[minqicho]

Hi Tetsuya

It depends on what kind of connectivity value you are looking for. If you want to simply extract the number of tracts connecting two ROIs, you can do it in BrainSuite. There is a button under Menu–>File–>Save Connectivity Matrix can do it.
You want any further metric of the tracts (eg. Mean/Median of FA/MD), you need to use TractConnect tool (http://neuroimage.usc.edu/neuro/Resources/TractConnect)

Let me know if you have any further questions.

[iidaka]

Hi, experts.

Thank you for response.
I installed TRACTCONNECT module and tried to extract MD value from fibers.
To do so, I computed the tract and (File -> Save Fiber Tracks -> Save All Fiber Tracks)
saved it as fibers.dft.

Which of all or filtered tarcks should be saved?

Then on the command line, I entered
tractconnect (‘fibers.dft’, ‘–label’, ‘name.svreg.label.nii.gz’, ‘–allROI’, ‘–data’, ‘name.dwi.RAS.correct.MD.T1_coord.nii.gz’, ‘–desc’, ‘brainsuite_labeldescription.xml’, ‘–pvc’, ‘name.pvc.frac.nii.gz’);

Is this corerct?
Particularly, are RAS and reoriented… same meaning?

Then memory error occurred but it was fixed.

The computation lasted 9 hours!! and following message appeared.
Does message “Exception….” indicate error?

*******************************
Load tract file successfully
Load arguments succesfully
Calculating connectivity
Calculation done and saved succesfully
Time is 118846.479214 sec
Exception in thread “Timer-36505” java.lang.OutOfMemoryError: GC overhead limit exceeded
at com.mathworks.mlwidgets.explorer.model.FileDecorationModel.<init>(FileDecorationModel.java:59)
at com.mathworks.mlwidgets.explorer.util.FileSystemUtils.getDecorationSynchronously(FileSystemUtils.java:113)
at com.mathworks.mlwidgets.explorer.model.table.FileSystemExpansionProvider.getDecoration(FileSystemExpansionProvider.java:124)
at com.mathworks.mlwidgets.explorer.model.table.FileSystemExpansionProvider.getDecoration(FileSystemExpansionProvider.java:131)
at com.mathworks.mlwidgets.explorer.model.table.PathAffordance.shouldEverShowAffordance(PathAffordance.java:330)
at com.mathworks.mlwidgets.e
*****************************

I found results.csv that appeared to have tract MD values , but it contained so many
data, 60000 rows! Are these connectivity between voxels, but not ROIs?

One row is shown below.
612 and 2163 126 914 0.000499 0.000785 0.000717 0.001763 0.000205

Is this corerct value for MD of particular fiber?

I expected to have connectivity between ROIs as defined by Surface Registration,
but not between each voxels

Your help is veyy much appreciated.

tetsuya

[minqicho]

Hi Tetsuya

How many ROIs do you have?
If you set the flag ‘allROI’, you should have all pair-wise ROI results, which means if you have 300 ROIs you will get 300*300 rows.

It may be more appropriate if you only focus on certain ROI pairs’ connectivities if you having a too specific parcellation.

[iidaka]

Hi, BrainSuite

I need connectivity between 90 ROIs defined as ROI_ID from 120-641 by SurfReg.
This may create 90 x 90 = 8100 rows.

tetsuya

[minqicho]

Hi tetsuya

From the MATLAB error message, it seems that you ran into a memory issue.
As I do not have access to your label file, would you elaborate it a bit more? Is it the direct output from BrainSuite?

[iidaka]

Hi,

BrainSuite ends at *Time is 118846.479214 sec*
and probably following messages have been from matlab.

tetsuya

***********************************
Exception in thread “Timer-36505” java.lang.OutOfMemoryError: GC overhead limit exceeded
at com.mathworks.mlwidgets.explorer.model.FileDecorationModel.<init>(FileDecorationModel.java:59)
at com.mathworks.mlwidgets.explorer.util.FileSystemUtils.getDecorationSynchronously(FileSystemUtils.java:113)
at com.mathworks.mlwidgets.explorer.model.table.FileSystemExpansionProvider.getDecoration(FileSystemExpansionProvider.java:124)
at com.mathworks.mlwidgets.explorer.model.table.FileSystemExpansionProvider.getDecoration(FileSystemExpansionProvider.java:131)
at com.mathworks.mlwidgets.explorer.model.table.PathAffordance.shouldEverShowAffordance(PathAffordance.java:330)
at com.mathworks.mlwidgets.e
**********************************

[Author]

Posts