Computing the orientational-average of diffusion-weighted MRI signals: a comparison of different techniques

Area of Science:

• Diffusion Magnetic Resonance Imaging (dMRI)
• Quantitative MRI
• Image Analysis

Background:

• Orientationally-averaged diffusion-weighted signal is crucial for dMRI applications.
• Standard averaging methods assume uniform spherical gradient sampling, which is not always achieved.
• Non-uniform sampling poses challenges for accurate signal estimation.

Purpose of the Study:

• To compare the accuracy of different orientationally-averaged diffusion MRI signal estimation methods.
• To evaluate the performance of these methods under varying signal-to-noise ratio (SNR) and sampling densities.
• To assess the robustness of Mean Apparent Propagator MRI (MAP-MRI) against conventional averaging techniques.

Main Methods:

• Simulation of diffusion MRI signal averaging techniques, including weighted averaging, spherical harmonic representation, and MAP-MRI.
• Comparison of methods under diverse SNR levels and sampling configurations.
• Validation of simulation findings using in vivo dMRI data.

Main Results:

• All methods yield comparable results with dense, isotropic sampling.
• MAP-MRI demonstrates significantly higher accuracy with reduced SNR and sampling density.
• MAP-MRI estimates show slightly increased bias at higher b-values but maintain superior overall accuracy.
• Orientationally-averaged signals are predominantly Gaussian distributed.

Conclusions:

• MAP-MRI offers a more robust and accurate approach for estimating the orientationally-averaged diffusion MRI signal, particularly in challenging acquisition scenarios.
• The findings support the use of MAP-MRI for improved quantitative analysis in diffusion MRI.
• The Gaussian nature of averaged signals provides insights into underlying diffusion properties.