3dMVM and 3dLME can take an input table like 3dMVM … -dataTable @myfile.tsv. Both are very particular about the format of the input file.

• Subj must be the name of the first column (and thus the first word of the file)
• InputFile must also exist and be a valid 3D file. You will get an error if any row has an inputfile that is more than one volume (4D; file with many “briks”; 3dinfo -nt > 1) and does not have a sub-brik specifier ( ❌deconve_output.nii.gz; ✅ deconv_output.nii.gz[Coef#0])

Here's example GNU R code to add InputFile to an existing dataframe df ¹⁾

path_prefix <- '/Volumes/Hera/Datasets/adept/data/tat2/sub-'
df <- df %>%
        mutate(InputFile = paste0(path_prefix, subjid, '_ses-0', visit, '_tat2.nii.gz')) %>%
        filter(file.exists(InputFile))

• 4dConcatDataTable (an lncdtool) reuses the same dataTable format to create 4D images useful for the “graph” view in AFNI's GUI.
• maskvolume uses 3dROIstats w/mask as input and mask to get ROI sizes in mm^3
• 3dmaskave_grp uses 3dmaskave or 3dROIstats to get roi averages in parallel to a single file