In SPM, you can use proportional scaling to deal with whatever global
variable you choose, and I'm guessing that you want to deal with total
intra-cranial volume.
An affine transform does not give an especially good match for the whole
head/brain. Because it does not have the flexibility to deal with
nonlinear deformations, it has to match some parts of the image at the
expense of not matching other parts so well. The determinant of an
affine transform is a bit arbitrary as the result will depend on which
parts of the image are aligned by the algorithm. Maybe the affine
transform matches the whole head - in which case the result will depend
on what part of the head is within the FOV. Alternatively, it may match
the brain, or some part of the brain. Depending what is in the image,
and things like skull thickness etc, will all influence the estimated
affine transform, and therefore its determinant.
Best regards,
-John
Dear All,
I've been thinking about the issue of modulation (non-linear
versus affine+nonlinear) and it makes sense to only use the
non-linear component alone. However, in DARTEL, flow fields are
used and you it seems that SPM only allows you to warp by the
determinant of the total.
Could one "remove" the affine modulation effects by using the
affine component of the transform from the segmentation
procedure (assuming the end result is in MNI space) that same
way its removed in VBM toolbox OR does the "affine" portion
change when DARTEL is used? If so, is there a way to extract it?
Best Regards, Donald McLaren
=================
D.G. McLaren
University of Wisconsin - Madison
Neuroscience Training Program
Office: (608) 520-0586
=====================
--
John Ashburner <[log in to unmask]>
|