Test-retest reliability of diffusion tensor imaging scalars in 5-year-olds

Area of Science:

• Neuroimaging
• Developmental Neuroscience
• Biomedical Engineering

Background:

• Diffusion Tensor Imaging (DTI) offers insights into brain microstructure.
• DTI data reliability is crucial, especially in pediatric populations due to potential artifacts.
• Understanding the repeatability of DTI scalars is essential for longitudinal and cross-sectional studies.

Purpose of the Study:

• To assess the intrascan test-retest repeatability of four DTI scalars: fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD).
• To compare the reliability of these scalars between two preprocessing pipelines: volume censoring and outlier replacement.
• To evaluate DTI scalar repeatability in 5-year-old children.

Main Methods:

• Investigated 67 five-year-old children using DTI with 96 diffusion encoding directions.
• Applied two preprocessing pipelines: volume censoring and outlier replacement.
• Utilized region of interest (ROI) and voxelwise analyses with tract-based spatial statistics, assessing repeatability via Intraclass Correlation Coefficient (ICC(3,1)) and variability.

Main Results:

• Both preprocessing pipelines demonstrated good to excellent reliability (ICC(3,1) > 0.75) for most ROIs with low variability (<10%).
• Voxelwise analysis showed higher ICC(3,1) for FA and RD compared to AD and MD, with consistent low variability (<12%).

Conclusions:

• The study indicates high intrascan repeatability for DTI scalars in pediatric populations.
• Findings support the use of DTI data in future pediatric cross-sectional and longitudinal research.
• The results provide confidence in the reliability of DTI metrics for studying early brain development.