Reducing variability in along-tract analysis with diffusion profile realignment.

Area of Science:

• Neuroimaging
• Diffusion Magnetic Resonance Imaging (dMRI)

Background:

• Diffusion weighted magnetic resonance imaging (dMRI) tractography reconstructs white matter pathways non-invasively.
• Analyzing dMRI measures along white matter tracts offers greater specificity than region-of-interest or averaged measurements.
• Group analyses using along-tract strategies require precise point correspondence across subjects, often achieved through resampling.

Purpose of the Study:

• To address potential misalignment issues in dMRI tractography analysis, even after resampling.
• To introduce a novel method, diffusion profile realignment (DPR), for improving streamline correspondence in 1D space.
• To enhance the specificity of group analyses in neuroimaging studies.

Main Methods:

• Developed diffusion profile realignment (DPR) to realign representative streamlines in a 1D space.
• Validated DPR using synthetic datasets to assess its impact on diffusion measures.
• Tested DPR on 100 in vivo human connectome project datasets, simulating anatomical changes.

Main Results:

• DPR reduced the coefficient of variation for mean diffusivity, fractional anisotropy, and apparent fiber density on synthetic data compared to unaligned data.
• Regional changes in diffusion measures were identified using DPR on simulated altered in vivo datasets.
• DPR preserved previously detected differences while revealing new regional effects.

Conclusions:

• Diffusion profile realignment (DPR) is a novel strategy to correct for misalignment in dMRI tractography.
• DPR improves the detection of subtle, localized changes in white matter microstructure.
• This method has the potential to enhance specificity in longitudinal population studies and advanced neuroimaging analyses.