# [Eeglablist] Performining the t-test and highiliting the area

cyril pernet cyril.pernet at ed.ac.uk
Wed May 16 11:16:00 PDT 2012

```Hi Vijay

Alex proposal is fine although that assumes that you are happy loosing
information related to trial variance and also you are not bothered by
multiple comparisons (doing that many t-test will automatically lead to
some false positives)

as an alternative, you can do the statistical analysis with LIMO EEG
(EEGlab toolbox to download) - that will allow you to model congruent
and incongruent conditions per subjects and then test across subjects
for a significant differences (in fact across all electrodes and time
frames and using clustering to control the false positive rate).
--> there is a tutorial which show a case very similar .. it should be
easy enough to do on your data

Cyril

Subject:
[Eeglablist] Performining the t-test and highiliting the area
From:
Vijay Narne <vijaynarne at gmail.com>
Date:
11/05/2012 05:54

To:
eeglablist at sccn.ucsd.edu

Dear List,
I am Vijaya Kumar Narne, PhD. We are running the
experiments on N400. We would like to compare ERP of the congruent and
incongruent condition. We are intrested in running t-test and highliting
the area of N400 in EEGLAB.  As we are new to EEGLAB, if any one help us
do this.
Thanking you
Vijay

Subject:
Re: [Eeglablist] Performining the t-test and highiliting the area
From:
Alex Davila <axel.1963 at hotmail.com>
Date:
14/05/2012 20:26

To:
<vijaynarne at gmail.com>, <eeglablist at sccn.ucsd.edu>

Dear Vijaya,

I assume that your two conditions are run on the same participants
(repeated measures). Then, what you need to do is just this:

1. Define the variable D = C - I

where C: your variable for the congruent condition and I: your variable
for the incongruent condition.

2. Calculate the mean M and the sample standard deviation S for D.

3. Calculate t = M/(S/rootsquare(N))

where N: your number of participants.

4. Depending on your hypothesis, compare your calculated t with a one
tail or two tail table t-student value. I'd suggest to use an alpha
value identical to those available in the relevant literature.

5. As you're interested in the N400 value, I assume you may extract it
from your ERP data of potentials across the time domain where you
may identify local maxima and minima for each participant.

6. Of course, before getting this data, you need to apply the standard
averaging procedure to extract the ERP signal from the noise.

All the best,

Alex.

--
Dr Cyril Pernet,

Brain Research Imaging Center
http://www.bric.ed.ac.uk/
Division of Clinical Neurosciences
University of Edinburgh
Western General Hospital
Edinburgh
EH4 2XU
Scotland, UK

cyril.pernet at ed.ac.uk
tel: +44(0)1315373661
http://www.sbirc.ed.ac.uk/LCL/
http://www.sbirc.ed.ac.uk/cyril

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20120516/e73c0b51/attachment.html
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: not available
Url: http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20120516/e73c0b51/attachment.pl
```