I am doing preprocessing of different kind of images, all of them are T1 3D MPRAGE or MPRAGE + GRAPPA...and I have seen them previously using imglob and some of them are more obscure and others are more bright
I would like to do two kind of analyses:
1) VBM
2) FAST segmentation to get WM and GM, apply a parcellation using AAL template and then calculate the volume of some subcortical structures of interest.
I thought about using FIRST for the second option...but one colleague adviced me to first run segmentation and then parcellation and to improve segmentation use SUSAN to denoise data and correct for inhomogeneities..what made me to search and read about SUSAN but I am not sure....about brightness threshold...it seems something arbitrary you apply...you can apply it to each whole image (3D) or to every sclice (2D) what it is estrange for me....and lately, about the Mask SD: it is set by the user.....
For me it seems that you are applying a prior smoothing to data that will be smoothed again by VBM process or by FAST process...and I think this is not correct as it is known that different kinds of smoothings can affect the results...so I would like to be sure what I am using and if it is justified from a methodological point of view.
Thank you a lot for your help.
Yours sincerely,
Rosalia.