# Issues reproducing results with freesurfer based ACT

**URL:** <https://community.mrtrix.org/t/issues-reproducing-results-with-freesurfer-based-act/5868>\
**Category:** Uncategorized\
**Created:** [July 8, 2022, 2:53pm UTC](https://community.mrtrix.org/t/issues-reproducing-results-with-freesurfer-based-act/5868 "2022-07-08T14:53:06Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![CallowBrainProject](https://community.mrtrix.org/user_avatar/community.mrtrix.org/callowbrainproject/32/2370_2.png) [@CallowBrainProject](https://community.mrtrix.org/u/CallowBrainProject)\
**Post date:** [July 8, 2022, 2:53pm UTC](https://community.mrtrix.org/t/issues-reproducing-results-with-freesurfer-based-act/5868/1 "2022-07-08T14:53:06Z")

</div>

Hello,

I previously rant ACT using the following commands however, when asked to clear up space on the linux system I use I removed the 10 million streamline tract files from the previous analysis. However, while I kept the connectome.csv output files, a few months later I needed to replicate/rerun tckgen and I accidentally overrode the assignment.csv and weight.csv file from the previous run. Meaning I can’t reproduce the tractogram from the previous analysis. This wouldn’t be a huge problem if the second time around the results were similar to the first analysis, however, my connectome.csv files appear to be quite different despite (I believe, a near-identical pipeline). Additionally, I know I have not altered the wmfod.mif and aparc+aseg.mgz file has not been modified since the first iteration. Additionally,I noticed I used the nodes\_old.mif file instead of the nodes\_fixSFM\_old.mif file for tck2connectome the first time around so I kept that consistent this time around.  
Attached are images of the new and old connectome (couldn’t fit the whole thing in an image unfortunately)

 ![New_connectome](https://community.mrtrix.org/uploads/default/original/2X/8/83850a947166cc0a8dba324df7cb785a96a632eb.png)  
 ![Old_connectome](https://community.mrtrix.org/uploads/default/original/2X/3/351c0f719583708ada176977aaff7de200c0bbe8.png)

Below is the script I ran the second time and thought it was essentially identical to what I ran the first time. Just wondering if anyone might have an idea of what could be causing the differences or if this is just to be expected with -dynamic\_seed. For reference, I think this is something else since the global and local efficiency measures I am getting are pretty drastically different between the two iterations.

First, I took the aparc+aseg file from freesurfer into diffusion space.

Then run 5ttgen with the freesurfer option

Convert and fix labels

and then run ACT with dynamic seeding, probabilistic tractography, and SIFT2

**mri\_vol2vol --mov lowb\_brain\_corrected.nii.gz --targ /data/bswift-1/dcallow/aging/freesurfer/${subj}/mri/aparc+aseg.mgz --inv --interp nearest --o cort\_subcort.nii.gz --reg register.dat**

**5ttgen -sgm\_amyg\_hipp freesurfer cort\_subcort.nii.gz 5TT.mif**  
**labelconvert cort\_subcort.nii.gz FreeSurferColorLUT.txt fs\_default.txt nodes\_old.mif**  
**labelsgmfix nodes\_old.mif cort\_subcort.nii.gz fs\_default.txt nodes\_fixSFM\_old.mif -premasked**

**tckgen -nthreads 4 -act 5TT.mif -backtrack -crop\_at\_gmwmi -seed\_dynamic wmfod.mif -maxlength 250 -select 10M -step 1 -angle 45 -cutoff 0.06 wmfod.mif 10M.tck**

**tcksift2 -nthreads 4 10M.tck wmfod.mif -act 5TT.mif weights.csv**  
**tck2connectome -force -nthreads 4 10M.tck nodes\_old.mif connectome\_old.csv -tck\_weights\_in weights.csv -out\_assignments assignments\_old.txt -symmetric -zero\_diagonal**

Thank you for any help or any suggestions on what I could look into/what could be leading to such different results!