Diffusion MRI simulation in thin-layer and thin-tube media using a discretization on manifolds.

Area of Science:

• Medical Imaging
• Computational Physics
• Biomedical Engineering

Background:

• The Bloch-Torrey partial differential equation models transverse magnetization evolution in diffusion MRI.
• Simulating diffusion MRI signals is crucial for understanding biological tissues.
• Standard 3D finite element methods face challenges with thin or tubular structures.

Purpose of the Study:

• To develop an efficient finite element discretization on manifolds for simulating diffusion MRI signals.
• To accurately model diffusion MRI in domains with thin layer or thin tube geometries.
• To reduce computational resources and mesh generation complexity for complex structures.

Main Methods:

• Proposed a finite element discretization method on manifolds.
• Incorporated variable thickness of 3D domains into the weak formulation on manifolds.
• Simulated diffusion MRI signals from extracellular space (thin layer) and neurons (thin tube).

Main Results:

• Achieved good agreement between proposed method's simulated signals and reference signals from standard 3D FEM.
• Demonstrated improved approximation accuracy with increasing diffusion time.
• Significantly reduced simulation time, computational memory, and mesh generation difficulties.

Conclusions:

• The proposed manifold-based finite element method offers an efficient and accurate approach for diffusion MRI signal simulation.
• This method enables cost-effective simulation of complex biological structures, aiding brain imaging research.
• Facilitates a better understanding of diffusion MRI in complex microstructural environments.