Performance of orientation distribution function-fingerprinting with a biophysical multicompartment diffusion model.

Area of Science:

• Diffusion MRI
• Neuroimaging
• Biophysics

Background:

• Standard Orientation Distribution Function (ODF) peak finding methods struggle with reconstructing crossing fibers at angles below 40 degrees.
• This limitation leads to significant errors in diffusion MRI tractography, hindering accurate brain white matter mapping.

Purpose of the Study:

• To introduce a modified Orientation Distribution Function-Fingerprinting (ODF-FP) approach using a biophysical multicompartment diffusion model.
• To overcome the limitations of existing methods in reconstructing complex fiber architectures, particularly at shallow crossing angles.

Main Methods:

• A randomized mechanism was developed to generate a multidimensional ODF-dictionary covering biologically plausible ranges of intra- and extra-axonal diffusivities and fraction volumes.
• The modified ODF-FP approach was evaluated using numerical simulations and in vivo diffusion images of major brain fascicles.

Main Results:

• The modified ODF-FP successfully identified crossing fibers at angles as shallow as 10 degrees in simulated data.
• In vivo, the approach achieved 56% true positives in fiber direction determination, yielding superior reconstruction of major tracts like the pyramidal tracts, arcuate fasciculus, and optic radiations.
• Estimated microstructure parameters in the corpus callosum aligned with literature values.

Conclusions:

• The modified ODF-FP method significantly outperforms conventional fiber reconstruction techniques for shallow fiber crossings.
• This advancement leads to improved deterministic tractography outcomes for major white matter fascicles.
• The proposed approach also facilitates the linearization of microstructure parameter fitting.