Modeling a Shared Reality of Tractography through Varied Structural Imaging

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Medical Image Analysis

Background:

• Diffusion MRI (dMRI) is standard for white matter tractography, but its specificity is debated.
• Understanding if dMRI captures unique diffusion properties or general structural information is crucial.

Purpose of the Study:

• To investigate if white matter pathways can be extracted from FLAIR images without dMRI.
• To clarify what dMRI specifically captures in white matter architecture.
• To explore structural-functional relationships of anatomical bundles.

Main Methods:

• Developed a framework to extract white matter pathways from FLAIR images.
• Employed a teacher-student model: dMRI-trained teacher guides FLAIR-based tractogram creation by a student model.
• Validated using 14 subjects from BLSA, with 9 withheld for robustness checks.

Main Results:

• FLAIR-template generated streamlines showed statistically similar performance to gold standard diffusion streamlines.
• Compared to other non-diffusion methods, the proposed approach demonstrated comparable results.
• The method captures prior probabilities of tractography without diffusion data.

Conclusions:

• Tractography patterns may originate from a shared latent space of structural information, not exclusive to dMRI.
• FLAIR imaging can potentially capture essential white matter architecture information.
• This opens avenues for tractography using non-diffusion-based imaging sequences.