[Eeglablist] baseline for response-locked ERPs

[Elise Radtke]

Dear all,

I was wondering, whether anybody has experience in / ideas about baseline corrections concerning response-locked ERPs.We are primarily interested in processes happening directly before participants come to a decison, i.e. press a key. 

Our experimental setup was the following: After a fixation cross, a prime word (positve/negative) was presented for 300 ms, followed by a target, where subjects had to make a binary decision and indicate it by keypress. Participants responded without having any time restrictions. The averaged Response Times varied from 1200-3000 ms.

Having these varying RTs, 3 possibilities come to my mind, how to set the baselines, when epoching the data:

1) Using the 500ms before target presentation or prime as baseline.
2)  Determining a baseline relative to response e.g. 1500 ms to 1000 ms before button press. This would raise the problem, that the baseline varies between trials, it can be either post- or pre-target, or even on the target.
3) Using the average voltage of the whole epoch and subtracting it from the epoch data, as proposed by Steven Luck in his book (2005).
It would be nice if you have some opinion or an idea how to deal with the problem.

Best

Elise
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20120202/7854f56f/attachment.html>