Many head injury indices and finite element (FE) head models have been proposed to predict traumatic brain injury (TBI). Although FE head models are suitable methods with high accuracy, they are computationally intensive. Head motion-based brain injury criteria are usually fast tools with lower accuracy. So, the objective of this study is to propose new criteria along with an artificial neural network model to predict TBI risks, which can be fast and accurate.

For this purpose, 250 FE head simulations have been carried out at 5 magnitudes and 50 rotational impact directions using the SIMon model. The effects of directions and magnitudes of rotational impacts were assessed for cumulative strain damage measure (CSDM) values. Next, statistical analysis and neural network were applied to predict CSDM values.

The results of the present research showed that the direction of rotation in the sagittal and frontal planes had a considerable effect on the CSDM values. Furthermore, new brain injury indices and a radial basis function neural network have been proposed to predict CSDM values which having high correlation coefficients with SIMon responses.

The results of this research demonstrated that rotational impact directions should be used to develop new head injury criteria being able to predict CSDM values. However, findings of present research proved that head motion-based brain injury criteria and RBF network can be used to predict FE head model responses with high speed and accuracy.