Whiplash associated disorders are among the most common injuries reported for automotive rear end impacts. Although these injuries are typically considered minor, their high incidence rate and often long-term consequences lead to significant societal costs. The rationale of this research is that a mathematical model of the human head and neck can contribute to a better understanding of neck injury mechanisms and can be used in injury prevention research. The first objective is to develop and validate a detailed three dimensional mathematical model, that describes both the global dynamic behaviour of the human head and neck and the local loading of the neck tissues in accident situations. The second objective is to apply the model to provide insight in factors that might influence the risk of neck injury, such as the amount of activation of the neck muscles, the initial seating posture and the head restraint position.

A detailed multibody neck model has been developed. The model consists of a rigid head, rigid vertebrae, (non)linear viscoelastic discs, frictionless facet joints, nonlinear viscoelastic ligaments and segmented contractile muscles. These muscles follow the curvature of the neck, resulting in realistic muscle force lines of action. Stiffness properties of the tissues are based on literature data. The neck model can be applied separately or integrated into a model of an entire human body. The global kinematics like head, translational and angular, movements and accelerations as well as local kinematics such as vertebral rotations and tissue loads can be predicted with this neck model.

The neck model is validated quasi-statically as well as dynamically. Published quasi- static experimental data were used to test the segment models for 6 degrees of freedom and to test the entire ligamentous cervical spine model for flexion and extension. Frontal, lateral and rear end sled experiments using volunteers as well as post mortem human subjects (PMHSs) were simulated to validate the model dynamically. Neck model simulations of 15 g frontal and 7 g lateral volunteer experiments were performed. The rear end validations were performed with the neck model included in a total human body model. The rear end impact validation ranged from high severity (12 g) for PMHS experiments to mid (4-5 g) and low severity (0.7 g) for volunteer experiments. Accurate muscle activation properties were missing for the volunteer validations. Therefore, simulations were performed with several settings of the reflex delay and activation levels in the muscle model. These settings were based on muscle activation scenarios presented in the literature and derived from low severity rear end experiments performed at Maastricht University. Additionally, simulations with varying head restraint position and initial seating posture were performed for the rear end impacts.

1. The main conclusions of this study are:​
2. The quasi-static passive mechanical behaviour of the model agreed reasonably with the experimental data.
3. The combined findings of existing experimental data presented in the literature and the modelling results presented in this thesis indicate that neck muscles can alter the head and neck kinematics during rear end impacts, as well as during frontal and side impacts.
4. The model with muscle activation predicted reasonably to well, the head and neck kinematics for the frontal, lateral and rear end impact simulations without head contact. Without muscle activation, the model showed larger head motions compared to the experimentally defined volunteer corridors. In case of simulations with head impact with a head restraint less accurate results were obtained.
5. For varying conditions of muscle properties (muscle reflex time, activation level, co-contraction and initial activation), head restraint position and initial posture different trends were observed for several global and local loads and for the head-neck kinematics. Varying muscle activation showed larger influences on the internal loads for the discs, facet joints and ligaments than varying head restraint position or initial seating posture. This indicates that defining accurate model input for the muscle activation is not only necessary to predict the global response, but is also necessary to predict the local response correctly. The correlation found between numerically predicted global injury criteria and local tissue loads with the simulations performed within this research is not strong. Therefore, it is questionable whether global injury criteria may be used to predict local injuries.
6. The model presented in this thesis is suitable for studying further the neck injury mechanisms and neck injury criteria, since it reveals the loads and deformations of individual tissues of the neck.