Evaluation of a deformable musculoskeletal model for estimating muscle-tendon lengths during crouch gait

Abstract

The hamstrings and psoas muscles are often lengthened surgically in an attempt to correct crouch gait in persons with cerebral palsy. The purpose of this study was to determine if, and under what conditions, medial hamstrings and psoas lengths estimated with a "deformable" musculoskeletal model accurately characterize the lengths of the muscles during walking in individuals with crouch gait. Computer models of four subjects with crouch gait were developed from magnetic resonance (MR) images. These models were used in conjunction with the subjects' measured gait kinematics to calculate the muscle-tendon lengths at the body positions corresponding to walking. The lengths calculated with the MR-based models were normalized and were compared to the lengths estimated using a deformable generic model. The deformable model was either left undeformed and unscaled, or was deformed or scaled to more closely approximate the femoral geometry or bone dimensions of each subject. In most cases, differences between the normalized lengths of the medial hamstrings computed with the deformable and MR-based models were less than 5 mm. Differences in the psoas lengths computed with the deformable and MR-based models were also small (<3 mm) when the deformable model was adjusted to represent the femoral geometry of each subject. This work demonstrates that a deformable musculoskeletal model, in combination with a few subject-specific parameters and simple normalization techniques, can provide rapid and accurate estimates of medial hamstrings and psoas lengths in persons with neuromuscular disorders.