OK, I can replicate with this data. Let me see what's going on.
Sebastian Moeller wrote:
Hi Doug,
On Feb 19, 2010, at 7:16 AM, Douglas N Greve wrote:
Hmmm, da stimmt etwas nicht.
Ganz genau, das ist was faul im Staate Daenemark... :)
I cannot replicate this. How was the original f.nii created? Can you send me f.nii?
The f.nii is in the filedrop (https://www.nmr.mgh.harvard.edu/facility/filedrop/showgroup/18898/1/a0629c2e...). I this case it was created by mri_vol2vol using a short as inputs. I initially thought that maybe mri_vol2vol does something funky (I use --mov --targ --o and --reg switches) even though the volumes look fine with jip-display. But I can reproduce this with importing DICOMs as .nii (short) and then run mc-sess (afni2) the resulting output is of .nii (float) and does not display as --targ in tkregister2 but as moveable. Also if I mri_convert the f.nii (the short, working one) to -odt float the resulting .nii float has the reported problem. All the data is from a v17 TIM trio maybe there is something wonky in the headers... Anything I should try to figure this out?
Best Sebastian
doug
Sebastian Moeller wrote:
Dear Freesurfer Experts,
I am a happy user of freesurfer for a number of years now. And finally I managed to switch over from v3.0.5 to 4.5.0. Most of this switch was pretty straightforward only slowed down by a few self-inflicted complications. One of the consequences of switching to V4 is that most of my functional 4D volumes now are floats instead of the short we used before. And regarding the floats, I just encountered a weird issue with tkregister2 (from freesufer-linux-centos4_x86_64-stable-pub-v4.5.0-full). I try to register 2 uncompressed single file nifti files, and the target does not appear only the moveable (when I hit the target button on the gui I just the image window turns black). Interestingly this also happens when I load the same volume as target and moveable:
tkregister2 --regheader --reg tmp.reg --targ ./bold/003/f.nii --mov ./bold/003/f.nii -> nicely visible mov invisible targ, also fmo-targ does not help.
however if I convert the file to int or short the targ becomes visible again: mri_convert -i /bold/003/f.nii -o /bold/003/fi.nii -odt int tkregister2 --regheader --reg tmp.reg --targ ./bold/003/fi.nii --mov ./bold/003/fi.nii -> nicely visible mov and targ, also: tkregister2 --regheader --reg tmp.reg --targ ./bold/003/fi.nii --mov ./bold/003/f.nii -> nicely visible mov and targ,
but note: mri_convert -i /bold/003/fi.nii -o /bold/003/ff.nii -odt float tkregister2 --regheader --reg tmp.reg --targ ./bold/003/ff.nii --mov ./bold/003/ff.nii -> nicely visible mov invisible targ,
So it looks like tkregister2 does not really like floats. Is there any thing I can do to work around this problem short of mri_converting all templates to temporary ints? I would be happy to help debug this issue, so please let me know what I can do to help.
best regards & thanks for all the cool tools in freesurfer Sebastian
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html