[Eeglablist] comparing normal and clinical populations

[Tarik S Bel-Bahar]

Hello Jose,

1. Search on google scholar for previous ERP articles that have looked at
P100 between one group and another.
 You can search for normal ERP articles, or EEGLAB articles, etc.. There is
where you should find examples for your basic questions.

2. A relatively normal analysis is to use repeated measures anova, using
single subject averages.

3. single subject average ERPs
 are usually computed as a mean, peak, or adaptive mean during a time
window.
Please look for yourself in previous aritcles by searching on google
scholar, and reading the methods sections of what you find.

4. See Luck's and also Handy's  two handbooks of ERP. See Steve Luck's
webiste for some sample chapters.
One basic idea that seems relevant to your general questions,
the more trials you have, the better your signal-to-noise ratio should be.

5. When working with eeglab, please review  the online eeglab tutorial and
articles to get a sense of what
you can do using single trials with eeglab. For starters, erpimage allows
for visualization of all trials at once.
IC demposition allows you to examine a more "clearly-separated" P100
component. Also look into
ERSP and ITC metrics, and note how they differ from normal "single-value
per participant" ERPs.

6. a caveat, Work on using single trial metrics is still in it's infancy,
and few researchers have developed
methods to use single trials effectively as  metric instead of ERP.
However, there is promising work
in this area, google scholar can help you there also to get oriented
correctly.
in addition, you'll find yourself on somewhat uneven ground as you try to
develop your own single trial metrics.

7. Why don't you first show  that there is an effect on P100 between the
two groups,
in the normal traditional way [testing the null that there is no difference
between the groups]
and just do a t-test or anova using "one value" from each participant.

8. Note that latency is another measure you did not mention, nor
lateralization, both which may be applicable.
9. Good luck, and let us know of how things go.



On Sun, Dec 4, 2011 at 11:02 AM, Jose Rebola <jrebola at gmail.com> wrote:

> Hi
>
> I am running a study to investigate differences between  williams syndrome
> patients and controls in a visual task. I have eight subjects of each
> population.
>
> How do I compare the amplitude of the P100 between the populations?
>
> Should i include only one value (the peak around 100ms on each of the
> subject's average) per subject ?
> It seems to me that if i do this i will only have one value per subject
> and i am "throwing away" the 100 trials per condition that i have
>
> Isn't there a way that i can compare between populations while retaining
> within-subject variability?
> Otherwise, how does it compensate to perform 1000 trials instead of 10?
>
> I only know how to do two-level analysis when all subjects perform two
> conditions for example, thus getting one value of "effect" per subject and
> moving to the next level...
>
> Isn't there any paralle of this to two populations?
>
> Hope somebody can help me,
>
> José Rebola
>
> _______________________________________________
> 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/20111205/2cb48d20/attachment.html>