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
3dvolreg is not OpenMP-ized -- it is so fast already that it wasn't worth the effort to retrofit it.
Try typing something like
3dAllineate -help
(or use 3dLocalstat, or 3dREMLfit, or a bunch of others that ARE OpenMP-ized). The end of the help output should tell you if the program is compiled for OpenMP.
The Xft errors are coming from a missing XFreeType library, which might act
by
Bob Cox
-
AFNI Message Board
When running the Clusterize pseudo-plugin, and when the Rpt table is open, if you press the 'j' key while the input focus goes to an image viewer window, then the crosshairs will jump to the location of the peak (or cmass) of the cluster in which the clusters currently reside. The purpose of this is to let you get to that location once you click in a blob, without having to switch your
by
Bob Cox
-
AFNI Message Board
I have fixed these errors in the source code, and will be rebuilding binaries tonight.
by
Bob Cox
-
AFNI Message Board
Someone, I forget who, asked for the ability to externally "drive" (e.g., via plugout_drive) the intensity range for AFNI image viewer windows. I just added this feature to the source code, so that a command like
afni -com 'OPEN_WINDOW A.axialimage range=20,80'
should do the trick. This will be built into the next set of binaries that we compile.
by
Bob Cox
-
AFNI Message Board
Most of the code to do this exists internally, but there is no "drive" command to invoke it. I could add such a thing, but I'm home sick today, and probably not up to it at the moment. My experience is that writing code when ill is likely to cause more problems that it solves.
by
Bob Cox
-
AFNI Message Board
3dNwarp (Calc and Apply) are works in progress, and not for the faint-of-heart at this time. They have yet to be polished up for spaces and stuff like that. I'm glad to get feedback, and in fact I'm working on a better nonlinear registration algorithm now -- seems to work OK on my first try, but there's some peculiar bug I'm having trouble figuring out.
I'm sorry you
by
Bob Cox
-
AFNI Message Board
http://andysbrainblog.blogspot.com/2012/09/afni-part-1-introduction.html
The part about "We shall become as gods among fMRI researchers" is especially impressive.
by
Bob Cox
-
AFNI Message Board
Well, if you really want, I can add an option to let you explore the window function (beyond the binary choice now between 3-term and 2-term cosine). But is it really worth it?
by
Bob Cox
-
AFNI Message Board
Actually, I meant setting up the whole run with all the data and sitmuli included, and then using '-censor' or '-CENSORTR' to remove the time points of no interest. This strikes me as simpler than the approach you have crafted.
by
Bob Cox
-
AFNI Message Board
Including the full timeseries and censoring seems like an option, but also seems to be not censoring's intended purpose.
"Purpose" is one thing, but actual implementation is another -- your solution quoted above seems to me to be the ideal one.
I never did modify 3dDeconvolve to allow for negative stim times because the original poster didn't come back to ask for it.
by
Bob Cox
-
AFNI Message Board
The binaries that are building tonight will allow the astute user to control the operation of the '-final wsinc5' interpolation in detail. The control is done via Unix environment variables, and the documentation is in the output of '3dAllineate -HELP' (notice the all caps, to get the advanced options).
In particular, you can control:
* the radius of the interpolation m
by
Bob Cox
-
AFNI Message Board
If you want LOTS of possibilities, I'll have to modify the code a little more. That will take a couple days to roll out, since I'm busy with other things at the moment.
by
Bob Cox
-
AFNI Message Board
What binary package do you want re-built? Let me know, and I'll compile it and then post instructions on how to modify the wsinc5 interpolation parameters via environment variables.
by
Bob Cox
-
AFNI Message Board
It is possible to alter the sinc interpolation parameters in a couple of ways with Unix environment variables, at least in the version of 3dAllineate that I made for myself today to experiment with these things. I could compile this up for you, if you want.
However, the changes don't make much different, at least in the images I played with (1 mm resolution, interpolated to 0.5 mm grid).
by
Bob Cox
-
AFNI Message Board
When I try this, I don't see a LOT of Gibbs ringing. Depending on your purposes for the image, you could partially suppress this effect by post-filtering the image with something like
3dWinsor -irad 3.5 -prefix anatW anat+orig.HEAD
for example. To get a quick feel for if this might help, you can open the 'Disp' control panel in an AFNI image viewer, and then under the '
by
Bob Cox
-
AFNI Message Board
AFNI's blurring programs (3dmerge and 3dBlurInMask) just add smoothness to datasets, so if there is some pre-existing smoothness, you will end up with more smoothness than you might anticipate. I imagine that SPM (etc.) does the same.
Assuming the spatially uniform Gaussian model for smoothness is correct (which is what 3dFWHMx estimates), then if the pre-existing smoothness were describ
by
Bob Cox
-
AFNI Message Board
In fact, if you have a number of seed locations (x,yz) coordinates, you can run 3dGroupInCorr with the '-batch' option to run it entirely without AFNI -- it then just reads the seed locations from a file and creates a dataset for each given location.
by
Bob Cox
-
AFNI Message Board
When you run into such a problem, you need to post the entire 3dDeconvolve command and list of output messages. Otherwise, it's impossible to help.
by
Bob Cox
-
AFNI Message Board
The -allzero_OK option is designed specifically to let you have an all zero regressor, usually in the situation when a singular subject has a "perfect" set of responses but most subjects do not. Please be sure you understand what you are doing when you stray from the usual analysis in such devious ways!
by
Bob Cox
-
AFNI Message Board
Unfortunately for you, Daniel Glen (who is in charge of @AddEdge and other registration tools) is on vacation for all of July. And Rick Reynolds, who might be able to help, is still stuck at home without electricity after our big storm a week ago, and with a sick baby to boot. We are kind of short-handed at the moment, and usually don't have a lot of spare capacity anyhow.
by
Bob Cox
-
AFNI Message Board
This is due to a new security requirement of the NIH, and we are working on altering the Web site to make access to the files possible through Web pages, rather than the now-forbidden directory listings.
by
Bob Cox
-
AFNI Message Board
You can use negative times in TENT or CSPLIN -- this is allowed for just the situation you are in. You might want to do TENTzero(-12,12,13), which should give one "knot" (break-point for the TENT) every 2 seconds, with the "zero" meaning that the first and last points in the TENT-modeled HRF will be set to zero -- thus you will have 11 basis functions.
by
Bob Cox
-
AFNI Message Board
Rick, that backquotes within quotes trick is pretty neat -- I never thought of that!
by
Bob Cox
-
AFNI Message Board
Koene, your last point about unlocking having an undesired side effect, can probably be fixed with a little code tweaking. But that will have to wait until at least next week, after we return from Beijing.
by
Bob Cox
-
AFNI Message Board
I had completely forgotten that I allowed the user to set the color pbar to a completely new set of values via the driver interface. I'm glad you found this, since my amnestic state means I couldn't have advised you on this matter!
by
Bob Cox
-
AFNI Message Board
Daniel's terse answers may be confusing.
By design, the image viewers within a single AFNI controller have the crosshairs locked together. Originally, I had an option to allow them to be unlocked, but I found that this feature could be very confusing, so I disabled it a long long time ago.
You can open a second AFNI controller 'B' via the 'New' button. In that con
by
Bob Cox
-
AFNI Message Board
First off, none of us AFNI drones have ever seen or tried to analyze data such as you describe, so we don't really have any idea AT ALL if these registration steps will work well enough for your purposes. You need to somehow establish a test-bed system where you know there is "activation" and then see how well you can find it with this approach. In the early days of FMRI, the tes
by
Bob Cox
-
AFNI Message Board
This resizing of a cropped window when the dataset is switched should only happen when the new dataset is on a different 3D grid than the old one. If that is the case, the program doesn't know how to crop the same way as before, so it starts over with the un-cropped image view. If the 3D dataset grids are the same, then it can keep the same cropping.
The reason for this "feature&qu
by
Bob Cox
-
AFNI Message Board
To be a little more precise:
You can get the 12 parameters of the transformation in 2 different formats from 3dAllineate.
-1Dmatrix_save gives the matrix that takes (x,y,z) in the base image to (x,y,z) in the source image. This matrix has the form
[ a b c d ]
[ e f g h ]
[ i j k l ]
where the letters stand for numbers. The formula for x{source}, for example, is then
x{source} =
by
Bob Cox
-
AFNI Message Board
IMHO, it is too conservative to use this correction for FMRI data, since the correction is made to allow for arbitrary non-independence in the random values -- and is not needed if the variables are positively correlated -- which smoothing usually enforces.
However, if you (or anyone else) wants to use this correction, I've added a new environment variable AFNI_NON_INDEPENDENT_FDR -- if y
by
Bob Cox
-
AFNI Message Board