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 4111 - 4140 of 4524
That all sounds good (and expected). The Makefiles are based
on ease of distribution, more than ease of a local use.
Glad it's working for you,
- rick
by
rick reynolds
-
AFNI Message Board
Uncomment the line after "in case user wants to override with system dynamic libs":
XLIBS = -lXm -lXt
These are reasons why I do not suggest compiling, unless you really need to.
It sounds like compiling is not going to gain you much, unless you just want a
better understanding about it.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Gabe,
When logged in, go into the 'Control Center' and edit the
'Forum Settings'. Under that panel, change "Message
reading display options" to whatever suits you.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Gabe,
The Makefiles are set up for us to compile and distribute binaries easily.
But depending on your OS and version, things might vary quite a bit, and
there may be many libraries to install. Having an openmotif-devel package
available is not even a given.
You started with one of the Makefiles that uses CCOLD to get around
an old compiler problem, using -V. Since your compiler do
by
rick reynolds
-
AFNI Message Board
Hi Cynthia,
As Peter states, there is a -num_glt option. Without that option, the
internal limit is 10 GLTs. You probably have more than 10 but are not
using the -num_glt option.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Bryce,
This has been fixed. Thanks for bringing it up.
- rick
by
rick reynolds
-
AFNI Message Board
I would not expect them to add too much beyond the others.
Note that the outlier fractions used by afni_proc.py might
make more sense in a similar way. The motion derivatives
also might make more sense.
One aspect of the rms time series that I do not like is that
they will generally have a big spike (downward) at the
registration base TR, since the rms value there is alawys 0.
One would
by
rick reynolds
-
AFNI Message Board
Hi Swaroop,
I guess that would be: 3dBandpass -band LP HP ...
But in any case, those should be basically the same if
the motion_params time series had also been bandpassed.
If not, some amount of those removed frequencies would
be reintroduced from the motion file into the result.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Gaurav,
Sorry for being slow. It looked like your post had been replied
to, but it was by you.
I'm going to pass on addressing the validity of this much. It
seems okay to me though, providing you make clear that these
are residual correlations during task.
Are you planning to regress out all of the tasks and baseline
before running the correlation analysis?
Rather than c
by
rick reynolds
-
AFNI Message Board
Hi Ricky,
I will make a note to at least try this. Once we get a new machine to
set up virtual servers on, we will try to make a Fedora Core 3 build of
those binaries. But it will be a while, and sometimes it is difficult to
install old operating systems on new hardware. We can see.
Another option is to compile yourself. If you have all of the library
packages, including openmotif,
by
rick reynolds
-
AFNI Message Board
Hi Daniel,
1) Yeah, I keep forgetting to add tshift to example 9. That example
was based on an analysis in a case where the slice timing did not
exist. I just never remembered to add it...
2) It is up to you what to do with the low frequency vs polort. They
are not quite the same, so I left both in. It probably does not matter
much either way. Maybe I will revisit that and see how
by
rick reynolds
-
AFNI Message Board
Hi Nicole,
Actually, just your script calling afni_proc.py is enough. You
just have a few scripting problems in there. The errors have
nothing to do with afni_proc.py.
1. The -copy_anat line has a comment at the end of it.
That is not legal because you are trying to use the '\' line
continuation character. The result of this should be a
'-do_block: Command not found&
by
rick reynolds
-
AFNI Message Board
Hi Nicole,
This is likely not a problem with afni_proc.py at all.
I expect that the script was corrupted somehow, and it is no
longer a Unix text file. Try running dos2unix on it.
In any case, it would be nice to see the actual text output,
from the beginning down to the first errors. If you are running
it the way it suggests, then you have an output.* file with all of
the text out
by
rick reynolds
-
AFNI Message Board
Hi Steve,
Try 3dTsort or 3dRank. Once you rank the voxels, use 3dcalc
to set anything larger than 100 to 0.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Lydia,
The TENTzero() basis functions are identical to those of TENT,
except that the first and last basis function are not included (so
the endpoints are *assumed* to be zero, and are not estimated).
So TENTzero(0,14,8) is identical to TENT(2,12,6). I suggest you
stick with TENT(2,12,6) because it is more clear what you are
doing. It is not terribly appropriate to plot the zero e
by
rick reynolds
-
AFNI Message Board
Note that if t < T0 (or T1), the curve is considered 0.
But what is probably throwing you off is the default
scaling of the curves. Unless AFNI_CONVDIFFGAM_DO_SCALE
is set to 'NO', the base curves are scaled to be unit
height (so if A0=1.0, then the peak is 1.0). If you turn
off that scaling and re-compute and re-plot the results,
they should match.
The peak of the curve
by
rick reynolds
-
AFNI Message Board
I'll take a closer look, but I don't like the 'i' in the matlab
output. That function should send real numbers to real
numbers.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Noah,
If you are trying to compare average head motion between groups,
then the question becomes how to compute the average head motion.
Note that when using afni_proc.py to process the resting state data,
it will output the average motion (both uncensored and censored) at
the end of the processing. Those averages come from the enorm
dataset, which is the Euclidean norm time series
by
rick reynolds
-
AFNI Message Board
Yes, that is the formula given in the help output, and is what
is applied. But it is difficult to say much without seeing what
you are doing.
How did you generate the fitts? More importantly, what is
your reverence time series? And what values did you test
with in your R function?
The gamma_model() function looks good to me, including
scaling of the peak to A.
However there may b
by
rick reynolds
-
AFNI Message Board
Note that you should be able to update your binaries via:
@update.afni.binaries -d
Please let me know if you run into any related questions after
the update.
- rick
by
rick reynolds
-
AFNI Message Board
Hi SL,
You probably just need to update your AFNI binaries.
I suspect the -regress_stim_types option is not known
in the (presumably old) version that you have, and it
is reading that option, plus the subsequent types, as
additional -regress_stim_times parameters.
That option was added May 21. What does "afni -ver"
say?
- rick
by
rick reynolds
-
AFNI Message Board
Hi Gabe,
Is there some aspect that does not seem to be working?
The outline is pretty straightforward. First, make a list of
all extrema (depending on what the test for that is), then
remove those that are too close together, saving the more
extreme ones.
Doug Ward wrote 3dExtrema. I am much more comfortable with
3dmaxima (though it is about 14 years old). Unfortunately,
3dmaxima
by
rick reynolds
-
AFNI Message Board
Note that -automask should not be used on residual (or any
detrended) time series. The automask operation basically
requires the brain region to be much higher intensity than the
background.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Isaac,
Just out of curiosity, why are you even looking at iszero(i),
for example? Are you adding a (partial) border to the image?
I would have expected the median to be sufficient.
- rick
by
rick reynolds
-
AFNI Message Board
Hi Henry,
The most details are in the example usage in the -help
(example 15a).
The option is meant to calculate the maximum difference
in the subject's head locations across all pairs of time
points. If there were only 1 motion parameter, it would be
the difference between the min and the max. With 6 motion
parameters, it is the euclidean norm (sqrt(sum squares)) of
the diff
by
rick reynolds
-
AFNI Message Board
Hi Lauren,
It ought to be reassuring that you keep getting the same
result, particularly after using something relatively automated
like afni_proc.py. Note that you can test the contributions of
tissue-based regressors in afni_proc.py using the -regress_ROI
option (takes any of brain, GM, WM, CSF, where 'brain' is the
whole brain mask, from which the global signal is generated)
by
rick reynolds
-
AFNI Message Board
Hi David,
This turned out to be pretty straightforward.
First, note that there is originator (location of origin)
information in the ANALYZE datasets. That was apparently
used in the conversion from ANALYZE to NIfTI. If you tell
AFNI to do the same thing, then the NIfTI files are not
needed.
To make the ANALYZE images come out "correctly" as AFNI
datasets, I did:
by
rick reynolds
-
AFNI Message Board
Hi David,
Sure, just upload a pair of datasets and I will take a look
to see if our evaluations match.
Do you already have the upload link?
- rick
by
rick reynolds
-
AFNI Message Board