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
There are 2 ways.
#1 is interactively in AFNI, from the Datamode control panel, using the SaveAs OLay button -- presuming you are displaying the GIC results as the overlay dataset.
#2 is via the -batch command line option in 3dGroupInCorr, which does the work without AFNI interaction.
by
Bob Cox
-
AFNI Message Board
No such plans. We don't have the time to deal with ICA as well as other stuff. And there are packages well advanced in this area -- such as Vince Calhoun's GIFT -- http://mialab.mrn.org/software/gift/index.html
by
Bob Cox
-
AFNI Message Board
Not really. You could use 3dpc to do a PCA on a dataset, but it doesn't do any statistics or clustering. We don't have any ICA tools.
For viewing "networks", you can use SUMA -- but it is too complex to outline the steps on the message board.
by
Bob Cox
-
AFNI Message Board
Okay, the problem is obvious. The max filename allowed currently is 511 characters. Somewhere in the screen output before the crash should be a cryptic message like
** ERROR: Path name too long. Returning './'
We will increase the max filename length, and even improve the error message. But ultimately there is a limit to the absurdity of filename paths that we will be able to h
by
Bob Cox
-
AFNI Message Board
This is peculiar -- the filename lengths (about 570 characters) are not overly long for AFNI. Can you get the crashlog for this from /home/gose66/.afni.crashlog and email it to me?
by
Bob Cox
-
AFNI Message Board
I am running linux_openmp_64 on the Linux box in my office right now -- at build Oct 13 2016 (Version AFNI_16.3.03). The command
3dClustSim -acf 0.5 7 9 -mask askull+orig
is working well. If I do 'top -U rwcox' it shows up.
Perhaps you need to update your AFNI binaries? I don't know what else to suggest.
by
Bob Cox
-
AFNI Message Board
What kind of system are you running on? This does not happen on my Mac.
by
Bob Cox
-
AFNI Message Board
There is a bug. The problem was that when '-BminusA' is used, the program could end up changing the sign of the results TWICE -- probably some leftover old code wasn't removed by some unnameable august personage.
I have fixed this problem in the source code, which is now on GitHub. I will start a rebuild of the binaries tonight, so using @update.afni.binaries tomorrow morning
by
Bob Cox
-
AFNI Message Board
You can set environment variable AFNI_DONT_LOGFILE to YES to turn off the logging. This trick is useful in scripts if you are running a program thousands of times to test something (as I am doing right now).
by
Bob Cox
-
AFNI Message Board
1) Since you don't give the error message, it is hard to say what the problem is. In general, I would recommend avoiding very complex command lines like this, and do it in 3 steps: 3dcalc as needed, 3dttest++, then delete the temporary files
2) In general, it is not true that regressing with the average of 2 regressors gives the result that is the average of the separate regressions. Th
by
Bob Cox
-
AFNI Message Board
With 3dMVM, there's a difficulty in that (at present) it does not write out residuals to be analyzed. If you analyze the statistics sub-brick(s), the ACF will be biased by the presence -- you hope! -- of activations.
One option would be to compute the 3 ACF parameters for each subject's residuals (errts) time series dataset, then take the mean or median of each of these, and use tho
by
Bob Cox
-
AFNI Message Board
If you are using 3dMVM, I'd say use 3dFWHMx -acf followed by 3dClustSim -acf -- there really isn't another option. However, we don't have any simulations to back this usage up, since running 1000 3dMVMs on RS data would take approximately forever.
To be clear about the covariates issue in 3dttest++ -- the main effect had proper FPR control when the analysis included a covariate
by
Bob Cox
-
AFNI Message Board
3dttest++ -Clustsim has now been tested against 1-sample problems. The false positive rates (FPRs) were pretty close to the nominal 5%. Sometime "soon" I will add those tables to our bioRxiv manuscript.
I also ran 2-sample tests with the -unpooled option, and these FPRs were also were close to the nominal 5% false positive rate.
With -covariates (1- and 2-sample), the main effec
by
Bob Cox
-
AFNI Message Board
(1) you don't need or want to use -resid with -Clustsim, since the program will do that for you when you use -Clustsim.
(1) 3dttest++ -Clustsim does NOT use -acf or any smoothness estimates at all. It doesn't model the spatial structure of the noise directly. It uses a randomization of residuals approach to compute noise-only data, and then runs that stuff through 3dClustsim
(2)
by
Bob Cox
-
AFNI Message Board
(1) the important thing is that 3dClustSim ran and seemed OK
(2) you must be using an older version, since the output text format has changed somewhat, and more iterations are now the default
As far as the 'Bad fd number' error goes, it comes from the shell that is invoked to run the 3dttest++ -randomsign iterations. My guess is you are using Ubuntu, since that error message only s
by
Bob Cox
-
AFNI Message Board
Modifying 3dClustSim to work with surfaces will not happen in a "couple of weeks". It is a major undertaking, and I don't know when it will happen.
by
Bob Cox
-
AFNI Message Board
I suspect you are using -resid as well. Remove that option and try again.
I did make a fix to 3dttest++ yesterday that should fix this problem (conflict between -resid and -Clustsim). You can @update.afni.binaries and try that as well.
by
Bob Cox
-
AFNI Message Board
I would do it before bandpassing.
However, I'm currently doing work to replace 3dBlurToFWHM with something better, but don't know when that will become available.
by
Bob Cox
-
AFNI Message Board
I think I see the problem. There's a bug if you use the '-resid' option with the '-ClustSim' option. Remove the '-resid' option and try it again.
I'll try to fix the bug soon, but you can do what I just suggested now and report back.
by
Bob Cox
-
AFNI Message Board
Also, if any of your input filenames or labels have spaces in them (that you used quotes to input to 3dttest++), then the -Clustsim option will probably fail.
by
Bob Cox
-
AFNI Message Board
Can you please update your AFNI binaries and try this command again? Even if it crashes, the newer version of 3dttest++ should output more information about what it's doing as it progresses through the -Clustsim option, and that may help debug.
by
Bob Cox
-
AFNI Message Board
At present, there isn't a way to do this in 3dttest++. I would have to derive the formula for t-testing the covariate slope estimates (taking into account covariance between the covariates), and then implement it.
At the present moment, I am deeply involved in developing a response (both in the AFNI code and in a manuscript) to http://www.pnas.org/content/113/28/7900.full and so don
by
Bob Cox
-
AFNI Message Board
The "-warp" option for 3dfractionize has nothing at all to do with the much-later-written nonlinear warping programs in AFNI. You'll have to manually transform the ROI dataset back to individual subject space, and then 3dfractionize that.
I've almost forgotten the 3dfractionize program, which I wrote about 20 years ago. And I'd completely forgotten its "-warp&qu
by
Bob Cox
-
AFNI Message Board
What do you (or the reviewer) mean by "A > B"?
That the covariates values themselves are significantly different?
Or that the effect of covariate A is greater than the effect of covariate B in the region of interest?
Or something else?
by
Bob Cox
-
AFNI Message Board
Our posters for the Geneva HBM 2016 meeting are available in the links below:
Cox #2082 (Monday pm)
https://afni.nimh.nih.gov/pub/dist/HBM2016/Cox_Poster_HBM2016.pdf
Chen #2112 (Monday pm)
https://afni.nimh.nih.gov/pub/dist/HBM2016/Chen_MVM_Poster_HBM2016.pdf
Taylor #3534 (Wednesday pm)
https://afni.nimh.nih.gov/pub/dist/HBM2016/OHBM2016_Taylor_FATCAT_new.pdf
by
Bob Cox
-
AFNI Message Board
When I'm done with my current round of simulations and make my poster, I'll put it up for perusal. Late this week, inshallah.
by
Bob Cox
-
AFNI Message Board
This is something I'm working on right now. Hope to have something better to say about this at HBM in Geneva in a couple weeks -- at my poster #2082 (Mon Jun 27 12:45-14:45).
by
Bob Cox
-
AFNI Message Board
We can't schedule the next bootcamp in Bethesda because the one auditorium that is usable for us (Lister Hill) is scheduled for renovations, and until it is known when those will be finished, they won't allow anyone to book the room. Since we need it for a solid week, we usually have to book it well ahead of time. I was assured we will get priority, but what that probably means fairly
by
Bob Cox
-
AFNI Message Board
Here's an undocumented trick:
You can replace the spaces in '1D: 0 151 101 75' with '1D:0,151,101,75' and then you can eliminate the quotes, since they are only needed to pass the spaces through to the program (which the shell normally will not do).
Why do you want to try the -gridlist option, anyway?
by
Bob Cox
-
AFNI Message Board