difference between Cedrus response pad and exp.time()

Diego

Hi,

I'm designing a priming study using OpenSesame and I have a difference from around 11ms to around 20ms between the responses collected using a button box and the recording of exp.time(). Below more infos.

The task is to decide if the 2nd word appearing on the screen is word or not pressing a botton (if it's not a word, the participant has to wait for 3000ms for the new stimulus). I'm using the Cedrus RB-530 response pad to collect the reaction times. As discussed in this forum, I've downloaded the pyxid library and I'm using it to get the time measurements of the button box.

the design of the experiment is the following:
1stWord(prime) --> 50ms blank --> 2ndWord(target) --> next trial

If no button is pressed the target word will stay on the screen for 3000ms, otherwise, it will move to the next stimulus when a button is pressed.

In order to control for this, I set up a coroutine around the target sketchpad. In the function associated with this coroutine, I set the duration of the corutine (by default set to 3000ms) to 0 when a button is pressed. In the same corutine function, I'm doing assigning t1=exp.time() in the 2nd line of the function (immediately after def) and t2=exp.time() when a button clic is detected. At that point, I compute tt=t2-t1 as the time between the visualization of the stimulus on the screen and the button press. Unfortunately, there is always a difference of btw 11 to 20ms from the button box results and tt (tt is always higher than the buttonbox response). Does anyone experienced the same problem? Which measurement should I consider more accurate?

thanks a lot for your help,

Diego

Diego

I'm also trying to print exp.time() before and after a sketchpad with duration 96ms(for the refresh rate reason) inside a sequence (x10), and I would expect that the difference between the 2 exp.time() in the 10 iterations should be always the same (aka 100ms) but this is not the case. Same story, if I consider exp.time_sketchpad. Am I doing something completely wrong or do I have a big problem here?

thanks a lot again,

Diego

Josh

Hi Diego,

You should consider the time logged by the buttonbox as the accurate one. It is not weird that your tt computation is a bit off: the variable t2 is created slightly after the response collection, and t1 is created slightly before the start of the trial.

Cheers,

Josh

Diego

Hi Josh,

thanks for you answer. I agree with you, I was just concerned about the warning in the pyxid library:

"At the time of this library release, there is a known issue with clock drift in XID devices. Our hardware/firmware developer is currently looking into the issue. "

I've run some tests, capturing the buttonbox time, the system time and opensesame time. They are all kind of off, but a correlation between these values gives me very high coefficients (>.97). This makes me feel better!

I was just curious to know, if someone else was using the Cedrus RB-530 response pad in Windows (10 in my case), and how they were measuring reliable reaction times.

Cheers,

Diego

ellie

Hi there, I know it's been quite a long time but I'm using an RB-540 and I can't get it to connect so I was wondering whether someone could share the code to download the pyxib2 library and get it to work (I have no idea how to do that and the "Installing packages, plugins and extensions" page is not helping - I might be reading the wrong stuff though).