Up until now we always used response boxes connected to the LPT-port for measuring reaction times. This gives us an accuracy of 0.1 milliseconds. Since LPT-ports are not that common anymore we are looking for alternatives. One of the alternatives we have tested is the RB-x30 response box, using Presentation with the provided extension. We have tested this box using a “hardware participant” that always gives the same response time. We noticed a jitter of 16ms (1 frame?) in reaction times (and 2 spikes with more jitter). See the attached jpg.
Is this the best possible accuracy we can get in Presentation, or should we get better results? The same setup / Presentation script is used testing the LPT-port, and this gives us a 0.1ms jitter.
Hello Jacco, thank you for sharing the test results. With USB, I would expect a delay in the order of a few milliseconds (5 or 6ms), but not 16ms. I do not know the source of this jitter. It could be the interaction between your OS (which version, BTW?) and the USB driver, or interaction between Presentation and the USB driver.
Also, I am personally curious about your “hardware participant”. Can you please share more information about it?
We are using a pc running Windows 7 (32-bit). Our Presentation experiment only presents a visual stimulus, and measures the reaction time of the participant through an input device. Our “hardware participant” is basically a combination of a photodiode (for detecting the visual stimulus) and a function generator for generating the fixed response-time.
I read about the XID timing features, and I was hoping that the Presentation extension would use the Reaction Time Timer to get rid of most of the USB uncertainties?
A combination of the built-in RT timer and using the light sensor would indeed get rid of the USB uncertainties. We are working on updating the Presentation XID plug-in.
