| Author |
 Topic  |
|
tankersley
BIAC Alum
USA
143 Posts |
 Posted - Aug 25 2006 : 10:36:42 AM
|
Couldn't find any topics on FSL and wasnt sure where to post this, but here goes . . .
I have 4 exams in study Reward.01 (20060726_46003; 20060802_46027; 20060803_46030; 20060809_46050) where the functionals are not properly regsitering on the anatomicals.
I've looked at ORIENT.mat, FEAT_{run}.fsf, and bxh header files for these exams, and they all seem normal.
Any other ideas for whhat might cause registration problems?
Thanks :)
Dharol |
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Aug 25 2006 : 11:54:20 AM
|
| I looked at Dharol's data. The .bxh files were generated when the daemon was using an old version of pfile2bxh. The X and Y spacing that small by a factor of 4 (z direction was correct), resulting in a squashed brain. This can be resolved by regenerating the .bxh files from the pfile headers using the currently installed pfile2bxh. |
 |
|
|
Jeff_Browndyke
Junior Member
35 Posts |
Posted - Aug 25 2006 : 12:49:09 PM
|
Do you think the same issue may be going on with our ADCCN.01 data? See recent T1 to inverse spiral thread...
Jeff |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Aug 25 2006 : 12:56:26 PM
|
I haven't looked at your data, so I don't know for sure, but your .bxh files will show an X and Y dimension spacing different than you expect if it is the same as Dharol's.
However if it is recent data, it shouldn't be affected by this particular issue, because the daemon is using the updated tools (since Aug. 9). |
|
|
|
Jeff_Browndyke
Junior Member
35 Posts |
Posted - Aug 26 2006 : 12:56:48 PM
|
Yep. That was the problem. I have data from late 2005 and early this year that we are now pulling together for analysis.
The raw pfiles are not available. As it stands, in each subject directory, I have 334 .img files and one .bxh and one .pfh file. Is the .pfh file sufficient to handle the fix or can I manually edit the .bxh headers before running convert_b?
Thanks,
Jeff |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Aug 26 2006 : 1:29:06 PM
|
The .pfh file is the p-file header, and that (plus the image files) is enough to re-create the .bxh files:pfile2bxh runXXX_XX.pfh V*.img runXXX_XX.bxh Manual editing will probably work too -- but I don't know what else in the file was affected, so it might be safest to regenerate the .bxh files.
(BTW, I'm not familiar with "convert_b", so I don't know what its requirements are) |
|
|
|
Jeff_Browndyke
Junior Member
35 Posts |
Posted - Aug 26 2006 : 2:08:09 PM
|
Syam,
Can pfile2bxh run on systems not networked with BIAC? I'm assuming it's a Matlab routine, right? If so, where can I download it?
If it must be run remotely on the BIAC servers, can you please provide info on how to handle this?
BTW - will this fix the z-dimension origin issues as noted in another thread?
I appreciate your time and assistance.
Jeff |
Edited by - Jeff_Browndyke on Aug 26 2006 2:09:43 PM |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Aug 26 2006 : 4:00:35 PM
|
If you have access to BIAC file servers, you can grab the BXH tools
for Windows from \\gall\programs\bxh\bin . On windows, you have to use bxhabsorb:bxhabsorb --fromtype pfile runXXX_XX.pfh V*.img runXXX_XX.bxh If you don't have access to BIAC file servers, or if you need Linux
versions, let me know, and I can find a way to package it up for you. |
|
|
|
Jeff_Browndyke
Junior Member
35 Posts |
Posted - Aug 26 2006 : 4:37:58 PM
|
Thanks, Syam. I'll give a try.
Jeff |
|
|
|
Jeff_Browndyke
Junior Member
35 Posts |
Posted - Aug 26 2006 : 7:31:25 PM
|
Syam,
Running the most recent pfile to bxh conversion yielded the following:
Original .bxh header -
<rasorigin>118.125 107.135642158531 -98.1547370431426</rasorigin>
<dimension type="x">
<units>mm</units>
<size>64</size>
<origin>118.125</origin>
<gap>0</gap>
<spacing>3.75</spacing>
<direction>-1 0 0</direction>
</dimension>
<dimension type="y">
<units>mm</units>
<size>64</size>
<origin>107.135642158531</origin>
<gap>0</gap>
<spacing>3.75</spacing>
<direction>0 -0.977246545397932 0.212106552251005</direction>
</dimension>
<dimension type="z">
<units>mm</units>
<size>34</size>
<origin>-98.1547370431426</origin>
<gap>0</gap>
<spacing>3.79862833132597</spacing>
<direction>0 0.21210353076458 0.977247297763824</direction>
</dimension>
<dimension type="t">
After running updated version, the resulting .bxh header -
<rasorigin>118.125 107.135642200708 -86.3281379155815</rasorigin>
<dimension type="x">
<units>mm</units>
<size>64</size>
<origin>118.125</origin>
<gap>0</gap>
<spacing>3.75</spacing>
<direction>-1 0 0</direction>
</dimension>
<dimension type="y">
<units>mm</units>
<size>64</size>
<origin>107.135642200708</origin>
<gap>0</gap>
<spacing>3.75</spacing>
<direction>0 -0.977246522903442 0.212106555700302</direction>
</dimension>
<dimension type="z">
<units>mm</units>
<size>34</size>
<origin>-86.3281379155815</origin>
<gap>0</gap>
<spacing>3.79862837345016</spacing>
<direction>0 0.212197795510292 0.977226734161377</direction>
</dimension>
<dimension type="t">
------------
So, it certainly changes the origin in the z-dimension, but the origin on this plane is still quite distant from what it should be. Any thoughts on how to tackle this? Whatever solution you worked out with the Cabezalab is likely to work with my data as it is collected in collaboration with them.
Also, I would appreciate any help in running the pfile2bxh remotely, as it was very cumbersome to attempt the bxhabsorb within the Windows "run..." command structure. Can you run it on the subjects in our /func directory from your location?
Jeff
|
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Aug 26 2006 : 9:45:09 PM
|
| You may wish to contact James Kragel. My only contribution to the "solution" is to provide pfile2bxh/bxhabsorb, which provides (one of) the origins provided by the scanner. I do not believe I have easy access to your data. If you need to run this on several data sets on a Windows machine, perhaps you can create a DOS batch script using cut and paste/search and replace in a text editor. |
|
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Sep 25 2006 : 11:49:55 AM
|
I'm trying to run batch scripts for first level FEAT analysis in fsl. I enter the command in Golgi and the following happens:
[tankersl@golgi: ~/net/katz/data/ExecInfo.09/Data/FSL] $ ./batchFSL 31238
Input Vox2RAS-----------------
-3.99903988838196 0.0402844697237016 0.0779120706219313 121.95130161196
-0.0433365702629088 -3.99645566940308 -0.162675802157025 152.620057493448
0.0761745125055312 -0.163456976413727 3.9959329546664 -63.1834804043174
0 0 0 1
Input FSL Vox2RAS-----------------
4 0 0 0
0 4 0 0
0 0 4.000001569315 0
0 0 0 1
Ref Vox2RAS -----------------
-2 0 0 91
0 2 0 -109
0 0 2 -91
0 0 0 1
Ref FSL Vox2RAS -----------------
2 0 0 0
0 2 0 0
0 0 2 0
0 0 0 1
-----------------
0.99975997209549 -0.0100711174309254 -0.0194780100136995 -30.95130161196
-0.0108341425657272 -0.999113917350769 -0.0406689345836639 261.620057493448
0.0190436281263828 -0.0408642441034317 0.998982846736908 27.8165195956826
0 0 0 1
/mnt/katz/data/ExecInfo.09/Data/FSL/31238/run1_output
And then it just sits there for a couple of hours, producing no files or error messages.
Any one know what this means?
Thanks,
Dharol
|
|
|
|
petty
BIAC Staff
USA
453 Posts |
Posted - Sep 25 2006 : 12:01:08 PM
|
looks like the numbers you posted are from the orientation matrix production ...
if you copy your batchFSL file to your data directory i can look at it: /data/users/tankersl/
looks like the FSL paths aren't loaded, or you may not be logged into the server. |
|
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Oct 31 2006 : 11:25:51 AM
|
I'm getting this registration problem:
My log report error is:
** ERROR: nifti_image_open(highres): bad header info
Error: failed to open file highres
ERROR: Could not open image highres
Image Exception : #22 :: Failed to read volume highres
/usr/local/packages/fsl/bin/convert_xfm -inverse -omat standard2highres.mat highres2standard.mat
Could not open matrix file highres2standard.mat
Cannot read input-matrix
I don't understand exactly where the error is, but the file standard2highres.mat is usually created and copied to the output folder, and it is not being created when I run on this exam.
Any suggestions?
Thanks,
Dharol
|
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Oct 31 2006 : 11:37:55 AM
|
quote:
Originally posted by tankersley
I don't understand exactly where the error is, but the file standard2highres.mat is usually created and copied to the output folder, and it is not being created when I run on this exam.
If it got as far as creating that registration report, I would have expected it to have created the .mat files too. Which registration does that picture refer to? Is that functional to highres?
Could you also post the contents of the .mat files (that exist)? And if you send FSL an initial matrix to get it situated that would be helpful too. |
|
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Oct 31 2006 : 11:44:57 AM
|
I use the ORIENT.mat file you created.
The posted image is a func2standard; func2highres are not being created. I've put everything that is being called on or created when I run this script into
\\Gall\Docs\Temp_Tankers\10-31-Problem
Dharol |
|
|
|
Topic |
|
FSL registration
