[FieldTrip] fieldtrip Digest, Vol 8, Issue 40

Eric Maris e.maris at donders.ru.nl
Thu Aug 4 14:08:39 CEST 2011


Hi Anna,

 

 

 

thank you for your first answer. Actually my question is really on the right
way to implement a group comparison in fieldtrip (patient vs control group).


Is it ok to run an independent t-test with the following design:

design = zeros(2,n_group1+n_group2); 
design(2,:) = 1;
design(1,1:n_group1)     = 1:n_group1; 
design(1,n_group1+1:end) = 1:n_group2;   



This is not correct. I advise you to have a look at the Fieldtrip tutorial
on cluster-based permutation tests. A between-subjects experiment (e.g.,
patients versus controls) is formally equivalent to a between-trials
experiment (explained in detail in the tutorial), and therefore should be
analyzed in the same way. You can even copy-paste part of the tutorial code
for the analysis of your between-subjects data.

A subsequent question is whether it is possible to implement a mixed design
or we always have to proceed by combining conditions and compare it between
groups.



I do not see the contrast between these two options. What is wrong with the
second option? In fact, it is the option that is advocated in many
statistics books (e.g., Maxwell & Delaney, Johnson & Wichren).

 

Best,

 

Eric Maris

 

 

 


Cheers, 
Anna.




Message: 2
Date: Sat, 30 Jul 2011 20:45:12 +0200
From: "Eric Maris" <e.maris at donders.ru.nl>
To: "'Email discussion list for the FieldTrip project'"
       <fieldtrip at donders.ru.nl>
Subject: Re: [FieldTrip] Group comparison statistics - ERF study
Message-ID: <03a601cc4ee8$d152b760$73f82620$@maris at donders.ru.nl>
Content-Type: text/plain; charset="us-ascii"

Hi Anna,



You can compare the two groups with respect to any linear combination that
can be formed using the observations in the 2 (conditions) x 2 (response
types) within-UO design. For testing interaction effects, these linear
combinations are the usual contrasts.





Best,



Eric Maris





dr. Eric Maris
Donders Institute for Brain, Cognition and Behavior

Radboud University
P.O. Box 9104
6500 HE Nijmegen
The Netherlands
T:+31 24 3612651 <tel:%2B31%2024%203612651> 
Mobile: 06 39584581

F:+31 24 3616066 <tel:%2B31%2024%203616066> 
 <mailto:e.maris at donders.ru.nl> mailto:e.maris at donders.ru.nl

 <http://www.nphyscog.com/> http://www.nphyscog.com/







From: fieldtrip-bounces at donders.ru.nl
[mailto:fieldtrip-bounces at donders.ru.nl] On Behalf Of Anna Lambrechts
Sent: donderdag 28 juli 2011 11:32
To: fieldtrip at donders.ru.nl
Subject: [FieldTrip] Group comparison statistics - ERF study



Hi,

I am trying to run a group comparison analysis on event-related fields data
in a 2 (groups) x 2 (conditions) x 2 (response types) design. Is this
possible at all with any fieldtrip script? As far as I know implemented
statistics look at within-groups comparison.

Thanks,
Anna.

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20110730/1f410
391/attachment-0001.html>

------------------------------

_______________________________________________
fieldtrip mailing list
fieldtrip at donders.ru.nl
http://mailman.science.ru.nl/mailman/listinfo/fieldtrip

End of fieldtrip Digest, Vol 8, Issue 40
****************************************

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20110804/594a259c/attachment-0002.html>


More information about the fieldtrip mailing list