Calliberating keyboard response

Builder
1

Hi everyone,

I’m designing a PsychoPy experiment where two participants press different keys on the same keyboard (e.g., A and L) at the same time. The timing difference between their keypresses is one of my main behavioural measures.

I wanted to ask how PsychoPy handles two almost simultaneous keypresses.

  • If both participants press their assigned keys at exactly (or nearly) the same time, will PsychoPy reliably record both keypresses?
  • Does PsychoPy introduce any delay between recording the first and second key, or is any difference only due to the keyboard/operating system?
  • Is there a recommended way to validate or benchmark the timing accuracy for simultaneous keypresses (for example, to estimate the system’s timing error before collecting data)?
  • Would using the new psychopy.hardware.keyboard.Keyboard class be sufficient, or would you recommend separate USB response devices instead of a shared keyboard for this type of task?

Any advice or experience with measuring millisecond-level differences between two simultaneous keypresses would be greatly appreciated.

2

Hello

I recommend using a dedicated response pad to collect such responses. There is quite a lot of variance in RT measurements with the standard keyboard, which might be too large for your experiment. However, I do not know the size of the effect you are trying to measure.

Best wishes Jens

3

I’ve no reason to believe keypresses would be missed if they happen at a similar time, even in a textbox on a webpage you can press two keys at the same time and see both letters appear. However, using a regular office keyboard will certainly have less accurate timestamps than what you probably want if keypress timing is one of your main measures. I second what @JensBoelte says above: Don’t worry about using two devices, just get yourself one good response pad.

Black Box Toolkit (who make response pads) published a paper which you may be interested in, they describe how they tested the latency of a standard keyboard compared to their button boxes. Doing this yourself is likely overkill if you have a good response pad (the manufacturer will likely have done these kind of tests and I expect would be more than happy to tell you about them, I doubt they’ll turn down an opportunity to brag :wink: ) but if you really want to be sure then this is how you’d measure it.