Hi Dorit,

So your reponse to #3 looks good.

And I'm sorry for the confusion, but my comment about the option
"-stim_A_preceeds_B 7 4" had nothing to do with counterbalancing.
I should have added another topic divider '---' in there...

The purpose of that option (which I am not using as -ordered_stimuli,
allowing for more than 2 stimuli in a set order) is for the case that
you gave in your first message, showing:

fixation (ITI) - stimulus A (150ms) - blank ('ISI') - stimulus B (2s)...

There simulus B is attached to stimulus A in some sense, in that it
would _always_ follow it with nothin but 'ISI' between them. Is that
correct?

----

So my intention is that you could run a command like this:

make_random_timing.py -num_stim 12 -num_runs 6 -run_time 300 \
-stim_dur 1.65 2.15 1.65 2.15 1.65 2.15 1.65 2.15 1.65 2.15 1.65 2.15 \
-stim_labels A1 B1 A2 B2 A3 B3 A4 B4 A5 B5 A6 B6 \
-ordered stimuli 1 2 \
-ordered stimuli 3 4 \
-ordered stimuli 5 6 \
-ordered stimuli 7 8 \
-ordered stimuli 9 10 \
-ordered stimuli 11 12 \
-show_timing_stats -seed 314159 -t_gran 0.05


That way the 'A*' stimuli can happen randomly, but the 'B*' stimuli
must be the first to follow them, after a random amount of rest.

Note that I specified durations of 1.65 seconds for the 'A*' regressors
and 2.15 seconds for the 'B*' regressors. That represents 0.15 seconds
for the actual stimulus, and either 1.5 or 2 seconds of minimum rest
following them (as your minimum ISI and ITI, respectively).

Note that -show_timing_stats will spit out some useful information
(similar to what 'timing_tool.py -multi_show_isi_stats' would show),
and that -t_gran is applied to reduct the time granularity from 0.1
seconds to 0.05 seconds (since your durations are at that level).

How does that seem?

- rick