A generative model of realistic brain cells with application to numerical simulation of the diffusion-weighted MR signal

• Published in: NeuroImage (Orlando, Fla.) Vol. 188; pp. 391 - 402
• Main Authors: Palombo, Marco, Alexander, Daniel C., Zhang, Hui
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.03.2019; Elsevier Limited
• Subjects: Brain; Brain - cytology; Brain - diagnostic imaging; Cell structure; Computational neuroscience; Computer Simulation; Cytology; Diffusion; Diffusion Magnetic Resonance Imaging - methods; Diffusion-weighted MRI; Humans; Intracellular space; Investigations; Magnetic fields; Magnetic resonance imaging; Metabolites; Models, Theoretical; Monte Carlo simulation; Morphology; Nervous system; Neuroimaging - methods; Permeability; Tissue microstructure; Cell structure; Brain; Tissue microstructure; Diffusion-weighted MRI; Intracellular space; Diffusion; Monte Carlo simulation
• ISSN: 1053-8119; 1095-9572
• DOI: 10.1016/j.neuroimage.2018.12.025

To date, numerical simulations of the brain tissue have been limited by their lack of realism and flexibility. The purpose of this work is to propose a controlled and flexible generative model for brain cell morphology and an efficient computational pipeline for the reliable and robust simulation of realistic cellular structures with application to numerical simulation of intra-cellular diffusion-weighted MR (DW-MR) signal features. Inspired by the advances in computational neuroscience for modelling brain cells, we propose a generative model that enables users to simulate molecular diffusion within realistic digital brain cells, such as neurons, in a completely controlled and flexible fashion. We validate our new approach by showing an excellent match between the morphology (no statistically different 3D Sholl metrics, P > 0.05) and simulated intra-cellular DW-MR signal (mean relative difference < 2%) of the generated digital model of brain cells and those of digital reconstruction of real brain cells from available open-access databases. We demonstrate the versatility and potential of the framework by showing a select set of examples of relevance for the DW-MR community. The computational models introduced here are useful for synthesizing intra-cellular DW-MR signals, similar to those one might measure from brain metabolites DW-MRS experiments. They also provide the foundation for a more complete simulation system that will potentially include signals from extra-cellular compartments and exchange processes, necessary for synthesizing DW-MR signals of relevance for DW-MRI experiments. •We propose a generative model for controlled and flexible modelling of brain cells.•We simulate molecular diffusion inside digital cells with an unprecedented level of realism.•Relevant examples for simulation of diffusion-weighted MR signal are shown.•Excellent match between the morphology of real and digital brain cell models.•Excellent match between the DW-MR signal from real and digital brain cell models.