Hi Stephen,

1) Yes, this is probably a better idea if you are collecting in different scan sessions. In general I think this is to be avoided as you will have other differential distortions that will add blurring (e.g. gradient nonlinearities, B0 distortions, different tongue/jaw positions, etc...).

2) Historical. We haven't really tried to compare N3 to anything else. Let us know if you think the FSL stuff works better for you.

cheers, Bruce On Mon, 30 Oct 2006, Stephen Towler wrote:

Greetings all,

I'm new to freesurfer, so apologies if this has been covered. Two questions about early processing:

1. It looks like the recon-all script averages multiple T1's prior to

correcting inhomogeneity. If processing T1's acquired on different days, and/or on different scanners, might it be advantageous to manually run mri_nu_correct.mni on each individually before averaging? Other than CPU time, why average first and then correct for inhomogeneity?

2. Is freesurfer's use of MNI_N3 over FSL's fast inhomogeneity correction

algorithmic, empirical, historical, or other?

Thanks! Stephen