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
Page 1 of 12
Pages: 12345
Results 1 - 30 of 356
You should only use that Makefile if you're using the Intel C Compiler (icc). I got it to compile on 10.10 with Homebrew using the macosx_10.8_gcc Makefile, which took some doing, but it's less of a problem than buying a new compiler.
HTH,
Isaac
by
Isaac Schwabacher
-
AFNI Message Board
So am I! Thanks for the tip about the Fedora Makefile; my previous solution involved an automated script calling git rebase to forward-port local changes, but now I can just build from the most recent tag before master, which is so much less fragile.
Yes, it looks like SL 7 is based on RHEL 7.
Thanks!
by
Isaac Schwabacher
-
AFNI Message Board
Scientific Linux 7. I can build from the Fedora Makefile with no problems.
by
Isaac Schwabacher
-
AFNI Message Board
Hi gurus,
I'm not quite as clear on how dynamic linking works as I'd like to be, but this seems fishy. The @update.afni.binaries was from a built-from-source install with other_builds/Makefile.linux_fedora_19_64 (is that discrepancy intentional?).
Could this just have happened because libpng15.so was what was there?
Isaac
$ @update.afni.binaries -package linux_fedora_21_64
by
Isaac Schwabacher
-
AFNI Message Board
Here's a link to the earlier discussion of this topic:
by
Isaac Schwabacher
-
AFNI Message Board
Did you also just upgrade to OS X 10.11? One of the security features of that release is that DYLD_FALLBACK_LIBRARY_PATH is ignored.
by
Isaac Schwabacher
-
AFNI Message Board
I had been using a little thing I brewed up, but unless you're doing something nifti can't handle (like complicated slice timings), it's probably easier to convert to nifti and use nibabel. If you want it, though, here's what I had before I threw it out and started using nibabel:
(Be aware that this is python3; you might need to backport it.)
ijs
by
Isaac Schwabacher
-
AFNI Message Board
Well, it could be residual signal, or it could be a math error on my part. I'm going to try to reduce to a submodel that can be examined with 3dttest++ so that I can examine residuals I know are calculated correctly.
This was the most spectacular of the residual images I could find, but almost every single one of them had something that looked like one or more known networks, and these we
by
Isaac Schwabacher
-
AFNI Message Board
Hi Gang,
Thanks for those two points. I had thought that it should be relatively easy to deal with the fixed/random split, but it occurs to me that I don't even have any idea how to get 3dLME to spit out enough information to let me separate the within-subjects residuals from the between-subjects residuals. I don't even know how to extract the beta weights I would need to use to comp
by
Isaac Schwabacher
-
AFNI Message Board
Hi Gurus,
In response to the recent foofaraw about cluster size thresholds, I want to estimate smoothness from my group-level residuals to see if I get drastically different results. Is there an easy way to do this? My current plan involves 3dcalc, but the glts that I used aren't orthogonal, so I have to think about whether I'm using the right parameters (good!) or the dual to the ri
by
Isaac Schwabacher
-
AFNI Message Board
The seaborn python package with a bunch of custom crap thrown in. I'm trying to figure out a way to ship this back upstream, but right now it's a total kludge. How much of a kludge? At the bottom of the post is the horrible code I used to create it.
If and when I beat this into a reusable shape, I'll post a reply.
#!/usr/bin/env python3
import numpy as np
import scipy as
by
Isaac Schwabacher
-
AFNI Message Board
Be careful with the values you extract from such a cluster-- see for instance Fig. 3 of Kriegeskorte, et al. (2009) and Question 9 of Kriegeskorte, et al. (2010)!
Here is an example of a plot I'm in the process of creating, based on the second reference I gave:
I'm trying to show the shapes of the distributions (with the violins), the hierarchical structure of the data (each s
by
Isaac Schwabacher
-
AFNI Message Board
Did this appear in a terminal, or was it saved to an output file? If your label accidentally starts with a bunch of backspace characters, and if I counted correctly, you could get that output in a terminal. Try rerunning your command with "&>temp.out" and opening temp.out in an editor that displays special characters. Then if this is the problem, you'll see this:
> **
by
Isaac Schwabacher
-
AFNI Message Board
Looks like it's comparing 1.10 to 1.5 as numbers. I've submitted a fix on GitHub. ...And a fix for the fix. Derf.
P.S.: If you want my version before it's been checked over and incorporated by the AFNI gurus, you can get it at . I recommend diffing the new version against the old and making sure you understand the differences (they're pretty minimal) before running somethi
by
Isaac Schwabacher
-
AFNI Message Board
Here's an example:
In this image, there are two black disks on a white background. One of the disks is half a pixel higher than the other, and this is clearly visible. If I wanted to align the two, I could upsample both images to get higher-resolution but blurry (or in some other way degraded) versions of both. Aligning these would be easier, and then I could report what I did ("
by
Isaac Schwabacher
-
AFNI Message Board
Now that I've gotten around to actually making use of this, I've put my script up as a GitHub gist.
When you have multiple basis functions (e.g. with -stim_times_AM2 or TENT or a multi-row GLT), the script produces pairwise histograms of p-values from each basis function. I'm still trying to figure out whether they're useful for anything, but they were just so darn easy...
by
Isaac Schwabacher
-
AFNI Message Board
I'm not an expert, and especially not in tcsh, but I see two things that look suspicious:
- The regurgitated command has "imagery_stim_times.*.1D" in it, which probably should have been expanded by then. Do those files exist?
- "$: Command not found." suggests you have a line that starts with "$", and indeed, I see "|$" in your command. I think tha
by
Isaac Schwabacher
-
AFNI Message Board
Hi Gang,
I ran the analysis, and it looks like study/scanner has a big impact everywhere (voxelwise p < .005):
This map looks too meaningful, and I want to wrap my head around it.
To give some background on the processing stream, I followed Ashburner's 2010 VBM tutorial (warning: PDF), which specifies the following preprocessing steps:
Segment the images into GM, WM and CSF
by
Isaac Schwabacher
-
AFNI Message Board
I had this problem too and just put in a PR on github. It looks like it's just a missing initialization; R_io.so takes care of filling in the field if it's not provided.
by
Isaac Schwabacher
-
AFNI Message Board
OK. I'll try it and see what happens. Thanks!
ijs
by
Isaac Schwabacher
-
AFNI Message Board
Yes, that's how I planned it out. It's precisely that assumption that makes me uneasy, but I'm struggling to imagine how to assess it.
by
Isaac Schwabacher
-
AFNI Message Board
I'd look at line 1460 of 3dMVM.R for a place to put debugging statements, starting with removing the suppressMessages() wrapper and the silent=TRUE from the try().
by
Isaac Schwabacher
-
AFNI Message Board
I mean we did two separate studies several years apart, but so many subjects participated in both studies that I can't just lump all the data together, since I expect that the repeated scans from the same subject would violate some sort of sphericity assumption.
The other variables that we're interested in testing are mostly things like average minutes of moderate / vigorous physical
by
Isaac Schwabacher
-
AFNI Message Board
Hi Gang (and other gurus),
I have two studies whose subject pools overlap significantly and I'm trying to run a VBM analysis using both sets of data. I want to be able to take into account a number of complications:
Our scanner was replaced between the two studies. I have no idea what the effect of this improvement is on the GM probabilities estimated by SPM. The second study was carr
by
Isaac Schwabacher
-
AFNI Message Board
You could be swapping. If you're running on linux, figure out the PID of your 3dMVM process, then follow it to its child R process (3dMVM itself is a tcsh script that calls R, so it won't be using much memory itself). I'd use pidof or pstree for this, though you can do it with just ps. Then you can `grep VmSwap /proc/$pid/status` and see if your R process is using a lot of swap
by
Isaac Schwabacher
-
AFNI Message Board
Every time I need to know this stuff I get confused and have to work it out from scratch-- and then I only know it for as long as I can hold it in my head. You could try `3dcalc -a fred+orig -expr t` to get a dataset containing the time at each voxel; this takes slice timing into account. Other than that, there's the help in README.attributes under "Time-Dependent Dataset Attributes&q
by
Isaac Schwabacher
-
AFNI Message Board
The first thing I try when I see errors involving ___R.read.AFNI files is to add -cio to the command line of the offending program.
by
Isaac Schwabacher
-
AFNI Message Board
That is from Fforde! I recognized him, but I could have sworn your use preceded his, so you must have been referring to the original source, which I could never find despite much effort!
by
Isaac Schwabacher
-
AFNI Message Board
You could try using 3drefit to multiply all the dimensions in the dataset by 10, then clusterize and divide by 10...
by
Isaac Schwabacher
-
AFNI Message Board
Page 1 of 12
Pages: 12345