Diffuse brain injury is a form of severe brain injury which occurs primarily in vehicular accidents, and is responsible for 35% of the deaths in.severely head injured patients. A coordinated series of animal experiments, physical and analytical simulations, and isolated tissue tests have been used to formulate a tolerance criterion for concussion and more severe forms of diffuse axonal injury in man, but the relationship between these tolerance criteria and measurable .vehicle crash parameters has not been studied. In this report, the kinematics of the occupant in an idealized side impact are studied using. the Crash Victim Simulator/Articulated Total Body (CVS/ATB) program. An emphasis was placed on relating peak coronal plane rotational acceleration and rotational velocity of the head to the change in velocity (ΔV_y) describing the crash. Results using a three ellipsoid model of the head-neck-upper torso (no head contact) indicated substantial changes in torso velocity (ΔV=36 mph) were necessary to attain the maximum non-injurious loading conditions measured in human volunteer tests. Higher crash velocities were needed to exceed the tolerance for concussion (ΔV=46 mph), and yet higher velocities for mild to severe DAL. These results suggest the importance of head contact to generate the inertial loading conditions to cause diffuse brain injuries in minor to moderate collisions. The conclusions regarding the importance of head contact in producing diffuse brain injuries is supported by recent epidemiological studies.