Review the afni_proc.py command.  Display the command file in a terminal
window.  Also consider opening the file in an editor or using less.

% cat s01.ap.simple

      ---------------------------------------------
      #!/bin/tcsh

      afni_proc.py -subj_id FT                \
              -dsets FT/FT_epi_r?+orig.HEAD   \
              -copy_anat FT/FT_anat+orig      \
              -tcat_remove_first_trs 2        \
              -regress_stim_times FT/AV*.txt  \
              -regress_stim_labels Vrel Arel  \
              -regress_basis 'BLOCK(20,1)'    \
              -regress_opts_3dD               \
                  -gltsym 'SYM: Vrel -Arel'   \
                  -glt_label 1 V-A            \
              -regress_est_blur_errts
      ---------------------------------------------

option: -subj_id FT

   This subject ID will be used as part of the output directory name (default
   is subject_ID.results), as well as part of any important file name, created
   by the processing script.

option: -dsets FT/FT_epi_r?+orig.HEAD

   This lists the EPI datasets to be processed.  Note that they can be listed
   fully as AFNI datasets:

      (a)  -dsets FT/FT_epi_r1+orig FT/FT_epi_r2+orig FT/FT_epi_r3+orig

   or fully as files (e.g. with a .HEAD or .BRIK suffix):

      (b)  -dsets FT/FT_epi_r1+orig.HEAD FT/FT_epi_r2+orig.HEAD    \
                  FT/FT_epi_r3+orig.HEAD

   Since the option in the example uses the '?' wildcard character to match a
   single character,  The -dsets option in the command is equivalent to giving
   the full .HEAD file names as example (b) shows.  When using wildcards, the
   expression must expand into complete file names (so the suffix is needed).

   ** If there were a spelling mistake when using wildcards, the error message
      from the shell (not from afni_proc.py) will be "No match".

   ** To test whether the names are correct (when using wildcards), try to list
      the files using the 'ls' command, or maybe 'echo'.

        works:  ls -l FT/FT_epi_r?+orig.HEAD
        error:  ls -l FT/FT_epi_run?+orig.HEAD

   Cut and paste these commands and verify the expected output.

   ** Keep in mind that when using wildcards, the files will appear in 
      alphabetical order.  That must match the run order.  For example, a
      file called epi_r9+orig.HEAD would come _after_ epi_r10+orig.HEAD,
      which presumably not be desired.  That is why zero-padding files is
      important, i.e. using epi_r09+orig.HEAD, instead.

option: -copy_anat FT/FT_anat+orig

   Specify the anatomical dataset to be copied into the results directory.

option: -tcat_remove_first_trs 2

   Request to remove the first 2 TRs from each run.  Note that this may affect
   the stimulus timing files.  If the scanner started at time t=0 seconds and
   the TR is 2.5 seconds, then removing 2 TRs means the stimulus times will be
   off by 5 seconds.  The timing files should be adjusted, if necessary.

option: -regress_stim_times FT/AV*.txt

   Using a wildcard expression, provide the stimulus timing file names.
   Keep in mind

   Test that 2 files are seen:

        % ls FT/AV*.txt

option: -regress_stim_labels Vrel Arel

   This option is to provide labels that match the stimulus timing files
   given in the previous option.  Be sure that if wildcards are used above,
   the order matches what is expected.  That is why the files start with AV1
   and AV2, so that the order is clear and alphabetical.

   So files FT/AV1_vis.txt FT/AV2_aud.txt match labels Vrel and Arel (the
   visual-reliable and audio-reliable stimulus classes).
   
option: -regress_basis 'BLOCK(20,1)'

   This is where the fact that each stimulus lasted 20 seconds is applied.

   We know from file FT/AV1_vis.txt that the first visual-reliable stimulus
   occurred 60 seconds into run 1.  Using this basis function, we are telling
   3dDeconvolve that the stimulus lasted from time t=60 to time t=80, and that
   we want this 20 second boxcar function convolved with an incomplete gamma
   function to become (part of) the ideal response function lasting from t=60
   to t=95 (until 15 seconds after the end of the stimulus).

   We will review the ideal response functions later, after they are generated
   by 3dDeconvolve.