[FieldTrip] doublecheck significance cfg.alpha and cfg.clusteralpha
e.maris at donders.ru.nl
Fri Jan 28 14:05:07 CET 2011
> I am relatively new to fieldtrip, and just starting to make some
> serious use of the cluster-based permutation routines, and I would just
> like to (double)check something before I continu.
> I use ft_timelockstatistics with the following configuration structure:
> cfg.parameter = 'individual';
> cfg.method = 'montecarlo';
> cfg.statistic = 'depsamplesT';
> cfg.correctm = 'cluster';
> cfg.clusteralpha = 0.05;
> cfg.clusterstatistic = 'maxsum';
> cfg.minnbchan = 2;
> cfg.tail = 0;
> cfg.clustertail = 0;
> cfg.alpha = 0.025;
> >From my understanding of the tutorial the "cfg.clusteralpha" stands
> for the alpha-value used for the initial test-statistic (in my case a
> paired t-test) for thresholding the data,
This is correct.
and that the "cfg.alpha"
> field contains the alpha-value for the second step, i.e. selecting
> clusters after the permutation tests.
This is not correct. Actually, cfg.alpha is superfluous. In the output,
every cluster has a p-value assigned to it, and if there is one or more
cluster with a p-value less than your critical alpha-level (mostly 0.05 for
a one-sided and 0.025 for a two-sided test), then you have found a
> However, the naming of the parameters seems rather contraintuitive to
> me. I would have expected it to be the other way around, i.e
> "cfg.clusteralpha" for my clusters (second step), and not the initial
> Could anyone confirm (or deny) the way it is written in the tutorial?
> Thanks a lot,
> Hanneke Meeren
> Hanneke K.M. Meeren, PhD.
> Researcher and Lecturer
> Medical Psychology and Cognitive Neuroscience
> Tilburg University
> P.O. Box 90153
> 5000 LE Tilburg
> The Netherlands
> email: hanneke.meeren at uvt.nl
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
More information about the fieldtrip