The effect of feature image on sensitivity of the statistical analysis in the pipeline of a tractography atlas-based analysis

Area of Science:

• Neuroimaging
• Diffusion Tensor Imaging (DTI)
• White Matter Analysis

Background:

• Tractography atlas-based analysis (TABS) requires accurate registration of diffusion tensor images (DTI) from native to standard space.
• The choice of feature image for calculating deformation fields impacts TABS accuracy and sensitivity.

Purpose of the Study:

• To evaluate the impact of different feature images (FA, T1, hFA) on the accuracy and sensitivity of TABS.
• To determine the optimal feature image for robust white matter property analysis using TABS.

Main Methods:

• Calculated deformation fields using fractional anisotropy (FA), T1-weighted, and Hessian-based FA (hFA) images for DTI registration.
• Performed voxel-wise statistical analysis along fiber tracts using the TABS method.
• Investigated gender differences in diffusion measurements along the corpus callosum.

Main Results:

• The tensor template showed highest consistency with individual subjects when using FA images for deformation field calculation.
• TABS analysis using FA images as features detected significant gender differences in the right corpus callosum.
• FA images provided more accurate tensor information and analysis results compared to T1 or hFA images.

Conclusions:

• Fractional anisotropy (FA) images are the optimal feature images for accurate and sensitive tractography atlas-based analysis (TABS).
• Using FA images enhances the reliability of white matter property analysis and the detection of subtle group differences.