# Inquiry regarding the extraction of Diffusion Tensor metrics

**URL:** https://community.mrtrix.org/t/inquiry-regarding-the-extraction-of-diffusion-tensor-metrics/8731
**Category:** Uncategorized
**Created:** [April 1, 2026, 8:05am UTC](https://community.mrtrix.org/t/inquiry-regarding-the-extraction-of-diffusion-tensor-metrics/8731 "2026-04-01T08:05:09Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![Dr.quocvan](https://community.mrtrix.org/user_avatar/community.mrtrix.org/dr.quocvan/32/5435_2.png) [@Dr.quocvan](https://community.mrtrix.org/u/Dr.quocvan)
#### Post date: [April 1, 2026, 8:05am UTC](https://community.mrtrix.org/t/inquiry-regarding-the-extraction-of-diffusion-tensor-metrics/8731/1 "2026-04-01T08:05:09Z")

</div>

Dear Expert,  
I have a technical question regarding the processing of DTI data for anatomical consistency.

Assuming I have already obtained the unique 3x3 Diffusion Tensor matrix (D) solved via Least Squares from 64-direction DWI signals. Currently, the standard components (D\_xx, D\_yy, D\_zz are reported relative to the MRI scanner’s laboratory frame, which leads to inconsistent values when a subject’s head orientation varies.

If I have pre-defined a fixed anatomical coordinate system—specifically the unit vectors v\_x, v\_y, v\_z of a known fiber bundle (e.g., projection fibers)—can I directly extract a new set of metrics (D’\_xx, D’\_yy, D’\_zz) using the quadratic form d = v^T .D.v for each respective direction?