(F-203) A Multiscale Model of Cyclic Head Rotations in the Neonatal Piglet

Introduction:: Repetitive shaking head accelerations are considered to be a central mechanism responsible for traumatic brain injury (TBI) in some cases of abusive head trauma (AHT) in young children.^1,2 Finite element (FE) head models have been developed to assist with understanding and predicting TBI in pediatric and other populations, but most FE models drastically simplify the structures and mechanics of the pia-arachnoid complex (PAC), a thin layer between the brain and the skull containing connective tissues, vasculature, and cerebrospinal fluid (CSF). Accurately modeling the mechanics of the PAC has been shown to significantly improve the accuracy of FE head model predictions.³ But directly computing the mechanics of the complex structures and CSF in the PAC is often prohibitively computationally expensive.

Here, we implement a new constitutive model of the forces developed in the PAC from cyclic sagittal head rotations in an FE model of the neonatal piglet brain, resulting in a reduced-order, multiscale FE neonatal piglet head model. The model allows us to carry out FE simulations of the brain only and represents the forces the skull and PAC impose on the surface of the brain at a low computational cost. The aim of this study is to carry out a qualitative comparison between the cortical stresses and displacements predicted by the new model and an existing FE head model of the neonatal piglet that uses sophisticated three-dimensional models of the PAC and skull and to thereby determine the feasibility of this modeling approach.


Materials and Methods:: The FE brain model was constructed from a geometry file assembled from MRI images of a Göttingen minipig brain.⁴ We added a falx cerebri structure in order to capture the correct gross strain behavior. The brain tissue was modeled as a time-dependent Ogden hyperelastic material, following Prange et al.⁵ and using their materials property values for 5-day-old porcine tissue. Following Scott et al.,³ in order to accurately model the mechanics of the PAC, we defined 12 regions on the top surface of the brain with different average densities of arachnoid trabeculae (AT), strands of connective fibers that extend between the dura mater lining the inner surface of the skull and the pia mater covering the brain, and used their “low-volume fraction” model distribution values (Figure 1a).


The constitutive model of the forces developed in the PAC models the AT in a region as uniform strings with constant diameter and spacing. It applies a theoretical model⁶ of the response of a thin fibrous gap to oscillating shearing impacts. In the model, fibers in the gap are displaced in response to relative motion between the brain and the skull in the sagittal plane, while the flow of CSF through the fibers is governed by an effective medium approach and the Darcy-Lapwood-Brinkman model. The key dimensionless parameters that govern the oscillating shear process are the Womersley number, representing the influence of frequency, the Brinkman number, representing the influence of the porous structure, and the Bingham number, representing the influence of the fiber stiffness.








Results, Conclusions, and Discussions:: To validate our FE model, we compared the predictions of our piglet head model to those of the piglet head model developed by Scott et al.³ While the brain geometries and rotation conditions were different, we found qualitatively similar stress maps, with the highest stress concentrations appearing along the falx and in laterally oriented bands in the mid-anterior region.

We then carried out simulations of repetitive shaking head accelerations using our model. Figure 1b shows the von Mises stresses (color map, units: N/m²) and the normalized in-plane brain tissue displacements (black arrows) that develop on the top surface of the brain for a shaking frequency of 3 Hz at 0° and 90° phase angles. The inset is a close-up of in-plane brain tissue displacement vectors. The magnitudes of the displacements are represented by the lengths of the arrows. All simulations were carried for an angular velocity value of ω = 28 rad/s, a typical value of angular velocity in a pediatric abusive head trauma setting.⁷

We presented a reduced order, finite element head model for the neonatal piglet that aims to accurately represent the forces from the pia-arachnoid complex (PAC) that act on the surface of the brain at a low computational cost via constitutive modeling. Simulations of cyclic head rotations at parameter values typical of pediatric abusive head trauma predicted that the peak stress on the surface of the brain decreases with increasing shaking frequency, and that peak brain tissue displacements happen at 0° phase angle and are concentrated at the lateral edges of the brain hemispheres. While the qualitative comparison with an existing piglet head model is promising, further validation studies must be carried out to demonstrate the accuracy of the model presented here.


Acknowledgements (Optional): : This work was partially supported by the Regenerative Medicine Interdisciplinary Graduate Education Program at Virginia Tech and by NSF Grant #2003077. AES thanks B. Coats for helpful discussions.


References (Optional): : 1. Coats, B et al., J Neurotrauma, 34(1):235-247, 2017. 2. Joyce, T et al., Pediatric Abusive Head Trauma, StatPearls Pub, 2022. 3. Scott, G et al. Biomech Model Mechanobiol, 15(5):1101-19, 2016. 4. thingiverse.com/thing:1696714. 5. Prange, M, and Margulies, S, J Biomech Eng, 124:244-252, 2002. 6. Lang, J, Wang, L, and Wu, Q, Phys. Fluids, In submission. 7. Prange et al., J Neurosurg, 99:143-150, 2003.