[Eeglablist] Identifying induced responses

[Keith Yoder]

Hi all,
I know that this problem has been discussed to some extent already (e.g.
http://sccn.ucsd.edu/pipermail/eeglablist/2008/002262.html), but I've been
unable to find a define solution.  What I'm attempting to do is the portion
of the response that is induced, rather than evoked.
Our paradigm involves a computer screen with five sections, each flickering
at a different frequency.  At any given time, one of the sections is cued
(the section is surrounded by a a bright red line).  We are attempting to
identify induced changes time-locked to cue shifts.  I know that other
authors have had some success using the following general procedure:
   1. Compute the time-domain average (ERP)
   2. Perform a Fourier transform on the time-domain average
   3. Subtract the Fourier transform coefficients from the Fourier transform
coefficients of each single trial
   4. Calculate the amplitude spectrum of the subtracted Fourier coefficents
at each trial
In theory, the resulting amplitude spectrum from step 4 should isolate the
induced response.  Before I implement the above steps, I wanted to ask this
list for opinions.
Have others on this list found the procedure to be the most effective method
for isolating induced responses?  If not, what methods have proved to be
more effective?  Finally, are there EEGLAB functions that can be manipulated
(either by the appropriate parameters or by small changes to the code
itself) to perform some or all of the above steps?
Thanks in advance,
Keith
--
Research Aide
Belmonte Autism Lab
Cornell University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20090923/1545f1de/attachment.html>