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
Emily, that's pretty much exactly what InstaCorr does with the "seed radius" -- it does a flat average over the sphere around the seed. This is in addition to any blurring ordered earlier, which applies to all voxel time series, not just the seed voxel.
This feature was added at the behest of Steve Gotts, who thinks it produces clearer results.
by
Bob Cox
-
AFNI Message Board
It would be relatively easy to transform the y-axis, since the facility for scalar transformations ("0D") of the data already exists -- a new function for this just would have to be added.
But what would the x-axis be in this case? That facility also exists, but not on a voxel-wise basis -- that is, each graphed voxel time series gets the same mapping from index to x-axis position.
by
Bob Cox
-
AFNI Message Board
At this time, NO programs allow for missing data in .1D (text) files.
by
Bob Cox
-
AFNI Message Board
The problem with -prefix in 3dNwarpFuncs has been fixed (oopsie). Source code is available now, and binaries will be built by Wednesday morning sometime (Insh'allah).
by
Bob Cox
-
AFNI Message Board
I just returned from a backpacking vacation. I'll look at these issues soon.
by
Bob Cox
-
AFNI Message Board
I just checked in the fix (I hope it fixes it, anyhoo), and the updated source code afni_src.tgz is now available.
by
Bob Cox
-
AFNI Message Board
This is a bug introduced when separating these macros from another header file. We'll fix it in the source code today.
by
Bob Cox
-
AFNI Message Board
If one of your labels in the covariates file contains the +tlrc part of the filename, the system is unhappy with you!
by
Bob Cox
-
AFNI Message Board
Ham sandwiches -- not going to happen. I'm personally against eating mammals.
Also, you could generate the random numbers with the 1deval program instead of R.
by
Bob Cox
-
AFNI Message Board
What error message do you get?
Try running the program with the option '-DAFNI_DEBUG_TABLE=YES' as the first option -- this will print out some information as it tries to read the covariates table, and that may help find the problem.
If the file was created in Word or some other word processor (vs. a plain text editor), then you'll probably have non-good characters that AFNI
by
Bob Cox
-
AFNI Message Board
Rather than ask an abstract or out of context question, it is better to post a command that fails and ask what the problem is. You can suggest what you think the problem is, which can be helpful to us.
In this case, since I don't know what you actually did or what type of error message you got, I don't understand at all what the issue is.
by
Bob Cox
-
AFNI Message Board
A warning might be possible. A fatal error would be too much. Someone might want to run the program with (say) just the first 500 out of 1000 time points, using an input dataset syntax like fred+orig'[0..499]'. They shouldn't have to modify all the other inputs to match.
by
Bob Cox
-
AFNI Message Board
That's the rub -- the software and the script is included in the latest AFNI, but there are few guidelines available yet. i have to write those down, and the tedious-ness of that task has kept me away from it.
I'm willing to write something down for you, IF you are a user who can take a csh script, read it, and then modify it for your own needs. At the present stage, this capabilit
by
Bob Cox
-
AFNI Message Board
The "best" way?
If you want that, then you can use a script I wrote that nonlinearly warps the brains together to create
an average template that will have a lot of detail. However, this requires
(a) a fair amount of time (hours to days)
(b) the use of nonlinear warping to bring each subject's data to the template space
We have software for that, but to use it well you&
by
Bob Cox
-
AFNI Message Board
It is possible, but is it desirable? If the dataset is of any size, the resulting text file will be huge. A 64x64x30 volume has 122,880 voxels, so you would get a 1D file with 122,880 lines of text.
If you insist, you can try something like
3dmaskdump -noijk FILE.hdr > FILE.1D
and see if you like the result. 3dmaskdump can also give you voxel coordinates, or just give part of the d
by
Bob Cox
-
AFNI Message Board
Dataset #N was a quick hack for Patrick Bellgowan, formerly of the NIMH, now Director of Cognitive Neuroscience at the Laureate Institute. That's why it is limited in abilities -- not because of Pat, but because it was done quickly.
by
Bob Cox
-
AFNI Message Board
By default, the coordinate display in the upper left corner of the AFNI window is Right, Anterior, and Inferior = negative (RAI coordinate order). However, you can flip that to the SPM/FSL convention (LPI) by right-clicking in this coordinate sub-window, and then choosing the appropriate menu option from the hidden popup menu.
As an aside, you can tell if a location in the AFNI GUI has a hidd
by
Bob Cox
-
AFNI Message Board
Since you're here at the NIH, the simplest thing is for you to walk them over to me on a flash drive. Then I can look at them right away and see if I have any questions. Monday should be OK.
by
Bob Cox
-
AFNI Message Board
It is possible to input a distortion field to 3dNwarpApply, computed from a field map, to transform the EPI datasets. I have tried this, and gotten semi-reasonable results.
However, I'm more pleased with results computed from using reversed blip direction EPI scans, as is done in the Human Connectome Project. There, the 2 sets of EPI datasets are distorted in opposite directions, and th
by
Bob Cox
-
AFNI Message Board
We have not processed T2-weighted images much -- the people we work with at the NIH don't use them very often.
How are you evaluating the uniform-ization? By histogram?
One option would be to use 3dUnifize on a T1 volume, and save the scale factor with the '-ssave' option, then apply that scale factor to the T2 volume.
Another option would be to send me a T2 volume, plus a des
by
Bob Cox
-
AFNI Message Board
Are you using black blood (GRE T2*) or white blood (time-of-flight) imaging methods?
We don't have a direct tool for making venograms, since there never has been any demand from our user base. We have some tools that are useful for this task, but not a complete package.
For black blood imaging (which I think is the most common form for venography), you can use the Project function in
by
Bob Cox
-
AFNI Message Board
Ziad is away for the next month-ish period, so it will be a while before he can respond. You might better re-post this around the beginning of August, when he'll have a chance to think about it.
by
Bob Cox
-
AFNI Message Board
You can always subtract 2 volumes (e.g., with 3dcalc), but you can't get any estimates of statistical significance.
by
Bob Cox
-
AFNI Message Board
There is one circumstance under which you can do something like
3dttest++ -setA FILE1 -setB FILE2
and that is when the 2 files each contain multiple volumes. That may be what you did in the earlier case.
However, AFNI error messages pre-suppose that you understand what you are doing, and the messages are designed to help you patch up simple errors. It is a major error on a user's
by
Bob Cox
-
AFNI Message Board
You can also put the layout directly in the .afnirc file with the following trick
In the ***ENVIRONMENT section, put the definition
AFNI_LAYOUT_FILE = .afnirc
At the bottom of the .afnirc file, put the following (for example)
***LAYOUT
A geom=+1024+29
by
Bob Cox
-
AFNI Message Board
3dUnifize may work for T1 weighted images, depending on your needs, but probably will be a disaster for T2 weighted images. I haven't tried it on such images yet, actually.
by
Bob Cox
-
AFNI Message Board
The magic function you want is 'ispositive' inside the program 3dcalc; for example
3dcalc -a Yourmask+tlrc -b Atlasmask+tlrc -expr 'ispositive(a)*b' -prefix WhateverYouWant
ispositive(x) returns 0 for x=0 or negative, and returns 1 for x > 0. Et voila!
Answers are slow this week because (a) a bunch of us were at HBM in Seattle, and (b) a bunch of us are sick, and
by
Bob Cox
-
AFNI Message Board
Something like
1dmatcalc '&read(INFILE.1D) &Psinv &write(OUTFILE.1D)'
would do the job.
by
Bob Cox
-
AFNI Message Board
My poster for HBM 2013 ("Nonlinear Warping in AFNI") is now available (as a PDF) at this directory:
http://afni.nimh.nih.gov/pub/dist/HBM2013/
As the other members of this Core finalize their posters (etc.), they will be made available in the same place.
Posters numbers and stand-by days are:
Cox #1829 Mon
Glen #1431 Tue
Chen #1926 Mon
Saad #2016 Mon
Stop by and say &q
by
Bob Cox
-
AFNI Message Board