| Author |
 Topic  |
|
syam.gadde
BIAC Staff
USA
421 Posts |
 Posted - Oct 31 2006 : 11:54:45 AM
|
| Your design.fsf file indicates that you have enabled a main structural image (for registration to the standard brain), but the field indicating the structural image file is empty. |
 |
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Oct 31 2006 : 2:47:58 PM
|
| Found it; thanks :) |
|
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Nov 14 2006 : 3:22:19 PM
|
Could problems with old pfiles cause func_2_highres and highres_2_standard to be inferior-superior flipped? Should the same fix (pfile2bxh runXXX_XX.pfh V*.img runXXX_XX.bxh) be applied?
Thanks,
Dharol |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Nov 14 2006 : 3:28:13 PM
|
| It's possible, but it's hard to say without seeing the data. I suggest re-running pfile2bxh but creating a new .bxh file (not overwriting the old one) and seeing if the "direction" vectors change at all. If not, then this is not the cause of the problem. |
|
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Nov 14 2006 : 5:18:20 PM
|
Okay. We ran pfile2bxh and the problem wasn't fixed.
So the func to highres is flipped on the inferior-superior axis, as is the highres to standard. Below is highres to standard run without using the pfile fix.
Interstingly, the func to standard ends up looking right.
So is this a problem that FLIRT ends up fixing by the last step, or do we need to fix it, and any suggestions as to how?
Fyi, this data was collected April 2004 to March 2005 on 4T.
Thanks,
Dharol
|
Edited by - tankersley on Nov 14 2006 5:20:13 PM |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Nov 14 2006 : 5:27:42 PM
|
| I would have to have access to the data to know for sure what is going on. It sounds like maybe the anatomical might be mislabeled (if the func2standard registration is correct). Does the anatomical come up with the correct labels if you open it in showsrs2? You can also try viewing it using fslview (one of the options in the FSL GUI). |
|
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Nov 14 2006 : 5:52:32 PM
|
I have lost access to Nernst from Golgi and so can't look at the fslview;
in showsrs, the x-dimension values get larger as you go from right to left. I don't know if you can tell from this.
 |
|
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Nov 15 2006 : 3:43:59 PM
|
FYI:
The standard brain that FSL normalizes all your data to is
avg152T1_brain.hdr
and can be found in:
\\Gall\Programs\User_Scripts\fsl_standard |
|
|
|
Dichter
BIAC Faculty
190 Posts |
Posted - Nov 15 2006 : 3:53:59 PM
|
I am also having trouble with FSL registration:
This is the "example_func to standard " registration from data acquired this morning (i.e., with the most recent pfile to bxh conversion). I am using an ORIENT.mat file.
My functional bxh header has this orientation information:
<?xml version="1.0"?>
<!-- This is a BXH (BIAC XML Header) file. -->
<bxh xmlns:bxh="http://www.biac.duke.edu/bxh" xmlns="http://www.biac.duke.edu/bxh" version="1.0">
<datarec type="image">
<!--AUTOGEN: Orientation is oblique axial -->
<rasorigin>-118.125 104.508865371346 -83.8343312218785</rasorigin>
<dimension type="x">
<units>mm</units>
<size>64</size>
<origin>104.508865371346</origin>
<gap>0</gap>
<spacing>3.75</spacing>
<direction>0 -0.982088208198547 0.188421785831451</direction>
</dimension>
<dimension type="y">
<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="z">
<units>mm</units>
<size>34</size>
<origin>-83.8343312218785</origin>
<gap>0</gap>
<spacing>3.80121279988374</spacing>
<direction>0 0.188137769699097 0.982142627239227</direction>
</dimension>
<dimension type="t">
<units>ms</units>
<size>340</size>
<origin>0</origin>
<gap>0</gap>
<spacing>1</spacing>
</dimension>
My anatomical bxh header has this orientation information:
<?xml version="1.0"?>
<!-- This is a BXH (BIAC XML Header) file. -->
<bxh xmlns:bxh="http://www.biac.duke.edu/bxh" xmlns="http://www.biac.duke.edu/bxh" version="1.0">
<datarec type="image">
<!--AUTOGEN: Orientation is oblique axial -->
<rasorigin>119.531254 105.8899244375 -61.81858225</rasorigin>
<dimension type="x">
<units>mm</units>
<size>256</size>
<origin>119.531254</origin>
<gap>0</gap>
<spacing>0.937492</spacing>
<direction>-1 -0 0</direction>
</dimension>
<dimension type="y">
<units>mm</units>
<size>256</size>
<origin>105.8899244375</origin>
<gap>0</gap>
<spacing>0.9375</spacing>
<direction>-0 -0.982094 0.188392</direction>
</dimension>
<dimension type="z">
<units>mm</units>
<size>68</size>
<origin>-61.81858225</origin>
<gap>0</gap>
<spacing>1.90121430638578</spacing>
<direction>0 0.188410614216783 0.982090342305867</direction>
</dimension>
Thanks for any help...
-Gabriel
|
Gabriel S. Dichter, PhD
UNC Departments of Psychiatry & Psychology
http://www.can.unc.edu/ |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Nov 15 2006 : 4:00:34 PM
|
quote:
Originally posted by Dichter
My functional bxh header has this orientation information:
<direction>0 -0.982088208198547 0.188421785831451</direction>
<direction>1 0 0</direction>
<direction>0 0.188137769699097 0.982142627239227</direction>
That orientation info looks unusual (and perhaps wrong). I think we've had examples of this in the past. Can you let me know what exam number this is? |
|
|
|
Dichter
BIAC Faculty
190 Posts |
Posted - Nov 15 2006 : 4:03:09 PM
|
| 20061115_46256 and 20061115_46258, both from today. Thanks. |
Gabriel S. Dichter, PhD
UNC Departments of Psychiatry & Psychology
http://www.can.unc.edu/ |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Nov 15 2006 : 4:23:07 PM
|
quote:
Originally posted by Dichter
20061115_46256 and 20061115_46258, both from today. Thanks.
The orientation labels are definitely wrong, and this
matches with data we've seen in the past. A temporary
solution until we figure out what is going on here is
to manually set the orientation in the .bxh file to what
you suspect it really is using bxhsetorient. The current
,bxh file says PRS, but I think it is either LPS or RPS
(probably LPS):
bxhsetorient LPS run005_01.bxh run005_01-new.bxh
And then use the new .bxh file.
|
|
|
|
Dichter
BIAC Faculty
190 Posts |
Posted - Nov 15 2006 : 4:25:43 PM
|
| Thank you. How would we verify that it should indeed be LPS? I wouldn't want to guess with something like this. |
Gabriel S. Dichter, PhD
UNC Departments of Psychiatry & Psychology
http://www.can.unc.edu/ |
|
|
|
syam.gadde
BIAC Staff
USA
421 Posts |
Posted - Nov 15 2006 : 4:32:06 PM
|
quote:
Originally posted by Dichter
Thank you. How would we verify that it should indeed be LPS? I wouldn't want to guess with something like this.
Other evidence in the P-file indicates LPS (i.e. the
first coordinate is elsewhere labeled as
[ 120 124.27070617675 -53.4404602050781], so starting
on the right side of the brain). |
|
|
|
tankersley
BIAC Alum
USA
143 Posts |
Posted - Nov 20 2006 : 12:29:19 PM
|
We're trying to run FSL on this old data set and having registration problems. I think the ANAT header file is messed up, but am not sure how to fix it.
<?xml version="1.0"?>
<!-- This is a BXH (BIAC XML Header) file. -->
<bxh xmlns:bxh="http://www.biac.duke.edu/bxh" xmlns="http://www.biac.duke.edu/bxh" version="1.0">
<datarec type="image">
<!--AUTOGEN: Orientation is axial -->
<dimension type="x">
<units>mm</units>
<size>256</size>
<origin>119.53125</origin>
<gap>0</gap>
<spacing>0.9375</spacing>
<direction>-1 -0 0</direction>
</dimension>
<dimension type="y">
<units>mm</units>
<size>256</size>
<origin>119.53125</origin>
<gap>0</gap>
<spacing>0.9375</spacing>
<direction>-0 -1 -3.1e-05</direction>
</dimension>
<dimension type="z">
<units>mm</units>
<size>68</size>
<origin>-73.59635053125</origin>
<gap>0</gap>
<spacing>1.89999998507463</spacing>
<direction>0 0 1</direction>
</dimension>
Can I go in and manually put
1 0 0
0 1 0
0 0 1
for direction specifications for x y and z respetively, and maybe this fix it, or is something else going on?
Sorry if this is the second time I've posted this.
The registration is getting mixed up on the func to highres stage, but I've posted the highres to standard because it's easier to see.
Also, the registration is backwards from func to highres and from highres to standard, but then loooks good from func to standard. But FSL probably needs the correct func to highres for intermediate processing, right?
Thanks,
Dharol |
|
|
|
Topic |
|
FSL registration
