Surrogate model-assisted numerical analysis of biomechanical properties for carpal tunnel syndrome

Carpal tunnel syndrome (CTS) is a common neurological disorder with major occupational risks including forceful repetitive work. This study aimed to develop a surrogate modeling framework for rapid prediction of carpal tunnel biomechanical properties under different loading conditions, thereby providing a methodological framework for future CTS research.

A high-fidelity three-dimensional wrist model was reconstructed from CT scans of a healthy volunteer using ANSYS and SOLIDWORKS. Subsequent finite element analysis was conducted to quantify the mechanical responses under static and dynamic loading scenarios. To facilitate rapid prediction, two surrogate modeling approaches-Polynomial Chaos Expansion (PCE) and Support Vector Regression (SVR)-were established, enabling efficient estimation of the structural mechanical properties.

Under static palmar tensile load, the transverse carpal ligament was the main deforming structure, with proximal displacement (5.66 mm) significantly larger than distal displacement (0.63 mm, p < 0.05). Under sinusoidal dynamic loads, proximal and middle parts of transverse carpal ligament showed peak displacement responses at 450-550 Hz, while the distal end peaked at 650-750 Hz; proximal responses were higher than middle and distal ends (p < 0.05). The relative average absolute error for static response prediction was 1.23%-2.80%. Both surrogate models had significantly shorter calculation times than finite element analysis (p < 0.05).

As a proof-of-concept demonstration, this study suggests that surrogate models can serve as efficient numerical tools for predicting carpal tunnel biomechanics under varying loading conditions, with the potential to support large-scale simulations in future translational applications.