.fsf and .fts files are completely different.  .fsf stores lots of overall design details and is not needed for randomise. -fts stores F-contrasts.

Contrasts are stored in the .con file.  Are you doing F-tests or t-tests?  If just t-tests (ordinary contrasts) then you don't need the -f or -F options

It might be worth reading parts of the FEAT manual to get a better feel for design matrices

Cheers,

Eugene



--


On 27 January 2014 12:26, olof lindberg <[log in to unmask]> wrote:
Hi I sent you the Amir file so that why I put it into the command else my files have the name 

randomise -i A_Correlatin_left_HC.nii.gz -m A_Correlatin_left_HC_mask.nii.gz -o Correlation_Amir_L_HC_AGE -d New.mat -t New.con -f New.fsf --fonly -D -F 3

but I not sure I understand how you mean by -f is incorrect - is it the contence of the .fsf file or should I just not use the -f New.fsf  at all in the command? And in that case how do I then specify my contrast for the correlation

I am new beginner so would be a great help if you could explain this!




2014-01-27 Eugene Duff <[log in to unmask]>

The -f input is incorrect: see the input spec - this is meant .fts files specifying f-contrasts.  Your .mat, .con and .fts files will usually all come from the same design spec - why do yours have different names?  Otherwise OK I think assuming you are not interested in t-tests. 

Cheers,

Eugene
--


On 27 January 2014 12:08, olof lindberg <[log in to unmask]> wrote:
randomise -i A_Correlatin_left_HC.nii.gz -m A_Correlatin_left_HC_mask.nii.gz -o Correlation_Amir_L_HC_AGE -d Amir.mat -t New.con -f AmirDesign.fsf --fonly -D -F 3

like this i think i tried doing it




2014-01-27 Eugene Duff <[log in to unmask]>

Just to be clear: you gave randomise the .fsf file or the .mat file?  Can you give your full command? 


--


On 27 January 2014 11:54, olof lindberg <[log in to unmask]> wrote:
OK, I run the First as usual, just providing a .mat file.

then I want to run correlation analysis with randomise - so I am creating  the necessary files in glm_gui

I attached this file.

I probably do some simple misstakes but basically I try to input age as one EV do see if I have any shape correlation with age


2014-01-27 Eugene Duff <[log in to unmask]>

I'm a bit unclear exactly what you're trying, but design.fsf files are used with FEAT.   FIRST itself is for segmentation and doesn't take design matrices (design.mat).

Eugene


--


On 27 January 2014 08:44, Olof Lindberg <[log in to unmask]> wrote:
Hi FSL experts. I am trying to run correlation by age in First. I just put in one EV containing age of subjects in the same order as I created the concat-file.

getting: design.fsf is not a valid vest file" when I am trying to run this


what am I doing wrong?