The human brain develops from a smooth cortical surface in early stages of fetal life to a convoluted one postnatally, creating an organized ensemble of folds. Abnormal folding patterns are linked to neurodevelopmental disorders. However, the complex multi-scale interactions involved in cortical folding are not fully known yet. Computational models of brain development have contributed to better understand the process of cortical folding, but still leave several questions unanswered. A major limitation ...
The human brain develops from a smooth cortical surface in early stages of fetal life to a convoluted one postnatally, creating an organized ensemble of folds. Abnormal folding patterns are linked to neurodevelopmental disorders. However, the complex multi-scale interactions involved in cortical folding are not fully known yet. Computational models of brain development have contributed to better understand the process of cortical folding, but still leave several questions unanswered. A major limitation of the existing models is that they have basically been applied to synthetic examples or simplified brain anatomies. However, the integration of patient-specific longitudinal imaging data is key for improving the realism of simulations. In this work we present a complete computational pipeline to build and validate patient-specific mechanical models of brain development. Starting from the processing of fetal brain magnetic resonance images (MRI), personalised finite-element 3D meshes were generated, in which biomechanical models were run to simulate brain development. Several metrics were then employed to compare simulation results with neonatal images from the same subjects, on a common reference space. We applied the computational pipeline to a cohort of 29 subjects where fetal and neonatal MRI were available, including controls and ventriculomegaly cases. The neonatal brain simulations had several sulcal patterns similar to the ones observed in neonatal MRI data. However, the pipeline also revealed some limitations of the evaluated mechanical model and the importance of including patient-specific cortical thickness as well as regional and anisotropic growth to obtain more realistic and personalised brain development models.
Statement of Significance: Computational modelling has emerged as a powerful tool to study the complex process of brain development during gestation. However, most of the studies performed so far have been carried out in synthetic or two-dimensional geometries due to the difficulties involved in processing real fetal data. Moreover, as there is no correspondence between meshes, comparing them or assessing whether they are realistic or not is not a trivial task. In this work we present a complete computational pipeline to build and validate patient-specific mechanical models of brain development, mainly based on open-source tools.
+
