External Email - Use Caution
Dear experts, I have three groups: - pre treatment (baseline) - post treatment (follow-up after 6 month of the baseline visit) - healthy controls.
The three groups have PET data. I ran surface based paired t test between pre and post treatment. The analysis failed to show significant change in the signal between baseline and follow-up visits. As an alternative approach I tried: - Surface-based analysis between pre treatment vs healthy controls. This analysis found cluster A of significant difference - Surface-based analysis between post treatment vs healthy controls. This analysis found cluster B of significant difference
I subtracted the two clusters A and B then I used the resultant mask (the difference between the clusters A and B) to measure the required signal.
Colleague of mine said that this might be criticized for being circular and it may not be acceptable? My understanding is that this analysis is not circular because the signal was measured in a completely new region represents the difference between two clusters.
I appreciate any suggestion or highlights!
John