Show all posts by user
Dear AFNI users-
We are very pleased to announce that the new AFNI Message Board framework is up! Please join us at:
https://discuss.afni.nimh.nih.gov
Existing user accounts have been migrated, so returning users can login by requesting a password reset. New users can create accounts, as well, through a standard account creation process. Please note that these setup emails might initially go to spam folders (esp. for NIH users!), so please check those locations in the beginning.
The current Message Board discussion threads have been migrated to the new framework. The current Message Board will remain visible, but read-only, for a little while.
Sincerely,
AFNI HQ
History of AFNI updates
Rick,
One of our studies has exactly this problem, that one of the populations moved significantly more in the scanner than either of the other two. We've still been able to draw conclusions from the other two groups, but we've had to ignore this entire group because we didn't have enough usable subjects, and because we were concerned about the effects of the group differences
by
Isaac Schwabacher
-
AFNI Message Board
Gang,
We have two nearly-identical studies using two different populations of interest, and one of them is in more of a time crunch than I'd realized, so objective c) is that I need a solution I can hack together and run in a few days to a week, at least for one of the two studies. Since I don't have any experience with surface analyses other than playing with FreeSurfer a little, I
by
Isaac Schwabacher
-
AFNI Message Board
Hi all,
I would like to use PPI to determine how a painful stimulus changes connectivity between various regions of interest that are active during that stimulus. A little playing around has shown that I get the best picture by taking each seed ROI to be a very small sphere around the peak voxel in the relevant anatomically-determined mask, but unfortunately I only have one run of the painful
by
Isaac Schwabacher
-
AFNI Message Board
Andrew,
Yeah, I thought about it and it's almost certainly not a problem. Since you're using bash, I think that what will happen is that a string of the form "foo[12]" (quoted here for clarity; quotes in the shell would stop this behavior) will be replaced by "foo1" or "foo2" or (more problematically) "foo1 foo2" (interpreted by the shell as t
by
Isaac Schwabacher
-
AFNI Message Board
Andrew,
Are you sure this isn't a shell issue? I notice you have unescaped quotes inside the variable list1, so that your subbrik indices are not being escaped the way you want them to.
Isaac
by
Isaac Schwabacher
-
AFNI Message Board
Rick,
I can upload both the mask dataset from the original question and the anatomical from this one, as well as various datasets produced from these. I'll need to deface the anatomical first, though, and I think it would be a good idea to use xxd for that rather than an AFNI program that might disturb the header.
Isaac
PS:
AFNI version is
Version AFNI_2011_12_21_1014
[]
by
Isaac Schwabacher
-
AFNI Message Board
Well I think I figured out at least one thing that @Align_Centers was warning me about. It follows symbolic links, even if they point to datasets it's already processed!
by
Isaac Schwabacher
-
AFNI Message Board
Hi again,
The strange mask resampling issue ended up on the back burner because the datasets did look like they were in the right place, but I just discovered a much more problematic variation of this discrepancy in an actual subject's data, and I'm still trying to make sure that my fix won't break anything else. This subject's dataset origin is way off in the wrong place,
by
Isaac Schwabacher
-
AFNI Message Board
Hi,
I'm trying to figure out how to do cluster correction on group data, and I'm confused by the admonition not to use 3dFWHMx on the output of 3dDeconvolve or 3dREMLfit. My inclination is to run 3dMEMA on my data, subtract the results from each beta map (i.e., subtract the estimate for the appropriate group), and then run 3dFWHMx on the residuals and feed the results to 3dClustSim.
by
Isaac Schwabacher
-
AFNI Message Board
Liana,
I'd be willing to bet that there are exactly 94377 voxels that are zero across all subjects.
by
Isaac Schwabacher
-
AFNI Message Board
Rick,
We created these datasets with the intention of having them aligned properly, but we're not sure we succeeded in doing that. One of the datasets is created from our data and given 4x4x4 voxels with adwarp; the other is a ROI mask created by resampling an atlas to 4x4x4 resolution with 3dresample -dxyz 4 4 4. I tried redoing the mask by using -master and it gave me a dataset that a
by
Isaac Schwabacher
-
AFNI Message Board
Andrew,
3dDeconvolve -help has this to say about your problem:
***N.B.: If NO run at all has a stimulus of a given class, then you
must have at least 1 time that is not '*' for -stim_times_*
to work (so that the proper number of regressors can be set
up). You can use a negative time for this purpose, which
will produce a w
by
Isaac Schwabacher
-
AFNI Message Board
Honored and exalted Emperor Zhark,
We have received great bounty but also ill omen from Your most magnificent software, 3dcalc:
*+ WARNING: Template space of dataset /study/fmfmri/data/group_processed/PPI_PAG_pain_MEMA+tlrc does not match space of first dataset
Continuing anyway, but be sure the datasets really do match
In our pathetic attempts to rectify our errors, we have learned tha
by
Isaac Schwabacher
-
AFNI Message Board
Andrew,
I wanted to do this a while ago because I noticed that 3dREMLfit seemed to be numerically unstable in the presence of zero columns, and I eventually ended up writing a little python script to strip zero columns out of the X-matrix. Since it was a quick-and-dirty fix, you'll probably need to clean it up a bit before it works (for instance, I hard coded in the path to the AFNI bina
by
Isaac Schwabacher
-
AFNI Message Board
Not that I know of, but you can click the "New" button (near "Done") and create a second AFNI window. If in this new window you choose an underlay in the same space as the original (possibly the same image), the two windows will be locked together and clicking around in one window will move the crosshairs in both. This way you can look at two overlays over the same underlay,
by
Isaac Schwabacher
-
AFNI Message Board
Thanks!
...I feel really sheepish about missing a file named README.
by
Isaac Schwabacher
-
AFNI Message Board
Apropos of PESTICA: I haven't been able to find good documentation on the process. Is there a help file or FAQ hiding somewhere?
Isaac
by
Isaac Schwabacher
-
AFNI Message Board
Ziad,
Having more knots in your TENT(b,c,n) than there are TRs in its duration does in fact kill 3dDeconvolve IF all the stimuli fall on integral multiples of TR, because you are using n regressors to model a space of spanned by floor((c-b)/TR)+1 translates of your stim file. A demonstration is appended-- note the warning about the inter-knot timing with TENT. I also tried this with '1D
by
Isaac Schwabacher
-
AFNI Message Board
Jon,
That's a different error from the first one you posted (which you in fact fixed by adding -num_glt 21), and it has a different cause. In this case, I would bet that your regressors are linearly dependent. I can't tell what your stim files contain from the command above, but this can happen if you are modelling the baseline as a separate stimulus or have in some other way accou
by
Isaac Schwabacher
-
AFNI Message Board
Vanessa,
Sorry, I didn't mean that you should use -stuff input+orig literally, I meant that you should put in whatever options and input datasets you were originally going to give to 3dDeconvolve, then add ">& 3dDeconvolve.log" at the end.
Isaac
by
Isaac Schwabacher
-
AFNI Message Board
Dave,
My investigations of TENT and CSPLIN suggest that they are really only equivalent if you have exactly one stimulus per TR; that is, when they're being used to imitate the old-style 3dDeconvolve where each TR following the stimulus got its own regressor. For instance, the command below yields an instructive X-matrix. Notice how the first three stimuli produce identical regressors,
by
Isaac Schwabacher
-
AFNI Message Board
You could use
3dDeconvolve -stuff input+orig 2>&1 | tee 3dDeconvolve.log
(if you're using bash)
or
3dDeconvolve -stuff input+orig |& tee 3dDeconvolve.log
(if you're using tcsh)
which would save all the screen output to a file 3dDeconvolve.log in the current directory. If you don't care about seeing the output on the screen and only want the file, you could
by
Isaac Schwabacher
-
AFNI Message Board
Just to check, since 3dSeg still dies with an error and still produces output on every dataset I've tried so far (except one with negative values from interpolation... that it did not like): am I getting all the output files I'm supposed to be getting? Here's what 3dSeg is producing before it dies:
$$$ ls Segsy/
Anat+tlrc.BRIK
Anat+tlrc.HEAD
AnatUB+orig.BRIK
AnatUB+orig.HEA
by
Isaac Schwabacher
-
AFNI Message Board
Ziad,
Very strange... I hadn't checked the output directory since 3dSeg died with an error message, but I was just poking around and discovered a segmented volume. I am confused, but pleased.
Isaac
P.S.: The AFNI version set up by the sysadmin works fine for all other purposes, and is the copy used by the rest of the lab, so I intend to keep using it for actual analysis until it ge
by
Isaac Schwabacher
-
AFNI Message Board