Abstract:
Biological systems are inherently complex, often necessitating computational methods such as finite element modeling (FEA) for comprehensive biomechanical analysis. However, most computational analyses traditionally fall short in accounting for the inherent variability and uncertainty present in model inputs and boundary conditions, which severely limits their ability to predict a probability of injury within a given biological system. This study directly addresses this limitation by aiming to calculate the probabilistic biomechanical response of a validated and verified parametric cervical spine finite element model. This is achieved through the crucial incorporation of variability into the model's inputs, specifically focusing on the biomechanical properties of soft tissues and geometric parameters. The intricate geometry of the finite element model was meticulously created using a set of geometry parameters derivable from Computed Tomography (CT) scans, measured from both male and female volunteers to capture inter-subject variability. Crucially, the FE model's mesh was generated using the Truegrid preprocessing software, which built the mesh based on these geometry parameters defining the surface geometry of each vertebra. Material properties for the soft tissues of the cervical spine, essential for accurate biomechanical simulation, were meticulously determined from a combination of existing literature and experimental data. The NESSUS probabilistic engineering analysis software was then utilized to perform the probabilistic analysis, perturbing both geometry (via Truegrid) and material properties to feed into an LS-Dyna model for simulations. The work highlights the importance of transitioning from deterministic biomechanical models to probabilistic frameworks to better represent the range of human responses to mechanical loads. By introducing this variability, the model offers a more realistic and robust prediction of spinal injury risk, moving beyond single-point estimates. This approach contributes significantly to injury biomechanics by providing a framework that can quantify the likelihood of injury, ultimately enhancing the predictive power of human body models and informing injury prevention strategies. The methodology encompasses thorough verification and validation, ensuring the model's biofidelity and reliability for real-world applications in crash simulations and other high-impact scenarios.
