Finite element (FE) models of the human head have been used extensively to assess engineering response parameters (pressure, shear strain, etc.) that constitute the basis of various injury criteria including the Head Injury Criterion (HIC). There is a wide range of brain material properties reported in the literature, and linear elastic models are frequently used for the brain tissue which is viscoelastic in nature. The uncertainty around the properties of brain tissue affects the perceived importance of those response parameters under consideration.

In this research, the effect on head injury criteria of five reported viscoelastic brain models and their linear elastic alternatives is studied using both a simple spherical head model and a geometrically realistic head model. It is concluded that even the most proximate linear elastic model, which employs a Young’s modulus equal to the short term modulus of the viscoelastic model, does not provide realistic representations of shear responses possibly associated with head injuries.

The calculated HIC value is found to be insensitive to the choice of brain material model, whether based on acceleration histories taken at the center of gravity or the side of the head, despite the fact that the shear strain histories in the brain vary drastically from model to model. Furthermore, on the basis of a hypothetical shear strain criterion for the threshold of tissue damage, the percentages of injured brain tissue predicted using different models vary greatly. The pressure response of all models is found to be virtually identical provided that the same bulk modulus (that of water) is employed. Parametric studies indicate that decreasing the bulk modulus leads to lower peak pressures and higher HIC values.

The fact that the widely varying internal responses associated with the various models considered in this study does not lead to variations in calculated HIC exposes the limitation of HIC. Therefore, HIC may not be a reliable indicator of actual internal dynamics and tissue damage. The FE method could be used to address this deficiency and provide guidance for the experimental determination of realistic constitutive models for brain tissue.