Hi Jamie,
it's hard to tell from the pics. The superrestore might be erodeing the borders of the thalmus and the pallidum. Can you flip back and forth between the uncorrected and corrected to see if that is the case. Also, check the gray/white junction in primary motor cortex. Basically you want as much correction as you can get as long as it doesn't cause these borders to shift position. Note that our T1.mgz *does* erode the pallidum and thalamus borders intentionlly, as we only use it for cortical surface models.
cheers, Bruce
On Mon, 27 Aug 2007, Jamie Hanson wrote:
Hello Freesurfers-
I wondered what the word was on preprocessing before you get into the land of freesurfer and recon-all.
I have some Inversion Recovery-T1s, where bias / field inhomogeneity is a slight issue. The standard N3ing that is done autorecon1 really wasn't cutting it (it was fine, but it required additional time in the back end, e.g., correcting skull-strips, adding control points). And for all non-freesurfing, I have been using either: 1) a combo of N3 and MFAST (b/c we have T2 where the same bias isn't present) or 2) a tweaked MFAST routine.
I however didn't want to "over-correct" my data, so I wanted to get some feedback before I go down either rabbit-hole. Anyone have some feedback / past experiences?
Pics from both routines are at: http://brainimaging.waisman.wisc.edu/~hanson/freesurf1/
Prefix of n3mfast- Combo of N3 and MFAST Prefix of superrestore- Tweaked MFAST routine
The superrestores looked cleaner, but I was worried it might be "too much" preprocessing.
Thoughts? Thanks much, jamie.