BrainVISA

hi to all,

I imported the dicom files. I set manually the AC/PC and I tried to run the rest steps in the pipeline of Morphologist 2012. For my first subject everything worked correctly and I got a head.mesh, Rhemi.mesh and Lhemi.mesh as I wanted. I viewed them with Anatomist and they are very good.
However, with the next subject I took the following error when I run the Mopphologist 2012:

in 4 - Split Brain Mask 3 System command exited with non null value : 111
RuntimeError
neuroProcesses.py (2400) in _processExecution:
result = process.execution( self )
SplitBrain.py (116) in execution:
apply( context.system, call_list+option_list )
neuroProcessesGUI.py (2380) in system:
ret = apply( ExecutionContextGUI.system, (self,) + args, kwargs )
neuroProcesses.py (2522) in system:
raise RuntimeError( _t_( 'System command exited with non null value : %s' ) % str( ret ) )

in Morphologist 2012 3 System command exited with non null value : 111
RuntimeError
neuroProcesses.py (2400) in _processExecution:
result = process.execution( self )
neuroProcesses.py (1064) in execution:
return self._executionNode.run( context )
neuroProcesses.py (1723) in run:
return self._run( context )
neuroProcesses.py (1952) in _run:
result.append( node.run( context ) )
neuroProcesses.py (1723) in run:
return self._run( context )
neuroProcesses.py (1868) in _run:
return context.runProcess( self._process )
neuroProcesses.py (2173) in runProcess:
result = self._processExecution( _process, None )
neuroProcesses.py (2400) in _processExecution:
result = process.execution( self )
SplitBrain.py (116) in execution:
apply( context.system, call_list+option_list )
neuroProcessesGUI.py (2380) in system:
ret = apply( ExecutionContextGUI.system, (self,) + args, kwargs )
neuroProcesses.py (2522) in system:
raise RuntimeError( _t_( 'System command exited with non null value : %s' ) % str( ret ) )

What can I do?
thanks in advance,
Chrysa Lithari

Hi,

First you should check the brain mask, if it is ok, that really the process "Split Brain Mask" which failed. The process depends a little of the quality of your image. He needs to find three separate components (cerebellum, left and right hemispheres) and to do that he thresholds to get the white matter and erodes. If your image's contrast is bad or if your image is a little bit noised, it can be better to reduce the "bary factor" parameter, which will reduce the threshold between the grey matter and white matter. You should get a result with that, if not, you can may be send me your image to help you better.

If your brain mask is not ok, you should check the histogram analysis and see if the values of the gray matter mean and variance and the white matter mean and variance are realistic. If not, tell me and I will tell you what you could try.

Best regards,

Clara Fischer

Hi Clara,

thank you very much for your reply.
Finally, the problem was that when I set manually the AC/PC points I had set as 'true' the label 'allow flip initial MRI' and indeed there was a transformation because my data were totally wrongly oriented.
Somehow, I did this twice and therefore the initial data were wrong again. It is not clear to me how this happened and how BrainVisa solved this double transformation, but I solved the problem by deleting all the files and importing them again and run each procedure only once on each subject. Till now works well and I have enjoyed my meshes!

Thanks again for your reply,
Chrysa

chrysa wrote:Hi Clara,

thank you very much for your reply.
Finally, the problem was that when I set manually the AC/PC points I had set as 'true' the label 'allow flip initial MRI' and indeed there was a transformation because my data were totally wrongly oriented.
Somehow, I did this twice and therefore the initial data were wrong again. It is not clear to me how this happened and how BrainVisa solved this double transformation, but I solved the problem by deleting all the files and importing them again and run each procedure only once on each subject. Till now works well and I have enjoyed my meshes!

Thanks again for your reply,
Chrysa

Chrysa, I have a same problem now but my image orientation seems OK. Previous steps are completed correctly (at least they are green checked) and I investigate Brain Mask which seems also correct... I cannot find out where problem is.

Hi Thornpyros,

I guess you tried both deleting the dicom files and inserting them again, and also changing the bary factor that Clara proposed.
If you want I can try the procedure with your image also.
I am a newbie too, but 2 newbies sometimes are better than one.

Chrysa

Hi
I am going to segment a structural MRI with Analyze format. I used both "Morphologist 2011 " and "T1 Pipeline 2007 " but both of them failed in "Split Brain Mask" step. The error which is shown in window is :
________________________________________________________________________________________________
in 4 - Split Brain Mask 1 System command exited with non null value : 111
RuntimeError
neuroProcesses.py (2400) in _processExecution:
result = process.execution( self )
SplitBrain.py (116) in execution:
apply( context.system, call_list+option_list )
neuroProcessesGUI.py (2380) in system:
ret = apply( ExecutionContextGUI.system, (self,) + args, kwargs )
neuroProcesses.py (2522) in system:
raise RuntimeError( _t_( 'System command exited with non null value : %s' ) % str( ret ) )
_________________________________________________________________________________________________
I tried with different parameters, but still I give the error. I appreciate if any one can help me to resolve this problem.

Hi,

Did you look at the results of the previous steps to see where is the problem ? Did the bias correction and brain mask seem ok ?
Are you sure you correctly specified the AC/PC points ?
You could also try the last version of Morphologist which is named Morphologist 2012. It is included in Brainvisa version 4.2.1.

Dominique

Hi Dominique

Thank you very much for your answer. There is green check mark in front of all steps. I think it means that the processing in these steps are successful. I am not sure about the AC/PC points. I selected "Prepare subject for anatomical Pipeline" and in the drop box in front of "Normalized" option, "MNI from SPM" was selected. Actually I used these options because for other selections in AC/PC point, I got error. Do you think that my problem is related to these step?

Thanks
Saideh

Hi Dominique

I tried again with Morphologist 2012. The process fails in "Compute Brain Mask". In Morphologist 2011 I had selected "Vip Get Brain". When I selected "3_Compute Brain Mask" the error was similar with the error that I have got in "Compute Brain Mask" in Morphologist 2012.

Best

Hi Saideh,

The green check mark in front of the steps only means that the process terminated without raising an error. But it does not always mean that the results are correct or good enough to get the next steps successful.
You should check the results of the previous steps anyway by clicking on the eye button near the commissure_coordinates and mri_corrected parameters.
But if you choose the parameter normalized="MNI for SPM" whereas you're image is not normalized it may be the reason of the error. Indeed, clicking o the Anterior and posterior commissure points gives Morphologist reference points in the image that helps the algorithms for the segmentation. When you select MNI for SPM, Brainvisa suppose that the image is normalized in SPM template and set the coordinates of AC and PC points in the MNI referential.
For more details about how to spot AC/PC points, read the following documentation, there's even a video that shows how to do that:
http://brainvisa.info/doc/axon-4.2/bv_t ... re_subject

Dominique

Hi all,

I'm a brand new user of BrainVISA trying to use Morphologist pipeline. I'm running:

- OS: Ubuntu 11.04 64bits
- BrainVISA 4.1.1 (older/newer - 64/non 64bits versions give me the same error)

Morphologist pipeline gets stucked at AC/PC Or Normalisation step showing this error (SPAM models correctly installed):
------------------------------------------------------------------------------
Process Morphologist 2011 2 started on 2012/04/20 17:22

Skip unselected node: Check SPAM models installation
X is flipped

 Flip needed, with matrix:

[ [1, 0, 0, 0]
[0, 1, 0, 0]
[0, 0, -1, 181]
[0, 0, 0, 1] ]
Image needs to be flipped, but you did not allow it, so it won't change. Expect problems in hemispheres separation, and sulci recognition will not work anyway.
-------------------------------------------------------------------------------
Nothing changes:

- changing coordinates and in particular Left_Hemisphere_Point
- setting allow_flip_initial_MRI = True (it's anyway wrong because my image is oriented according to radiological convention)
- using any of the Normalised field option

Using my colleague computer (Mac) same image, same parameters it works without any problem.
What should I try next?

Thanks in advance,
Claudia

Hi Claudia,

There's no reason it should work on a computer and not on another one... Are you sure you correctly clicked on the AC and PC points ?
The message you mentioned generally means that the option allow_flip_initial_MRI should be set to true to allow Brainvisa applying a transformation to the image.
If you wish, send me the image by private message, I will have a look at it.

Dominique

chrysa wrote:Hi Thornpyros,

I guess you tried both deleting the dicom files and inserting them again, and also changing the bary factor that Clara proposed.
If you want I can try the procedure with your image also.
I am a newbie too, but 2 newbies sometimes are better than one.

Chrysa

Chrysa, sorry for delay. Here is the image,
http://dl.dropbox.com/u/3482709/ImageYilm.zip

I appreciate any help : )