Dear Priyantha,
>Dear all,
>
>When I needed to look for cortical areas commonly activated by two
>contrasts, I would make the conjunction of the two contrasts.
>When I do this in spm99b, I get contrast histograms depicting two
>simultaneous contrasts being computed. The results are displayed and a t
>image is saved.
>Later, if one only looks at the contrast that was saved from the above
>operation, then there is only one contrast histogram and the results are
>very different. In this case the resultant contrast is called "Contrast A
>(otrthogonalized with respect to {some other contrast}) "
>I am not clear about what happens here. Can some one explain to me please?
>
Let us say you have to valid contrasts named c1 and c2, which have been
estimated and saved already. Let us now say you want to look at the
conjunction of c1 and c2. First assume c1 and c2 are orthogonal (you may
think of them as independent), then the conjunctions is simply created by
displaying for each voxel the minimum t value across the two contrasts.
This does NOT involve SPM creating and saving any new .img files.
Assume now they are not orthogonal. SPM will then proceed to orthogonalise
one of the contrasts with respect to the other (note that this is not
equivalent to orthogonalising the contrast weight vectors) and the
resulting "new" contrast will be saved, and will be tagged something like
"Contrast c1 orthogonalised wrt c2". Finally SPM will display the minimum
t-field from c2 and the orthogonalised c1, equivalent to above. Also, the
display will show you the resulting contrast weight vector of the
orthogonalised contrast.
Be warned that the resulting contrast weight vector may be quite different
from that which you entered.
>
>Thanks so much,
>
>
Good luck Jesper
Jesper Andersson
Wellcome Dept. of Cognitive Neurology
12 Queen Square
London WC1N 3BG
phone: 44 171 833 7484
fax: 44 171 813 1420
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|