[Eeglablist] Fitting dipoles to grand averaged ERP data

Tarik S Bel-Bahar tarikbelbahar at gmail.com
Sat Oct 8 20:08:02 PDT 2011


Greetings, hope this helps a bit

1a. As far as I understand, ICA requires many time samples (i.e., many
epochs) to "detect" independent components in the data.
Also generally speaking, I think decomposition approaches take
"single-subject grand averages" as the minimum input, so as
to derive components for each subject.

1b.  it seems that the
the current normative method is to decompose EEG data one subject at a time,
and somehow account for or examine
inter-individual variation (or similarity) across subject components.

2. If you want to go with ICA, I suggest using your single trial data for
each participant. But ICA requires at least a good portion of data,
more depending on more channels or sampling frequency. One rule of thumb is
((number of channels squared) X 30 ))  should be less than  or near to
(number of time points or time samples in the EEG data).

3. Joseph Dien's ERP/PCA Toolkit does nice work with spatial and temporal
PCA (see recent applications by Donchin, Hajcak, and Dien).
and should be able to happily ingest and digest your Netstation-prepared
single trial or, more easily I believe, your single-subject average files.

4. See some of the extant eeglab discussions of Netstation to eeglab  to
ERP/PCA Toolkit import-export inter-operability
http://sccn.ucsd.edu/pipermail/eeglablist/2011/003913.html




On Thu, Oct 6, 2011 at 10:09 AM, Leanna Cruikshank <leannac at ualberta.ca>wrote:

> Hi there,
>
> I am attempting to source localize my ERP data. I have a grand averaged ERP
> file that has been processed in Netstation, which I exported for EEGlab. I'm
> wondering whether I am able to run ICA on the grand averaged file, or
> whether I need to run ICA on each individual subject's data first? Thanks in
> advance!
> *
> *
> --
> Leanna
>
>
> _______________________________________________
> 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20111008/69e811a2/attachment.html>


More information about the eeglablist mailing list