open-discussion > problems running on mac
Showing 1-2 of 2 posts
Display:
Results per page:
May 21, 2018  04:05 PM | Scott Sorg
problems running on mac
Hello,
I've been trying to test out DTIprep on some data, but I haven't been able to get it to run successfully. I'm not sure if it is due to errors I'm making or if the data may just be unsuitable for DTIprep (my assumption is that it should be fine). One thing that I noticed is that it doesn't immediately recognize where slicer is located or fsl. I set the paths to both programs (fsl_bet2 and slicer) in the protocol path section. I am running it with the default settings. Last time I ran it (on a simple 6 direction scheme with 1 b0 image) the message: "Single b-value DWI without a b0/baseline!" appeared within the Interlace-wise Check Artifacts section. I installed slicer (4.8.1), FSL (5.0) and DTIprep (DTIPrep1.2.8_MacOSX). Any assistance here would be most appreciated. 
-Scott
May 24, 2018  12:05 PM | Martin Styner
RE: problems running on mac
Hi Scott, the default settings can be a bit aggressive (they will be changed to be less aggressive in future releases), so you probably want to change the parameters (e.g. increasing the thresholds that are related to standard deviations, and displacements/rotations or decreasing those that are related to correlation).

A 6-directional scheme is very sensitive in our QC, as a single remove DWI will fail the whole dataset (as one cannot recover the tensor from only 5 directions). If you look at the txt report in the output folder, you can see which DWIs are kicked out in which stage (you can also see it in the output xml, but that one is much harder to read). Maybe try with a dataset that has more than 6 gradients.

Slicer and FSL are only used for the brain masking step and thus should not matter for the QC of the DWI (only for the brain mask computation and the DTI reconstruction). If that is an issue you may also simply disable the brain masking and do that step outside DTIPrep (as a work around)

Martin
Originally posted by Scott Sorg:
Hello,
I've been trying to test out DTIprep on some data, but I haven't been able to get it to run successfully. I'm not sure if it is due to errors I'm making or if the data may just be unsuitable for DTIprep (my assumption is that it should be fine). One thing that I noticed is that it doesn't immediately recognize where slicer is located or fsl. I set the paths to both programs (fsl_bet2 and slicer) in the protocol path section. I am running it with the default settings. Last time I ran it (on a simple 6 direction scheme with 1 b0 image) the message: "Single b-value DWI without a b0/baseline!" appeared within the Interlace-wise Check Artifacts section. I installed slicer (4.8.1), FSL (5.0) and DTIprep (DTIPrep1.2.8_MacOSX). Any assistance here would be most appreciated. 
-Scott