Development of a full flexion 3D musculoskeletal model of the knee considering intersegmental contact and deep muscle activity during high knee flexion

Abstract

Habitual kneeling in high knee flexion postures is a risk factor for knee joint dysfunction yet critical parameters for modeling this range of motion remain unknown or untested in three dimensions. High flexion is defined as postures exceeding 120° at the knee joint to a maximum of approximately 165°. Specific occupational and ethnic populations that regularly use high knee flexion postures have increased prevalence of degenerative knee diseases. This could suggest a causal relationship between habitual kneeling and disease prevalence resulting from repeated exposures. Therefore, this thesis was designed to explore two critical components for high knee flexion biomechanical modeling: intersegmental (thigh-calf and heel-gluteal) contact forces and lower limb muscular activation patterns across the full range of knee flexion. The global objective of this work was to develop a 3D musculoskeletal (MSK) model of the knee to estimate tibial contact forces in high knee flexion postures for determining the effect of intersegmental contact on these calculations. Two experimental studies, verification against a ‘gold-standard’ dataset, and an application study supported this global objective. Study 1: The purposes of this study were: 1) to measure total intersegmental contact force magnitude and centre of force (CoF) location during six high knee flexion movements and 2) to define regression models, based on anthropometrics, for the estimation of intersegmental contact parameters. Fifty eight participants completed six high knee flexion movements while motion capture and pressure data from the right lower limb were recorded. High knee flexion movements had average peak total intersegmental contact force magnitudes ranging from ~50-200N or ~8-30 %BW. Intersegmental CoF locations were segregated between thigh-calf and heel-gluteal regions with CoF, at peak total force, being ~6.2 cm and ~32.7 cm distal from the functional knee joint center about the long axis of the femur respectively. Five parameters of intersegmental contact (onset, maximum knee flexion angle, total contact force, thigh-calf CoF, and thigh-calf contact area) were then assessed for anthropometric based regression model fit. Strong correlations and linear regression models were found for maximum knee flexion angle and thigh-calf CoF, but only moderate to weak results were found for all other intersegmental contact parameters. The overall poor fit and variance explained by the linear regression models for onset, total force, and contact area suggest further work is needed to provide estimations of these parameters for use in future modeling efforts. Study 2: The purposes of this study were: 1) to measure surface and fine-wire EMG activation profiles in six high knee flexion movements and 2) to establish if surface EMG sites can be used as a proxy for fine-wire activation profiles. Sixteen participants completed the same high knee flexion movements, and level walking, as study 1 while activation waveforms from three deep muscles—vastus intermedius (VI), adductor magnus (AM), and semimembranosus (SM)—were recorded using fine-wire electrodes for comparison to easily accessible surface sites. Average peaks of VI, AM, and SM fine-wire activations during high knee flexion movements were approximately 30, 85, and 35 %MVC respectively. None of the surface sites recorded satisfied our criteria to successfully model fine-wire recordings. This was largely due to the considerable variability of surface-indwelling comparisons between participants. Our findings would suggest that the use of fine-wire EMG to obtain representative activation waveforms from VI, AM, or SM may be required if isolated muscle/motor unit activity is needed. MSK model: A full range of knee motion MSK model was developed for the estimation of tibial contact forces. Verification of the MSK model was completed by calculating the error between tibial compressive force estimates and measurements from an instrumented knee implant (gold standard). Vertex based object files of participant bones and CAD files of implant components were obtained from a public repository for gold standard data with muscle geometry scaled from our MSK model. Tibial compression estimates strongly fit implant data shape during walking (R2 0.83), squatting (R2 0.93), and ‘bouncy’ walking (R2 0.74) with an RMSD of 0.47, 0.16 and 0.58 BW respectively. Qualitative assessments of recorded EMG and muscle force estimations showed poor agreement between time-series data. Therefore, the strong fit of MSK tibial compression estimates to gold standard data would suggest this model is phenomenological in nature and does not accurately represent neuromuscular control. Application: The purpose of this study was to quantify the effect of including intersegmental contact on external knee joint moments and tibial contact force estimations. This study used participant data collected from study 2. There was an average RMSD of 3.56, 0.16, and 0.06 %BW*HT in flexion/extension, ab/adduction, and int/external external moments respectively when considering intersegmental contact parameters. Reductions in external moments caused changes to mean RMSD tibial contact force estimates: 0.14 BW lower compression, 0.2 BW lower posterior shear, and 0.03 BW higher lateral shear. Muscle force estimates generally followed EMG waveforms in shape for vastii, gluteus medius, and AM with SM having an improved agreement using its indwelling signal compared to surface measurements. General conclusions: Intersegmental contact forces must be considered when reporting tibial contact forces during high knee flexion movements as significant reductions to tibial posterior shear and increases in lateral shear were observed. Further work is required to refine MSK models in these ranges of knee motion as pressure sensor technology and soft tissue artifact are considerable limitations. Measurement of populations who habitually perform these activities needs to be completed to assess the translation of these findings to appropriate individuals.