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 4381 - 4410 of 4524
Hi Isaac,
Since the Emporer is preparing for a presentation, let me make a
few comments.
Like all voxelwise programs, 3dcalc assumes that all voxels are
located consistently in space, and performs one computation per
voxel across datasets. So if there is reason to believe that the
voxels are not consistent, a warning is issued, at least.
The IJK_TO_DICOM_REAL attribute was put there
by
rick reynolds
-
AFNI Message Board
Unfortunately, I think the help is accurate in that dmBLOCK is the only basis
function family that allows for duration modulation at the present time. The
SPMG functions take a duration parameter, but that duration would need to
be constant across stimulus presentations.
If you married a parameter in SPMG2, it would result in an amplitude
modulation, not a duration modulation.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Guillaume,
The program suggests rotating the color bar to get other color patterns.
Just below and to the right of the 'autoRange' text are the 'Rota' buttons,
to rotate colors in the bar in one direction or the other.
Does that help?
There do not seem to be options for altering the default colors, probably
because of that rotate ability.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Becky,
I will PM you with instructions. See "Private Messages" in the upper right.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Yiyu,
That seems right (3dANOVA2 -type 3). Basically, each of those
contrasts would be done as a t-test of the contrast weight
against zero, with one value per each of the 13 subjects.
---
Since we are on the topic of 3dANOVA2, let me suggest that you
look at "example C 2" from "gen_group_command.py -help". That
can help write such a command.
A simple case
by
rick reynolds
-
AFNI Message Board
Hi Liana,
For duration modulation alone (which is what I was commenting on),
AM1 is the only appropriate choice.
For amplitude modulation alone (which is what the 3dDeconvolve help
refers to), then AM2 is usually the more appropriate choice (so one
models the modulation separately from the average response).
When doing both, the second point applies.
Basically, duration modulation
by
rick reynolds
-
AFNI Message Board
Try running "3drefit -markers" on the dataset first.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Marwa,
That was not exactly my point. If the motion parameters were also
regressed out (which they should have been), then maybe they are
still in your "clean" data.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Marwa,
Bob suggested other options that would make 3dTfitter
significantly faster. Please see this thread:
http://afni.nimh.nih.gov : msg-83351
Regarding timing_tool.py, please consider example 6a in the
-help output as well, and maybe the help for option -timing_to_1D.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Colm,
That looks like a problem with OpenMP. Since you have a 10.7 system, try
either the 10.7 binaries, or maybe 10.6.no.omp (which would be slower without
OpenMP).
- rick
by
rick reynolds
-
AFNI Message Board
Re: PPI - 11 years ago
Hi Marwa,
If your regression had motion parameters, which it should have, then those motion
curves are still in your "clean" data and are probably causing the spikes.
In 3dSynthesize, instead of selecting 'polort', consider selecting 'baseline' (which
should include motion, certainly if you used afni_proc.py for the analysis).
- rick
by
rick reynolds
-
AFNI Message Board
Hi Becky,
How well is the EPI aligning with the anatomy? If you go to a good voxel location
in the volume, and you look at how the surfaces sit on the volume there (i.e. with
afni and suma talking), is that voxel between the smoothwm and pial surfaces?
If you are unsure of the registration or anything, perhaps you could upload all of
the data needed to run the afni_proc.py command and t
by
rick reynolds
-
AFNI Message Board
Hi Karen,
Note that you can edit the afni_proc.py command that is displayed
by uber_subject.py when it is created via the left action button.
You can edit and save directly in that window.
Some advantages of editing there is that uber_subject will take care
of removing the old results directory before running your new script,
as well as saving the processing output in a text file.
--
by
rick reynolds
-
AFNI Message Board
Hi Melanie,
Understanding why the mask extends so far out would require looking at
the EPI data. A mask that large is unusual in humans.
For details about masking, including why we do not suggest it at the
single subject level, look at the afni_proc.py -help output, and see the
section titled "MASKING NOTE" (there are lots of NOTE sections :).
To summarize, it makes some se
by
rick reynolds
-
AFNI Message Board
So there are 2 things needed for running the 3dClustSim operation,
an EPI mask (via the mask block) and blur estimates (via either
-regress_est_blur_errts or _epits).
Maybe I should have AP output a comment about whether it will
be estimating the blur. These little peculiarities can be confusing
(though it mentions the requirements in the help for the
-regress_run_clustsim option, but on
by
rick reynolds
-
AFNI Message Board
Hmmm, -regress_apply_mask should be obsolete.
I will check on that when I get the chance.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Luke,
It is not appropriate to add stim timing files that collapse across
stim classes. For example, I see the file stimes.ALL.1D included.
If this is a combination of all other timing files, 3dDeconvolve will
certainly fail.
This is not an AFNI thing. There is nothing wrong with a factorial
design.
What you are probably trying to do is use contrasts in 3dDeconvolve.
Consider t
by
rick reynolds
-
AFNI Message Board
Hi Liana,
Check your proc script for 3dClustSim. Note that there must
be a 'mask' block applied in order for the cluststim step to run.
If you are unsure, please post (or email me, or 'PM' me) your
afni_proc.py command.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Eliana,
Right-click on the continuous color bar (the blue to red bar) and
choose one of the ROI color bars, such as ROI_i128. There
are many others there, as you will see.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Liana,
The mask block is needed for some computations, but it is not
actuallly applied to the single subject results (since we think
that is a bad idea).
If you really want it applied to the single subject results, add
the -mask_apply option, with one of the paramters: epi, anat,
group, extents, e.g. to apply the EPI mask:
-mask_apply epi
But we suggest making your own grou
by
rick reynolds
-
AFNI Message Board
Hi Matthew,
Since you are showing a group result, the stdev for the bar graph
should come from the variance across subjects, not across voxels.
Basically you are showing what 3dttest++ might show, but just for
one set of values.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Guillaume,
Looking into it, that basically means maybe your binaries need to be
updated. The @AddEdge script now looks for open port blocks, so
that other instances of afni, etc. do not interfere.
Anyway, if you are able to, update your binaries and try again.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Nicole,
Try installing the macosx_10.6_Intel_64.no.omp.glib package
and see how it goes.
Do you know how to set your path and such? What have you done
so far?
Note the HowTo #0 page for 10.6 (even though you are 10.7). But
for the binaries I have noted, you can skip steps 2 and 3 (Xcode, fink).
- rick
by
rick reynolds
-
AFNI Message Board
Hi Mingbo,
Please let us know of individual directories that you are interested in.
We have made listing for some of the main directories, but not all.
- rick
by
rick reynolds
-
AFNI Message Board
Hello,
Your basic model seems okay to me.
In order to compare the modulators in the contrast, you should
specify index #1 (since index #0 is the unmodulated response).
Consider this:
gltsym 'SYM: n_A[1] -n_B[1]'
As for alignment, try adding options to the align_epi_anat.py
command to 1. allow for a large shift between the unaligned
volumes, and 2. use a cost funct
by
rick reynolds
-
AFNI Message Board
Hi Pengmin,
The warning does not imply there is any problem.
If the datasets are oblique and you do normal afni_proc.py processing, aligning to
anat and going to standard space, then do not worry about it. Just make sure that
everything is aligned in standard space, as usual.
If you are going to do a single subject analysis in orig space, then you should make
sure about the coordinat
by
rick reynolds
-
AFNI Message Board
Hi Lisa,
I have no trouble with that same command, even if I set AFNI_DECONFLICT to NO
as you have it. My guess is that if you were to update your binaries, it would work.
Since your package is year old, it is probably a good idea anyway (unless you are
in the middle of something).
- rick
by
rick reynolds
-
AFNI Message Board
Hi Liana,
Let me just make a few points:
1. Motion will tend to cause outliers, so the TRs above each of the
limits might entirely overlap.
2. Motion censoring censors the previous TR, as well as the current,
since we do not know when the motion occurs. Ah, you noted this.
3. Even with censoring the previous TR, high motion often occurs over
multiple TRs, e.g. with 4 cons
by
rick reynolds
-
AFNI Message Board
Try our direct path (to pub/dist/tgz):
http://afni.nimh.nih.gov/pub/dist/tgz/macosx_10.7_Intel_64.tgz
- rick
by
rick reynolds
-
AFNI Message Board