Dear Lydia,
there are 2 different ways that you may want to use a contrast to define a seed region VOI for PPI.
Firstly, you typically want to find a peak voxel, i.e. a voxel with the highest activation for your contrast, around which you usually draw a sphere of a certain radius. Which contrast you choose depends on your research question of course - usually it will be a contrast of 2 experimental conditions, e.g. task A > B. The "peak voxel" could be defined at the group level, or alternatively at the subject level in which case you usually want to restrict the search for the subject-specific peak somehow to end up in the "same" region across subjects (e.g. by using an anatomical mask of, say, the left IFG).
Secondly, you may want to restrict the sphere centered on your peak voxel to "activated" voxels. There just was another discussion on this topic: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=SPM;453c6b1d.1903
As mentioned there, you use a contrast with a certain p-value threshold to define "activated" (i.e. supra-threshold) voxels within the sphere (at the subject level). The contrast will usually be the same that you used to determine the peak voxel. You could also just avoid this step all together and simply take all voxels inside the sphere.
Hope this helps!
All the best,
Phil
|