How to quantitatively compare two FOD reconstructions (subset reconstruction vs. full data)? — metrics & MRtrix commands

1

Hello everyone,

I have two FOD reconstructions from the same dataset:

  • FOD_A: reconstructed from a subset / alternative reconstruction (method X / fewer directions).

  • FOD_B: reconstructed from the complete original dataset (reference).

I want to compare them voxelwise and across the volume to answer questions such as: are the same fiber populations detected, do peak directions match, and how different are the FOD shapes/amplitudes?

What I’m asking

  1. Which quantitative metrics are commonly used to compare two FOD images?

  2. Recommended MRtrix commands / workflow to compute those metrics.

  3. Practical advice for matching peaks between FODs (how to pair peaks reliably when number of peaks differs).

  4. Any pitfalls or visualizations that help interpret differences (e.g. fixel maps, histograms).

Capture d'écran 2025-10-22 130258
Capture d'écran 2025-10-22 1302581163×888 180 KB

Thanks in advance!