Dear Freesurfer experts,
I would like to perform a longitudinal analysis with freesurfer 5.1 and I have some questions regarding the whole processing stream. I have BL and 1Y follow-up from all patients (110), and moreover I have 2Y follow-up from only some patients (more or less 50). Do you recommend using the 2Y when available in the whole stream? Or, on the other hand, it would be better to perform a BL and 1Y analysis in the whole sample and then, separately, do a second analysis only with those subjects that have 2Y follow-up? My idea is to study rates of change (cortical and ventricles, mainly).
I have another more technical question. When running the base (recon-all -base <templateid> -tp <tp1id> -tp <tp2id> ... -all), does the order of the time points matter? In Reuter et al., 2012 says that all scans are treated the same, so I guess it doesn't matter if tp1 is the 1Y and tp2 is the BL. Am I wrong?
Finally, to look at ventricle changes, do you also recommend using the longitudinal stream or rather the cross-sectional one? It seems that the longitudinal stream is more cortical specific, but maybe I would be wrong...
Thanks
Best,
Eduard