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 2161 - 2190 of 2632
Thanks for reporting the bug, Justin! I forgot to change one place after a recent modification. The next AFNI build should have the bug fixed. Let me know if you want the program now.
by
gang
-
AFNI Message Board
Justin,
The binary 0/1 demarcation is a very unfortunate (and maybe sad) presentation of reality considering how arbitrary the thresholding is. I would suggest that you directly compare the groups on the connectivity measure (e.g., z-score), not the conjunction maps.
by
gang
-
AFNI Message Board
In addition to the problem pointed out by Peter, variable 'task' was not specified in the script. You need the following line:
-wsVars 'task' \
by
gang
-
AFNI Message Board
In addition to Peter Molfese's suggestion, five places where you used '-gltcode" should change to '-gltCode'
by
gang
-
AFNI Message Board
> So the question we'd like to ask in this case is really if there is any connectivity
> (resting, seed-based) across this age range that does not change with age but
> is significant for everyone (e.g. what, if any, resting correlations are established
> earlier than childhood that do not undergo age-related changes but are maintained
> through this developmental period
by
gang
-
AFNI Message Board
John,
You can use 3dttest++ to
1) run one-sample t-test and find the average effect of each condition within each group;
2) run paired t-test and compare any two conditions within each group;
3) run two-sample t-test and compare the two groups under each condition; or
4) un two-sample t-test and compare the two groups for any liner combination of multiple conditions (from the GLTs)
by
gang
-
AFNI Message Board
Laurel,
Your model is like this,
y = a + b*x + e,
where x and y are age and connectivity measure respectively, a and b are effect estimates, and e is the residual term.
If x is *centered* in the analysis, a is interpreted as the group average connectivity measure that corresponds to the center age. If x is *not* centered, a would be the group average connectivity measure presumably wh
by
gang
-
AFNI Message Board
Hi Yi,
Which components are you focusing on, the instantaneous or the lagged effects? With former, use 1) and 2) for group analysis. Similarly 3) and 4) can be used for group analysis of the lagged effects.
by
gang
-
AFNI Message Board
Hi Yi,
The specification matrix is the culprit!
With a network of n regions, the SVAR or SEM can only allow for up to n(n-1)/2 (half the off-diagonal elements in the specification matrix). With n = 4 in your case, the upper bound is 6. So you have to trim at least one path.
by
gang
-
AFNI Message Board
Hi Yi,
> what does "order" mean here? lags between ROIs? what is the optimal number for fMRI data?
The order (or lags) in the VAR or SVAR system indicates the maximum delay in time intervals (e.g., number of TRs) among two regions. See more details here:
> No. 1 specification matrix for instantaneous effects file name: xxx.1D
>
> In this matrix file, if I be
by
gang
-
AFNI Message Board
Chuck,
My best guess is that you have an extra space right after the backslash (\) on those lines involving subject 10128 or the line of "Subj Group Condition Time Inputfile \". Are the input data from BOLD response estimated with multiple basis functions?
by
gang
-
AFNI Message Board
Hi Matt,
I've found the bug, but I'll need some time pondering the solution.
> This seems to be working by using the groups as a covariate. In this situation,
> I would use -set instead of -groups, yes?
That's right. You pretend there is only one group of subjects.
> Also, how would I model the interaction? I had originally coded gender as 0 or 1
> (same
by
gang
-
AFNI Message Board
There might have a bug in the program, but I need you to upload all the input files to help me debug it.
On the other hand, I have two comments about your 3dMEMA script:
1) Since you're treating subjects' sex as a covariate, why not also treat the two groups (SCN and CON) as a covariate? That way you would have two advantages: (a) get around the current problem; (b) capability to
by
gang
-
AFNI Message Board
Annie,
You have a relatively simple data structure in which all the fixed-effects factors have two levels. If you know how to dummy code those two between-subjects factors, you can use 3dttest++ or 3dMEMA. Otherwise go with 3dMVM.
by
gang
-
AFNI Message Board
Matt,
Could you verify your AFNI and 3dMEMA version?
afni -ver
3dMEMA -help | head -10
Also providing a copy of your 3dMEMA and the covariate file would be more helpful.
by
gang
-
AFNI Message Board
> I'm not sure why there are two values (i.e. T-stat) for each sub brik.
There is only *one* value per sub-brick, but there are *two* sub-bricks associated with each effect: one for the effect estimate (e.g., percent signal change) and the other for the t-statistic of the effect estimate.
Interpretation of the results:
-- At sub-brick #0 'LOW-HIGH_mean' datum type is fl
by
gang
-
AFNI Message Board
You just provided the information about the output from 3dmerge, not directly from 3dttest++. Use the output from 3dttest++ to visualize the data, and you should be able to check *each* group's effect, group difference, and the covariate effects as well.
I can't see the full command of your 3dmerge command, and can't judge what's meant for.
by
gang
-
AFNI Message Board
Hi Elissa,
Could you provide the output of the following command on terminal?
3dinfo -verb outputOf3dttest++
by
gang
-
AFNI Message Board
Hi Elissa,
> The results changed a lot when I did not center scanner - we saw a lot less
> activation in areas of interest and a reversal of direction. Further, the clusters
> we did find were in completely different regions.
The most recent result is very different compared to which analysis? You should compare the most recent result with the analysis with 3dttest++ but withou
by
gang
-
AFNI Message Board
Carissa,
I can't think of an easy way to directly contrast the three correlations, but you may try this: run the following model
Y = a0 + a1*X1 + a2*X2 + a3*X3 + e
and you can directly see their differences in one model.
P.S.:
1) it seems possible to compare the correlations with 3dMVM, but that would probably involve some significant modulations in the progam.
2): it
by
gang
-
AFNI Message Board
Hi Carissa,
If I understand it correctly, you've run three models:
Y = a1 + b1*X1 + e1
Y = a2 + b2*X2 + e2
Y = a3 + b3*X3 + e3
Is this correct?
> I'm wondering if you can recommend a program within AFNI or otherwise
> that could be used to test this specific contrast.
By contrast do you mean b1-b2, b1-b3, and b2-b3?
by
gang
-
AFNI Message Board
I assume that your trials are event-related. If that is the case, how to model reaction time is a moot issue. There are two straightforward approaches, each of which carries slightly different assumption and provides you different results and interpreations:
1) Making a crude assumption of linearity between the BOLD response and RT. This is typically done in amplitude (or parametric) modulatio
by
gang
-
AFNI Message Board
The -WAV option in waver provides a response function with an undershoot at the end, but it might be possible to tweak the innards (http://afni.nimh.nih.gov/afni/doc/faq/17) to obtain an initial dip?
by
gang
-
AFNI Message Board
> My stimuli are presented for variable durations (RT), ranging from 1 to 4 seconds.
> By removing duration modulation, and assuming the Gamma function (even amplitude
> modulated by RT), wouldn't I be losing this information? Or is the difference small
> (I'll look at the design matrix for each).
This is why I asked before how you would like to run group analysis re
by
gang
-
AFNI Message Board
It depends on what question you're addressing: compare the three groups at each group's respective average age or the overall average age across all subjects? Once you make up your mind about this issue, you center the age values accordingly in 3dMVM (or 3dttest++/3dMEMA).
by
gang
-
AFNI Message Board
Hi Hugo,
I would treat RT as a modulation variable, and don't use dmBLOCK. In other words, treat RT the same way as the other three modulation variables.
Further more, work on the following two lines:
-stim_times_AM2 1 SS.1D GAM -stim_label 1 SS \
-stim_times_AM2 2 LL.1D GAM -stim_label 2 LL \
In SS.1D and LL.1D, incorporate the values of all the four modulation variables. See
by
gang
-
AFNI Message Board
You may use 3dttest++, 3dMEMA, or 3dMVM. The last one is the easiest in terms of scripting.
Do the three groups have roughly the same average age? If not, you have to be careful about the analysis.
by
gang
-
AFNI Message Board
Emily,
What program did you use to run the group analysis? What is the output from the following command?
3dinfo -verb Group_MoodInd_ANOVA_AntDMN_9.nii
by
gang
-
AFNI Message Board
What kind of factors (between- or within-subject) do you have? And what type of covariate: between- or within-subject?
by
gang
-
AFNI Message Board
Hi Hugo,
It seems to me that you need to better define the goal of your group analysis before laying out the model strategy at individual level.
It looks to me that you have four modulation effects, right?
1. sum(subjective values),
2. difference(subjective values),
3. delay(LL-choice), and
4. RT
> -stim_times_AM2 3 ../stimuli/${subj}stimlist2.1D 'dmBLOCK' -stim_labe
by
gang
-
AFNI Message Board