Microstructure-Informed Myelin Mapping (MIMM) from routine multi-echo gradient echo data using multiscale physics...

Area of Science:

• Neuroimaging
• Biophysics
• Medical Physics

Background:

• Myelin quantification is crucial for studying demyelination in neurodegenerative diseases.
• Existing methods may have limitations in accuracy or require specialized MRI sequences.

Purpose of the Study:

• To introduce and validate Microstructure-Informed Myelin Mapping (MIMM), a novel noninvasive MRI method.
• To quantify myelin volume fraction (MVF) using a routine multi-gradient echo sequence (mGRE) and a biophysical model.

Main Methods:

• Developed a multiscale biophysical signal model incorporating myelin (Hollow Cylinder Fiber Model) and iron effects.
• Generated a dictionary of mGRE signal evolution and susceptibility using finite elements.
• Employed stochastic matching pursuit for voxel-wise MVF and iron susceptibility mapping, evaluated with and without diffusion tensor imaging (DTI) or atlas-based fiber orientation.

Main Results:

• DTI- and atlas-guided MIMM effectively reduced MVF overestimation in white matter tracts compared to basic MIMM.
• MIMM demonstrated good agreement with T2-relaxometry myelin water fraction (MWF) reference maps.
• MIMM MVF maps corrected myelin overestimation in the globus pallidus seen with three-pool complex fitting (3PCF) MWF.

Conclusions:

• MIMM processing of mGRE data provides accurate MVF maps.
• This method enables myelin quantification from routine clinical MRI scans, eliminating the need for specialized sequences.