Contrast weights for between-sources comparisons

Hi Alfonso and others


I'm unsure about contrast weights in between-sources comparisons - thanks for any pointers!

If I compare e.g. 4 frontal against 2 parietal ROIs in a t-test, my weights vector should be [1 1 1 1 -2 -2] in order to sum to zero. The CON* images in the folder results/secondlevel/(myanalysis) folder should then be the linear combination of the BETA* images for the corresponding ROIs (from the first-level analysis), weighted by my contrast vector.

Is this correct?

Best wishes, 

Eugeno