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
Page 1 of 2
Pages: 12
Results 1 - 30 of 34
Just want to give you a very Canadian appology for taking so long to say THANK YOU! I'm constantly impressed at how quickly your team implements changes like this.
All the best,
Joseph
by
jdviviano
-
AFNI Message Board
Hi there AFNIs,
Was wondering if there was a way to silence the creation of the 3dDeconvolve.err file, or specify a directory for them to appear in? We've using 3dDeconvolve to analyze data en masse and these files have a habit of mucking up our filesystem.
Thanks,
Joseph
by
jdviviano
-
AFNI Message Board
Hi Rick,
rick reynolds Wrote:
-------------------------------------------------------
> Hi Joseph,
>
> Using AM1 simply applies the modulators to the
> individual
> responses, rather than allowing for a mean
> response along
> with modulation responses. The single regressor
> in AM1
> should be the sum of the modulation regressors
> (including
> t
by
jdviviano
-
AFNI Message Board
Hello AFNIS,
I've been working on an analysis for a group and am having a lot of trouble replicating some published results. Briefly, we have two behavioural measures (A, B) that might amplitude modulate some blocks and we want to see if this is true.
I originally did my analysis with AM2 (looked something like this):
3dDeconvolve \\
-input {input_data} \\
-mask {datadir
by
jdviviano
-
AFNI Message Board
Hi Rick, that worked... in nibabel, I don't reconstruct the header and then have nibabel set the bare minimum fields automatically (updated code pushed to github). On the downside, I lose the extended header (I imagine that is the file history from 3dinfo?) This isn't a dealbreaker, but just curious if there is a better way (where I can retain that history). Also, why is retaining the o
by
jdviviano
-
AFNI Message Board
Hi All,
I've written a python program that does TR-wise scrubbing, code here:
This edits the dimensions of the output file in the time dimension, and changes the relevent fields in the header, before writing out a NIFTI file.
I've used this code in the past without much trouble, but now I am noticing that AFNI programs do not like the output of this program. When I try to do s
by
jdviviano
-
AFNI Message Board
Hello Gentlemen,
I was wondering if any verion controlling of AFNI is being done and how we can best take advantage of it. Currently we have a few versions of AFNI floating around, and we would like to know if there is an archive we can point users to if they would like to try reproducing their results with the exact same software.
How would you recommend going about this best?
Thanks!
by
jdviviano
-
AFNI Message Board
Hi Nick,
We provide onset:length values (no amplitudes). We therefore used stim_times_IM, but the stim_times_AM options with a fixed amplitude give us the same results.
This experiment was not originally designed as a mixed block/effect paradigm, so quite a few of the 'blocks' are pretty short (about 15 seconds). Otherwise, we just have people doing a battery of cognitive tasks a
by
jdviviano
-
AFNI Message Board
Hello Team AFNI,
We are attempting to analyze a mixed effect/block design with 3dDeconvolve and are finding some strange things. When we contrast our condition events with control events, we obtain positive contrasts, but when we contrast the same condition's blocks with control blocks, we obtain negative contrasts.
The patterns of activation are largely overlapping, but the block con
by
jdviviano
-
AFNI Message Board
Thanks for clearing this up. I'll change all of our scripts to use 3dCopy by default... maybe we have had some undetectable problems with the nifti format.
by
jdviviano
-
AFNI Message Board
Hello, is this a new 'feature'? I've been using align_epi_anat.py with binaries compiled two days ago on 10.8 with nifti files.
by
jdviviano
-
AFNI Message Board
Hi Rick, thanks. Yes I am on 10.6. Volumes are very close pre alignment.
Turning off the -useweight option and using -pears gives pretty good results. -pears stopped the freezing, and -useweight caused strange 'pulls' at the edge of the brain.
Now we only have one aberrant sulci floating off into space. Celebration! I will immediately try TT_N27.
Cheers, Joseph
by
jdviviano
-
AFNI Message Board
Dear Zhark,
I love the idea of 3dQwarp, but am having trouble with it. It appears to get stuck pseudo-randomly during the registration process. Two cores sit at 47%, while the rest go to zero. No amount of time brings it out of this sad state.
Here is the complete output from a sample run with -verb. Note that I am using the distribution compiled yesterday. Any help would be amazing, and I
by
jdviviano
-
AFNI Message Board
Hi Bob, don't mean to hijack this thread, but I feel I have 2 similar enough questions that they doesn't need their own.
1) This ringing from the filter... am I correct in saying that if you force the filter to have an FFT length equal to the length of your run, that this ringing will not occur?
2) What to you mean, exactly, when you say that this type of filter does not play well
by
jdviviano
-
AFNI Message Board
Great, thanks so much for your help with this!
Cheers, Joseph
by
jdviviano
-
AFNI Message Board
OK, great, thanks. So thanks to your suggestion and some luck, I've discovered that 3dAutomask gives you all the slices with all 0s.
So I can call this:
3dWarp -matvec_out2in mat_regTarget.aff12.1D -prefix anat_EPI_regTarget.nii.gz -cubic -newgrid $XRES -zpad 0 anat_EPI_brain_1.nii.gz
3dAutomask -prefix anat_EPI_regTarget_mask.nii.gz -clfrac 0.5 -peels 3 anat_EPI_regTarget.nii.gz
A
by
jdviviano
-
AFNI Message Board
Hello team AFNI,
So, I've been using your software with (generally) great success aligning very thin slices of EPI up to T1 space. Normally these slices are oblique and are therefore deobliqued, then motion corrected within sessions, then registered up to T1 space.
These are high resolution files. In order to prevent absolutely massive (and therefore useless) outputs, I'm forced
by
jdviviano
-
AFNI Message Board
Hmm, don't think so. I'm on OSX, and the data is all stored on my internal hard drive.
by
jdviviano
-
AFNI Message Board
Hi, no, I monitor my RAM usage very closely and this typically happens when I still have about 12 GB free. I still can't figure out what is causing this issue. I thought I had resolved it by installing python 2.7.3 and rebooting, but the problem has crept up on me again, even if run on a machine doing nothing else after a fresh reboot and, as I said, there is still many gigs of RAM free.
by
jdviviano
-
AFNI Message Board
Hi there, just an update, script has been running since last night, it is almost finished step one without a hitch (almost all 150 images have been co-registered). You are right, the script always froze (before) when the terminal header said: 3dAllineate.
The only thing that has changed is I reset my computer after updating to python 2.7.3. RAM usage has been approximately 11-12 GB according t
by
jdviviano
-
AFNI Message Board
OK, so I can confirm that upgrading to the new version of python did NOT fix this bug.
To be clear, I am running Python 2.7.3, and align_EPI_ANAT.py 1.35
The problem does seem to be a problem with python threading, as I described earlier.
Any ideas? Thanks!
by
jdviviano
-
AFNI Message Board
Wait a second... before anyone starts attempting to debug this, I might have found the solution, I just can't try it out right now. I followed the python traceback call to here: and I am indeed using Python 2.7.2. I'm going to update my python to 2.7.3 and report back ASAP.
This is what I get when I kill the script:
File "/Applications/afni/align_epi_anat.py", line 3042
by
jdviviano
-
AFNI Message Board
Hello there,
I have a script that utilizes align_epi_anat in a loop to register a large number of high resolution anatomical images together (about 150). It takes a very long time to do everything but the results are as close to perfect as I have seen.
However, the script seems to randomly get stuck. Every 15 or so registrations, the script will freeze at some point during the paramater opt
by
jdviviano
-
AFNI Message Board
Daniel Glen Wrote:
-------------------------------------------------------
> 3dWarp doesn't have a save matrix option, but you
> can either redirect the matrix from stdout output
> (as done in align_epi_anat.py with "3dWarp -verb")
> or get the header attribute that is applied to the
> output dataset (WARPDRIVE_MATVEC_FOR_000000 and
> WARPDRIVE_MATVEC_INV_0
by
jdviviano
-
AFNI Message Board
unfortunately, afni_proc.py is not an option, because I am not really doing a conventional analysis. Hopefully I can get everything done in one interpolation step using align_epi_anat. I somehow missed the 3dwarp interpolation option. Does 3dWarp allow you to save a transformation matrix?
by
jdviviano
-
AFNI Message Board
Great, thanks, that sounds simple. I was doing everything in steps to retain control, but I'm not sure there's any really good reason to do that.
Is there much of a disadvantage to deobliquing as one step and then doing the motion correction + registration as a second step (as we are currently doing)? Do you know what kind of interpolation 3dWarp uses?
Really appreciate the guidan
by
jdviviano
-
AFNI Message Board
Hi there,
I'm struggling to find a way to co-register multiple (tiny) slabs to one another. The solution I have come up with involves deobliquing all the runs, volreg within-runs and then allineate to a common target. I know it is best to only interpolate an EPI dataset once.
1) Most of this, if not all, can be done with align_epi_anat.py. I am currently using it to create transformat
by
jdviviano
-
AFNI Message Board
No, I don't think it is worth our time. I suppose I am just letting the world know why one would find a discrepancy between the software packages, for those doing very high resolution work. Furthermore, the window AFNI implements preserves more detail... so the psychological experience of FLIRT being "better" due to reduced ringing was nothing more than that. I doubt the difference
by
jdviviano
-
AFNI Message Board
Hi there, just to follow up:
This was inspired by our desire to replicate the (apparent) interpolation quality afforded by FLIRT -sinc. It appears that there remain subtle differences between the two programs with respect to the window type used. I have better success with 3dAllineate in general as a registration tool, so I was hoping to elucidate what would cause the appearance of more Gibbs
by
jdviviano
-
AFNI Message Board
I understand. Thanks for your help with this. I'll be sure to keep in close touch with you as we find what works best.
JDV
by
jdviviano
-
AFNI Message Board
Page 1 of 2
Pages: 12