Injury prediction and mitigation are common overarching goals of modern biomechanical research. This research is fundamental to preventing and mitigating injuries sustained by those exposed to dangerous conditions including but not limited to occupational hazards, warfighter risks, automotive accidents, etc. Unlike traditional mechanical system research, biological systems are difficult and costly to test resulting in a need for robust and accurate numerical simulations. Models of the cervical spine are complex, nonlinear systems that must accurately model dynamic loading, large deflections, elastic, and viscoelastic behavior. In addition to individual complexities, population variance in both material properties and shape must be taken into account for accurate injury prediction.
As part of a hierarchical validation and verification (V&V) methodology, lateral impact cadaveric cervical spine experiments were compared to a high fidelity statistical shape finite element model (SSFEM) of the cervical spine and head. Specimens were mounted to a sled and accelerated using a pendulum impact with 1, 2, and 3 m/s impact velocities. The kinematics of the head and all individual cervical vertebrae were recorded with a Vicon motion capture system along with sled acceleration data. Sled accelerations were used as input boundary conditions for the probabilistic study using the SSFEM. Head and vertebrae rotations between the experimental and model responses were then compared.
A latin hypercube probabilistic analysis was performed for each impact velocity to determine the probabilistic response of each rotation metric. When comparing these responses, both the average and variation must be taken into consideration. This is accomplished using a quantitative validation metric based on the area between the cumulative distribution functions (CDF) of experimental response and the computed probabilistic response. Our results showed a very good match between the model and experiment at the higher impact velocities and a slightly stiffer response at lower rates. These results are consistent with previous validation studies performed with this SSFEM.
By expanding the validation data set with lateral impact loading, greater confidence in the model is obtained under different loading modes. This confidence allows the model to be used for probability of injury predictions as well as to identify important system variables in preventing injuries. High fidelity numeric modeling allows for rapid and cost effective assessment of hazardous loading conditions and safety equipment compared to experimental modeling. The knowledge gained from these modeling studies is fundamental and necessary for safe and effective design and injury mitigation.
