Hi all,
We use freesurfer 5.1.0 and we ran the recon-all on 100 subjects. We checked the talairach transform, skull strip the pial surfaces and the segmentations as suggested on Freesurfer Wiki. Everything seems to be fine, however, the ICV values from the stats/aseg.stats seem to be incorrect. The average value of ICV's is 1280 cc for men and 1108 cc for women, which seems pretty low. In 21 subjects from the 100 the ICV was even smaller than the supratentorial volume. The talairach transforms seems to be fine by visual inspection.
1. I think the wrong estimation of ICV value (because of the talairach transformation is not the best, although it seems right by visual inspection) won't affect the cortical/subcortical segmentations. Is it right?
2. What volumes should I sum to get an ICV, which is not estimated from the talairach transformation?
3. Is it a volume (other than ICV) reported by freesurfer, which can be used for correcting for overall head size?
I'd appreciate an early reply.
Thanks, Gabor