For the past 20 years or so, performance requirements intended to protect motor vehicle occupants against head injury have invariably been based on the use of the Head Injury Criterion (HIC). The most evident difficulty with the Criterion is simply that the assumed inverse relationship between the tolerable level of head acceleration and its duration leads to predictions that contradict practical experience.

The data and analysis most widely cited in support of the continued use of HIC (with a time limit of 15 ms) are reviewed. The assumed equivalence of HIC values from cadaver and ATD head impacts is discussed and several difficulties identified. The validity of the derived "threshold risk curve" is tested against the data by numerically simulating the original experiments. The curve is found substantially to misrepresent the experimental data on which it is based. Extensive data from tests by Transport Canada and other agencies are presented which show that, regardless of either the time duration of the calculation or the criterion level, HIC is incapable of distinguishing potentially injurious events from those that are known to be harmless. For the protection of motor vehicle occupants in frontal collisions, the same data show that an 80 g limit on ATD head acceleration constitutes a rational, attainable and effective performance requirement.