Hi Nicole,

Sorry this is so much trouble. This command is more readable with the file names. In fact, I can run your exact command without any errors (using local files that match yours). Is there anything else to the script? Is this the exact command in your script?


Regarding ggc.py, I did not think it through. That tool is set up to take inputs where all sub-bricks for one subject are in a single dataset. So it might be easiest to start with the more original datasets. But just to demonstrate, consider this tcsh script to first copy the data to a new directory and new files, then run gen_group_command.py on it.

# -------------------------------------------------------
# make bucket datasets in a new directory
set newdir = data.buckets
mkdir $newdir

set slist = ( `\ls -1 p[0-9]*+tlrc.HEAD | awk -F_ '{print $1}' | sort | uniq` )
set ltext = 'DN,DW,DL,LN,LW,WN'

foreach sid ( $slist )
   3dbucket -prefix $newdir/${sid}_all ${sid}_{$ltext}+tlrc.HEAD
end

cd $newdir

# -------------------------------------------------------
# create a 3dANOVA2 command script
gen_group_command.py -command 3dANOVA2          \
         -write_script cmd.3dANOVA2             \
         -dsets *.HEAD -subs_betas 0 1 2 3 4 5

You can use -options to add on to that in the ggc.py command, or you can just edit the script to add -adiff and related options.

Note that in your command, the order of the datasets does not quite match the order applied in -amean and -adiff options. The files come as DN DW DL LN LW WN, so some of those need adjusting.

Does that sound easier? At least it would give you a script to start with. You can use labels for -subs_betas, to use something more clear than 0 1 2 3 4 5.

- rick