Header

Tortoise 3.1.4 EstimateTensorNLLSRESTORE - after DRBUDDI

Submitted by Chris Panks on
Category

Greetings,

I was looking at the previous post about generating the mask nifti. The line ' ExtractImage -i  subjID_DMC.nii -v 0 ' made me think that this was the process to try if DRBUDDI had not been run, because the DMC.nii is a DIFFPREP output. If that is correct, what is the process for generating the mask nifti from DRBUDDI output?

Thank you,

Chris Panks
MacDonald Lab
University of Washington
Seattle, Washington

 

Submitted by nayaka on Wed, 05/08/2019 - 16:27

Permalink

Hi Chris,

You can provide any nifti file as input to the ExtractImage command. In this specific case, it would be the 'DRBUDDI_ouput.nii' file  (using a generic example name here) that is generated in the DRBUDDI_proc folder.

ExtractImage    -i   <full_path_to_input_nifti_file>      -v   <volume_number_you_want_to_extract>  and then you can mask that extracted volume.  

  Alternate solution for masking DRBUDDI output: mask the structural.nii that is in the DRBUDDI_proc folder. DRBUDDI output is in the structural.nii space and with using the structural file for masking, you can bypass the ExractImage step.  

 

Thanks,

Amritha

Submitted by Chris Panks on Wed, 05/08/2019 - 17:04

In reply to by nayaka

Permalink

Hi Amritha,

Thanks for the prompt reply! So this is what you're saying?

bet2  ~/DRBUDDI_proc/structural.nii mask -m -f 0.3

and then

EstimateTensorNLLSRESTORE -i ~/DRBUDDI_proc/id_DRBUDDI.list -m <output of bet2>.nii.gz

 

Chris

bet2  ~/DRBUDDI_proc/structural.nii mask -m -f 0.3 does appear to make a mask successfully.

However, 'EstimateTensorNLLSRESTORE -i ~/DRBUDDI_proc/id_DRBUDDI.list -m <output of bet2>.nii.gz' takes less than 15 minutes to run, and ComputeFAMap fails when using the output.

Any help would be appreciated, thanks.

Submitted by nayaka on Tue, 05/21/2019 - 18:30

Permalink

Hi Chris,

sorry for the delay as we were away at a conference. Can you please try computing the tensor using EstimateTensorNLLS and then try ComputeFAMap to see what result you get? This test is mainly to separate if its a problem with EstimatetensorNLLSRESTORE or ComputeFAMap.

Also, have you downloaded TORTOISEV3.1.4 that was released on 05/06 and try the EstimateTensorNLLSRESTORE from the current version? we recently re-released TORTOISEv3.1.4 (an email was sent with the bug fixes), where the beta version of the RESTORE command was been replaced with a more stable version. Please use the command from this new re-released version and see if you get better results.

Thanks,

Amritha