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 2041 - 2070 of 2632
Yes, the results should be fine regardless of that mysterious error.
by
gang
-
AFNI Message Board
Kevin,
This is strange because I couldn't duplicate the error here. Did you get the output?
by
gang
-
AFNI Message Board
Based on your random effect specification
-ranEff '~1+RT' \
you have a quantitative variable RT. However, it's not seen in your model, nor in the data table. If you don't have such a variable, the data might be better handled with 3dMVM.
by
gang
-
AFNI Message Board
> what is the difference between doing a PPI analysis and doing 2 separate simple
> correlation analysis (1 for correct and 1 for error)
Simple correlation looks for the signal similarity between the seed and the rest of the brain. It's typically for resting state because you don't need to worry about teasing apart among the different sources of the signal. Also the correlatio
by
gang
-
AFNI Message Board
Some of your input files may have different dimensions. Use 3dinfo on each of those files and see which one(s) are the troublemaker.
by
gang
-
AFNI Message Board
My suspicion is that in those GLT coding lines you forcefully separated the '-' sign from the number '1'. For example, line
-gltLabel 1 'mtbi-hc' -gltCode 1 'Group : 1*mtbi - 1*hc' \
should be
-gltLabel 1 'mtbi-hc' -gltCode 1 'Group : 1*mtbi -1*hc' \
by
gang
-
AFNI Message Board
What is the output (two lines) of the following command at the terminal?
afni -ver
by
gang
-
AFNI Message Board
Could you copy and paste the following in R?
first.in.path <- function(file) {
ff <- paste(strsplit(Sys.getenv('PATH'),':')[[1]],'/', file, sep='')
ff<-ff;
#cat('Using ', ff[1],'\n');
return(gsub('//','/',ff[1], fixed=TRUE))
}
source(first.in.path('AFNIio.R'))
(ll <- find
by
gang
-
AFNI Message Board
It's hard to diagnose the problem without access to the data.
The script only shows data with two subjects. You have more than two subjects, right? In addition, the data structure seems to be better modeled with 3dMVM instead of 3dLME. You only need to change the first few lines.
by
gang
-
AFNI Message Board
Sounds like a permission issue. At the terminal, try this:
touch hello
ls hello
See what you get with these two command lines.
by
gang
-
AFNI Message Board
How are $path and $treat01 defined?
Put the following two lines before the script and see what their values are?
echo $path
echo $treat01
by
gang
-
AFNI Message Board
The error message indicates that 1dUpsample was trying to read file '1797555712' or '-108684202', which of course was not the real file you intended. How did you run the script? You can add the option '-x' in tcsh to spill out how the shell interprets:
tcsh -x YourScript
and see if you can identify the problem.
by
gang
-
AFNI Message Board
> Is there any way to do a 3dLSS type of analysis while using TENTs for the condition that gets individually modulated?
As long as you use 3dDeconvolve to set up the design matrix for 3dLSS, theoretically you can run the analysis 3dLSS. I would be a little concerned about the number of regressors and the multicollinearity issue.
by
gang
-
AFNI Message Board
> is there a way to convert the Full F-statistic that is spit out from 3dDeconvolve to show which
> regions are significant activation and which are deactivation?
The F-stat tests the following hypothesis:
beta1 = 0, beta2 = 0, beta3 = 0
Therefore as an omnibus test the F-stat does not directly show the activation directionality. However, if you believe that beta1 captures the m
by
gang
-
AFNI Message Board
> In another dataset all the groups have different number of subjects, how can I > design an ANOVA experiment with this condition
Use 3dMVM.
by
gang
-
AFNI Message Board
Jim, the message board does not allow file attachment. Could you post your 3dMEMA script here?
by
gang
-
AFNI Message Board
Two potential issues:
1) file media/sf_D_DRIVE/Freesurfer_Data/Processed/CNT/CNT02/SUMA/std.60.lh.thickness.niml.dset cannot be found. Does the file really exist? Spelling?
2) The following error
-dset: Command not found.
is caused by some extra space right after a backslash.
by
gang
-
AFNI Message Board
> The output for each subject includes a test of the coefficient for each condition
> that I then take to an ANOVA. Does this coefficent (beta) test whether the fit for
> any subject/condition is different than ZERO or different than the BASELINE?
If you have a baseline condition (e.g., fixation) that is not included in those six conditions of interest, the answer is YES to both:
by
gang
-
AFNI Message Board
Since you have a 2 by 3 between-subjects ANOVA, change
3dANOVA2 -type 3
to
3dANOVA2 -type 1
by
gang
-
AFNI Message Board
> Would it be possible to calculate the t-value for a given q-value, even though there is no
> voxel in the dataset that passes this threshold and return that value instead? Or would it
> be possible to return a t-value that is very high instead?
Jonathan, the threshold for the t-stat values associated a specific FDR q-value is determined by the distribution of all the t-values in
by
gang
-
AFNI Message Board
Model comparison is a little bit controversial. In the full model with all the explanatory variables (regressors), there are two statistics that can provide you some useful information: the significance (t-stat) for each explanatory variable and the amount of variance it accounts for (F-stat).
by
gang
-
AFNI Message Board
Could you provide more details about your model? What are the factors A and B? How many subjects?
by
gang
-
AFNI Message Board
Could you explain what you mean by "Beta Series"? Documents except images would not be attached in posting here.
by
gang
-
AFNI Message Board
I have a script that de-correlates 1D files. Send me an email.
by
gang
-
AFNI Message Board
Piero,
Try function wtc() in the R package biwavelet. In R:
install.packages(‘biwavelet’)
require(‘biwavelet’)
by
gang
-
AFNI Message Board
Rick,
That inconsistency does seem to be troubling. The data structure is better modeled with 3dMVM with minor modifications on your current script. So it's probably a good idea to try out 3dMVM, and see where 3dLME misbehaves. Could you let me know how it pans out?
by
gang
-
AFNI Message Board
Sobana,
Based on your description, everything looks normal to me. If you want to run post post hoc tests, use options -gltLabel and -gltCode.
by
gang
-
AFNI Message Board
Different in the omnibus F-statistics or GLT t-stats? How big are the differences?
by
gang
-
AFNI Message Board