EEG Task Events & Recording First Key Press

[Channing Hambric]

Hi Everyone,

I am currently working on programming a verbal fluency task w/ EEG in E-prime 3. I'm pretty familiar with E-prime, but am new to EEG research. I am currently running into two separate issues, and would welcome any thoughts/suggestions!

In our task, a category is presented on the screen and participants type in examples of that category using a keyboard. A textbox is displayed for every entry. Their typed response is shown live using the display echo, and each entry is submitted using Enter as the termination response. We have two task event triggers being recorded, the onset of the textbox and the first key press for that trial. When doing some post-recording coding of the responses, I noticed there were a few extra key-press triggers in the EEG files here and there (in comparison to the trials recorded in the edat file), some of which were marked back-to-back in the span of 100ms. When looking at the set up of the task events more carefully, I noticed that there was only a single event for the key press response with no delay. I assume the lack of delay is responsible for the extra triggers. I then tried adding a second key press event with a delay of 50ms, but it marked every single key press in the response, not just the first. Both the "real" event and the delayed event were set to WriteByte.  I then tried setting the delayed version of the event to ResetBit, and everything seems to be working properly now. Does that sound like a viable solution, or is there something I'm missing?

The second problem relates to the reaction times collected in the edat file. Ideally, we would like to record the time between the onset of the textbox and the first key press. Currently, I can only seem to get the RT for when the participant presses the termination response, not the first key. For example, if a participant types "cat," I would like the time it took to press 'c', not 'cat' + Enter. I've tried a couple of inline coding plug ins, like the one from here (https://support.pstnet.com/hc/en-us/articles/115001908708-Multiple-Response-Collection-30022), but haven't had any luck. Is it possible to get the RT for the first key press in an open-ended, multi-key response with an explicitly defined termination key? I have verified that the EEG trigger does correspond correctly to the first key press of a response, so I'm not quite sure why the task event is picking this up correctly but not the RTs/edat file.

Currently, I am using the sampling/latency info from the vmrk EEG file as a work around to get the first key press latency. Right now, my approach is to subtract the textbox onset latency from the first key press latency and then adjust for the sampling rate (500 Hz). So for example, if S 1 is the textbox onset and S 4 is the first key press of that trial, I calculated the RT as:

  

  ((6205-5437)/500)*1000 = 1536 ms

Does this seem like a reasonable work around, or has anyone ever done anything similar?

Thanks in advance! Let me know if I can provide any clarifying info.

Channing Hambric

Postdoc, Bowdoin College