Hello everyone
Warning for a confused post.
I have a question regarding which kind of normalization procedure to use. In my preprocessing I'm using New segment + Dartel and the "normalize using Dartel" options. Is this correct or are there certain criterias important to fulfill before using the Dartel normalization procedure? What is the main reason for using DARTEL overall?
The reason I'm asking is that I think my results look strange. I get very very high FC for all my seeds when I do a one sample t-test in SPM8. I guess this is not that weird because the whole brain is connected, but we're talking FWE corrected thresholds of .0001 here and pretty much the whole brain lights up. Maybe the normalization procedure could have something to do with this? I've looked though my normalized brains and they look fine.
Is there a way to tell whether or not normalization has been done using DARTEL (some special prefix perhaps)?
Kind regards and thank you for a wonderful toolbox.
evast001
Hi Eva,
Hi Eva,
Using "New segment + Dartel" and the "normalize using Dartel" options are correct combination if you want to use DARTEL for spatial normalization of your functional images.
"New segment + Dartel": This will perform "new segment" procedure on your structural images, thus you can get the tissue maps. In addition, DARTEL will be performed to normalize the structural image and tissue maps to template space and MNI space. The normalization parameters will be generated as well.
"Normalize using Dartel": This refers to the functional images. The parameters generated in "New segment + Dartel" steps will be applied to your functional images, thus normalize them into MNI space.
As compared in a paper, DARTEL got pretty good normalization performance: Klein, A., Andersson, J., Ardekani, B.A., Ashburner, J., Avants, B., Chiang, M.C., Christensen, G.E., Collins, D.L., Gee, J., Hellier, P., Song, J.H., Jenkinson, M., Lepage, C., Rueckert, D., Thompson, P., Vercauteren, T., Woods, R.P., Mann, J.J., Parsey, R.V., 2009. Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration. Neuroimage 46, 786-802.
For the files under the directory have a postfix "W", these files were normalized into MNI space, e.g., FunImgARCW.
For one sample t-test on FC maps, usually the whole brain are highly correlated (especially if global signal regression is NOT performed). I don't think this is caused by normalization or smooth procedure. There is a reference: Scholvinck, M.L., Maier, A., Ye, F.Q., Duyn, J.H., Leopold, D.A., 2010. Neural basis of global resting-state fMRI activity. Proc Natl Acad Sci U S A 107, 10238-10243.
Best,
Chao-Gan