Hi Hwamee,
I don't use SPM2, and don't seem to have the file spm2_conj under my
SPM2 directory anyway. But I'll try to help...
The way conjunctions are implemented changed in SPM5 I believe, based
on the controversy and response in:
http://dx.doi.org/10.1016/j.neuroimage.2004.12.005
http://dx.doi.org/10.1016/j.neuroimage.2005.01.013
You want to find regions activated in both A and B, i.e. the
conjunction null, or u<=1 effects under the null hypothesis, using the
terminology of the above papers.
I think a valid approach for this is to save a binary mask of the
significant (for FDR if you wish) results for each of A and B
separately (e.g. using the save button from the results GUI), and then
to take the intersection of these two masks, e.g. using imcalc with
the expression 'i1 & i2'.
Hopefully someone will shout if the above is invalid, or very
different from using the SPM5 conjunction machinery with FDR
correction...
Best,
Ged
> I've posted the following message a few weeks ago, but couldn't get any
> reply yet. If more exlanation about the problem is needed, please let me know.
>
> We have 2 conditions (A & B) and we want to identify regions that are
> significantly activated in both. However, when we apply spm2_conj and use
> the FDR option, we got the following error:
> ??? Error using ==> spm_uc_FDR
> n & number of mapped images doesn't match (1,2)
>
> It would be really appreciated if you let me know how to solve this problem.
>
> Thank you,
>
> Hwamee
>
|