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 2191 - 2220 of 2305
3dclust and 3dROIMaker were written by separate folks, so I guess that's why the functionality is a bit different.
Two snippets from 3dclust help file:
* Clusters are defined by a connectivity radius parameter 'rmm'
Note: by default, this program clusters on the absolute values
by
ptaylor
-
AFNI Message Board
Hmm, if you're not using a '-refset', then that shouldn't be happening... Are you sure that the blobs are *really* independent, or do they share any voxel edges/nodes?
--pt
by
ptaylor
-
AFNI Message Board
That's interesting about the -flip_z that appears to be ingrained currently. At the Bootcamp, I believe that their data hadn't needed to be flipped.
The main problem with flipping of gradients is that it is a scanner-by-scanner issue. From some of my data sources, I have to use -flip_y, from others, -flip_z (rarely do I have to use -flip_x). That's part of the main annoyan
by
ptaylor
-
AFNI Message Board
Ok, I'm glad that worked out. The whole 'flipping vector' thing is a pesky annoyance, but I'm not sure how to eliminate/automate the process. Unfortunately, it involves a bit of guessing/testing.
One thing that might help is a program that's just been added to AFNI/FATCAT: 3dTORTOISEtoHere. You can take a DT in TORTOISE format and convert it to that of AFNI, while
by
ptaylor
-
AFNI Message Board
Hi-
Different programs use different orderings of columns of the bmatrix (row- or diagonal first; also, there's a difference of a factor of 2 in the off-diagonals). Assuming that you are using the correct orderings for a given program, then the gradient and bmatrix are mathematically equivalent.
The difference could involve the 'flipping of the gradients' difficulty... In
by
ptaylor
-
AFNI Message Board
Ah, magical-- that has done it. (I did have to include also a '-ljpeg' in the same manner.)
A very large thanks,
pt
by
ptaylor
-
AFNI Message Board
Am sending an email with the long output text. I will report back later here on findings for the benefit of future generations.
--pt
by
ptaylor
-
AFNI Message Board
I installed libpng in my /sw/lib directory, but I still get the same error message
Listing my lib{Xm*,png*}:
/sw/lib/libXm.3.0.0.dylib /sw/lib/libXm.dylib /sw/lib/libpng.dylib /sw/lib/libpng15.15.dylib /sw/lib/libpng16.la
/sw/lib/libXm.3.dylib /sw/lib/libXm.la /sw/lib/libpng.la /sw/lib/libpng16.16.dylib
/sw/lib/libXm.4.dylib /sw/lib/libpng.3.dylib /sw/
by
ptaylor
-
AFNI Message Board
Ah, the obvious library then... (I had been trying libXm.a and other ones).
No libpng there:
bash-3.2$ sudo fink install libpng
Information about 12309 packages read in 1 seconds.
Failed: no package found for specification 'libpng'!
Maybe I'll try pip or something to get it then.
Cheers,
pt
by
ptaylor
-
AFNI Message Board
I updated some packages with fink recently on my Mac OS 10.6, and now I can't build AFNI from source on it anymore. I don't see how to 'undo' the update, nor to view the history of what was done. The break gives me the following error message (below), and any suggestions on what I should now update/install/remove would be quite welcome. I've tried a few packages based
by
ptaylor
-
AFNI Message Board
Would it be possible to have two separate 'Define Overlay' panels? For example, right now, if there is no overlay, there is not define overlay panel. When I add an overlay, I get the panel. If I were able to add *another* overlay, then there could be a 'Define Overlay2' panel pop out over the other one, whose viewing I could toggle on/off-- much in the same way I can have a
by
ptaylor
-
AFNI Message Board
I assume it would require having another 'Define Overlay' panel, called, for example, 'Define Overlay2', which could be a duplicate of the current one. I would imagine in most cases, a user would manipulate the separate colorbars/thresholding in each panel. One might make one colorbar have a yellow-red family, and the other with 'cool'/blue colors, and overlaps could
by
ptaylor
-
AFNI Message Board
What a way to start the day.... Fantastico, thanks.
Re. Milano, it's looking less likely at the moment, but hopefully Hamburg?
--pt
by
ptaylor
-
AFNI Message Board
Ah, beautiful, thanks. All that functionality *and* something to read about p-values.
But it still seems like I should have that functionality from my ~/.bashrc settings with:
export AFNI_SLAVE_FUNCTIME=YES # time index changes overlay AND underlay
export AFNI_SLAVE_THRTIME=YES # time index changes threshold, too
I've tried putting these same settings into my ~/.afnirc as well (wi
by
ptaylor
-
AFNI Message Board
I used to have the functionality of having the Olay and Thr brick in a dataset linked together when I use the 'index' up/down button on the left of the main controller-- that is, I can switch Olay brick and Thr brick simultaneously. This is not default behavior, I don't believe, and I accomplished this with some the settings in my ~/.bashrc file (the AFNI*SLAVE* ones):
export A
by
ptaylor
-
AFNI Message Board
There are many, many tools out there. The problem with just registering volumes for motion correction is that the contrast in DWIs is different than that of reference B0s, and also that the contrast in each DWI effectively changes as the gradients change (ie., the corpus callosum signal may be large from one gradient angle, but small from another).
Probably the most comprehensive tool for d
by
ptaylor
-
AFNI Message Board
Ok, thanks. At the moment the 3dThreetoRGB sounds the most appealing, given the final control in the window.
--pt
by
ptaylor
-
AFNI Message Board
If I could promote a 'feature request' for AFNI, it would also be to allow two overlays to be selected dynamically in the main controller window. My reasons aren't purely functional but DTI-tractographic. It would be really nice to have the ability to overlay both a network of target ROIs and the tractographic results associated with them.
I see there are ways of doing this
by
ptaylor
-
AFNI Message Board
Right. For 3dDWUncert, the expectation was (originally) that the gradient input would be the normal form when if getting data from 3dDWItoDT. The bmatrix input was included mainly for people using TORTOISE, in particular, and that's why it matched with that notation.
So, at present, I would suggest just using the gradient format, via 1dDW_Grad_o_Mat'ing if it's not already. (
by
ptaylor
-
AFNI Message Board
Peter-- thanks, yep, that must be it.
Isaac-- nice to know about the new feature(s)-- quite elegant. I think I might have 'cheated' before by having the [0] and [1] brik be the same set.
--pt
by
ptaylor
-
AFNI Message Board
I think I might have gotten this from one of Rick or Daniel's previous posts, but howz about: concatenating your ROIs into a multi-brick data set, and then using 3dTstat to give each voxel a value based on the brik value at which the max (or, ideally, only nonzero?) voxel value occurs?
$ 3dTcat -prefix ROI_BRICKS.nii.gz ROI_FILE_01 ROI_FILE_02 ROI_FILE_03 ....
$ 3dTstat -argmax -pref
by
ptaylor
-
AFNI Message Board
I'm not sure what units you are working in; a standard DTI scale for diffusivity in human parenchyma is that MD is approximately (or of order) 0.7 x 10^{-3} (mm^2)/s.
In 3dDWItoDT, the b-value is not used in the tensor fit, and so the numbers output are really unitless. Thus, if you want unit-bearing diffusivity measures, you should divide by your bvalue (using both the number and t
by
ptaylor
-
AFNI Message Board
I think you will probably have to be more specific about the 'errors' you get with melodic-- can melodic not run at all? does the command hang? are your output ICs not shining examples of standard resting state networks?
Also, did you check your data for motion/artifacts visually before the ICAing?
It would probably also be good to include your method of converting to NIFTI fi
by
ptaylor
-
AFNI Message Board
Hi-
I think whatever processing you apply to your resting state data (I generally use something like afni_proc examples 9-10b), once you get your LFFs/errts at the end, you can use them in 3dReHo.
Is that specific enough?
--pt
by
ptaylor
-
AFNI Message Board
Hi-
As Rick stated, I believe that your second method of using anaticor in afni_proc and using separately 3dRSFC afterwards is fine, you just don't want to bandpass in afni_proc. And as Rick also stated, if you're going to calculate spectral parameters (ALFF, fALFF, RSFA, etc.), then you don't want to do censoring.
Therefore, you would leave the following three options out
by
ptaylor
-
AFNI Message Board
I would try blending the examples using each from the afni_proc.py help. Precisely how to do so depends on your larger afni_proc'ing philosophy. So I guess starting with 10b, or not 10b-- that is the question.
For example, 10b is currently the following:
afni_proc.py -subj_id subj123 \
-dsets epi_run1+orig.HEAD
by
ptaylor
-
AFNI Message Board
I think warping the standard space ROIs to native diffusion space is the way to go-- I don't see how it's feasible otherwise, given all the warping involved and the need to keep gradients consistent with the data (a difficulty you point out in your message above).
You can make a map from standard -> native space using either 3dQwarp or 3dAllineate, depending on which you feel is m
by
ptaylor
-
AFNI Message Board
Hmm, nope, 3dROIMaker shouldn't care whether it was a Z-score, a T-stat or any other data sets-- it's all just numbers.
One thing to check *might* be if there were NAN or bad values somehow in the T-map that aren't in the Z-score set.
However, the skeletoning part certainly shouldn't cause badness, either, so I'd like to see what's going on there. I don'
by
ptaylor
-
AFNI Message Board
Hmm, if you want to send me a copy of the data set, I can check what I get on my end, otherwise I'm not sure; just as a test, if you try running without the options such as thresholding, skeleton, etc. does that make a difference (it shouldn't...)?
--pt
by
ptaylor
-
AFNI Message Board
I'm a bit perplexed. I've been using ROIMaker on a few different sets and not had issues with it. It looks like that's from a recent build as well? There's a new-new version from yesterday.
Have you run the FATCAT_demo scripts on your computer, and does Do_03*tcsh give the same error?
--pt
by
ptaylor
-
AFNI Message Board