A multicomponent T2 relaxometry algorithm for myelin water imaging of the brain

Area of Science:

• Magnetic Resonance Imaging
• Biophysics
• Computational Science

Background:

• Multicomponent T2 relaxometry is crucial for applications like myelin water fraction estimation.
• Estimating relaxation parameters and amplitudes becomes challenging with increasing components.
• Non-negative least squares (NNLS) is a common but limited approach.

Purpose of the Study:

• To compare the performance of NNLS with a novel algorithm, Exponential Analysis via System Identification using Steiglitz-McBride (EASIS-M).
• To evaluate algorithm efficiency and accuracy in estimating multicomponent T2 relaxation parameters.
• To assess the utility of these methods for myelin water fraction mapping in vivo.

Main Methods:

• Simulated T2 relaxation data were used to evaluate both algorithms.
• Performance was benchmarked against the Cramér-Rao bound for statistical precision.
• The algorithms were applied to an in vivo human brain multi-echo spin-echo dataset.

Main Results:

• EASIS-M demonstrated superior performance on simulated T2 relaxation data compared to NNLS.
• Application to in vivo brain data yielded a myelin water fraction map with a more concentrated distribution.
• The EASIS-M map exhibited reduced noise levels compared to the NNLS-derived map.

Conclusions:

• EASIS-M offers an efficient, user-parameter-free alternative to NNLS for multicomponent relaxation analysis.
• This algorithm provides a novel method for mapping spatial variations of myelin in the brain.
• The findings suggest EASIS-M is a valuable tool for quantitative MRI analysis.