# Converting FD, FC and FDC from fixel to voxel space

**URL:** https://community.mrtrix.org/t/converting-fd-fc-and-fdc-from-fixel-to-voxel-space/8858
**Category:** Uncategorized
**Created:** [October 5, 2026, 7:45pm UTC](https://community.mrtrix.org/t/converting-fd-fc-and-fdc-from-fixel-to-voxel-space/8858 "2026-10-05T19:45:00Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![rpraveen](https://community.mrtrix.org/letter_avatar_proxy/v4/letter/r/cc9497/32.png) [@rpraveen](https://community.mrtrix.org/u/rpraveen)
#### Post date: [October 5, 2026, 7:45pm UTC](https://community.mrtrix.org/t/converting-fd-fc-and-fdc-from-fixel-to-voxel-space/8858/1 "2026-10-05T19:45:00Z")

</div>

Hi MRtrix3 Experts,

I’m trying to convert FD, FC and FDC from an FBA analysis into voxel-wise maps so that they can eventually be compared with other voxel-based diffusion measures.

I found the previous discussions on fixel2voxel, and my understanding is that voxel-wise FD can be obtained by summing the FD of multiple fixels within a voxel.

For a test subject, I tried:

fixel2voxel fd.mif sum FD\_voxel.mif

This gives me FD in the population-template voxel space.

What I’m mainly unsure about is what should be done after this if I want the final maps in the same standard/MNI space as my other diffusion measures.

Would it be appropriate to:

template fixels  
→ fixel2voxel  
→ template voxel map  
→ warp back to subject space  
→ resample to the subject’s native DWI/FA grid  
→ apply the existing FA-derived TBSS transformation

Or, since the FBA measures are already in the population-template space, would it be better to register the population template to MNI and transform the voxelized maps directly to MNI space?

I also wanted to clarify two related points:

For FC, is an FD-weighted mean of log(FC) the recommended way to obtain one value per voxel?

For FDC, what is the recommended fixel-to-voxel operation when there are multiple fixels within a voxel?

And finally, is it reasonable to apply TBSS skeleton projection to voxelized FD/FC/FDC, or would that change the interpretation of these FBA-derived measures in a way that makes this approach inappropriate?

I wanted to confirm the correct approach before proceeding further.

Thanks!