Fast Realistic MRI Simulations Based on Generalized Multi-Pool Exchange Tissue Model.

Area of Science:

• Medical Imaging
• Computational Biology
• Biophysics

Background:

• Accurate magnetic resonance imaging (MRI) simulations are crucial for developing advanced quantitative methods.
• Existing simulators often use simplified tissue models, limiting their applicability to complex biological systems.
• Computational demands of realistic MRI simulations hinder widespread adoption on standard hardware.

Purpose of the Study:

• To introduce MRiLab, a novel, comprehensive MRI simulation platform.
• To enable large-scale, realistic MRI simulations on PCs with graphical processing units (GPUs).
• To facilitate the assessment of advanced quantitative MRI methods for inferring sub-voxel tissue microstructure.

Main Methods:

• Developed MRiLab, a simulator integrating realistic tissue modeling with numerical virtualization of MRI systems.
• Employed a generalized tissue model with multiple exchanging proton pools for flexible microstructure representation.
• Utilized parallelized GPU execution to achieve high computational speeds for complex 3D models.

Main Results:

• MRiLab successfully simulated diverse voxel compositions and demonstrated the limitations of simplified models.
• GPU acceleration provided approximately a 200x increase in computational speed compared to standard CPUs.
• The simulator accommodates a broad range of MRI approaches, including advanced quantitative techniques.

Conclusions:

• MRiLab offers a powerful, open-source, and extensible environment for MRI research.
• The platform streamlines the development and validation of novel MRI methods.
• MRiLab enhances the quantitative inference of tissue composition and microstructure from MRI data.