Re (2), the c* files were certainly crisper for me, though obviously
you will want to run your own tests.
Re (3), I might have been unclear in formulating my question. I always
generate flow fields from rc* files, but I was unsure if flow
fields should be applied
to images (eg T1 or Jacobians from HDW) aligned with rc* files or
images aligned to c* files. The
guide seems to imply rc* files, but it seems that in actuality it is c*
MCLAREN, Donald wrote:
[log in to unmask]"
type="cite">(1) The rc*.nii images are rigidly aligned (6 parameters)
to the MNI template space. This is done by applying the rigid
component of the transformation stored in the *_sn.mat files.
(2) In theory, the c*.img will produce crisper segments after
warping because they have one less resampling step in the data
processing stream. I have not yet inspected the differences between
using rc* and c* files, yet; but it is on my to do list.
(3) Creating flow field steps should always use the rc* files.
On Fri, Jul 24, 2009 at 7:50 PM, Neil
Chatterjee <[log in to unmask]>
I'm trying to understand the Initial Import step for DARTEL a little
better, and I'm hoping someone here can help me out.
Compared to the gray matter from the segment algorithm (c1*.nii), the
imported segments (rc1*.nii) seem to be aligned slightly differently
(see attached "misalignment.png"). The DARTEL guide indicates that
some sort or rigid tissue alignment is happening in the import step, so
this isn't too surprising, though I am kind of lost as to what is being
aligned to what. Are the tissue classes rigidly aligned to some sort
of SPM template?
Referring to the imported images, the DARTEL guide also says that "the
headers of these files encode two affine transform matrices, so DARTEL
tools are still able to relate their orientations to those of the
original T1-weighted images". Am I then correct in understanding that
the flow fields map the transformation from the original space of the
T1 images to the group template? To test it out, I warped a set of T1
scans before and after coregistration to the rc1*.nii images (see
"avg_anat.png" attached) using flow fields (u_rc1*.nii) from the dartel
imported GM segments (rc1*.nii). anat.nii shows the mean of the warped
original T1 images, and ranat.nii shows the mean of T1 images that had
been coregistered to the rc1*.nii images prior to warping (rc1*.nii =
ref, c1*.nii = source, T1 = other) . anat.nii looks much cleaner than
ranat.nii, implying that flow fields map from original space to group
Am I to take it then, that when creating modulated GM segments for
t-tests, I should use the original segments and NOT the segments from
DARTEL's initial import? Comparing averages of warped GM segments from
before (c1*.nii) and after (rc1*.nii) import, it looks like it's the
c1*.nii that warp together better (see attached "avggray.png"). If so,
then perhaps there is a slight mistake in the DARTEL guide, which
indicates that the rc1*.nii files (not the c1*.nii) should be selected
when running "Create Warped". Or is it that "Create Warped" and the
Deformations toolbox are not doing the same thing?
Any clarification would be much appreciated. This is using an updated
SPM8 on Matlab 2008b.
Stanford Systems Neuroscience and Pain Lab
[log in to unmask]
Best Regards, Donald McLaren
University of Wisconsin - Madison
Neuroscience Training Program
Office: (608) 265-9672
Lab: (608) 256-1901 ext 12914
This e-mail contains CONFIDENTIAL INFORMATION which may contain
PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and
which is intended only for the use of the individual or entity named
above. If the reader of the e-mail is not the intended recipient or the
employee or agent responsible for delivering it to the intended
recipient, you are hereby notified that you are in possession of
confidential and privileged information. Any unauthorized use,
disclosure, copying or the taking of any action in reliance on the
contents of this information is strictly prohibited and may be
unlawful. If you have received this e-mail unintentionally, please
immediately notify the sender via telephone at (608) 265-9672 or email.
Neil R. Chatterjee
Stanford Systems Neuroscience & Pain Lab
[log in to unmask]