Hi - it's hard to tell what the problem is. I would first try
removing all the first-level reg_standard directories and try running
the higher-level FEAT again; this will re-create the higher-
resolution upsampled data in reg_standard from scratch.
It may be that some of your files didn't come back from backup
correctly if they were symbolic links (e.g. to highres or standard
space images). You should check inside the reg subdirectories for
broken links.
If this doesn't work then probably the best thing to do is to re-run
things from scratch.
Cheers, Steve.
On 19 Oct 2006, at 07:11, Nils Schneider wrote:
> Hi,
>
> This is the situation: The data was already analyzed with FSL 3.2,
> but some
> of the higher analyses got lost because of a hard drive failure.
> All the
> first level analyses were recovered from backups, but two got
> corrupted. I
> was able to repair one first level feat, but the second one I had
> to run
> again with 3.3. It looks good as far as I can see.
>
> Now this first level feat(produced with 3.3 as the only one) throws
> an error
> when sticked into a group analysis together with 3.2 first level
> feats. The
> first featwatcher says:
>
> ------
> /usr/local/fsl/bin/overlay 0 0 ../bg_image -a maskunique 0.9 15
> maskunique_overlay
>
> /usr/local/fsl/bin/slicer maskunique_overlay -S 2 750
> maskunique_overlay.png
>
> /bin/cp /usr/local/fsl/etc/luts/ramp.gif .ramp.gif
>
> /usr/local/fsl/bin/avwmaths mask -Tmin mask
> ++ WARNING: nifti_read_buffer(mask.nii.gz):
> data bytes needed = 54157740
> data bytes input = 25865881
> number missing = 28291859 (set to 0)
> -------
>
> And the second featwatcher says:
> -------
> Log directory is: stats
> Setting up:
> No f contrasts
> nevs=1
> ntpts=15
> ngs=1
> nvoxels=0
> Running:
>
>
>
> An exception has been thrown
> Logic error:- detected by Newmat: incompatible submatrix dimension
> -------
>
> All other group analyses are working fine (3.2 first level on 3.3),
> only
> this one is not working. It was working with 3.2 before the hd
> failure. And
> the raw data (which also had to be recovered from backups) looks fine,
> including the highres I'm registering to. I tried both to setting
> up a new
> design.fsf in FSL 3.3 (first level feat looks good) and load the 3.2
> design.feat (first level feat with 3.3 is still looking good).
>
> Any suggestions? Should I switch back to 3.2 for that data? I could
> also
> rerun all the first levels in 3.3, it is just a lot of work...
>
> Thanks,
> Nils
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|