[FieldTrip] ft_freqstatistics for between subject design

"Jörn M. Horschig" jm.horschig at donders.ru.nl
Wed Mar 19 15:28:04 CET 2014


Hi BeomJun Min,


> I think that each 5 values correspond with the frequency data point.
yes, that;s correct.

>
> Is that script suitable for the test for the significance, especially 
> avgovertime option 'yes'?
Yes and no. Yes, because you are testing for significance and, depending 
on what hypothesis you have, averaging over time is perfectly fine. Use 
avgovertime if you are looking for an effect independent of time, 
otherwise set it to 'no'. No, because you should pay attention to why 
selecting this time-frequency region. Now it seems that you selected 
frequency and time window based on eyeballing the data. If you have no 
apriori hypothesis about these frequencies and time window, it's a 
faulty approach to just select the region that shows the strongest 
approach and do a statistical test. You should take the whole 
time-window and whole freuqency spectrum (narrowed down as much as 
apriori hypotheses allow). If this is a real effect it is likely that a 
proper statistical test will show up only around that time-frequency.

>
> If I change the avgovertime option to 'no', then can I found the 
> p-values for each frequencies(5 values), and each time point(51 
> values), in total 255 values that could have multiple comparison issue?
Yes, you have this problem, is there a reason why you are not using the 
cluster-based permutation test (or any other correction for multiple 
comparison which is implemented by setting the cfg.correctm parameter 
accordingly)?

>
> And I thought the 'mask' values of 0 indicates 'statistically 
> non-significance'. Is that right?
Yes, that is correct.

Best,
Jörn

>
> -- 
> BeomJun Min, M.D.
>
> Department of Medical System Engineering (DMSE)
> Gwangju Institute of Science and Technology (GIST)
> 261 Cheomdan-gwagiro(Oryong-dong), Buk-gu, Gwangju
> 500-712, Republic of Korea (South)
> Phone: +82-62-715-3266 / Fax: +82-62-715-3244
> E-mail: mbj0310 at gmail.com <mailto:mbj0310 at gmail.com>, 
> http://bmssa.gist.ac.kr <http://bmssa.gist.ac.kr/>
>
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip


-- 
Jörn M. Horschig
PhD Student
Donders Institute for Brain, Cognition and Behaviour
Centre for Cognitive Neuroimaging
Radboud University Nijmegen
Neuronal Oscillations Group
FieldTrip Development Team

P.O. Box 9101
NL-6500 HB Nijmegen
The Netherlands

Contact:
E-Mail: jm.horschig at donders.ru.nl
Tel:    +31-(0)24-36-68493
Web: http://www.ru.nl/donders

Visiting address:
Trigon, room 2.30
Kapittelweg 29
NL-6525 EN Nijmegen
The Netherlands




More information about the fieldtrip mailing list