| Author |
 Topic  |
|
|
Bethany
BIAC Alum
37 Posts |
 Posted - Feb 06 2006 : 7:22:14 PM
|
I've had errors with bputil on subjects 31399 and 31412 on the RiskTime.01 experiment. When I try to run bputil it returns this error:
WARNING: not a supported Analyze orientation --
labeling as transverse unflipped (0) anyway...
and the data is no longer usable after normalization.
Apparently bputil can't figure out what orientation the scan is in. Can anyone tell me how to fix this?
Thanks,
Bethany |
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Feb 07 2006 : 12:45:39 PM
|
The warning is somewhat of a red herring -- SPM requires Analyze 7.5 format files, but Analyze 7.5 technically only supports 6 out of the 48 possible combinations of R/L A/P and S/I, so the conversion program is saying that the input data is in one of those 42. However, the .bxh files maintain the correct orientation information, so, assuming you started out with a .bxh file with the correct orientation, I believe bputil reorients the data to the orientation that SPM expects before it runs any of the preprocessing routines. So you can get this warning even in correctly processed data.
In what way is the data unusable after normalization?
|
 |
|
|
Bethany
BIAC Alum
37 Posts |
Posted - Feb 07 2006 : 3:38:10 PM
|
On the theory that a picture is worth a thousand words:
This is what we get after preprocessing. (This is subject 31412, sncrarun005_01.bxh).
|
|
|
|
petty
BIAC Staff
USA
453 Posts |
Posted - Feb 07 2006 : 3:50:00 PM
|
bethany ... you should check your orientation of the raw functionals to see if they were wrong from the beginning. there are a couple of ways to do this, here's one:
in matlab:
temp=readmr('name of func.bxh');
getorientation(temp.info)
-- it most likely should be "lps"
if this is not the case, there is a way to fix it. in golgi:
bxhsetorient --inplace LPS run004_01/run004_01.bxh
(this will set the orientation in that specific bxh to LPS, you can substitute RAS,etc if yours is different) |
|
|
|
jimmy.dias
BIAC Alum
USA
210 Posts |
Posted - Feb 07 2006 : 4:24:19 PM
|
Hey all,
Apparently, the orientation specified in the BXH files are wrong for EPI. Has anyone acquired EPI datasets NOT in the 'lps' orientation? Would it be incorrect for me to use the setorientation flag in pfile2bxh to hardcode it to 'lps'?
Jimmy |
|
|
|
Bethany
BIAC Alum
37 Posts |
Posted - Feb 07 2006 : 5:12:53 PM
|
If I'm reading this right, the raw orientation is apparently "prs" and bputil changes it to "ras" with (apparently) bad results.
I can try resetting it to "lps" orientation although I admit I'm not totally sure what orientation they were aquired in. (It was EPI, btw.) I suppose if I try lps and it still doesn't look right, I'll know. :-) |
|
|
|
Bethany
BIAC Alum
37 Posts |
Posted - Feb 08 2006 : 09:57:47 AM
|
So:
It looks like the anatomicals were in LPS orientation. I changed the functionals to LPS orientation using the commands above.
The resulting (RAS) files look the same as before, though.
Would it help to change BOTH the functionals and anatomicals to RAS before I run bputil, as that seems to be the format it wants to output them in? Or can I change the output format? |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Feb 09 2006 : 11:46:55 AM
|
Just to close out this issue, there was a difference in the way EPI and spiral scans represented the corner voxel locations in the P-files, so the orientations in the .bxh files for EPI data were incorrect. A fix was put in and deployed, so future EPI data should have the correct orientations (in Bethany's case, it was indeed LPS). So if you have affected data, you can use Chris' suggestion (bxh_setorient). The data shown above, after running bxh_setorient and running through bputil again exhibits the correct orientation after pre-processing, and is not grossly expanded outside the range of the image.
Another option is to regenerate your .bxh files on golgi. If you have the pfile header (ending in .pfh), and are in the run directory run004_01, the following:
pfile2bxh run004_01.pfh *.img run004_01.bxh
will recreate the .bxh file, with correct orientation vectors.
If you are still having problems, please let us know. |
|
|
|
Bethany
BIAC Alum
37 Posts |
Posted - Feb 16 2006 : 6:29:50 PM
|
Actually, I just got a batch of 6 runs of EPI files (31467) and of them, the first and the fourth were in prs and the second and the fifth were in lps. (The third and the sixth haven't come across yet.)
I went ahead and recreated the two .bxh files that were in prs. |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Feb 17 2006 : 11:32:06 AM
|
quote:
Originally posted by Bethany
Actually, I just got a batch of 6 runs of EPI files (31467) and of them, the first and the fourth were in prs and the second and the fifth were in lps. (The third and the sixth haven't come across yet.)
Apparently, not all the recon machines were updated with the latest tools, so that's why some of the runs were correct and the others were not. Jimmy's working on getting the right stuff installed. Sorry for the confusion. |
|
|
|
erin.douglas
BIAC Alum
USA
25 Posts |
Posted - Mar 01 2006 : 3:43:40 PM
|
When I type in bputil into golgi, I get the following error:
DBI connect('database=BIAC_Experiment_Information;host=pauli;port=3306','procdata',...) failed: Can't connect to MySQL server on 'pauli' (78) at /usr/local/packages/bite/lib/BIAC/Database.pm line 77
could not connect to database: Can't connect to MySQL server on 'pauli' (78) at /usr/local/packages/bite/lib/BIAC/Database.pm line 80
BIAC::Database::new called at /usr/local/bin/bputil line 90
ERROR! could not load database object
|
|
|
|
jamie.morris
BIAC Alum
59 Posts |
Posted - Mar 01 2006 : 3:45:42 PM
|
| When I try, I get nothing at all. The cursor just continues to blink. |
|
|
|
erin.douglas
BIAC Alum
USA
25 Posts |
Posted - Mar 01 2006 : 3:47:16 PM
|
| my curser blinked for about 30 minutes before it gave me the error... |
|
|
|
jimmy.dias
BIAC Alum
USA
210 Posts |
Posted - Mar 01 2006 : 3:57:43 PM
|
Hey all,
This error has been fixed.
Jimmy |
Edited by - jimmy.dias on Mar 01 2006 4:10:21 PM |
|
|
|
kchen
Starting Member
USA
1 Posts |
Posted - Apr 03 2006 : 6:21:12 PM
|
When I run bputil I get the following error:
1) Experiment Name Selected.
Experiment Name => PDfMRI.01
Experiment Location => /home/kchen/net/nernst/data1
2) Instruction Set Selected.
SPM99 Time Slice Correction
SPM99 Motion Correction
SPM99 Coreg and Normalize
SPM99 Smoothing
BIAC QA
3) Select Exam.
4) Select Functionals.
5) Select Anatomicals.
6) Select Processing Preferences.
7) Submit Processing Jobs and Exit.
8) Exit.
Enter option (1..8): 3
Checking Func/* for exam matches
Checking Anat/* for exam matches
ERROR! Exams Not Found for Experiment PDfMRI.01!!
Press return to continue...
|
|
|
| |
Topic |
|
Bputil Error
