Hi Francesca,
>So, I run 2 randomize procedures with all subjects (patients and control) >together (pre & post) and got the skeletons for each procedure... but if >i dont understand bad, i should not do that,
>but to run statistics for each subject separately first... is that >correct? so, shall i do all TBSS steps for each subject, that is, taking >2 files in each: pre and post? is that right?
>I have 25 control and 8 patients... so would I have like 33 files to >contrast doing the difference between the 2 timepoints?
>and then... how to proceed to do the difference between the two >timepoints? running randomize with a matrix like (...)
No sorry, not quite...
What you need to do is to run tbss on all the subjects (so 66 of them, pre_control1, pre_control2, ..., pre_patient1, pre_patient2..., post_control1... post_patient1...etc.). At the last stage, when you have the all_FA_skeletonised for the 66 subjects (hopefully organised in 4 nice blocks post_control, post_patient, pre_control, pre_patient, check the alphabetical order of your subjects!!), you just split this 4D image into these 4 blocks (two of 25 images, two of 8 images) with fslroi and substract them two by two, so something like:
fslmaths my4D_25_pre_control -sub my4D_25_post_control my4D_25_pre_minus_post_control
fslmaths my4D_8_pre_patient -sub my4D_8_post_patient my4D_8_pre_minus_post_patient
And then concatenate then to have a 4D of 33 images of the difference of FA for each subject:
fslmerge -t all_FA_skel_33pre_minus_post my4D_25_pre_minus_post_control my4D_8_pre_minus_post_patient
And run the analysis as I've described before.
Hope this makes sense,
Gwenaelle
--------------------------------------------------------------------
Gwenaëlle Douaud, PhD
FMRIB Centre, University of Oxford
John Radcliffe Hospital, Headington OX3 9DU Oxford UK
Tel: +44 (0) 1865 222 523 Fax: +44 (0) 1865 222 717
www.fmrib.ox.ac.uk/~douaud
--------------------------------------------------------------------
|