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
Hello,
From the macos installation guide, it looks like you still need to install the R library packages:
sudo rPkgsInstall -pkgs ALL
You might also want matplotlib, but that is not needed for the group analysis R programs.
- rick
by
rick reynolds
-
AFNI Message Board
Note that at least the REMLfit step is easy after -write_3dd_script. Running that 3dDeconvolve command will generate a new 3dREMLfit script, even if using -x1D_stop.
- rick
by
rick reynolds
-
AFNI Message Board
That's great, Adury!
Thanks for the follup-up.
- rick
by
rick reynolds
-
AFNI Message Board
Yes, it does. Assuming the HEAT_Late timing is 30s later than that of the Early, each would consider the other's time as off.
Note that BLOCK is a 15s response curve though. So the ends of the Early responses will overlap with the starts of the Late responses. Hopefully there are reasonably long gaps between Late events and the next Early ones. But you will see that overlap in the X-m
by
rick reynolds
-
AFNI Message Board
That is odd, but good news.
One thing to note is that your initial error message seemed to reference AFNI version 17.2.10 (since the error came from afni/17.2.10/plug_3dsvm.so), and it is a system package of AFNI (available to all users). I am not even sure if that was a CentOS build package, probably not.
The system_check output shows a working personal version of AFNI (in your own abin).
by
rick reynolds
-
AFNI Message Board
Hi Nick,
What is the output of: afni_system_check.py -check_all
Thanks,
- rick
by
rick reynolds
-
AFNI Message Board
Indeed, afni_proc.py can be used with "-combine_method m_tedana" to run the MEICA group's tedana as part of the step to combine echoes. See
APMULTI_Demo1_rest/scripts_desktop/do_26_ap_me_bT.tcsh for an afni_proc.py command script, including use of that combine method.
In any case, tedana creates the NIFTI datasets generally using nibabel, which writes an sform_code of 2, which
by
rick reynolds
-
AFNI Message Board
Sure, that is great. You could also rename the 1D files to include the run index.
- rick
by
rick reynolds
-
AFNI Message Board
If you want to save all of those files, just rename them in the loop. Move them to the
${global_path}5.1st_level_analysis/sub-0${i}/func/GLM_INPUT
directory, which is already being referenced for the subject in the loop.
The reason they are being overwritten is presumably because they are written to your current directory. So do something like:
for ((i=1;i<10;i++)); do
for ((j=1
by
rick reynolds
-
AFNI Message Board
Hi Ian,
Try adding the afni_proc.py option "-align_epi_strip_method 3dSkullStrip".
It looks like 3dAutomask is not doing a good job of removing the skull area of the EPI (which is pretty apparent in your data), and 3dSkullStrip will probably do a better job.
- rick
by
rick reynolds
-
AFNI Message Board
Hello,
These are reasonable basis functions, but there is a subtlety when combining them. Since they are all scaled to 1, all of the betas will be in terms of percent signal change. However if they were the exact same task, one event is 2.5 s and another is 3.0. If you perform a contrast between them, the longer event has an unfair advantage. If you are NOT going to compare negative or neu
by
rick reynolds
-
AFNI Message Board
Hi Catherine,
Thank you for the X-matrix file. The collinearity issues are not sooooo bad, but they seem to be coming from the motion parameters, not the regressors of interest. Try running these commands:
xmat_tool.py -load_xmat rf_matrix_tosend.1D -no_gui -show_conds
1d_tool.py -infile rf_matrix_tosend.1D -show_cormat_warnings_full
The first shows small condition numbers for the mai
by
rick reynolds
-
AFNI Message Board
Glad it's working for you, thanks for the update.
- rick
by
rick reynolds
-
AFNI Message Board
Indeed, they produce approximately the same matrices. So you could use 3dvolreg -1Dmatrix_save and apply it with 3dAllineate -1Dmatrix_apply and get similar output (subject to interpolation choices).
- rick
by
rick reynolds
-
AFNI Message Board
I have looked at this, and there are more subtleties. Basically, the param files between 3dvolreg and 3dAllineate are not appropriate for passing between the programs. Besides the column order, there are appear to be negatives between the shifts. Also, 3dvolreg uses the center of the volume for rotations, while 3dAllineate uses xyz = (0,0,0) for its center. So the sets of rotations/shifts wil
by
rick reynolds
-
AFNI Message Board
Great, I am glad the install is settled. I should have asked if you had admin rights...
To be sure, are you using 3dAllineate to apply the (converted) 3dvolreg parameters and are comparing the results?
What is the 3dAllineate command?
Can you be more specific about in what way (in a few places) the result looks unexpected? Does that mean a few locations in the volume? At a few time poin
by
rick reynolds
-
AFNI Message Board
Oh, I think one of the @uab commands did a partial job, and now those files are hiding in the auto_backup subdirectory.
Forget it though. Instead of hunting for the old script, just start with a downloaded version.
cd
curl -O https://afni.nimh.nih.gov/pub/dist/bin/misc/@update.afni.binaries
tcsh @update.afni.binaries -bindir ~/abin -package macos_10.12_local -defaults
- rick
by
rick reynolds
-
AFNI Message Board
I don't understand why the regular files were not copied, but maybe just run the version in the moved directory with explicit options:
tcsh ~/abin.corrupt/@update.afni.binaries -bindir ~/abin -package macos_10.12_local -defaults
- rick
by
rick reynolds
-
AFNI Message Board
Indeed, the output from that "ls -l" command shows that you do not own or have write permission to that library file (or any of the other 'denied' ones) It was possibly installed by an admin.
If you want to clean up, I would suggest a sequence like this:
cd
mv abin abin.corrupt
cp -r abin.corrupt abin
@update.afni.binaries -defaults
That will leave the current abin a
by
rick reynolds
-
AFNI Message Board
So that library is missing the symbol. Is it old for some reason? What is the output from:
ls -l ~/abin/afnipy/lib_afni1D.py
You could get a new version of that file, but it makes me doubt your installation. It might be worth updating your whole install, just to be sure.
@update.afni.binaries -defaults
Is that something you can do right now? Otherwise we could investigate more. But
by
rick reynolds
-
AFNI Message Board
Hi Edwin,
Hmmm, that is a bit strange. Do you know of anything peculiar in your python setup? I don't see anything strange in the system check output. But the error suggests maybe there is a discrepancy between versions of python files.
What is the output from this?
find ~/abin/ -name lib_afni1D.py
grep g_1D_write_styles ~/abin/afnipy/lib_afni1D.py
echo pypath = $PYTHONPATH
T
by
rick reynolds
-
AFNI Message Board
Hi Edwin,
I pondered this a few years back, and my take is expressed via example 23 from 1d_tool.py -help:
1d_tool.py -infile dfile_rall.1D -volreg2allineate -write allin_rall_aff12.1D
Give that a try and please let me know what you think.
- rick
by
rick reynolds
-
AFNI Message Board
The top part describes group and Symptoms but the dataTable specifies Severity.
Maybe that is causing the problem.
- rick
by
rick reynolds
-
AFNI Message Board
Is there some sort of group masking that is relevant here? The dark (higher magnitude) area looks a bit like a coverage intersection box. Are various subjects missing from coverage outside of that? Do you have a group mask? Do you have a group 'sum' mask (add up subjects per voxel)?
EDIT: For reference, it is nice to run something like:
3dMean -prefix QC/mask.mean sub*/*.results
by
rick reynolds
-
AFNI Message Board
Hi Catherine,
Yes, the all-zero columns will still result in those collinearity problem warnings, since the program cannot distinguish the cases.
I see you did have -GOFORIT in the original command
If you want, it would be okay to send me the matrix file via email (click on my name) so that I could investigate it locally.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Angelika,
The .afnirc file is in your home directory, ~/.afnirc or equivalently $HOME/.afnirc.
Alternatively, you can set that using an environment variable, though your ~/.afnirc file is probably a better place to do it.
To be sure, do you have a $HOME/.afnirc file?
Note that if you run the system check, it will suggest how to copy one to your $HOME directory in the final "plea
by
rick reynolds
-
AFNI Message Board
Hi Angelika,
Indeed, you can see that the sform_code values are shown as 2, which is the ambiguous or likely incorrect case.
Try setting AFNI_NIFTI_VIEW = orig either in your environment or in your ~/.afnirc file. Then test by running '3dinfo' on the NIFTI input and verify that without the variable, the space is +tlrc, but with it, the space is +orig.
For example, even before se
by
rick reynolds
-
AFNI Message Board
Great!
Sure, a command not found error is from the OS when you try to run a command that is not in your path. It usually means either a misspelling of the command, or as in this case, a missing line continuation character (\) on the previous line. The '-odir' line does not have a '\' at the end.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Austin,
This looks pretty good!
The R packages are needed for Gang's group analysis programs. Fixing that won't be important until you want those programs.
Oh, for that I should have suggested installing R-devel instead of R. You *might* need to ask for that.
Have you tried installing anything under ~/R? Is that directory empty? populated? old?
- rick
by
rick reynolds
-
AFNI Message Board