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
The particular foolishness in this case was an attempt to slice-time-correct a dataset in a way that didn't bias it towards a particular slice, so that I could run ICA on it. Then I realized that Fourier interpolation already does that.
by
Isaac Schwabacher
-
AFNI Message Board
If you have Xvfb, you might also have the shell script xvfb-run, which makes things easier. You can probably get away with "xvfb-run -a COMMAND", or "xvfb-run -a -s '-screen 0 800x800x24' COMMAND" if you want a bigger screen.
by
Isaac Schwabacher
-
AFNI Message Board
Hi gurus,
I was just using 3dUpsample to do something silly and I noticed that while the script correctly reduced the TR of the upsampled dataset by a factor of N, it also reduced the slice time offsets by the same factor. Since these offsets aren't affected by the upsampling, this probably should not happen. $ 3dinfo -tr elvis+orig
2.000000
$ 3dAttribute TAXIS_OFFSETS elvis+orig
1.9
by
Isaac Schwabacher
-
AFNI Message Board
In the example you gave, cp fails because there's no file named "${OLD_DIR}/to_3D_out"; the files are "${OLD_DIR}/to_3D_out+orig.HEAD" and "${OLD_DIR}/to_3D_out+orig.BRIK". Then AFNI can't find the copied files because they never existed. Try cp $OLD+orig.* .
rename out original ${OLD##*/}+orig.* instead and see what happens.
Isaac
EDIT: Whoops! R
by
Isaac Schwabacher
-
AFNI Message Board
Hi gurus,
This is not a big deal, but I just tried out @2dwarper.Allin because I'm working on data in which I expect motion and physiological noise to be important, and after it finished processing (and writing data, thankfully!), it crashed with the message "ver: Undefined variable.". A little digging showed that the script never executes the "ver = 6" in line 15; it
by
Isaac Schwabacher
-
AFNI Message Board
Funny you should ask. I've been meaning to convert this rough-around-the-edges python function I wrote to something that doesn't use numpy (since the AFNI python libraries for some reason explicitly avoid that wonderful and free piece of software) and contribute it, but if it helps, here's the function as is:
import re
import numpy as np
def read_CENSORTR_string(data, run
by
Isaac Schwabacher
-
AFNI Message Board
Hi gurus,
I recently modified my workflow to complain (output flashing red text, ring the terminal bell, e-mail me, set off all the car alarms in the parking lot, that sort of thing) whenever the string "error" (case insensitive) appears. Currently this happens whenever I use 3drefit to change the view on a NIfTI dataset, which I have to do because other programs don't handle t
by
Isaac Schwabacher
-
AFNI Message Board
Does 3dcalc -datum byte -prefix A_gt_B -a functional_data_set_A+tlrc -b functional_data_set_B+tlrc -expr 'step(a-b)' do it for you? This should give a mask that is 1 wherever A>B and 0 wherever A<=B.
Isaac
by
Isaac Schwabacher
-
AFNI Message Board
I don't think that will give you the cross terms, though. I wonder how much of a pain it would be to hack 3dcalc respectively 1deval to allow you to specify more than one -expr option, thereby creating more than one output subbrik respectively column. I guess other solutions are
3dTcat '3dcalc( ... )' '3dcalc( ... )' ...
or (bash only)
1dcat <(1deval ...) <(1dev
by
Isaac Schwabacher
-
AFNI Message Board
Is this what you want?
#!/bin/env python
import numpy
firstorder = numpy.loadtxt('motion_all_demean.1D')
secondorder = numpy.hstack(firstorder[:,(i,)]*firstorder[:,i:]
for i in range(firstorder.shape[1]))
numpy.savetxt('motion_all_demean_secondorder.1D', secondorder)
If your input columns are x, y and z, this will give output with colum
by
Isaac Schwabacher
-
AFNI Message Board
I guess that will do it. Yes, both the built-from-source copy and the linux_openmp_64 copy work (and give identical results).
by
Isaac Schwabacher
-
AFNI Message Board
Hi builders,
The linux_xorg7_64 AFNI distribution doesn't include program 3dMatch. I can build the program from source without any trouble (and without modifying any makefiles), so I'm confused as to how it's being omitted. I thought you'd like to know about this.
Isaac
by
Isaac Schwabacher
-
AFNI Message Board
That's the answer I was afraid of. :( But I don't see a way to code the general case while avoiding having race conditions where your data vanish out from under you.
by
Isaac Schwabacher
-
AFNI Message Board
Hi Daniel,
On the subject of using the first pre-steady-state EPI image for alignment, I had been doing that until we upgraded our scanner, after which the first frame became very stripy. (I'm not sure why it wasn't stripy before, since the acquisition was always interleaved...) Do you think it's still worth using that image for alignment, possibly after some sort of slice-wis
by
Isaac Schwabacher
-
AFNI Message Board
That sounds like the sort of devious, devilish device that only a destroyer of worlds would employ. Is that behavior part of afni.c, or is it in mcw_malloc.c where just anyone could use it?
by
Isaac Schwabacher
-
AFNI Message Board
One thing you can do if you don't have good physiological data is to use Erik Beall's PESTICA, which is a MATLAB package that uses independent component analysis to estimate the physiological regressors directly from your fMRI data. PESTICA can be found here:
by
Isaac Schwabacher
-
AFNI Message Board
...malloc() returned null and it tried three more times? o.O
by
Isaac Schwabacher
-
AFNI Message Board
I just downloaded the latest source. I had to make parser.o and parser_int.o (and libmri.a, but that was expected) in order to make ptaylor_all, but now I have 3dMatch.
by
Isaac Schwabacher
-
AFNI Message Board
Is 3dMatch still in development? It's not in my distribution (which I just updated looking for it) and it doesn't appear in the program help listings on the site.
by
Isaac Schwabacher
-
AFNI Message Board
Quote There are 3 trial types, each with 3 sequential(jittered) phases within each trial type, so 9 iresps. The appropriate amount of jitter was included in the task design (TR=2sec).
Wait a minute... I could be crazy, but it sounds like you're using separate regressors for the same stimulus type jittered to different time lags. If that is in fact what you're doing, is there a partic
by
Isaac Schwabacher
-
AFNI Message Board
I think it's caused by '%' followed by an identifier character in your post. Not sure why though.
EDIT: It looks like you can't PM any message that you couldn't post.
by
Isaac Schwabacher
-
AFNI Message Board
Would it make sense to graft an attribute into the header of a timeseries dataset that tells how many degrees of freedom it has? That way you could have to3d create a dataset with DoF = NT, and things like 3dBandpass and 3dDetrend could decrement the counter as needed. Any time the attribute wasn't present, you could assume that DoF == NT so that no existing pipelines would break.
by
Isaac Schwabacher
-
AFNI Message Board
I don't know whether this is a side-effect of an earlier fix to AFNI_1D_TRANOUT or not (where -prefix stdout: was writing to ./stdout:), but setting that environment variable to YES forces any 1D-ified outputs that aren't bound for stdout to be written to the current directory instead of the specified path:
$ 3dDetrend -DAFNI_1D_TRANOUT=YES -polort 1 -prefix temp.1D '1D:1,3,4,2
by
Isaac Schwabacher
-
AFNI Message Board
Got it! If you're logged in and looking at a post, there's an author field in the reply box that's autofilled (and not modifiable). When you search using the bar at the top of the page, the search function reads your login name in the author field of the reply box and feeds it to the search. I can also search for posts by you if I'm logged out: if I put "rick reynolds&
by
Isaac Schwabacher
-
AFNI Message Board
Right, so now I'm extremely confused. Relogging worked, but closing and reopening Firefox did not.
by
Isaac Schwabacher
-
AFNI Message Board
Aha! When I'm logged in and I just type something into the search bar at the top of the page, it only searches for posts of which I am the author.
...why on earth is that the default behavior?
by
Isaac Schwabacher
-
AFNI Message Board