Subject-Specific Musculoskeletal Modeling of Adolescent Scoliosis Patients: a Pilot Study

Abstract This paper presents developing and validating an algorithm-based protocol to build subject-specific musculoskeletal models using X-ray images and motion capture data. The MATLAB algorithm used bilateral X-ray images to determine the deviation angle of each vertebra in the frontal and sagittal planes and motion capture data to define the rest of the bone structure in a static posture. Previously validated musculoskeletal models were used as templates to avoid the necessity of scaling to each subject. Inverse kinematics and static optimization tools in OpenSim were used to define the joint angles and muscle activation for a scoliosis-specific exercise. Validation was done for the algorithm by comparing the vertebrae rotation angles with those obtained from a public database of scoliotic spine patients. The results showed that the Cobb angle deviation from the algorithm is smaller than the reported standard deviation for experimental measurements. Despite the successful simulation of muscle activation around the convex and concave sides of the scoliotic spine, future research could improve the protocol by including axial rotation of vertebrae and muscle scaling factors to enhance the individualization of each model. Nevertheless, the present protocol can create an individualized musculoskeletal model for a scoliosis patient.