arash-n / gbss Goto Github PK

Hi Arash,

I'd like to ask you what's the aim of the last line in gbss_3_fill.sh (fslmaths bin_mean.nii.gz -thr 0.995 mask). The reason is that I'm getting a nice skeleton in bin_mean.nii.gz but it gets very much disconnected when the threshold 0.995 is applied. Is it possible to lower that threshold to e.g. 0.95?

Thank you!

Dear sirs,
Thanks for your GBSS code, Now I meet a problem about two commands missing Atropos -d 3 -a. gbss_1_reg.sh (line 166: can not find command Atropos).And gbss_1_reg.sh (line 204 ImageMath command) missing.

FSL version5.0.
ANTs version is the latest.

Thanks for your help.

I think it would be nice to include some basic checks for the FSL and ANTS tools that GBSS depends on, and if they aren't found in the user's environment, we can spit out a "friendly" error message rather than just crash later on, cryptically. I've done this very basically in the my PR #2 by adding the check for FSLDIR. We could add specific checks for ANTS and other tools (e.g. fslmaths, etc..).

For instance,

command -v ANTS >/dev/null 2>&1 || { echo "ANTS required but not found. Aborting"; exit 1 }

Image Exception : #63 :: No image files match: /noddi_gbss/input/*GM*
terminate called after throwing an instance of 'armawrap::AWException'
  what():  No image files match: /noddi_gbss/input/*GM*
/noddi_gbss/GBSS/NODDI/gbss_2_skel.sh: line 110: 14432 Aborted                 (core dumped) fslmerge -t all_GM ${tmp_folder}/*GM*

I was wondering if the GM fraction maps should auto generate or we need to manually provide it?

Hi, after reading your original paper, and I tried to try your pipeline, but I found that the tutorial in the github is not so clear, I created the required output structure for the NODDI maps, and run the first command which is Ok, and then cd output_directory, and run the second script : gbss_2_skel.sh ., which gave me errors, do you have a better explanation for this different steps, because I found it very hard to understand.

Thanks in advance

Hao

Hello, Dear Expert,

As mentioned in #11, I noticed that you suggested to check the intermediate templates are saved for QC, because sometimes the template drifts partially out of the field of view. If that occurs, the option -y 0 of antsMultivariateTemplateConstruction2.sh can be used; however, GBSS uses an older version of ANTs (buildtemplateparallel.sh) . I look up the usage of buildtemplateparallel.sh, However, buildtemplateparallel.sh doesn't have the option -y 0 to avoid template drifts in translation or orientation. How to deal with template drifts (the frontal part of brain was lack in D1_template.nii.gz) in GBSS using buildtemplateparallel.sh?

Thank you very much!

Jay Chang

Hi Jon @pipitone

I was wondering if you can take a look at the new code added to the GBSS:

GBSS/NODDI/native/gbss_native.sh

It seems that the while loop is not working for some reason the j and k remain 1 instead of increasing by 1 each cycle.

Cheers,
Arash

Dear Arash,

we think, this would be great complement to the TBSS analysis on our data. However, I have a few questions an would like to kindly ask you for your advices.

First, I have more general question. With our dataset we aim to the DKI analysis. Is that possible to use your scripts for "GBSS-DKI" just by changing ODI and fIC parameters for DKI ones? Or do you see any reason, why this would fail?

If it would be OK, then I would have a question about the buildtemplateparallel.sh function (ANTs) in gbss_reg_1 script. In this issues forum someone wrote the function is outdated. You suggested to use the antsMultivariateTemplateConstruction2.sh with adjusted flags. I am not realy familiar with ANTs and differences between these two functions, could you please be more specific about changes needed?

Thank you for your advices.
Alzbeta

Hi Arash,

do you think it would be meaningful to use an input threshold e.g. 0.45 in gbss_2_skel.sh? The reason for using such a low value, is because otherwise the skeleton I get is very disconnected. In addition to that, I'm also using -thr 0.025 (instead of 0.05) in the following lines.

fslmaths fIC_non_lesion_bin_s_2 -thr 0.025 -bin -mul all_lesion.nii.gz -mul fIC_filler -add fIC_non_lesion all_fIC_filled
fslmaths ODI_non_lesion_bin_s_2 -thr 0.025 -bin -mul all_lesion.nii.gz -mul ODI_filler -add ODI_non_lesion all_ODI_filled

and increased -s 2 to -s 3

Would you please, have a look at the attached skeleton? I guess the reason I have troubles to get a good skeleton is the large voxel size (2.5 mm iso).

mask_thr045.nii.gz

Thank you very much in advance.

Hi Arash,
when we run the step of buildtemplateparallel.sh, we found that D1_template created could not match with the MNI_template, are there any parameters should I adjust?
looking forward to your reply
Yours
Hui Hong

Hi, I've just noticed that you use buildtemplateparallel.sh from ANTs to create the template. However, in ANTs they say that they'd rather recommend to use antsMultivariateTemplateConstruction2.sh instead of buildtemplateparallel.sh in general.
Do you think it would also work in the framework of your GBSS?
Thank you!

Hi Arash,
the threshold in gbss_2_skel.sh, line 139:
fslmaths all_fIC_skeletonised -mul mean_GM_skeleton_mask_general -thr 0.65 -bin all_lesion_fIC

does it refer to the same threshold defined in thresh (line 54)? if I choose another threshold, do I need to also change the one in line 139?

Thanks!

• mkdir /root/test/FA/D1
• cp /root/test/FA/FA/Sub001_pseudoT1.nii.gz /root/test/FA/D1/
• cd /root/test/FA/D1
• '[' 1 == 1 ']'
• echo 'OPTION 1: Creating Template Image'
  OPTION 1: Creating Template Image
• buildtemplateparallel.sh -d 3 -j 1 -o D1_ -n 0 -s MI -i 4 -m 30x50x20 -t GR -z /usr/share/fsl/5.0/data/standard/MNI152_T1_1mm.nii.gz Sub001_pseudoT1.nii.gz
  do you have qsub? if not, then choose another c option ... if so, then check where the qsub alias points ...
• '[' 1 == 0 ']'
• cd /root/test/FA/FA
• D1_folder=FA/D1
• ref=/root/test/FA/D1/D1_template.nii.gz
• for FAs in *FA.nii.gz
  ++ echo Sub001_FA.nii.gz
  ++ cut -f1 -d
• a=Sub001
• fsl_sub antsApplyTransforms -i Sub001_GM_frac.nii.gz -d 3 -e 0 -n NearestNeighbor -r /root/test/FA/D1/D1_template.nii.gz -t '/root/test/FA/D1/*Sub001_pseudoT1Warp.nii.gz' -t '/root/test/FA/D1/*Sub001_pseudoT1Affine.txt' -o /root/test/tmpspace/Sub001_GM.nii.gz --float
  file /root/test/FA/D1/D1_template.nii.gz does not exist .

Hi,
Trouble you again, from the debug info, The D1 folder is created automatically.
Where is D1_template.nii.gz ?

Best regards,
wang

Hi Arash,

do you know if it would make a difference to use NODDI AMICO instead of using the NODDI matlab package in GBSS?

Best

Hellow:
Thank for your work, I have a question when you create the template by using fsl_sub, you have a -r falg to point to the reference image(template), but I checked out the fsl fsl_sub wiki, there is no this flag, in my linux and mac, there are not neither, my fsl_sub versin is V1.1, can you please let me konw your fsl_sub version????

Thanks in advance

Hao

GM_skel_0.65.nii.gz

all_bin.nii.gz
all_fIC_filled.nii.gz
all_ODI_filled.nii.gz
all_fIC_skeletonised.nii.gz
all_fIC_filled.nii.gz
all_ODI_filled.nii.gz
all_ODI_skeletonised.nii.gz