Thanks Gwenaelle,

i did like this:

fslroi all_FA_skeletonised 25_pre_control 0 24

fslroi all_FA_skeletonised 25_post_control 25 49

fslroi all_FA_skeletonised 8_pre_MD 50 57

fslroi all_FA_skeletonised 8_post_MD 58 66

is there an error here? maybe I need to start from the beginning?

On Fri, Aug 27, 2010 at 4:37 PM, Gwenaëlle DOUAUD <[log in to unmask]> wrote:

Hi Francesca, there are actually 39 volumes in your all_FA_skel_33*. Are you sure you used correctly fslroi? Cheers, Gwenaelle --- En date de : Ven 27.8.10, Francisca Serrano <[log in to unmask]> a écrit : De: Francisca Serrano <[log in to unmask]> Objet: Re: [FSL] Re : [FSL] extract numerical FA values after TBSS in FSL À: [log in to unmask] Date: Vendredi 27 août 2010, 15h27 Hi, thanks a lot for the help the number is 457313 I sent a .rar file with all_FA_skel_33_pre_minus_post mean_FA_skeleton_mask design.mat & design.con and this was the command used randomise -i all_FA_skel_33_pre_minus_post -o tbss -m mean_FA_skeleton_mask -d design.mat -t design.con -n 500 --T2 my design is an anova 2x2 i have 25 control and 8 patients (33) in 2 timepoints (pre and post)... Thank you very much for this possibility, kind regards Francisca On Fri, Aug 27, 2010 at 12:10 PM, Matthew Webster <[log in to unmask]" target="_blank">[log in to unmask]> wrote: Hello,           If nothing else appears to be inconsistent with your input files, then if you tar them up and upload them to: http://www.fmrib.ox.ac.uk/cgi-bin/upload.cgi and let us know the 6-digit reference number, we will be able to check what's going wrong directly.. Many Regards Matthew hi, thanks... I'm sorry to insist... but is giving me the error all time. I'm sure there is no space, I went through all the text file... could it be another thing? On Thu, Aug 26, 2010 at 2:59 PM, Gwenaëlle DOUAUD <[log in to unmask]" target="_blank">[log in to unmask]> wrote: Hi Francesca, the design matrix and contrasts are the right ones, you simply should not have any space in the matrix (randomise then understands that you have 39 rows with these six carriage returns instead of 33) and contrasts. Cheers, Gwenaelle De: Francisca Serrano <[log in to unmask]" target="_blank">[log in to unmask]> Objet: Re: [FSL] Re : [FSL] extract numerical FA values after TBSS in FSL À: [log in to unmask]" target="_blank">[log in to unmask] Date: Jeudi 26 août 2010, 20h38 Hi Gwenaelle, I got to make work all the subjects in tbss and separate them... getting at the end all_FA_skel_33pre_minus_post... but when i run the analysis with randomize it gives me an error, I think because of the matrix, can you help me with them? so I used : randomise -i all_FA_skel_33_pre_minus_post -o tbss -m mean_FA_skeleton_mask -d design.mat -t design.con -n 500 --T2 with these matrix Loading Data: *************************************** Data loaded ERROR: Program failed An exception has been thrown Logic error:- detected by Newmat: incompatible dimensions MatrixType = Rect   # Rows = 33; # Cols = 33 MatrixType = Rect   # Rows = 39; # Cols = 121026 Trace: MatrixMult. The matrix I did: .mat /numwaves 2 /num point 33 /ppheights 1 1 /matrix 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 Contrast: /numwaves 2 /numcontrast 6 /ppheights 1 1 /matrix 1 0 -1 0 0 1 0 -1 1 -1 -1 1 Could you give me a hand again? thanks! On Mon, Aug 23, 2010 at 8:06 PM, Gwenaëlle DOUAUD <[log in to unmask]" target="_blank">[log in to unmask]> wrote: 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 --------------------------------------------------------------------