Motion analysis based on skin markers aims at evaluating the joint kinematics but the relative movement between the bones and the markers, known as soft tissue artifacts (STA), introduces large errors. Multi-body optimization (MBO) methods were proposed to compensate for the STA. However, the validation of the MBO methods using no or simple kinematic constraints (e.g., spherical joint) showed their inefficiency to estimate accurate in vivo kinematics. Anatomical constraints (i.e., contact surface and ligament constraints) were introduced in MBO methods and various ligament constraints were proposed in the literature with null, minimized or prescribed ligament length variations. The validation of these methods has not been performed yet. The objective of this study was to validate, against in vivo knee joint kinematics measured by intra-cortical pins on three subjects, the model-based kinematics obtained by a MBO method using three different types of ligament constraints. The MBO method introducing minimized or prescribed ligament length variations showed some improvements in the estimation of knee kinematics when compared to no kinematic constraints, to coupling curves, and to null ligament length variations. However, the improvements were marginal when compared to spherical constraints. The errors obtained by minimized and prescribed ligament length variations were below 2.5° and 4.1 mm for the joint angles and displacements while the errors obtained with spherical joint constraints were below 2.2° and 3.1 mm. These errors are generally lower than the errors previously reported in the literature. As the present study used a generic knee kinematic model, personalization of the geometry should be considered for further improvements of the estimation of the model-based kinematics from skin markers. As a conclusion, this study presented encouraging results for the compensation of the STA by MBO and for the introduction of anatomical constraints in MBO, with potential application to the development of personalized models and their inclusion in musculoskeletal models which should allow computation of the ligament and contact forces at the joint.
Gasparutto, X., Sancisi, N., Jacquelin, E., Parenti-Castelli, V., Dumas, R. (2015). Validation of a multi-body optimization with knee kinematic models including ligament constraints. JOURNAL OF BIOMECHANICS, 48(6), 1141-1146 [10.1016/j.jbiomech.2015.01.010].
Validation of a multi-body optimization with knee kinematic models including ligament constraints
SANCISI, NICOLA;PARENTI CASTELLI, VINCENZO;
2015
Abstract
Motion analysis based on skin markers aims at evaluating the joint kinematics but the relative movement between the bones and the markers, known as soft tissue artifacts (STA), introduces large errors. Multi-body optimization (MBO) methods were proposed to compensate for the STA. However, the validation of the MBO methods using no or simple kinematic constraints (e.g., spherical joint) showed their inefficiency to estimate accurate in vivo kinematics. Anatomical constraints (i.e., contact surface and ligament constraints) were introduced in MBO methods and various ligament constraints were proposed in the literature with null, minimized or prescribed ligament length variations. The validation of these methods has not been performed yet. The objective of this study was to validate, against in vivo knee joint kinematics measured by intra-cortical pins on three subjects, the model-based kinematics obtained by a MBO method using three different types of ligament constraints. The MBO method introducing minimized or prescribed ligament length variations showed some improvements in the estimation of knee kinematics when compared to no kinematic constraints, to coupling curves, and to null ligament length variations. However, the improvements were marginal when compared to spherical constraints. The errors obtained by minimized and prescribed ligament length variations were below 2.5° and 4.1 mm for the joint angles and displacements while the errors obtained with spherical joint constraints were below 2.2° and 3.1 mm. These errors are generally lower than the errors previously reported in the literature. As the present study used a generic knee kinematic model, personalization of the geometry should be considered for further improvements of the estimation of the model-based kinematics from skin markers. As a conclusion, this study presented encouraging results for the compensation of the STA by MBO and for the introduction of anatomical constraints in MBO, with potential application to the development of personalized models and their inclusion in musculoskeletal models which should allow computation of the ligament and contact forces at the joint.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
