# MRTrix3 on DTI data

**URL:** <https://community.mrtrix.org/t/mrtrix3-on-dti-data/970>\
**Category:** Uncategorized\
**Tags:** tractography, dwi2response\
**Created:** [June 12, 2017, 2:40pm UTC](https://community.mrtrix.org/t/mrtrix3-on-dti-data/970 "2017-06-12T14:40:36Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Vasudev](https://community.mrtrix.org/letter_avatar_proxy/v4/letter/v/a587f6/32.png) [@Vasudev](https://community.mrtrix.org/u/Vasudev)\
**Post date:** [June 12, 2017, 2:40pm UTC](https://community.mrtrix.org/t/mrtrix3-on-dti-data/970/1 "2017-06-12T14:40:37Z")

</div>

Dear all,

I have 50 subject DTI data set ( 30 directions ) , I would like to compute tractography analysis like to probabilistic tractography analysis , Could some please let me know if command line arguments of MRTrix can be applied of DTI dataset instead of DWI ?.

thanks  
Vasudev

---

<div class="post-metadata">

**Author:** ![mrburnst](https://community.mrtrix.org/letter_avatar_proxy/v4/letter/m/ed8c4c/32.png) [@mrburnst](https://community.mrtrix.org/u/mrburnst)\
**Post date:** [June 13, 2017, 9:01am UTC](https://community.mrtrix.org/t/mrtrix3-on-dti-data/970/2 "2017-06-13T09:01:03Z")

</div>

there are two possibilities here

1. you have DTI data, meaning you have a volume of tensors (prob 6 or maybe 9 channels in each voxel, because you have a 3x3 symmetric tensor in each voxel
2. you actually have DWI data, eg. 30 directions/channels in each voxel, plus maybe/hopefully gradient tables to know in which direction each channel points

in case 1, I don’t think mrtrix can help you, because it does its own tensor estimation during tractography. In case 2 however, you can run `tckgen` with a tensor-fitting algorithm, as described in [https://mrtrix.readthedocs.io/en/latest/reference/commands/tckgen.html](https://mrtrix.readthedocs.io/en/latest/reference/commands/tckgen.html)