Hi. I’m having fun with 3dNetCorr. (Of course, it’s very serious science. Serious. Serious. Harrumph, harrumph.) I appreciate its ease-of-use; I was able to get preliminary results very quickly. If I wanted now to get partial correlations with region size partialed out, should I stick with it, or might you recommend another AFNI program?
Glad that you are having some serious times with AFNI.
At present, there is not a way to include ROI size for partial correlation. I am not quite sure how that would go—the correlation is calculated between a pair of time series of length N, and the ROI size associated with each is “just” a scalar.
You’re right. Maybe I will use some sort of linear mixed model or select a random subset of voxels in the larger regions to match the smallest region.