How and how not to correct for CSF-contamination in diffusion MRI

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Gerontology

Background:

• Diffusion MRI is crucial for studying white matter changes in development and disease.
• Aging causes brain volume loss, leading to cerebrospinal fluid (CSF) partial volume artifacts in diffusion MRI metrics.
• These artifacts are particularly problematic in brain regions near ventricles, like the fornix.

Purpose of the Study:

• To model the impact of CSF contamination on diffusion MRI metrics.
• To evaluate post-acquisition correction strategies for CSF contamination: whole-brain volume control versus voxel-wise Free Water Elimination (FWE).

Main Methods:

• Simulated CSF contamination effects on diffusion MRI metrics.
• Applied tract-specific and voxel-based (TBSS) analyses to DT-MRI data from 39 older adults (53-93 years).
• Compared whole-brain volume correction with FWE-based voxel-by-voxel correction using the fornix as a model structure.

Main Results:

• Age-related decreases in whole brain and fornix tissue volume were observed.
• Diffusivity metrics (MD, AD, RD) were more susceptible to CSF contamination than fractional anisotropy (FA).
• FWE correction eliminated spurious age-related correlations with diffusivity metrics, while whole-brain correction had minimal effect.

Conclusions:

• Voxel-by-voxel correction for CSF contamination is essential for accurate interpretation of diffusion MRI findings in aging and disease.
• The Free Water Elimination (FWE) approach is a robust method for mitigating CSF partial volume effects.
• Findings have significant implications for understanding white matter changes in aging and neurological conditions.