Comparison of mouse brain DTI maps using K-space average, image-space average, or no average approach

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Medical Physics

Background:

• Diffusion Tensor Imaging (DTI) is crucial for visualizing white matter tracts.
• Improving image quality in DTI is essential for accurate quantification.
• Signal averaging techniques (k-space and image space) are used to enhance signal-to-noise ratio (SNR) in diffusion-weighted images (DWIs).

Purpose of the Study:

• To compare the impact of k-space averaging (k-avg), image space averaging (m-avg), and no averaging (no-avg) on DTI quantification.
• To evaluate the performance of these averaging methods across different brain tissues (white matter, gray matter, ventricles).
• To determine the optimal signal averaging strategy for in vivo DTI brain imaging.

Main Methods:

• In vivo DTI data were acquired from five normal mice.
• Simulated noisy DWIs with varying SNRs (5-30) were generated.
• DTI indices (RA, TR, λ║, λ⊥) were calculated for k-avg, m-avg, and no-avg datasets.
• Quantifications were compared in the corpus callosum, cortex, and ventricles.

Main Results:

• Both k-avg and m-avg improved DWI SNR with no significant difference between them.
• K-avg resulted in lower relative anisotropy (RA) in white matter and higher RA in gray matter compared to m-avg and no-avg.
• Image space averaging (m-avg) and no averaging (no-avg) yielded similar DTI quantifications.

Conclusions:

• K-space averaging is less suitable for DTI brain imaging due to its adverse effects on relative anisotropy measurements.
• Image space averaging or no averaging are preferred methods for DTI quantification, providing more reliable results.
• The choice of signal averaging technique significantly impacts DTI-derived metrics, particularly in different brain tissue types.