Computer Methods in Biomechanics and Biomedical Engineering edition:14 location:Tel Aviv date:20 - 22 September 2016
Acute subdural haematoma (ASDH) is a type of intracranial haemorrhage following head impact. Bridging vein (BV) rupture is considered a major cause of ASDH, which is why a biofidelic representation of BVs in finite element (FE) head models is essential for the successful prediction of ASDH.
Materials and methods
We investigated the mechanical behavior of the BVs in the KTH FE head model .
First, a sensitivity study was performed, quantifying the effect of varying loading conditions (pulse duration (PD), rotational (RA) and linear acceleration (LA) and peak rotational velocity (RV)) and mechanical properties (Young’s modulus (E) and outer diameter (OD)) on the BV strains using a multiple linear regression model.
Secondly, previously performed cadaver head impact experiments  were simulated with varying sets of mechanical properties. For each set of 12 simulations the prediction result of the simulations is cross checked with its corresponding experimental case and the number of successful prediction is tracked.
The correlation analysis showed that RV had a significant correlation with the maximum value of the RA and PD. These three parameters showed a pronounced positive correlation with the BV strains. RV had the biggest effect. In contrast, LA had no effect on the BV strains. For the mechanical properties it was found that both E and OD were negatively correlated with the BV strains. A multiple linear regression model using E, OD and AREA as independent variables to predict the BV strain yielded an adjusted R2-value of 0.81.
The second part of this study showed that depending on the choice of mechanical parameters, the success rate of the simulations fluctuated between 67% and 75%.
Discussion & Conclusion
An adjusted R2-value of 0.81 indicates that a big part of the BV behavior can be predicted with the E, OD and RV. It is also found that the prediction rates of the cadaver experiments are fairly good, which could be improved by either finding a higher number of relevant predictors or a more biofidelic representation of the BVs. Given the limited data available, it would also be very interesting to acquire a bigger database with accident or impact cases where BV rupture occurred. This would greatly benefit the improvement and validation of the BV representation in the FE head model.
Acknowledgments: This work was supported the Institute for the Promotion of Innovation through Science and Technology in Flanders (I.W.T.) and the Research Foundation – Flanders (FWO).
 B. Depreitere, et al., Journal of neurosurgery 104 (6) (2006) 950-956
 S. Kleiven, International Journal of Crashworthiness 11 (1) (2006) 65-79.