Any new NIfTI files that were created (from ANALYZE 7.5 data) with flip=1 will
be in the wrong orientation, if the correct setting would actually have been
0. The new files would be in NIfTI format, which encodes the orientation in
the header. The solution is to delete any files that were created, or
modified, while using the wrong settings and start again. If you use a Unix
system, then you can find these with:
find . -name '.*\.hdr$' -mtime +xxx -ls
find . -name '.*\.nii$' -mtime +xxx -ls
where xxx is some number in days. I have no idea how this would be done under
Windows.
Note that if NIfTI format is used throughout, then this is not an issue. If
DICOM files were initially converted using the DICOM button of SPM5, then
they will be in NIfTI format, and the orientation of all the images derived
from them should be fine. The problem only occurs when SPM reads an ANALYZE
7.5 header, and writes out a new header in NIfTI format.
Best regards,
-John
On Tuesday 05 February 2008 15:51, [log in to unmask] wrote:
> We changed the defaults.analyze.flip setting to 0 in our file. The scans
> look fine, the realigned and unwarped scans are fine, but as soon as I
> normalize them, they flip (-2x2x2 instead of 2x2x2, and since we have
> brains with lesions - the lesions are of course on the wrong side) again.
> So in the normalization process something must be going on we don't see or
> we overlook. Do you have any advice or ideas?
>
> Thanks so much again.
>
> Sarah
>
> "If the human brain were so simple
> That we could understand it,
> We would be so simple
> That we couldn't."
> (Emerson M. Pugh)
>
>
> Sarah Marchina
> Dept of Neurology, Palmer 127,
> Beth Israel Deaconess Medical Center / Harvard Medical School,
> 330 Brookline Ave, Boston, MA - 02215
> Phone: (617) 632 8923
> Cell: (617) 642 1191
>
>
> -----Original Message-----
> From: John Ashburner [mailto:[log in to unmask]]
> Sent: Tuesday, February 05, 2008 10:55 AM
> To: Marchina,Sarah T. (BIDMC - Neurology); [log in to unmask]
> Subject: Re: [SPM] Change in Voxel Size
>
>
> That bit of extra information helped thanks. The spm_defaults.m file would
> have been over-written when the updates were installed. If your original
> default settings were different from the usual (ie if you had originally
> set defaults.analyze.flip to 0), then this would have changed the apparent
> orientation of some of the data (those that were not in NIfTI format, or
> that did not have a .mat file). If all your data were in NIfTI format,
> then this would have made no difference. However, if your DATA are still
> in ANALYZE 7.5 format, then (because these headers do not encode the
> handedness of the images) then it would make a difference.
>
> Best regards,
> -John
>
> On Tuesday 05 February 2008 14:54, [log in to unmask] wrote:
> > Hello!
> >
> > I have exactly the same problem as described below. I have been using
> > SPM5 for almost one year now and never had that problem. I just occured
> > recently and we didn't change any of the settings. Could it be something
> > with the update?
> >
> > Thanks a lot for your help!
> >
> > Sarah
> >
> >
> >
> > "If the human brain were so simple
> > That we could understand it,
> > We would be so simple
> > That we couldn't."
> > (Emerson M. Pugh)
> >
> >
> > Sarah Marchina
> > Dept of Neurology, Palmer 127,
> > Beth Israel Deaconess Medical Center / Harvard Medical School,
> > 330 Brookline Ave, Boston, MA - 02215
> > Phone: (617) 632 8923
> > Cell: (617) 642 1191
> >
> >
> > -----Original Message-----
> > From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]On
> > Behalf Of John Ashburner
> > Sent: Tuesday, February 05, 2008 7:27 AM
> > To: [log in to unmask]
> > Subject: Re: [SPM] Change in Voxel Size
> >
> > > I am doing some fMRI analysis and have noticed that after normalization
> > > and smoothing I have a voxel size 2x2x2 but after estimation I notice
> > > when viewing my spmT image that the voxel size of some of my images
> > > have changed to -2x2x2. Has anyone else experienced this problem?
> > > Will it cause problems in my data or can I combine both image types
> > > together (ie treat 2x2x2 and -2x2x2 as the same)? Thanks
> >
> > This sounds worrying as it is the sort of thing that shouldn't happen
> > with correct use of SPM. How did you determine the voxel sizes? Did the
> > flip settings change at any point? Which version of SPM?
> >
> > If you do a search through the archives for keywords "neurological and
> > radiological" or maybe "flip", then you should find some of the related
> > answers. I would suggest narrowing the search to those more recent
> > answers (to avoid most of the old SPM99 stuff).
> >
> > Best regards,
> > -John
|