Head injury or head motion? Assessment and quantification of motion artifacts in diffusion tensor imaging studies

Area of Science:

• Neuroimaging
• Diffusion Tensor Imaging (DTI)

Background:

• The impact of head motion on diffusion imaging metrics like fractional anisotropy (FA) and mean diffusivity (MD) remains unclear.
• Previous simulations suggest motion can bias DTI values, but clinical quantification and evaluation of bias correction methods are lacking.

Purpose of the Study:

• To investigate head motion's effect on FA and MD across common DTI analysis pipelines.
• To assess the efficacy of removing diffusion-weighted images and current motion correction techniques.

Main Methods:

• Analysis of FA and MD in a large cohort of healthy controls using tract-based spatial statistics, voxelwise, and region of interest pipelines.
• Evaluation of motion artifact removal strategies, including image removal and gradient correction.
• Monte Carlo simulations to model the impact of random image removal on DTI metrics.

Main Results:

• Head motion introduced a positive bias in both FA and MD, with a more significant effect on MD.
• This bias was consistent across all three analysis pipelines and persisted despite standard motion correction protocols.
• Removing images with gross artifacts did not alter the motion-DTI scalar relationship; random image removal increased bias and reduced precision.

Conclusions:

• Current DTI processing techniques for motion bias correction appear insufficient.
• Head motion significantly impacts FA and MD, requiring robust quality assurance protocols for DTI studies.
• Quantifying head motion across populations is crucial for reliable DTI analysis.