Dear
all,
this is a question
to all you registration gurus.
For coregistration
between different MRI modalities for the same individual (eg registering for
exampe an EPI image to a T1 weighted structural), most of us routinely use rigid
body coregistration. For normalization to some template space, many non-rigid
normalization algorithms are available.
As for between
modality coregistration works reasonably well for lower field strenghts and/or
fast EPIs, it would also be handy to allow some non-lineairity, similar as
in normalization, possibly also taking into account spatial biases (for example
due to parallel imaging), and probably other effects.
For example, EPI
images are always more or less spatially distorted as compared to a T1 weighted
structural image, yielding a poor match when using rigid-body registration.
Especially
at higher field strengths (7T) this is real issue, or for high resolution fMRI
(<3 mm^3). Non-rigid coregistration would be quite handy in such cases, for
example modeling spatial distortions using DCT spatial basis functions, like in
normalization.
Are there toolboxes
availbale to do non-rigid coregistration for matlab/spm, or are some of you
willing to share some home brew code or experiences?
Thanks
in advance,
Bas
--------------------------------------------------
Dr. S.F.W.
Neggers
Division of Brain Research
Rudolf Magnus Institute for
Neuroscience
Utrecht University Medical Center
Visiting : Heidelberglaan
100, 3584 CX Utrecht
Room
B.01.1.03
Mail : Huispost B.01.206, P.O. Box 85500
3508 GA Utrecht, the Netherlands
Tel
: +31 (0)88 7559609
Fax : +31 (0)88
7555443
E-mail :
[log in to unmask]--------------------------------------------------
