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
Results 2071 - 2100 of 2305
Hi-
That looks like correct output with 2 ROIs. When you have N target ROIs, then every matrix will be of size 'N by N'-- in your case, 2x2.
These are like correlation matrices. Each element's coordinates tell you which targets' connections are being described.
So, for the mean FA matrix for example, you have
FA_{11} FA_{12}
FA_{21} FA_{22}
where 'FA_{12}
by
ptaylor
-
AFNI Message Board
Gotcha. It hasn't crashed in this case, but will likely switch to binaries when next we update the AFNI files here, just for simplicity.
Thanks,
pt
by
ptaylor
-
AFNI Message Board
Hi, Rick-
Ok, thanks. We ended up trying building from source using "Makefile.NIH.openSUSE.11.4_64", and that worked as well; I will look to the openmp_64 one though, on any other such machines.
Thanks,
pt
by
ptaylor
-
AFNI Message Board
Hi,
Is the above AFNI binaries version no longer updated? It had been put on an openSUSE cluster here, and we tried @update.afni.binaries on it to no avail. Is doing afni_src builds the only way to go?
--pt
by
ptaylor
-
AFNI Message Board
"macosx_10.7_Intel_64: Apr 1 2013" !!! I think the AFNI code was written in Latin back then!
Please download the newest version. That does predate many FATCAT programs, including 1dDW_Grad_o_Mat.
In general, AFNI is a continuing stream of consciousness, and it is a truth universally acknowledged, that every single researcher in possession of good fortune has a frequently updated
by
ptaylor
-
AFNI Message Board
Actually, the date just before that part is what I was looking for. For example, on my computer, this is the output:
$ afni -ver
Precompiled binary linux_ubuntu_12_64: Apr 28 2015 (Version AFNI_2011_12_21_1014)
(To be honest, I have no idea how to interpret that Version string-- I think it's some legacy thing.)
And what is the output of the following steps? Take the output of th
by
ptaylor
-
AFNI Message Board
That's odd. Could you please post the full command being used?
A few things to check:
+ what version of AFNI do you have? (Hopefully it's not sooo old that the program isn't there.)
+ is this being run in a script? If so, is there slicing/subbrick selection being used, like: FILE[0..5]? If so, sometimes not including apostrophes/quotations will confuse things, and I beli
by
ptaylor
-
AFNI Message Board
There is an output text file with extension '.grid' that contains mean and standard deviation of lots of standard quantities for each traced WM region.
If your 3dTrackID command had '-prefix OUTPUT', then the the file would be called 'OUTPUT_000.grid'. It contains matrices of structural properties (including mean FA values of each tracked region). You can open
by
ptaylor
-
AFNI Message Board
Hi--
There are various ways to put ROIs together in a single volume. Each target for tracking is the set of voxels with a given integer value, so you will want each target unique target to have a different integer value.
If you have just 2 in the same volume space (in your case, 'thalamus.nii' and 'mpfc.nii'), and you are sure that they do not overlap spatially, then y
by
ptaylor
-
AFNI Message Board
I am not an expert on the builds process by *any* means, but I have recently warped a Makefile to build on my laptop's Yosemite. I started with Peter Molfese's 'Cogneurostats' blog advice, and tried to hack my way through many errors that came up.
I use Homebrew now as my package manager, but I did have to do a couple things with Macports as well.-- I think in particular
by
ptaylor
-
AFNI Message Board
3dTrackID uses FACTID, not the FACT algorithm.
The original paper describing FACTID, and with validation and comparisons to TrackVis/Diffusion Toolkit, is available here:
Some further comparisons with other software for probabilistic tracking (they they are Euler or RK) can be seen in Taylor & Saad (2013, Brain Connectivity), with a version available under the 'FATCAT' section
by
ptaylor
-
AFNI Message Board
How about using 3dExtrema, if you have several ROIs? You would have options of dumping the ensuing mask with 3dmaskdump.
Or '3dTstat -max'?
--pt
by
ptaylor
-
AFNI Message Board
I will check with the file source, and email you.
Thanks,
pt
by
ptaylor
-
AFNI Message Board
Hola-
bash-3.2$ nifti_tool -disp_hdr -infiles FILE_01.nii | grep dim
dim_info 39 1 0
dim 40 8 3 256 256 256 0 0 0 0
pixdim 76 8 1.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0
bash-3.2$ nifti_tool -disp_hdr -infiles FILE_02.nii | grep dim
dim_info 39 1 0
dim 40 8 3 256 256 256
by
ptaylor
-
AFNI Message Board
Part of the problem, I suspect, is that the VOI files don't appear to have useful 'location' information-- they appear to be text files with lists of coordinates and some non-location-describing strings thrown around. It also doesn't look like a simple relative flip of one axis will solve the problem, either...
--pt
by
ptaylor
-
AFNI Message Board
(and, to clarify, by 'looking like having the same FOV...', I mean that the 3dinfo results are the same.)
--pt
by
ptaylor
-
AFNI Message Board
Hi--
I have them on my computer as well; the results of that nifti_tool command are:
bash-3.2$ nifti_tool -disp_hdr -infiles FILE_01.nii |grep -A 10 qform_code
qform_code 252 1 1
sform_code 254 1 1
quatern_b 256 1 0.0
quatern_c 260 1 0.0
quatern_d 264 1 0.0
qoffset_x
by
ptaylor
-
AFNI Message Board
Hi-
The issue has now been resolved; it was an oversight on my part when the insets had brick scaling factors (which are often used for short/byte insets). Now, none of the byte or short inputs lead to crashes, and the byte, short and float insets all give identical output.
In the next build (on/after April 27, 2015), 3dROIMaker should be good to go on byte/short/float insets with/without
by
ptaylor
-
AFNI Message Board
Glad that worked, and perfect timing-- I was just about to go to sleep, so now I will rest much easier!
--pt
by
ptaylor
-
AFNI Message Board
Hi, Bryan-
Can you please send me your data set and the exact command you used which caused the error? I made a SHORT data set and tried the command originally posted on this thread, and had no problems.
Thanks,
pt
by
ptaylor
-
AFNI Message Board
The forum is fine for airing programmatic grievances. Throwing bricks through windows *could* be a viable, but not recommended, alternative (for legal reasons, please don't think I'm advocating it!).
And in this case it appears doubly (if not trebly) appropriate, since apparently I didn't fix things 'back in the day'. Will take a look soon(ish) now.
--pt
by
ptaylor
-
AFNI Message Board
Ok, I was suspecting that was the case with the end-of-line characters.
It's possible that an older version had a sly bug in it-- I never really know what the Version numbers mean, the build date is the more useful measure.
You can equivalently do either of your suggested means to update-- '@update.afni.binaries' will get you the latest AFNI software.
Version of Mac should
by
ptaylor
-
AFNI Message Board
Hi-
Hmm, thanks for bringing that up, esp. if I never fixed it (I can't remember now-- the fog of programming). Is your inset which is providing woe also scaled shorts?
--pt
ps: and glad the '-only_conn_top' is useful; 'twas suggested by a user many moons ago.
by
ptaylor
-
AFNI Message Board
Hi-
I went to the FATCAT demo set, and tried running your command in two ways (NB: I changed the in_rois and inset filenames to appropriate ones in the demo):
First, I literally ran the code you pasted, which only had the one backslash (=continuation-of-line character) at the end of the first line, and I got the following error:
++ Reading in.
** ERROR: Need to load ROIs with >=1 subb
by
ptaylor
-
AFNI Message Board
If anyone is planning on being in Cape Town between April 20-24, 2015, and you would like to register for the Bootcamp at UCT, then you'd better email in a hurry... More information is available here:
--pt
by
ptaylor
-
AFNI Message Board
Sure, fair enough; I think all the other exit values for errors were >1, so exiting with '1' was unique but certainly not standard-- thanks for pointing that out. Hope it's happy tracking from here on out.
--pt
by
ptaylor
-
AFNI Message Board
"Can anyone tell me why?"
-> The reason is: purely an oversight on my part. It has been fixed now for the function (along with a few other ones, listed below) and should be correct in the next build.
The Infamous Five:
3dEigsToDT.c
3dDWUncert.c
3dDTtoNoisyDWI.c
1dDW_Grad_o_Mat.c
3dTrackID.c
Apologies for the inconvenience, hope that sorts the script-running.
--pt
by
ptaylor
-
AFNI Message Board
Hi-
For selecting the green monochrome map in AFNI:
Open an overlay in the AFNI controller. When it has opened, you will see the default colorbar (usually red...yellow...green...cyan...blue). Right-click on this colorbar, and you will see a small, three-line panel open up. Left-click on the bottom line, "Choose colorscale." Now you have a list of colors (probably the same list a
by
ptaylor
-
AFNI Message Board
Hi-
Just a couple things to keep in mind when running 3dTrackID:
+ TrackVis-style *.trk files will no longer be output by default. This is to save space in writing-- to output them, use the switch '-do_trk_out'.
+ a slightly new keyword can be given to the option to dump individual files of tract connections: '-dump_rois AFNI_MAP'. Instead of just outputting binary mask
by
ptaylor
-
AFNI Message Board
It looks from the command that two separate ROI files have been input?:
../masks/lAmygdala_roi+tlrc.
../masks/rAmygdala_roi+tlrc
However, the "-in_rois FILE" should be a single file with each ROI defined as a set of voxels with a given integer. Thus, you could combine your data into a 'network' first (assuming each file is a binary mask; otherwise, use 'step(a)+
by
ptaylor
-
AFNI Message Board