[Eeglablist] Re-Referencing Question

[Daniel Patrylak]

We are just getting started with EEG and in particular movement-related
cortical potentials. We work with epoch-averaged ERPs from -3000 to
+1000 ms re an event which corresponds to EMG onset for a given
movement.

Our question relates to a practical implementation manner in eeglab
(which we use for epoch extraction, baseline correction, and averaging,
not ICA) but also to the larger issue of selecting the most appropriate
site for the reference electrode.

We want to avoid the now well-known problems resulting from physically
linking Le and Ri ear references but the eeglab tutorial does not
provide much detail on offline re-referencing. If we use, in this
initial stage, only a limited number of recording sites (say for example
Cz, C3, C4, Fz, F3, F4) the average reference method certainly would not
apply. So that would leave the alternative to record to any reference
electrode and then re-reference in eeglab to mathematically-linked ears.
The two options then are (a) to use say the Le ear (lobe or mastoid) as
the reference and record from Cz, C3, C4, Fz, F3, F4, Ri ear as
channels, all referenced to the Le ear, or (b) to use Cz as the
reference (although the potentials of interest are typically largest at
Cz) and record from C3, C4, Fz, F3, F4, Le ear, Ri ear as channels, all
referenced to Cz.

Our specific eeglab question relates to how (or maybe if), using
eeglab's re-referencing pop-up window, both the above recording options
could be correctly rereferenced to mathematically linked ears while
obtaining/maintaining a recorded signal for all six locations Cz, C3,
C4, Fz, F3, F4. The re-referencing window allows one to enter one or
more (at least it appears that more than one can be used) channel
number(s) for the new reference but in the above scenario (a) the left
and right ears are not both available as channel numbers (for obvious
reasons because Le ear was the reference).

Or would it be better to record 12 channels with all six locations Cz,
C3, C4, Fz, F3, F4 referenced separately to the Le ear and Right ear and
then to average the Le and Ri ear referenced signals for each location?

The more general question is whether anyone has any more general input
on advantages and disadvantages of the various reference choices in a
case like this with a small number of electrode locations.

Linda Patrylak
Ph.D. student, UConn


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20030610/6bee70a1/attachment.htm