A review of problems in automotive head injury research has identified the fragmented and limited view of the severity of these injuries in terms of their mechanisms and outcome which we now posess. The lack of comprehensive interdisciplinary research which can integrate pre-crash, crash and post-crash events and extend studies of crash victims to include neuropsychologic and social factors observed serially over many years is shown to be a major deficit. Current findings in outcome research are reviewed and the severity of automotive head injuries is suggested to be significantly underestimated. A new paradigm for improved understanding of these problems and within which improved methodologies may be tested is proposed. It is argued that brain injuries cannot be viewed as static events, the outcome of which are wholly predictable in individual cases from the initial severity as guaged e.g. by the A.I.S. or even by acute phase prognostications based on clinical scales such as Glasgow coma and outcome scores with and without supplementation with acute phase diagnostic tests such as multimodal evoked potentials. Dynamic response processes of the brain to injury include continuous interactions with environmental factors for prolonged periods of time. These processes are based on the regenerative and reintergrative properties of the central nervous system, the key aspects of which are the cognitive and affective components of consiousness. The methodology we have developed and employed in an on-going study is outlined and a testable hypothesis to explain how outcome predictions may be improved is described.