ConFiG: Contextual Fibre Growth to generate realistic axonal packing for diffusion MRI simulation

Area of Science:

• Neuroimaging
• Biophysics
• Computational Neuroscience

Background:

• Generating realistic white matter numerical phantoms is crucial for developing and validating diffusion MRI microstructure modelling approaches.
• Current methods, often based on packing fibres, face limitations in achieving high densities and accurately representing complex fibre configurations.

Purpose of the Study:

• To introduce Contextual Fibre Growth (ConFiG), a novel approach for generating white matter numerical phantoms.
• To compare ConFiG's performance against state-of-the-art methods in terms of density, microstructural morphology, and realism of simulated diffusion MRI data.

Main Methods:

• ConFiG mimics natural fibre genesis by growing fibres individually based on axonal guidance mechanisms.
• Phantoms are generated with tuneable microstructural features, meeting targets for density and orientation distribution.
• Phantoms were generated in various configurations, including crossing fibre bundles and orientation dispersion, and compared to existing packing-based methods.

Main Results:

• ConFiG achieved up to 20% higher densities compared to the state-of-the-art, especially in complex crossing fibre configurations.
• ConFiG phantoms exhibit microstructural morphology comparable to real white matter tissue, with realistic diameter and orientation distributions.
• Simulated diffusion MRI signals from ConFiG phantoms closely matched real diffusion MRI data.

Conclusions:

• ConFiG is a feasible method for generating realistic synthetic diffusion MRI data.
• The approach enables the creation of tuneable white matter phantoms with high fidelity to real tissue microstructures.
• ConFiG facilitates the development and validation of advanced diffusion MRI microstructure modelling techniques.