most of the kinematic signals' vibration energy and are considered to be the predominant PRCs. Third, neural networks are then used as the classifier with feature vectors as the input to distinguish between ACLD and ACLI knees based on the difference of gait system dynamics between the two groups. Finally, experiments are carried out on forty-three patients to assess the effectiveness of the proposed method. By using the leave-one-out cross-validation style under normal and fast walking speed conditions, the correct classification rates are reported to be 95.12% respectively. In comparison to other state-of-the-art methods, the results demonstrate superior performance and the proposed method may serve as a potential assistant tool for the automatic detection of ACL deficiency in the clinical application.