Hi Dorit,

1. Is 'blank' really different from 'fixation', in terms of
subject stimulation? If so one of them would require a
regressor...

2. Should both the ITI and ISI time periods vary randomly?

3. Is there any fixed minimum or maximum on the ITI or ISI
period? Note that a minimum is okay, but a maximum might
be very difficult to deal with.

If your answer to question 3 is 'no', I may be able to alter
make_random_timing.py to deal with that. Esstentially, it
would mean adding restrictions like 'B' must follow 'A' to
what the program already does.

---

Dealing with "counterbalancing" is very difficult, because it
is opposite randomizing in some sense (imposing restrictions
on the order). For example, suppose you have 10 equal stim
classes and 10 runs. You might hope each class would be the
first stimulus in exactly one of the 10 runs. However the
probability of that is 0.00036, i.e. in 1000 generations of
random timing, it is likely that none of the sequences will
have that property.

Anyway, so I will look into adding an option like:

-stim_A_preceeds_B 7 4

So that would say that stim type 7 always preceeds type 4.

How does that seem?

- rick