An efficient framework for fully deformable finite element musculoskeletal spine modeling: development, sensitivity...

Area of Science:

• Biomechanics
• Computational modeling
• Spinal research

Background:

• Current computational musculoskeletal models simplify spinal structures and lack muscle integration.
• Accurate prediction of internal spinal loads under realistic conditions remains a challenge.

Purpose of the Study:

• To integrate active muscle architecture into a lumbar finite element (FE) model.
• To develop efficient surrogate models for predicting spinal kinematics based on muscle forces.
• To enable detailed load-sharing assessments across spinal components.

Main Methods:

• Utilized a multi-stage dimensionality reduction strategy (Plackett-Burman, Latin Hypercube, Farthest Point Sampling).
• Developed surrogate Generalized Linear Models to predict spinal kinematics.
• Employed an optimization procedure to identify muscle activations minimizing energy criteria.

Main Results:

• Model accurately predicted physiological trends, with increased abdominal activation and intradiscal pressure during flexion.
• Calculated L4-L5 compressive forces in standing (74% body weight) and detailed load distribution (disc: 44%, facet capsules: up to 26%).
• Demonstrated significant increases in disc load and facet capsule tension during flexion, with the disc bearing primary shear forces.

Conclusions:

• The integrated model accurately predicts complex spinal loading patterns.
• The approach reduces computational cost for musculoskeletal FE simulations.
• Provides a more comprehensive understanding of spinal biomechanics and load sharing.