Just wanted to verify:

I have ran my subjects through the longitudinal processing stream (two time points a couple weeks apart). 

I want to predict cognitive performance (from the first time point) with Freesurfer stats. As I have the longitudinal data available, would it be better to use longitudinally processed stats (as opposed to the cross sectional stats)?

I ask because I'm not aware of any publications that do this...

Thanks much,

Mike Kranz