A reproducible MRI-to-FE framework for generating population-specific and subject-specific finite element head models

Traumatic brain injury (TBI) remains a global health challenge with mechanisms that are still insufficiently understood. While neuroimaging has been used to probe microstructural alterations and their association with head kinematics, findings remain heterogeneous. Finite element (FE) head modelling offers a more robust alternative, demonstrating a superior correlation with observed microstructural changes compared to traditional impact exposure metrics. However, most existing FE models are derived from single-subject scans or generic atlases, which often fail to represent specific study cohorts and introduce significant output variability. This study presents a reproducible computational framework that generates a cohort-specific template brain from MRI scans of adolescent male rugby players to produce a representative FE head model. The model was validated against cadaveric head experiments, demonstrating strong agreement with observed nodal displacements. Furthermore, simulations comparing the template-based model to subject-specific FE models with the identical impact conditions revealed significant differences in brain response. These results underscore the critical necessity of subject-specific modelling for the personalised characterisation of brain biomechanics. Our framework utilizes open-access tools, ensuring full reproducibility for research groups seeking to develop population-, sex-, or ethnicity-specific models. By providing a more accurate representation of cohort-average and individual brain responses, this work contributes to the improved mapping of mechanical strain to clinical findings and neurological alterations. TRANSPARENCY, RIGOR, AND REPRODUCIBILITY SUMMARYThis study is part of an ongoing longitudinal study in New Zealand. All procedures conducted in this study are in accordance with the ethics approval from the New Zealand Health and Disability Ethics Committee (20/NTB/14). All participants aged 16 and older provided informed consent, while participants under 16 provided assent with parental consent. For this study, general exclusion criteria included contraindication to MRI, neurological/psychiatric conditions, and dental braces affecting imaging quality. A total of 78 male high school rugby players (aged 14-18 years old) participated in this study. Inclusion criteria required no mTBI within the past six months prior to start of study, no history of mTBI incident with loss of consciousness, no neurological disorders, no history of drug or excessive alcohol use and no diagnosis of dementia or delirium. Each scan included a multi-parametric MRI scan (i.e. structural, diffusion, functional MRI), and a cognitive and symptom assessment. More details of the parameters and tests used are reported in the manuscript. To record head acceleration exposure across the whole season, an instrumented mouthguard was provided for each rugby player. A control group (14-18 years old) composed of non-collision sport, male athletes, was recruited and scanned at a single timepoint following the same protocol as the rugby players. The same inclusion and exclusion criteria were applied for the control group, with the addition of no self-reported mTBI history or participation in collision sports within the past two years. The primary aim of this study is to establish a computational framework that enables the creation of an average brain from MRI scans of subjects and to develop an FE model. Moreso, this FE model will incorporate fibre dispersion parameter from diffusion MRI and be validated against human head cadaveric experiments reported in the literature. This study is among the few to present a complete framework from MRI to finite element modelling using open-access tools, making it reproducible.