Dear Max,
I suspect that you specified this second level job using your own script
and that the 'scans' fields (containing the filenames for each cell) are
1x17 cell arrays instead of 17x1 ones, ie you should transpose them.
They are displayed identically in the interface but handled differently
later on...
We'll try to change SPM so that this kind of mistake cannot happen
anymore as it's slightly mischievous.
The recursive option in the file selector ('Rec') can be quite handy in
this situation to select all, say, "con_0004*" files automatically,
removing the cumbersome task to select them one by one or writing a
Matlab script.
All the best,
Guillaume.
Maximilien Chaumon wrote:
> Hi,
>
> I'm currently trying to pass to the second level a HRF analysis with 2
> derivatives.
> For this, I created contrasts for each derivative in each subject
> between 2 conditions of interest.
> I enter these in a factorial design with one factor (Basis) with 3 levels.
> I thus have three cells specified in my batch editor saying
> Level 1 containing 17 files (named con_0004.img for each subject)
> Level 2 containing 17 files (con_0005.img for each subject)
> Level 3 containing 17 files (con_0006.img for each subject)
>
> I believe this is the correct way to pass this design to the second
> level, but when I launch this, the graphics window shows something I
> don't understand. The design matrix shows 3 columns (ok) and 17*3 rows
> (ok) but the labels in front of the rows seem incorrect. It shows that
> files con_0004 con_0005 and con_0006 are */not respectively/* associated
> with basis1, 2 and 3.
> This is probably clearer in the attached screen capture.
>
> Am I completely misunderstanding the design graphics?
>
> Thanks,
> Max
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|