Hi Rick,
I am enclosing the script because actually I did not use adiff but acontr and I'm wondering a) if your explanation still applies and b) how to use adiff in the script. There are 4 files for each subject and I'm collapsing across subjects. Each file contains sampled time points for tp (true positive), fn (false negative), tn (true negative) and fp (false positive). We were interested in obtaining the difference of tp - fn, tn - fp, tp - tn and fp - fn. Could you tell me how to write the argument for this using adiff?

Thanks very much, and, here's the script:

3dANOVA2 \
-type 3 \
-alevel 4 \
-blevel 7 \
-dset 1 1 202dw/dw_bucket12+tlrc'[21]' \
-dset 2 1 202dw/dw_bucket12+tlrc'[27]' \
-dset 3 1 202dw/dw_bucket12+tlrc'[23]' \
-dset 4 1 202dw/dw_bucket12+tlrc'[25]' \
-dset 1 2 205km/km_bucket12+tlrc'[23]' \
-dset 2 2 205km/km_bucket12+tlrc'[29]' \
-dset 3 2 205km/km_bucket12+tlrc'[25]' \
-dset 4 2 205km/km_bucket12+tlrc'[27]' \
-dset 1 3 212kw/kw_bucket12+tlrc'[23]' \
-dset 2 3 212kw/kw_bucket12+tlrc'[29]' \
-dset 3 3 212kw/kw_bucket12+tlrc'[25]' \
-dset 4 3 212kw/kw_bucket12+tlrc'[27]' \
-dset 1 4 218th/th_bucket12+tlrc'[23]' \
-dset 2 4 218th/th_bucket12+tlrc'[29]' \
-dset 3 4 218th/th_bucket12+tlrc'[25]' \
-dset 4 4 218th/th_bucket12+tlrc'[27]' \
-dset 1 5 227tv/tv_bucket12+tlrc'[23]' \
-dset 2 5 227tv/tv_bucket12+tlrc'[29]' \
-dset 3 5 227tv/tv_bucket12+tlrc'[25]' \
-dset 4 5 227tv/tv_bucket12+tlrc'[27]' \
-dset 1 6 239sf/sf_bucket12+tlrc'[9]' \
-dset 2 6 239sf/sf_bucket12+tlrc'[15]' \
-dset 3 6 239sf/sf_bucket12+tlrc'[11]' \
-dset 4 6 239sf/sf_bucket12+tlrc'[13]' \
-dset 1 7 241ft/ft_bucket12+tlrc'[9]' \
-dset 2 7 241ft/ft_bucket12+tlrc'[15]' \
-dset 3 7 241ft/ft_bucket12+tlrc'[11]' \
-dset 4 7 241ft/ft_bucket12+tlrc'[13]' \
-acontr 1 -1 0 0 cf7_tp_vs_fn \
-acontr 0 0 1 -1 cf7_tn_vs_fp \
-acontr 1 0 -1 0 cf7_tp_vs_tn \
-acontr 0 0 -1 1 cf7_fp_vs_fn \
-bucket ANOVA2_cf_contrast_IC