[Eeglablist] rmerp and evoked ERSP contribution

[Scott Makeig]

Yes, Agatha -- The 'rmerp' flag does rm the erp  from each epoch ('rm' is
unix for 'remove' because original unix development occurred in the days
before proper terminal keyboards were developed -- programmers used
teletype machine keyboards that required lots of force and travel to type a
lettter -- hence common unix commands were shortened to 2 letters each).

Note that this ERP removal may not be a good idea: The ERP potential
sequence or phase perturbation sequence (when viewed from another
mathematical angle) may have different strengths n each trial -- and
moreover, this may be true separately for each constituent independent
source ERP peak contribution or peak complex contribution!

A better way, therefore, to remove the effects of the processes
contributing to the channel ERP waveforms might be, for each contributing
ICA component process,  to regress out the ERP sequence from each trial,
then sum the resulting no-ERP trials to verify the ERP has been removed,
and then use the no-ERP trials in the time/frequency analysis. (I do not
now recall that 'rmerp' does that...). Otherwise, removing the
one-size-in-all-trials mean ERP from each trial will essentially add the
-ERP to some trials and leave part of the ERP in others...

But then also -- the latencies of the peak and peak-complex contributions
may also vary on each trial. Therefore, still more robust no-ERP component
epochs would require regression out at an appropriate lag for each
contributing component process -- and for epochs in which the ERP sequence
is 'hidden' under other ongoing component process activity, this might have
no best solution ...

Finally, there is no guarantee that the brain or individual independent
source processes respond to each event classified as 'identical' with the
same ERP sequence in every trial -- this is an assumption underlying use of
the ERP in the first place. Julie Onton and I found in our 2005 paper that
frontal midline theta-producing independent component processes during the
learning phase of trials in a modified Sternberg working memory task
sometimes produced a train of (5-6 Hz) theta activity and sometimes a burst
of (14-15 Hz) beta activity. The EEG response to events largely reflects
their implications re an expectancy model our brains are continuously
'computing' (to use a current metaphor)... Thus, the response
evoked/induced by an event depends crucially as well on the subject brain's
current expectancy as well on the stimulus characteristics... (e.g.,
'targets' in a given task should *not* be expected to be followed by the
same EEG response).

There is no reason to believe that brain processes shaping the ERP measure
sequence are in any way independent of the processes shaping the whole EEG
including the mean spectral power shifts measured by the ERSP.... That is
-- there may be no physiologically meaningful distinction between 'evoked'
and 'induced' activity -- certainly just using those terms does not make it
so !

Scott Makeig

On Mon, Jan 30, 2012 at 5:05 PM, Agatha Lenartowicz <alenarto at ucla.edu>wrote:

> Hi all - I'd like to verify that the 'rmerp' option in newtimef.m is the
> appropriate tool to calculate 'induced' rather than 'evoked' (ERP related)
> ERSPs. The help for this flag stats "Remove epoch mean from data epochs" -
> which is a little vague. I assume that the epoch mean is comparable to the
> ERP.
>
> Is this correct?
> Many thanks ~
> Agatha
>
>
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>



-- 
Scott Makeig, Research Scientist and Director, Swartz Center for
Computational Neuroscience, Institute for Neural Computation; Prof. of
Neurosciences (Adj.), University of California San Diego, La Jolla CA
92093-0559, http://sccn.ucsd.edu/~scott
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20120131/1949ca64/attachment.html>