Multi-scale simulation of axonal injuries

QC 2026-02-18

Abstract

Diffuse axonal injury (DAI) is a common and devastating form of traumatic brain injury caused by external forces applied to the head, which in turn damages axons within the brain’s white matter at the cellular level. This type of injury often remains undiagnosed, as symptoms may develop gradually over days, weeks or even months following the initial impact. Therefore, a deeper understanding of the mechanisms by which mechanical forces applied at the head level lead to axonal and molecular damage is crucial for improving early diagnosis and developing more effective treatment and prevention strategies. 

Because of DAI’s multi-scale nature, its exact mechanisms remain poorly understood and difficult to investigate experimentally. Computational modelling, however, provides a valuable tool to explore these mechanisms in greater detail. In particular, coupling different computational approaches enables the study of injury phenomena across multiple scales. This thesis addresses unanswered questions about DAI using finite element (FE) analysis and molecular dynamics (MD). In total, four studies were conducted to form this thesis and can be divided into two parts. The first two studies aim to investigate DAI at the cellular level using FE, and the other two studies focus on this injury at the molecular level using MD simulations. 

Axon’s behaviour under deformation depends largely on the matrix in which it is embedded. Namely, the axons in the brain are surrounded by various cells, the extracellular matrix, and myelin sheaths. However, the mechanical properties of the matrix that surrounds the axons in different regions of the brain are not known. To address this, we conducted Study I to determine the material properties of microstructural elements in the brain white matter across different regions. Namely, the tissue was modelled as a composite material consisting of an axon surrounded by a matrix. The matrix represented other components in the brain white matter that surround the axon. Then, the material properties of the axon and the matrix were calculated from available experimental data. The derived matrix material properties were used in Study II to investigate the response of the axon and its subcellular components to different loading modes. FE models of the brain white matter, consisting of axons embedded inside a matrix, were modified and simulated under tension, compression and shear, and the response of subcellular components of the axon was monitored. According to the results, the axonal compartments deform more under tension and compression than under shear. In addition, the membrane elements appear to be one of the main load-bearing compartments. Therefore, we conducted two molecular-level studies to investigate these components in more detail. 

Study III used coarse-grained MD simulations to model axonal membranes at different v regions and examine their deformation behaviour up to rupture. The results show that rupture strains across the membrane differ across different axonal regions, with the node of Ranvier having the lowest rupture strain. Moreover, by linking MD results to FE simulations, we estimated the axonal strains at the cell level associated with the membrane rupture at the molecular level. Since this study was used at a coarse-grained level where three or four large atoms are modelled together, it could not capture the atomistic changes that occur in the membrane. To address this limitation, Study IV employed atomistic MD simulations to model the axonal membrane as a protein embedded in a bilayer and to investigate the mechanical behaviour of both the bilayer and the embedded protein under strain at the atomic level. The results suggest that structural changes in the protein that might lead to its malfunction occur before the bilayer rupture. 

Collectively, these four studies provide new insights into the lesser-known aspects of DAI by combining computational methods across multiple scales. The findings establish more accurate injury thresholds for axons under various deformation modes and contribute to a more comprehensive understanding of DAI mechanisms at the tissue, cellular and molecular levels.

Link to DiVA