| T O P I C R E V I E W |
| Bethany |
Posted - Jul 09 2007 : 12:44:29 PM
In RiskTime.02, I have 5 task runs and a 6th run intended to localize a particular brain area. Right now I have FSL output combining across all 5 runs for a given subject, and seperate FSL output for the 6th (localizing) run. I'd like to use showsrs2 to put them all on the subjects brain and see whether they overlap.
I'm stymied. I've tried what seems like the most obvious combination (the subject's anatomical file anat1_brain.nii.gz as the base with \across_runs.gfeat\cope3.feat\stats\tstat1.nii.gz and mult_output.feat\stats\tstat1.nii.gz overlays -- the mult_output is the analysis of the 6th run). I've also tried some other combinations.
It seems like whatever I do, I get the following errors: First, a window pops up complaining, "Interpolation factor different in x & y dimensions! Did you really mean to do this???" Then if I click OK, showsrs2 crashes and the following error appears in the Matlab editor:
??? Error using ==> showsrs2
Error using ==> showsrs2
Error using ==> showsrs2>local_parseInputs
Overlay interpolation must use an integer scale factor
I'm not sure if I have to re-orient some files, use a different combination of files, or what. Help? |
| 6 L A T E S T R E P L I E S (Newest First) |
| Bethany |
Posted - Jul 09 2007 : 6:53:16 PM
Thanks, that seems to have fixed it. |
| syam.gadde |
Posted - Jul 09 2007 : 1:48:32 PM
Good catch, Chris. If this is the case, you can also have FSL automatically standardize all the statistics for that 6th run FSL output by running featregapply on the .feat directory. First make sure you are getting all the FSL variables set by running this:
. /usr/local/packages/fsl/etc/fslconf/fsl.sh
(don't forget the initial dot [.])
And then just run:
featregapply /my/dir.feat
Registered statistics go to the reg_standard subdirectory. |
| petty |
Posted - Jul 09 2007 : 1:42:54 PM
ahh, i get it. then the error is that your localizer is not in standard space, since you didn't have to combine it across any runs ... whereas this automatically happened when you made your single subject average across your other runs.
you could use flirt to output your stats in standard space.
/usr/local/packages/fsl/bin/flirt -ref ${INDIR}/reg/standard -in ${INDIR}/stats/tstat# -out ${OUTDIR}/newTstat -applyxfm -init ${INDIR}/reg/example_func2standard.mat |
| syam.gadde |
Posted - Jul 09 2007 : 1:40:34 PM
That's strange. Your FSL statistics outputs from second- and higher-level analyses should be in standard space, and should be labeled as LAS, and should have the same resolution (91x109x91) as the SPM brain Chris pointed to (the FSL template brains are similar). Did you specify a non-default brain as a standard brain during your first-level analyses? |
| Bethany |
Posted - Jul 09 2007 : 1:30:47 PM
No, I'd tried the bg_image.nii.gz file, too. When I do that, I get a window that pops up saying, "Overlay orientation flags "LPS", "LPS" are different from base orientation flag "LAS". Displayed orientation labels may be incorrect." When I click OK this is followed by the "Interpolation factor different in x & y dimensions!" window and then the same error message in the Matlab editor as before.
The spm brain gives me the same errors as the bg_image.nii.gz brain. |
| petty |
Posted - Jul 09 2007 : 1:08:52 PM
either use the subjects standard space brain: the bg_image.nii.gz
or the spm brain from the same space:
readmr(fullfile('\\Gall','Programs','MATLAB','spm5','canonical','single_subj_T1.nii')); |
|
|
