Electronic Thesis and Dissertation Repository

Development of a Combined Experimental-Computational Framework to Study Human Knee Biomechanics

Samira Vakili, The University of Western Ontario

Abstract

Despite decades of clinical, experimental, and computational research on knee joints, the lack of understanding of ligaments' contribution in knee biomechanics has hindered efforts to accurately predict the effects of diseases, injuries, and the long-term effectiveness of reconstruction surgeries. Therefore, a thorough understanding of the biomechanical contributions of knee ligaments is essential for developing successful rehabilitation and surgical planning after injury.

Contemporary, a growing number of studies have been conducted using cadaveric knees to quantify ligament contributions to normal and pathological knee biomechanics. In these studies, the anatomic variability in patient populations is present in cadaveric specimens, making study findings more applicable. However, studying the effects of injuries and surgical interventions on cadaveric specimens can result in irreparable tissue damage, requiring the use of new specimens, which is expensive and time-consuming. Alternatively, computational modeling has been used to parametrically analyze knee ligaments and evaluate their responses to various pathologies and treatments. However, these models are difficult to develop fully validate. Even after a model has been developed, the predictions based on that model may not be applicable to all subjects. Recent studies have focused on combining cadaveric experiment-based results with subject-specific multibody or finite element models. In these studies, experimental results are used as a basis for post-hoc tuning of computational models. While this approach may allow models to reproduce the system-level response of the joint, inaccuracies in how the ligaments are modelled may be buried amongst the other simplifications in the model. It is possible that these simplifications lead to unrealistic forces and inconsistent results.

The purpose of this thesis was to develop a combined experimental-computational technique for characterizing the biomechanical contributions of knee ligaments to normal and pathological joint biomechanics. In this approach, a computer simulation of soft tissues (virtual ligaments) was developed to mimic the force-length behavior of ligamentous structures. The designed virtual ligaments were used to stabilize knee joints on a joint motion simulator. Using virtual ligaments on knee joints bridged the gap between experimental and computational approaches, which can predict experimental data while parametrically simulating ligament repair and injury.