Severe-to-fatal head injuries in motor vehicle environments were analyzed using the United States Crash Injury Research and Engineering Network database for the years 1997-2006. Medical evaluations included details and photographs of injury, and on-scene, trauma bay, emergency room, intensive care unit, radiological, operating room, in-patient, and rehabilitation records. Data were synthesized on a case-by-case basis. X-rays, computed tomography scans, and magnetic resonance images were reviewed along with field evaluations of scene and photographs for the analyses of brain injuries and skull fractures. Injuries to the parenchyma, arteries, brainstem, cerebellum, cerebrum, and loss of consciousness were included. In addition to the analyses of severe-to-fatal (AIS4+) injuries, cervical spine, face, and scalp trauma were used to determine the potential for head contact. Fatalities and survivors were compared using nonparametric tests and confidence intervals for medians. Results were categorized based on the mode of impact with a focus on head contact. Out of the 3178 medical cases and 169 occupants sustaining head injuries, 132 adults were in frontal (54), side (75), and rear (3) crashes. Head contact locations are presented for each mode. A majority of cases clustered around the mid-size anthropometry and normal body mass index (BMI). Injuries occurred at change in velocities (DeltaV) representative of US regulations. Statistically significant differences in DeltaV between fatalities and survivors were found for side but not for frontal impacts. Independent of the impact mode and survivorship, contact locations were found to be superior to the center of gravity of the head, suggesting a greater role for angular than translational head kinematics. However, contact locations were biased to the impact mode: anterior aspects of the frontal bone and face were involved in frontal impacts while temporal-parietal regions were involved in side impacts. Because head injuries occur at regulatory DeltaV in modern vehicles and angular accelerations are not directly incorporated in crashworthiness standards, these findings from the largest dataset in literature, offer a field-based rationale for including rotational kinematics in injury assessments. In addition, it may be necessary to develop injury criteria and evaluate dummy biofidelity based on contact locations as this parameter depended on the impact mode. The current field-based analysis has identified the importance of both angular acceleration and contact location in head injury assessment and mitigation.