AFNI chauffeur - blue spots?

AFNI Message Board
1

AFNI version info (afni -ver): Precompiled binary macos_10.12_local: Jan 19 2026 (Version AFNI_26.0.03 'Pupienus Maximus')

Hi there!

I'm beginning to take advantage of afni_chaffeur, but have come across an issue that I can't seem to solve. When I generate the figure using chauffeur, I get these small blue dots that aren't there when I open it using the AFNI GUI (lower image). Any idea on what would be causing that/ how to get rid of them?

Also, WRT saving the colour bar, are there any options that show the range directly on the bar?

Thanks so much,

Carly

My command is: @chauffeur_afni
-ulay AnatAv_Underlay_PrevTRInc+tlrc
-olay "NA_AllvISI_INP_n135_prevTRincl+tlrc[1]"
-thr_olay_p2stat $P
-thr_olay_pside bisided
-prefix "{cond}_NN{NN}_c${clust}P${P}${GROUP}"
-set_dicom_xyz 17 30 32
-delta_slices 3 15 4
-set_xhairs OFF
-montx 4 -monty 4
-opacity 9
-func_range 20
-cbar Reds_and_Blues_Inv
-olay_boxed Yes
-pbar_saveim ${cond}
-left_is_left NO

image
image1920×2270 382 KB

image
image1014×1520 239 KB

2

Hi, Carly-

I'm glad that @chauffeur_afni is being useful.

Re. the blue spots: Those could be a couple different things. It might be interpolation leading to tiny negative values. But that would be surprising because I think the overlay+thresholding volumes are the exact same ones. I might also be surprised to see yellow values appearing, since those should be close to 0, on the positive side; but that would depend on the threshold value and the actual numbers in the dataset.

As to why it is different than in the GUI, I guess the range of colorbar values is different in the GUI. When you save out the colorbar with -pbar_saveim, you should get an accompanying JSON file that records the min/max values of the colorbar. Since you have -func_range 20 and no -pbar_posonly, that range should be [-20, 20]. You can compare that to the GUI range, by looking in the controller panel at the top/bottom of the displayed colorbar.

I have taken your chauffeur command and put a vertically spaced copy of it here. (NB: I got rid of some backslashes that were used above, I think to escape some characters; surrounding a multiline command in triple-backslashes ``` .... ```\ above and below the command.) I have also added the -cmd2script ... option, so you can get a driver script saved out that will open the GUI interactively in the same form as the image you are showing; in this case, that script will be called run_chauffeur.tcsh, which you can execute at the command line and hence see the GUI in vivo displaying the results. You can then click on a blue voxel, and see what value it has.

@chauffeur_afni                                                              \
    -ulay             AnatAv_Underlay_PrevTRInc+tlrc                         \
    -olay             "NA_AllvISI_INP_n135_prevTRincl+tlrc[1]"               \
    -thr_olay_p2stat  $P                                                     \
    -thr_olay_pside   bisided                                                \
    -prefix           "${cond}_NN${NN}_c${clust}*P${P}*${GROUP}"             \
    -set_dicom_xyz    17 30 32                                               \
    -delta_slices     3 15 4                                                 \
    -set_xhairs       OFF                                                    \
    -montx            4                                                      \
    -monty            4                                                      \
    -opacity          9                                                      \
    -func_range       20                                                     \
    -cbar             Reds_and_Blues_Inv                                     \
    -olay_boxed       Yes                                                    \
    -pbar_saveim      ${cond}                                                \
    -left_is_left     NO                                                     \
    -cmd2script       run_chauffeur.tcsh

Please let me know how that goes.

--pt

3

Hi there,

Thanks so much for getting back to me! This did indeed help! In launching it this way, I've realized that I had a 20 voxel cluster threshold applied in the GUI which is why that small blue cluster was being removed. Is there a way for me to apply an NN1 and cluster 20 within the chauffeur command?

Thanks so much,

Carly

4

OK, glad that helped.

Here is an example of applying clusterizing with the @chauffeur_afni command directly. You basically add a command with your clusterize parameters, like:

-clusterize        "-NN 1 -clust_nvox 200"

in that example (note the quotes).

Another very important feature in that script is the use of transparent thresholding, which is something we thing is really useful for understanding and presenting data. That is accomplished by including these opts:

-olay_alpha        Yes                                            \
-olay_boxed        Yes                                            \

We have written about it here:

  • Taylor PA, Reynolds RC, Calhoun V, Gonzalez-Castillo J, Handwerker DA, Bandettini PA, Mejia AF, Chen G (2023). Highlight Results, Don’t Hide Them: Enhance interpretation, reduce biases and improve reproducibility. Neuroimage 274:120138. doi: 10.1016/j.neuroimage.2023.120138 pubmed.ncbi.nlm.nih.gov
  • Chen G, Taylor PA, Stoddard J, Cox RW, Bandettini PA, Pessoa L (2022). Sources of information waste in neuroimaging: mishandling structures, thinking dichotomously, and over-reducing data. Aperture Neuro. 2: DOI: 10.52294/2e179dbf-5e37-4338-a639-9ceb92b055ea
  • Taylor PA, Aggarwal H, Bandettini PA, Barilari M, Bright M, Caballeros-Gaudes C, Calhoun VD, Chakravarty M, Devenyi G, Evans J, Garza-Villarreal E, Rasgado-Toledo J, Gau R, Glen DR, Goebel R, Gonzalez-Castillo J, Gulban OF, Halchenko Y, Handwerker D, Hanayik T, Lauren PD, Leopold D, Lerch J, Mathys C, McCarthy P, McLeod A, Mejia A, Moia S, Nichols TE, Pernet C, Pessoa L, Pfleiderer B, Rajendra JK, Reyes L, Reynolds RC, Roopchansingh V, Rorden C, Russ BE, Sundermann B, Thirion B, Torrisi S, Chen G (2025). Go Figure: Transparency in neuroscience images preserves context and clarifies interpretation. Nature Methods, https://rdcu.be/B4tikBETdQZa ([2504.07824] Go Figure: Transparency in neuroscience images preserves context and clarifies interpretation)

(the latter one was finally published just yesterday, so it is hot of the presses!)

--pt