3dMVM output issue

AFNI Message Board
,
1

AFNI version info: Precompiled binary linux_ubuntu_16_64: Jul 6 2026 (Version AFNI_26.2.01 'Gordian III')

Hi experts:

I want to use the 3dMVM tool to compare cortical thickness differences across several groups in surface space, after controlling for covariates like sex, genotype, and age. The groups include two baseline groups and three disease groups. I have the cortical thickness data in .gii format and have saved the covariates and .gii file paths in a .txt file. However, the output file is just a single .1D file containing 10 columns of F-tests, 10 columns of t-tests, and 10 columns labeled 'none'. I don't know what these 'none' columns mean, nor do I know how to visualize this .1D file using SUMA. My code is shown below. What should I do?

thank you!

3dMVM -prefix ./results/MVM_Left_Result -jobs 8 \ 

-bsVars "Group+GENDER+GENOTYPE+scanner_model_merge+Age+EDUCAT+EstimatedTotalIntraCranialVol+slice_thickness+slices_per_volume" \       

-qVars "Age,EDUCAT,EstimatedTotalIntraCranialVol,slice_thickness,slices_per_volume"        \

-num_glt 10        \

-gltLabel 1 AnBn_vs_AnBp -gltCode 1 'Group : 1*AnBn -1*AnBp       ' \

-gltLabel 2 AnBn_vs_pure -gltCode 2 'Group : 1*AnBn -1*pur     e' \

-gltLabel 3 AnBn_vs_worse -gltCode 3 'Group : 1*AnBn -1*wor     se' \

-gltLabel 4 AnBn_vs_resilience -gltCode 4 'Group : 1*AnBn -1*resilie     nce' \

-gltLabel 5 AnBp_vs_pure -gltCode 5 'Group : 1*AnBp -1*       pure' \

-gltLabel 6 AnBp_vs_worse -gltCode 6 'Group : 1*AnBp -1*       worse' \

-gltLabel 7 AnBp_vs_resilience -gltCode 7 'Group : 1*AnBp -1*resi     lience' \

-gltLabel 8 pure_vs_worse -gltCode 8 'Group : 1*pure -       1*worse' \

-gltLabel 9 pure_vs_resilience -gltCode 9 'Group : 1*pure -1*re     silience' \

-gltLabel 10 worse_vs_resilience -gltCode 10 'Group : 1*worse -1*r    esilience' \

 -dataTable @table_left.txt
4

图片
图片1263×114 8.78 KB

I tried to specify the -prefix as .gii, but the new output is different from the supposed name :sob:It seems only the header file was produced... help

5

First, if group comparisons are your primary effects of interest, some of the covariates may require careful justification. At the risk of a shameless plug, we discussed some of these modeling decisions and pitfalls in a recent paper of ours.

If you run 3dMVM with .gii as input files and specified -prefix with no extension, AFNI default to creating a standard .1D file. If you force a GIFTI output by specifying -prefix *.gii, 3dMVM does not natively write .gii as a final output format.

The solution is to write the output as-prefix *.niml.dset. SUMA natively reads .niml.dset files, and they are fully compatible with your surface geometries.

In case you need a .gii file for another software, you can easily convert it afterward using the AFNI program ConvertDset.

Gang Chen

6

thank you very much Gang! it works to specify -prefix .niml.dset :smiley: I will read this paper later