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
I stopped building the 10.5 icc-compiled version because 10.5 is so old now. I can start building a 10.8 icc-compiled set of binaries, if that will help. However, it will take me a day or two to get around to that.
The problem is a bug in the gcc compiler's implementation of OpenMP, and this problem doesn't seem to happen in the Intel icc compiler.
by
Emperor Zhark
-
AFNI Message Board
The Unix/Linux error message "Killed" means that the operating system stopped the program. This almost always happens because the system runs out of memory -- and notice your run required 3.6 Gbytes of memory (at least). There is no way for a program to intercept the "Kill" command, and Unix doesn't deign to actually say why the death penalty was invoked, so the situati
by
Emperor Zhark
-
AFNI Message Board
Sorry -- my explanation of how flashing a cluster works was somewhat garbled, but the overall implication is the same -- when the same dataset is being clusterized twice, the hidden volume being used may not correspond to the cluster table, and then weird things happen.
by
Emperor Zhark
-
AFNI Message Board
This problem is related to other issues I've seen with using Clusterize on the same dataset in 2 controllers, but in different ways (different sub-bricks, different thresholds, etc.). The somewhat rococo explanation follows:
The issue arises because of the way I implemented clusterization -- the volume to be overlaid is not the chosen OLay sub-brick, but instead is internally diverted to
by
Emperor Zhark
-
AFNI Message Board
Changing the subject slightly, I don't suppose you know anything about the village of Mingbo in Nepal? I'm hoping to go there next year.
by
Emperor Zhark
-
AFNI Message Board
What you are seeing is an oversight in the way the output parameters are labeled in the special case of '-EPI'.
Normally, the shear matrix is loaded with 3 parameters a,b,c:
[ 1 0 0 ]
=[ a 1 0 ]
[ b c 1 ]
and they get labeled as a='y/x-shear', b='z/x-shear', and c='z/y-shear'. However, in the case of '-EPI', the shear matrix is
by
Emperor Zhark
-
AFNI Message Board
rick reynolds Wrote:
-------------------------------------------------------
> Strangely enough, I do not see any driver command for
> saving the overlay dataset. We should really add that.
That has been added recently, with the new-ish WRITE_OVERLAY and WRITE_UNDERLAY commands.
by
Emperor Zhark
-
AFNI Message Board
I have fixed this problem (I think). The changes are in the source code, and will be reflected in the binaries when next we rebuild them -- this weekend, perhaps.
No guarantees there still aren't glitches in locking pbars. The possible sequences of redrawing events is too complex to hold in my head, and so something may still end up de-synchronized.
by
Emperor Zhark
-
AFNI Message Board
There are several things going on here, and I'll have to investigate how the code works and how to fix it. The locking operations are moderately complicated, and the ordering of events is confusing -- even to me, and I wrote the code.
by
Emperor Zhark
-
AFNI Message Board
It was 20 years ago, in early July 1994, that I started work on AFNI. An outline of the story can be found in this paper.
One reason the AFNI viewer is so different from what's available in FSL and SPM is that from the very beginning my goal was to let the user surf through data in 5 "dimensions" = xyz + imaging run + subject -- time was added in 1996, when computers had more
by
Emperor Zhark
-
AFNI Message Board
This is certainly a possible change to afni_proc.py, but will require the estimable Rick Reynolds to implement the necessary options. And he has some other commitments for a little while.
by
Emperor Zhark
-
AFNI Message Board
One of these days I'm going to retire the "version number", which doesn't really mean anything since we update (micro, minor, and major) all the time -- there is no release schedule.
It is the compile date that should be checked against our website.
by
Emperor Zhark
-
AFNI Message Board
Also, is the "center coordinate" the center of the 3D volume covered by the slice? That is, is in the middle in all 3 dimensions?
by
Emperor Zhark
-
AFNI Message Board
What do you plan to do with this "box"? It shouldn't be too hard to come up with a convoluted 3dcalc command to do this, but I'm curious.
When you say "normal coordinate", do you mean the vector normal to the slice?
It seems to me that your coordinates are such that R = +x, A = +y, and I = +z -- is this presumption correct?
by
Emperor Zhark
-
AFNI Message Board
And below is an image of the new Define Overlay control panel. Notice that the bkgd:xxxx controls, which let you show the ULay or OLay volume as the grayscale background, are now gone. Their functionality only remains in the image viewer 'u' keypress.
by
Emperor Zhark
-
AFNI Message Board
The changes below are being compiled now and will be available very shortly (Thursday 19 Jun 2014).
I have modified the AFNI GUI so that Clusterize is now outside of the dropdown menu that includes InstaCorr, etc. The result is that you can use Clusterize with InstaCorr and Group InstaCorr, which was not possible before. I've found this very useful in just the last 24 hours. I've cre
by
Emperor Zhark
-
AFNI Message Board
You have to be more clear. That is, you have described the end point of the problem that you see, but you have not described EXACTLY how you proceeded to get to this point. You give a vague description of what you did, and we can't proceed from that.
by
Emperor Zhark
-
AFNI Message Board
No, there's no elegant way to do what you want. Sorry.
by
Emperor Zhark
-
AFNI Message Board
The next AFNI bootcamp at the NIH will be held Mon-Fri, Sep 29-Oct 3, 2014. This class will last 5 full days, and will include a breakout day for DTI analysis and some advanced hands-on classes for group analysis.
Details and registration can be found at http://afni.nimh.nih.gov/bootcamp/
The schedule has not yet been finalized. The DTI breakout day will mostly likely be Mon or Tue, depen
by
Emperor Zhark
-
AFNI Message Board
Oh yes ... afni now has a -papers option, which will list out papers (including links to full text) about various components of the AFNI package.
by
Emperor Zhark
-
AFNI Message Board
Also, the afni program itself now has a -papers option, which will just list out some papers about AFNI that you can peruse for your enlightenment.
Some minor changes to 3dQwarp were made that might change the results a little. In particular, the warp is now defined outside of its bounding box by linear extrapolation. In particular, this made calculating the inverse warp more stable.
by
Emperor Zhark
-
AFNI Message Board
These changes are in the AFNI source code (afni_src.tgz) now and will be compiled up into the binaries by tomorrow morning (Tuesday 17 June 2014).
From within the Clusterize Report page, you can now save a dataset with a mask from a single cluster. To do this, you must use the right-click popup menu attached to the very top of the Report page (where it starts with "Voxels survived").
by
Emperor Zhark
-
AFNI Message Board
Please read the help section for 3dttest++ titled HOW COVARIATES WORK -- that explains exactly what is going on.
If you use the default centering option '-center DIFF' for a 2 sample test, then
#0 SetA-SetB_mean = difference of means will just truly be the difference of sample means (at each voxel). However, different options for centering will alter the "mean" re
by
Emperor Zhark
-
AFNI Message Board
I have made a change to the way the AFNI GUI Define Overlay displays the numerical range that is used to determine the scaling by which dataset numbers are converted to colors. At present, this change is optional and must be turned on via the environment variable described below.
I welcome people trying this out and seeing if any peculiarities arise. This modification is the first change to
by
Emperor Zhark
-
AFNI Message Board
Actually, the original program being referred to is FD2. It still exists, but has not been changed in years, and furthermore will not be changed -- even to fix bugs. Most of its abilities are incorporated into the AFNI Graph viewer window, as mentioned earlier.
by
Emperor Zhark
-
AFNI Message Board
This link http://hpc.sourceforge.net has binaries for "real" gcc for various flavors of Mac OS X. I use one of these for my own laptop.
by
Emperor Zhark
-
AFNI Message Board
Fixed. In the source code archive afni_src.tgz -- the binary rebuild will have to wait until next week, as I'm at home now, and this doesn't seem so urgent.
by
Emperor Zhark
-
AFNI Message Board
Hmmm. There is a single character typo in afni_driver.c about this. I'll fix it shortly. Obviously the work of code gremlins.
by
Emperor Zhark
-
AFNI Message Board
What is happening is that your first command line argument after '-setA' is a label 'AAA', not a dataset name (the program tests to see if it can open file 'AAA' as a dataset -- if that fails, then it is taken to be a label instead).
In this case (label immediately after '-setA' or '-setB'), then the arguments that follow are taken to occur in
by
Emperor Zhark
-
AFNI Message Board
You can also use the option
-expr 'atanh(a)'
which is much easier to type
by
Emperor Zhark
-
AFNI Message Board