Nick,
I thought I would send your very precise fix to the list in case anyone
else just might run into the a similar issue.
I'd also like to thank you and Bruce profusely for all your great
help!
The longitudinal subject ran through successfully last night.
Sky
At 04:01 PM 4/17/2008, Nick Schmansky wrote:
On Thu, 2008-04-17 at 15:20
-0700, Sky Raptentsetsang wrote:
> Hi Nick,
>
> Thanks for all the info.
> I'm fairly certain that both the subjects 005_S_0610X1a and
> 005_S_0610X2a have been run with FSv4.
the problem stems from:
1) using the -long flag, which skips the normal talairach code
and
instead copies the talairach.xfm file from tp1, and doesnt run the
newer
talairach_avi code, which is in v4 freesurfer, so if a subject is
processed with -long -talairach, then the file
<file>/mri/transforms/talairach_avi.log doesnt get created, and it
is
*that* missing file that is the source of the error you are
encountering. that file is created if you run recon-all -s
<subj> -
talairach (notice the missing -long flag).
2) v4.0.2 doesnt check for the existence of that
talairach_avi.log
file, where v4.0.3 does
so, you have your choice of solutions as specified in my prior
email.
are none of them working as solutions? if not, i will need the
recon-
all.log files indicating the error.
hopefully one of the options will work.
> At 01:48 PM 4/17/2008, Nick Schmansky wrote:
> > Thanks, but it looks like I don't need them, as I found the
problem.
> >
> > The solution is either:
> >
> > - include the -notal-check flag in recon-all
> >
> > OR
> >
> > - run recon-all -s subj -talairach
-tal-check on all the
> > subjects
> > which do not have the file
<subj>/mri/transforms/talairach_avi.log
> >
> > OR
> >
> > - modify the recon-all script to change this
line:
> >
> > # now run Avi's QA check on the results found in
talairach_avi.log
> > if ( ! $UseMincMritotal) then
> >
> > to this:
> >
> > # now run Avi's QA check on the results found in
talairach_avi.log
> > set avilog =
($subjdir/mri/transforms/talairach_avi.log)
> > if ( ! $UseMincMritotal && -e $avilog)
then
> >
> > so that a check is made for the necessary log file prior to
running
> > the
> > check (this change appears in v4.0.3).
> >
> > This twister in all this was a triple combo of:
> >
> > 1) using -long, which transfers the .xfm file from the tp1
subject
> > to
> > subj
> > 2) having subjects processed by an older freesurfer which
didnt
> > produce
> > the newer talaiarch_avi.log file
> > 3) the missing bit of code in recon-all which checks for
the
> > existence
> > of that log prior to trying to use it.
> >
> > These three things were necessary for this really tricky
problem to
> > take
> > place. Also, I've updated recon-all to output a better
error
> > message at
> > the stage where it was failing for you.
> >
> > Nick
> >
> >
> >
Hi Nick,
I'm in the midst of running ~100 longitudinal ADNI subjects through
Freesurfer and I am getting an error during the talairach registration. I
was going to post the error to the list but I realize that someone else
(ccongle) posted there with a similar issue. Although my error looks a
bit different:
[sky@sirius subjects]$ recon-all -long -tp1
005_S_0610X1a -subjid 005_S_0610X2a
-autorecon1
...
...
...
...
Fri Apr 4 11:35:57 PDT 2008
mri_nu_correct.mni done
#--------------------------------------------
#@# Talairach Fri Apr 4 11:35:57 PDT 2008
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X2a/mri
mri_concatenate_lta -invert1 -tal
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X1a/mri/orig.mgz
/opt/freesurfer4.0.2/average/mni305.cor.mgz
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X1a/mri/transforms/005_S_0610X1a_to_005_S_0610X2a.lta
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X1a/mri/transforms/talairach.xfm
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X2a/mri/transforms/talairach.auto.xfm
invert the first LTA before applying it
Talairach xfrm src file is
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X1a/mri/orig.mgz
Talairach xfrm dst file is
/opt/freesurfer4.0.2/average/mni305.cor.mgz
volume geometry:
extent : (256, 256, 256)
voxel : ( 1.0000, 1.0000, 1.0000)
x_(ras) : (-1.0000, -0.0000, 0.0000)
y_(ras) : ( 0.0000, 0.0000, -1.0000)
z_(ras) : ( 0.0000, 1.0000, 0.0000)
c_(ras) : (-2.3505, 27.9458, -10.1200)
file :
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X1a/mri/orig.mgz
volume geometry:
extent : (256, 256, 256)
voxel : ( 1.0000, 1.0000, 1.0000)
x_(ras) : (-1.0000, -0.0000, 0.0000)
y_(ras) : ( 0.0000, 0.0000, -1.0000)
z_(ras) : ( 0.0000, 1.0000, 0.0000)
c_(ras) : (-2.3505, 27.9458, -10.1200)
file :
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X1a/mri/orig.mgz
Read individual LTAs
Warning: dst volume of lta1 doesn't match src volume of lta2
Volume geometry for lta1-dst:
Volume geometry for lta2-src:
Combining the two LTAs to get a RAS-to-RAS from src of LTA1 to dst of
LTA2...
mri_concatenate_lta successful.
#--------------------------------------------
#@# Talairach Failure Detection Fri Apr 4 11:35:57 PDT 2008
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X2a/mri
talairach_afd -T 0.005 -xfm transforms/talairach.xfm
talairach_afd: Talairach Transform: transforms/talairach.xfm OK
(p=0.7532, pval=0.6675 >= threshold=0.0050)
awk -f /opt/freesurfer4.0.2/bin/extract_talairach_avi_QA.awk
/cind/00/ADNI_FSvSNT/Freesurfer/subjects/005_S_0610X2a/mri/transforms/talairach_avi.log
Linux sirius.v21.med.va.gov 2.6.20-1.2307.fc5 #1 SMP Sun Mar 18 20:58:09
EDT 2007 x86_64 x86_64 x86_64 GNU/Linux
recon-all exited with ERRORS at Fri Apr 4 11:35:58 PDT
2008
Sorry to pester you. I would really appreciate the help.