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 2 of 3
Pages: 123
Results 31 - 60 of 74
Thanks Rick, that's a good point. I think they were looking for an MVPA type effect but it's (just as?) likely that using AM2 would reveal interesting effects. I'll let them know.
thanks again.
also, what's going on with the message board? why is there a massive colorful brain?
James
by
jkeidel
-
AFNI Message Board
hi Michael
I can't answer all your questions but I do have a few things that might help in the meantime. If the preprocessing done in SPM included smoothing then you need to account for that in your 3dClustSim command.
The data come off the scanner with a bit of smoothness, and the smoothing kernel used adds to this. This smoothness increases the likelihood of finding sets of contigu
by
jkeidel
-
AFNI Message Board
hi Rick
Thanks for the quick response!
I agree that it is not a lot of stimulus per class. and actually it's worse because the stims are presented for 2 s and I guess I had added 1s as -post_stim_rest. The design is 29 color patches and the idea is to correlate the response to each color with some external variables (I don't know which ones as I am only helping out with the timi
by
jkeidel
-
AFNI Message Board
hi all (especially Rick)
I am helping out someone here at Sussex with using make_random_timing.py to generate a jittered rapid event-related design for data that will then be fed into an MVPA analysis. there are 30 stimuli, which currently are to be presented 6 times each in about 16 minutes total. 3dDeconvolve -nodata does not complain about matrix problems.
here is the output of make_ra
by
jkeidel
-
AFNI Message Board
hey Gang
no problem! but I am still curious about the results of 3dClustSim...how can such a small cluster of voxels at such a lenient threshold be significant?
James
by
jkeidel
-
AFNI Message Board
hi Gang (and Michael)
aren't the two tables quite different in that one is for a 3mm^3 mask and the other for 2mm^3? so really the necessary cluster size in microliters is very different.
also, for such a big mask isn't that a really low number of voxels? 9 voxels at .05 uncorrected? could I ask what the smoothness estimate was, or what the entire 3dClustSim command was?
Jam
by
jkeidel
-
AFNI Message Board
hi Gang
thanks so much for your response. the clips were randomized with respect to their order but had a fixed 16 s interval between them. there were 20 of one category and 10 of each of the others (these latter two types are the ones of most interest). as we are not interested in the statistics at the first level and were more interested just in the effect estimates I think this is fine--
by
jkeidel
-
AFNI Message Board
hi Gang
Sorry, I wasn't being clear enough. the stimuli in question are all the same, they are clips from TV shows. the manipulation is previous exposure, basically. it's just that because some of the stimuli are longer than others there is no way to model the whole video exposure for each stimulus. So in the case of TENTzero (-3, 25, 11) I would be modelling the first 25 seconds
by
jkeidel
-
AFNI Message Board
hi all
We have a study in which participants view three kinds of video in the scanner. Because of the nature of the stimuli it was not possible to equate them in length, so instead they range from about 25-40 seconds. I have modelled them using -stim_times_AM1 with dmUBLOCK following Bob's recommendation (as I understand it) in the 3dDeconvolve help.
While the results are fine, I am
by
jkeidel
-
AFNI Message Board
hi Jatin
I am sure that other people on the board can give you more extensive help, but the first thing that jumped out at me is that you have not specified a -master dataset in your 3dNwarpApply command. The help states that you can often use the warp dataset here, so for you the t1hQw_WARP+tlrc dataset.
hope this helps, if not I'm sure someone else can
James
by
jkeidel
-
AFNI Message Board
hi Will
as a quick workaround I think you could use the -aglueto option in 3dbucket and loop through the values in reorder.1D in a foreach loop.
JK
by
jkeidel
-
AFNI Message Board
you can put an asterisk in the place of runs with no events for that condition in a run. so if you had 4 runs, and only 1 and 3 had errors, your file would look like this:
time1 time2...
*
time1 time2...
*
also, probably worth adding -local_times to your 3dDeconvolve command, though AFNI is very smart and usually figures out what you mean anyway.
James
by
jkeidel
-
AFNI Message Board
hi Mingbo and Gang
I have tried to do this as well and it did not work for me either. Specifically, the SPMGx basis functions do take an optional duration argument, but this cannot currently be used in combination with -stim_times_IM (though it can with just -stim_times), while say dmBLOCK can be used with -stim_times_IM.
I don't know how much work it would be to make 3dDeconvolve all
by
jkeidel
-
AFNI Message Board
hi
I think your command will work if you transpose your timing file into a single column, or just follow it with a backslash and single quote:
-time sampleTimes.1D\'
I don't think the commas are necessary either, but probably would be handled correctly.
James
by
jkeidel
-
AFNI Message Board
hi Diane
3dMEMA takes, for each subject, a map of betas and their associated t-statistics. when you entered:
-set groupA subA1 A1pre+tlrc A1post+tlrc subA2 A2pre+tlrc A2post+tlrc
3dMEMA thinks that subject A1 has A1pre+tlrc as the betas for that subject and A1post as the t-statistics corresponding to those betas, and proceeds accordingly.
When you say you switched the order of the ti
by
jkeidel
-
AFNI Message Board
hi Gang
glad to hear it. I really did mean to send you the version I had changed but because I don't really know enough about R I kind of thought it might create more problems than it solved. Also, I do remember using 3dMEMA when it ran in R--did that use scan or readline? Because it certainly worked in batch mode back then. Anyway, great to hear that it's working.
JK
by
jkeidel
-
AFNI Message Board
hey Will
this is an issue that has been brought up before on the board. I think (though I am not sure) that R's behavior with respect to readline changed at some point, but from what I can tell all the readline() commands have to be switched to scan(). As I mentioned somewhere before on the board, I have made 3dGC.R work in batch mode this way, but have not sent Gang a fix for 1dSVAR, p
by
jkeidel
-
AFNI Message Board
if you search for
afni ica
on google it is the first result, at least for me. so yes. it's called 3dICA.R
JK
by
jkeidel
-
AFNI Message Board
hi Andrew
I'm sure that works fine. I have made 3dGC.R work in batch mode by substituting scan for reaadline, but 1dsvar is a much longer script so that is taking a bit more time, to ensure the prompts are still output, etc. I have a version that works for what I am doing but I am not sure it would work for all possible branchings of the script. I hope to send it to Gang and maybe it w
by
jkeidel
-
AFNI Message Board
hi Christine
my understanding from the following excerpts from the help is that it is more like 3dmerge, in that specifying -FWHM 6 does not mean that the resulting data will have 6mm smoothness but rather something larger based on the intrinsic smoothness of the data before it is processed by 3dBlurInMask:
* Works iteratively, similarly to 3dBlurToFWHM, but without
the extensive o
by
jkeidel
-
AFNI Message Board
hi Gang
I have two questions about 1dSVAR (actually I have more like 100, including "what is it doing", but I will save those for another time).
1) As I mentioned before, I am extracting the timepoints from each condition and concatenating them, then giving the program the information about where the breaks in the data are. How does this affect the selection of the order of the p
by
jkeidel
-
AFNI Message Board
hi Christine
I tried this out and found the same problem--if I gave:
3dBlurInMask -FWHM 8 -input epi+tlrc -prefix blur_test -mask test_mask+tlrc
the output dataset is the blurred EPI data within the mask and 0 everywhere else.
for the moment, I think you could use a 3dcalc command to get what you want (the original data everywhere outside the mask and blurred data in the mask) like so
by
jkeidel
-
AFNI Message Board
hi Gang
ok, well I know that it works in batch mode but I have not tried it in interactive mode. also, the prompts that are included in the readline commands are important, especially from the perspective of keeping a record of the parameters of the analysis, so they will need to be turned into print commands. I will play around with this tomorrow and send you a script to take a look at.
by
jkeidel
-
AFNI Message Board
hi Gang and Andrew
Did this ever get sorted? I have come across this same problem. From what I can tell, when readline() is used in non-interactive mode it simply returns an empty string, which is causing the above error. I found that by replacing the calls to readline() with calls to scan()
I can get the AFNI R programs to work with R CMD BATCH--using scan(n=1) for lines containing inte
by
jkeidel
-
AFNI Message Board
thanks Gang! I'll give it a try. we'll see how it goes.
JK
by
jkeidel
-
AFNI Message Board
hi everyone (especially Gang)
I have a question about using 1dSVAR. We have a design where participants watch 3 types of video clip, A, B and C. The idea is to identify a network using the A videos (using PPI or 3dGC.R) and then seeing whether certain connection strengths between regions in this network differ while watching clips of type B or type C. Would it be statistically justifiable /
by
jkeidel
-
AFNI Message Board
hi Gang
thanks a lot for that clarification. I have noticed that SPM does things one-sidedly, in that if you set the uncorrected p to .001 you get the t-value that corresponds to p < .002 on the AFNI viewer's slider. This makes sense in the terms of a one-sided vs. two-sided test.
My question is then, if you run AlphaSim or 3dClustSim with an uncorrected p of .001, and it says tha
by
jkeidel
-
AFNI Message Board
hi Gang
I am now a bit confused. The help for 3dClustSim states:
-2sided = Normally, 3dClustSim does '1-sided' testing. The random
dataset of Gaussian noise-only values is generated, and then
it is thresholded on the positive side so that the N(0,1)
upper tail probability is pthr. If you use this option, then
by
jkeidel
-
AFNI Message Board
aha! thanks Isaac. I got confused by Bob's post since he said it did not work with the '**' bar but I get it now, it only works with the continuous color bar. I should have just tried that to be honest. oh well, thanks again.
JK
by
jkeidel
-
AFNI Message Board
hi all
I have downloaded the most recent binaries but still no luck in getting the alpha feature to work. Also, I do not seem to be able to get a screenshot to upload (it is in jpg format and small enough but still no luck). Is there anywhere I can send it for someone to take a look so that it's clear what is happening? From what I can tell AFNI_EDGIZE_OVERLAY is working but setting al
by
jkeidel
-
AFNI Message Board
Page 2 of 3
Pages: 123