[FieldTrip] Defining a custom statfun for cluster statistics

Herring, J.D. (Jim) j.herring at fcdonders.ru.nl
Mon Jun 2 10:51:38 CEST 2014

Hi Steve,

If I understand you correctly you basically want to test a one-sided 
interaction effect, which boils down to doing a one-sided dependent samples 
t-test between the difference scores of condition I vs the difference scores 
of condition II.

So, first you would have to calculate the difference timeseries diff_I(Ia – 
Ib) and diff_II(IIa-IIb). Then you run ft_timelockstatistics using 
statfun_depsamplesT on diff_I vs diff_II testing one-sided only looking at 
positive clusters (If you expect that diff_I > diff_II ).



From: fieldtrip-bounces at science.ru.nl 
[mailto:fieldtrip-bounces at science.ru.nl] On Behalf Of Stephen Politzer-Ahles
Sent: maandag 2 juni 2014 9:40
To: fieldtrip at donders.ru.nl
Subject: [FieldTrip] Defining a custom statfun for cluster statistics

Hello all,

Are there any examples for defining a new statfun to use in 
ft_timelockstatistics? I thought there were, but can't currently find them 
on the wiki.

I am testing a 2x2 interaction (let's call the conditions Ia, Ib, IIa, and 
IIb), and I an only interested in a certain kind of interaction, one where 
abs(Ia-Ib) is large and abs(IIa-IIb) is small (i.e., an effect of the second 
factor at just one level of the first factor, and not the other level). So 
rather than using F as my test statistic for defining clusters, I would like 
to use something like abs(ttest(Ia,Ib))-abs(ttest(IIa,IIb)). I just don't 
know how to define the function to do this; if anyone could point me in the 
right direction I would greatly appreciate it.



Stephen Politzer-Ahles
New York University, Abu Dhabi
Neuroscience of Language Lab

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20140602/1808686f/attachment-0002.html>

More information about the fieldtrip mailing list