If I understand correctly, you could simply list each press as its own trial like this:

Code:
stim = ~x (or whatever visual you want)
loc = 0,0,0
time = rt:any (or whatever keycode you want)
After 75 trials, you stop and have experimenter intervention, and then do 150, 300 and so on. DirectRT will track the inidvidual press RTs and the cumulative time passed at each stage.

Does that help?

-Blair