AFNI gui visualization

AFNI version info (afni -ver): 24.0.03 Caracalla

My team has been looking at group level 3dMVM results on the AFNI gui and we have noticed a troubling issue. Our statistical maps from 2 weeks ago had regions of activation visible at a threshold of p= 0.05. After we updated AFNI about 10 days ago, we now see no results at thresholds below p = 0.8. This is incorrect, as we still see activation values when clicking around the underlay.

Is there a binary that is directly related to thresholding/visualizing?


We have not noticed an issues with the afni GUI, so whatever information you can provide might help.

What is the full "afni -ver" output?
To be sure, you are noticing a result in viewing the same results that you viewed 2 weeks ago, or did you re-run 3dMVM?
Are you seeing any results at all, or nothing?
How are you specifying the overlay, threshold, clustering, etc? Feel free to attach a screenshot.


  • rick

Yes, here are those details you're looking for plus a bit more context.

/$ afni -ver
Precompiled binary linux_openmp_64: Jan 29 2024 (Version AFNI_24.0.03 'Caracalla')

We first noticed this issue when looking through a recent re-run of the 3dMVM (03/21) on 03/22. We then looked at a previous version of the 3dMVM (03/17) and observed the same abnormal drop off in activation/signal. This was when we knew that the issue was not simply running the 3dMVM incorrectly, as when we looked at the 03/17 results on that same day we observed results at p= 0.05, 0.01.

We do not see any results at all once the p-value is set to anything below 0.9 (for any group level brain file). We did not apply clusterize. The underlay is an MNI template and the overlay is the output from the 3dMVM group level analysis. The threshold settings for the overlay is a contrast (mean vs nice condition of fMRI task). I will add a screenshot as soon as our server completes its reboot.

I would also like to add that on 03/21 we updated R within afni (3.5 to 4.3, big version jump), installed the latest version of afni, then subsequently rolled back this afni installation.

Thanks for the details. I will send you a private message to see if we can look into this more closely.

  • rick