Hi,

I think this might help you out

http://blogs.warwick.ac.uk/nichols/entry/flame_without_1st/

There's a follow-up for cases without varcope files

http://blogs.warwick.ac.uk/nichols/entry/flame_without_1st_1/

Another option is to use randomise.

Cheers,

Jeanette

On Thu, Jul 26, 2012 at 6:12 PM, Fran Kanayet <[log in to unmask]> wrote:

Dear FSL experts,

I want to run a paired sample t-test between two statistic-maps generated by a searchlight analysis using pymvpa (the searchlight procedure used betas from a 2ndlevel analysis in feat so I have all the registration parameters for all subjects). I included the statistic maps from the searchlight analysis inside my 2ndlevel.gfeat directory and tried to use the feat model using inputs are 3d images from feat directories. When I try to run the model it complains that:

"Registration has not been run for all of the FEAT directories that you have selected for group analysis. Please turn on and setup registration."

I understand that the problem is that fsl doesn't know where to look for the registration parameters I used before doing the searchlight procedure. My question is what can I do to solve this? Can I just include the searchlight 3D images inside a cope.feat directory from the standard univariate analysis done with the same subjects and the same initial first-level analysis? Or if I have to copy some reg files which files do I have to copy?

Thanks for any help I can get.

Frank