Progressive ratio task

I am designing an experiment where participants will be required to press a button (that Empirisoft is designing) a set number of times to earn a reinforcer. Within a session, the number of required presses will increase (e.g., first ratio requires 75 responses, next one 150, next one 300, ...).

As such, I need the experiment to simply display some instructions on the screen, record and count the button presses, and display a message when the first ratio is complete. Ideally, i would like for the experiment to then require experimenter input to continue to the next ratio...

Does anyone have experience with this type of experiment?

Thanks in advance for any tips.

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)[/code]

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

Hey Blair-

Thanks for getting back to me...we are not currently running Direct RT. Do you see any way to do this without it?


Scott

MediaLab does have a key monitoring function--see info on the [I]k[/I] parameter in the user's guide or click here: [URL]http://www.empirisoft.com/medialab/help/queparameters.htm[/URL]

Problem is that there is no function built in to automatically [I]end[/I] the tracking after a defined number of keystrokes such as 150, 300, etc. You could create a custom item with JavaScript or VBScript that counts keystrokes if you have access to a programmer. Let me know if you want more details on that.

Sorry again to say that the easiest solution would be to use DirectRT. This kind of task is really quick to put together in that environment.

Hope that helps,
Blair

OK. Making progress here. We got Direct RT loaded and I tried the simple 75 response trial set up...

However, I need to be able to capture a higher degree of temporal resolution with the response inputs...on the order of 2-4 responses/sec. Right now, if I respond at the rate I expect my participants to, the 75 trial session will end up having a large number of responses that are not recorded...does that make sense?

Try adjusting the trial interval to 50ms (rather than the default of 1000ms) and see if that helps.