Hi,

How are you running fslreorient2std ?

If you only use one argument (e.g. "fslreorient2std myimage") then this does nothing to the image.  You need to run it with two arguments (e.g. "fslreorient2std myimage myreorientedimage") and then use the outputs for the VBM analysis.

If you are already doing that then can you double-check that the segmentation quality is good and also see what you get if you run FLIRT using the GM segmentation to the GM template that you have so far.

All the best,

            Mark

From: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of Rosalia Dacosta Aguayo <[log in to unmask]>
Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
Date: Tuesday, 30 August 2016 at 17:16
To: "[log in to unmask]" <[log in to unmask]>
Subject: Re: [FSL] fslvbm_2_template orientation problem

Yes please,

It would be a good idea to look carefully at the link I have sent it to you and then try other options if it is the case with your images...but, be careful,  if Mark Jenkinson saw some of your bad images, as I have seen in your previous emails,  and he tried fslorient is because it was the right approximation to your problem. So, I if you could try just a flirt with 6dof after doing fslorient...I think it would be a good option...

Kind regards,

Rosalia

2016-08-30 18:08 GMT+02:00 Amanda Rodrigue <[log in to unmask]>:

Also the most common misorientation is that they are upside down, not just simply flipped along the vertical axis. I will check the link provided to see if there are any other options but this doesn't seem to be a problem with the files themselves since the subjects that are a problem now were not a problem in the previous iteration.

On Tue, Aug 30, 2016 at 11:55 AM, Amanda Rodrigue <[log in to unmask]> wrote:

When I first encountered this problem, I  viewed the images in fsl view and all labels are correct and the images are oriented correctly. The labels are also correct in all the previous steps of the VBM pipeline.  I have already submitted data to someone and the two suggestions were the ones I have already attempted.  This was the response to my data submission.

"Hi Amanda,

I’ve been able to reproduce the issue locally. A fix that worked for the two “bad” files was to run fslreorient2std on the input structural ( and manually regenerate their registrations ) which gave good registrations for each step ( non-flipped )! It might be worth running fslreorient2std on all the “bad” structurals ( and maybe even the “good” ones, since it shouldn’t make anything worse ) and then re-running fslvbm_2_template from scratch - if you try this and the new merged output still contains flipped images, you may need to manually register the few bad files, and then regenerate the final template."

On Tue, Aug 30, 2016 at 11:15 AM, Rosalia Dacosta Aguayo <[log in to unmask]> wrote:

Hi Amanda,

I know perfectly what is waiting until a month to get a solution to your problem. Do not worry. The solution to your problem depends whether your images are just flipped, or have any other problems: it would be nice if you could open the problematic ones with fslview and see if they conserve the labels (A-P, R-L, I-F). Also, did you use any marker during the scanner in order to have information about R-L orientation, this could also help. Finally,  fslreorient2std is not the only option, because the option depends on the real problem you have with your images. I would recommend you read carefully this link: http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Orientation%20Explained?highlight=%28fslreorient2std%29 and try to see which is the best option for your images.

Also, if you still can not cope with those images it should be nice you share some of them with a link to any place where you save them and so others can see where it could be the problem.

Hope this helps,

Rosalia

2016-08-30 16:59 GMT+02:00 Amanda Rodrigue <[log in to unmask]>:

Hello again, I have tried running fslreorient2std on all my images and ran the whole analysis from scratch, still getting flipped images during template generation (fslvbm_2_template -a).  

I have also tried registering the problem subjects (n=9) before hand to the MNI152_T1_2mm_brain.nii.gz using FLIRT and putting the registered files into the template process.  It worked for all but 1 subject, but now I have a completely different set of subjects (n=5) that are now in the incorrect orientation in the template.  I don't think this is a problem with individual subjects but a problem in the template generation itself.  When looking at the files generated from the template building command,  for 4/5 of the subjects the file ending in *_GM_to_T is in the correct orientation but the  file *_GM_to_T_init is not.  The reaming subject has correctly oriented data but is flipped for both *_GM_to_T and *_GM_to_T_init files.

I have been trying to resolve this problem for a month and I am at a loss of what else to try.

On Fri, Aug 19, 2016 at 4:44 AM, Mark Jenkinson <[log in to unmask]> wrote:

Hi,

You can do "manual" registrations with either the GUI or the command-line in a place outside of your VBM analysis folder.  Then, when you are happy with them, you can replace the appropriate files in the folder and start the analysis from the appropriate point.  The precise details depend on whether it is the creation of the template or the subsequent registrations to the template.  However, if you are not confident with this then it is a lot easier just to run fslreorient2std on all of your images and restart the whole VBM process from scratch.

All the best,

            Mark

From: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of Amanda Rodrigue <[log in to unmask]>
Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
Date: Thursday, 18 August 2016 at 16:28
To: "[log in to unmask]" <[log in to unmask]>
Subject: Re: [FSL] fslvbm_2_template orientation problem

Can someone please provide further instruction on how I would manual register my problem subjects?     When you say "manual" registration do you mean with FLIRT outside of the automated template creating process?  If so, there doesn't seem to be a GM ICBM-152 (what the segmented images are registered to in the regular automated template pipeline)  listed in the standards folder.  Also would I run that registration on the skull stripped file or segment it first as in the automated pipeline and then register?  Finally, how does it work when I put  registered files in with others that are unregistered into the automated template generation process?  Sorry I don't have much experience using FSL.

On Mon, Aug 15, 2016 at 7:53 PM, Amanda Rodrigue <[log in to unmask]> wrote:

Sorry, but I have just now gotten back to this problem.  Ran fslorient2standard on all images and still had images come back flipped.  When you say "manual" registration do you mean with FLIRT outside of the automated template creating process?  If so, there doesn't seem to be a GM ICBM-152 (what the segmented images are registered to in the regular automated template pipeline)  listed in the standards folder.  Also would I run that registration on the skull stripped file or segment it first as in the automated pipeline and then register?  Finally, how does it work when I put  registered files in with others that are unregistered into the automated template generation process?  Sorry I don't have much experience using FSL.

On Wed, Jul 13, 2016 at 10:55 AM, Matthew Webster <[log in to unmask]> wrote:

Hi Amanda,
                    I’ve been able to reproduce the issue locally. A fix that worked for the two “bad” files was to run fslreorient2std on the input structural ( and manually regenerate their registrations ) which gave good registrations for each step ( non-flipped )! It might be worth running fslreorient2std on all the “bad” structurals ( and maybe even the “good” ones, since it shouldn’t make anything worse ) and then re-running fslvbm_2_template from scratch - if you try this and the new merged output still contains flipped images, you may need to manually register the few bad files, and then regenerate the final template.

Hope this helps,
Kind Regards
Matthew

> On 13 Jul 2016, at 15:18, Amanda Rodrigue <[log in to unmask]> wrote:
>
> Hi Matthew, any luck?
>
> On Friday, July 8, 2016, Amanda Rodrigue <[log in to unmask]> wrote:
> Ok, it looks like it worked.  It's called Rodrigue_VBM.  HCC_1213 and LCC_1118 are the bad ones.  Also in my struc directory there are also a ton of files that start with fslvbm2a.e4 followed by a bunch of numbers.  Almost all these files (all but 2 or 3) have a file size of 0.  I'm not sure if this matters or not but it seemed odd.  Thanks for your help!
>
> On Fri, Jul 8, 2016 at 7:16 AM, Matthew Webster <[log in to unmask]> wrote:
> Hello,
>          I’m afraid the tar file didn’t appear on our server - maybe it might be easier to upload a subset, e.g. for two “good” and two “bad” subjects, tar up all the files referring to that subject, and all the template_<blah> files?
>
> Kind Regards
> Matthew
>
>
>
>> On 7 Jul 2016, at 18:17, Amanda Rodrigue <[log in to unmask]> wrote:
>>
>> Can you let me know if it has uploaded. It was taking a really long time and I am not sure it went through.  It should be named Rodrigue_struc,tar.gz.  Thanks!
>>
>> On Wed, Jul 6, 2016 at 3:58 PM, Amanda Rodrigue <[log in to unmask]> wrote:
>> Hi, I'm in the process of uploading.  Her are the subjects that had the orientation issue:
>> FSL #/Subid
>> 25/HCC_1213
>> 50/LCC_1118
>> 51/LCC_1121
>> 66/LCC_1191
>> 84/SZ_1107
>> 85/SZ_1111
>> 89/SZ_1187
>> 100/SZ_1254
>> 112/SZ_1310
>>
>> On Tue, Jul 5, 2016 at 9:54 AM, Matthew Webster <[log in to unmask]> wrote:
>> Hello,
>>       This sounds quite odd, I think we need to have a look at all the files - can you tar and upload the whole directory to the oxfile link ( https://oxfile.ox.ac.uk/oxfile/work/extBox?id=700839B35190E48BF )?
>>
>> Kind Regards
>> Matthew
>>
>>> On 5 Jul 2016, at 14:15, Amanda Rodrigue <[log in to unmask]> wrote:
>>>
>>> Hi Matthew any thoughts on what might be wrong?
>>>
>>> On Wed, Jun 29, 2016 at 4:56 PM, Amanda Rodrigue <[log in to unmask]> wrote:
>>> So the <blah>_GM_to_T_init is not oriented correctly when overlaid on the template_GM_init but the <blah>_GM_to_T is correctly oriented when overlaid on  template_GM_init.
>>>
>>> On Wed, Jun 29, 2016 at 10:44 AM, Matthew Webster <[log in to unmask]> wrote:
>>> Hello,
>>>          The registered outputs created during VBM may change orientation, this is fine providing the registration is correct. For one of the problematic subjects, can you check if the first-pass registration has worked - e.g.
>>>
>>> compare <blah>_GM_to_T_init to template_GM_init in FSLview
>>>
>>> and let me know if the two are registered?
>>>
>>> Kind Regards
>>> Matthew
>>>
>>> > On 29 Jun 2016, at 14:30, Amanda Rodrigue <[log in to unmask]> wrote:
>>> >
>>> > Hello,
>>> > I am following the instructions for FSL's VBM analysis and  having some difficulties with the fslvbm_2_template step.  The script runs fine but when I look at the results (the template_4D_GM.nii.gz in fsl view there are about 8 or 9 subjects that are misoriented.  I have looked at them all in fsl view and the labels are correct and all are in the same orientation as those subs that are correctly oriented as reported by fslorient.  The brain image is in the correct orientation in all the files produced by the template script until the Sub#_anat_struc_GM_to_T_init.gz.  Consequently this is when the orientation switches from neurological to radiological.  I have also tried fslorient2standard on these subjects, as suggested by some comments on the message board but it does not seem to make a difference. Any input would be appreciated.
>>> > Amanda
>>>
>>>
>>>
>>> --
>>> Amanda Rodrigue
>>> Doctoral Candidate
>>> Franklin Foundation Neuroimaging Fellow
>>> Psychology: Brain and Behavior Sciences
>>> University of Georgia
>>>
>>>
>>>
>>> --
>>> Amanda Rodrigue
>>> Doctoral Candidate
>>> Franklin Foundation Neuroimaging Fellow
>>> Psychology: Brain and Behavior Sciences
>>> University of Georgia
>>
>>
>>
>>
>> --
>> Amanda Rodrigue
>> Doctoral Candidate
>> Franklin Foundation Neuroimaging Fellow
>> Psychology: Brain and Behavior Sciences
>> University of Georgia
>>
>>
>>
>> --
>> Amanda Rodrigue
>> Doctoral Candidate
>> Franklin Foundation Neuroimaging Fellow
>> Psychology: Brain and Behavior Sciences
>> University of Georgia
>
>
>
>
> --
> Amanda Rodrigue
> Doctoral Candidate
> Franklin Foundation Neuroimaging Fellow
> Psychology: Brain and Behavior Sciences
> University of Georgia
>
>
> --
> Amanda Rodrigue
> Doctoral Candidate
> Franklin Foundation Neuroimaging Fellow
> Psychology: Brain and Behavior Sciences
> University of Georgia
>

--

Amanda Rodrigue

Doctoral Candidate

Franklin Foundation Neuroimaging Fellow

Psychology: Brain and Behavior Sciences

University of Georgia

--

Amanda Rodrigue

Doctoral Candidate

Franklin Foundation Neuroimaging Fellow

Psychology: Brain and Behavior Sciences

University of Georgia

--

Amanda Rodrigue

Doctoral Candidate

Franklin Foundation Neuroimaging Fellow

Psychology: Brain and Behavior Sciences

University of Georgia

--

Rosalía Dacosta-Aguayo, Post-Doctoral Researcher.
Biomedicine, PhD
Clinical Neuropsychologist

Department of Neurosciences.
Group of Neurodegenerative Diseases
Gaixotasun Neuroendekapenezko Taldea
Neurozientziak Arloa

[log in to unmask]
T. +34 943 006 128

IIS BIODONOSTIA - OSI DONOSTIALDEA
Pº Dr. Beguiristain s/n * 20014 Donostia-San Sebastián SPAIN * T.+34 943
006 012 * F. +34 943 006 250 * [log in to unmask] *
www.biodonostia.org

--

Amanda Rodrigue

Doctoral Candidate

Franklin Foundation Neuroimaging Fellow

Psychology: Brain and Behavior Sciences

University of Georgia

--

Amanda Rodrigue

Doctoral Candidate

Franklin Foundation Neuroimaging Fellow

Psychology: Brain and Behavior Sciences

University of Georgia

--

Rosalía Dacosta-Aguayo, Post-Doctoral Researcher.
Biomedicine, PhD
Clinical Neuropsychologist

Department of Neurosciences.
Group of Neurodegenerative Diseases
Gaixotasun Neuroendekapenezko Taldea
Neurozientziak Arloa

[log in to unmask]
T. +34 943 006 128

IIS BIODONOSTIA - OSI DONOSTIALDEA
Pº Dr. Beguiristain s/n * 20014 Donostia-San Sebastián SPAIN * T.+34 943
006 012 * F. +34 943 006 250 * [log in to unmask] *
www.biodonostia.org