Data-driven algorithm for myelin water imaging: Probing subvoxel compartmentation based on identification of...

Area of Science:

• Neuroimaging
• Biophysics
• Medical Physics

Background:

• Multicomponent analysis of MRI T2 relaxation time (mcT2) is crucial for estimating myelin content.
• Voxel-based approaches face challenges due to ambiguity in multi-T2 space and low MRI signal-to-noise ratio (SNR).

Purpose of the Study:

• To present a data-driven mcT2 analysis method that leverages global white matter motifs.
• To overcome limitations of traditional voxel-based mcT2 analysis for improved myelin water imaging.

Main Methods:

• Utilized statistical strength of spatially global mcT2 motifs in white matter segments.
• Employed a tailored optimization scheme for deconvolution without prior assumptions on component number.
• Generated voxel-wise myelin water fraction maps.

Main Results:

• Demonstrated excellent fitting accuracy and agreement with ground truth for numerical and physical mcT2 phantoms.
• Provided proof-of-concept in vivo validation using human brain white matter.
• Showed good interscan stability of myelin water fraction values.

Conclusions:

• Studying global tissue motifs prior to voxel-wise mcT2 analysis stabilizes the optimization.
• The novel approach effectively overcomes ambiguity in the T2 space for improved myelin water imaging.
• This method enhances the investigation of microstructural compartmentation in biological tissues.