Print

Print
Saad, thank you very much for your answer.

I did step 3 because i thought it was needed the standard2diff.mat output from registration. I didn't notice that Flirt created itself the matrix. Do i have to use this .mat matrix from Flirt in the option: "Select Seed for diff transform" in Probtrackx?

Also, i have tried to use fslstats -thr and -mas but:

1) Which threshold is good to use?
2) When common mask was created i would like to know a percentage value, how can i obtain it?

Cheers,

Patricia.

2011/1/31 Saad Jbabdi <[log in to unmask]>
Hi Patricia

Comments below.

> I have some questions concerning to probtrackx tractography. The steps i am carrying out are:
>
> 1) Flirt each original data in my sample (after eddy_current_correction and bedpostx) to the standard MNI152_T1_2mm_brain image.

       You mean register e.g. the b0 image of each subject to MNI space, not the whole data? yes that's ok.

>
> 2) Then i create a mask in the Flirted image of each subject.
       ok.


>
> 3)  Registration each image to the MNI152_T1_2mm_brain image.
       not sure what you mean here. Haven't you done this in step 1?


>
> 3) After that i run probtrack with this mask and the matrix converter standard2diff.mat
       ok.


> 4) The outputs are fdt_paths.nii.gz of each subjetc
       yes, and these should be in MNI space.


>
> (I would be glad if you confirm that this process is correct).
>
> My question is if is it possible to obtain an unque image which contains the % of the paths whose traces match in all the sample and how to get this value.

If you mean that you want to consider samples that are common across subjects, you could for example threshold the fdt_paths and consider only the intersection across subjects.
You can do these things with fslmaths ( options -thr and -mas )

Saad.

>
> Thanks for yor attention,
>
> Patricia.
>

--
Saad Jbabdi
University of Oxford, FMRIB Centre

JR Hospital, Headington, OX3 9DU, UK
(+44)1865-222466  (fax 717)
www.fmrib.ox.ac.uk/~saad