Dear Freesurfer experts,
I’m doing a longitudinal study on cortical thickness and subcortical volumes and got some questions. I used the longitudinal stream for the preprocessing
1. On cortical thickness, I’m doing a paired analysis as described in https://surfer.nmr.mgh.harvard.edu/fswiki/PairedAnalysis, and then a cluster-wise correction as described in https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysis, with 5000 simulations instead of 5, and a vertex-wise threshold of 3 . But I don’t get the ocn.annot map, instead I get a ocn.mgh and ocn.lut. The CSD file also looks weird (constantly about 150 clusters with max size one or 2) and the resulting cluster.summary also gives a huge list of very small but significant clusters. Am I doing something wrong? How should I go about this? I’m using FS 4.5. Files are included (jpg for sig.mgh and cluster.mgh for negative simulation, and csd (zipped), summary and simulation log)
2. For the subcortical volumetric analysis I use the aseg.stats. Is the ICV a reliable measure to correct for overall brain size? I heard ICV measurements are not very reliable on T1’s (I get a mean of 2,4% decrease over 16 months...). Are other methods preferable to this one.
3. Is there a logic or rule of thumb for choosing the vertex-wise threshold?
Thanks for your help!
Pieter
