[FieldTrip] timelock statistics

[Raghavan Gopalakrishnan]

Dear all,
I have a question regarding statistics. I am doing cluster analysis using
monte carlo.
I have 2 conditions (cond-1 and cond-2). Each condition has 3
sub-conditions that are temporally separated (lets say cond-1a/2a,
cond-1b/2b and  cond-1c/2c). I did the statistics
1. on cond-1 vs cond-2 as a whole.
2. cond-1a vs cond-2a, cond-1b vs cond-2b and cond-1c vs 2c separately
after splitting the conditions cond-1 and cond-2 using ft_selectdata.

I was expecting the answers from #1 and #2 above to be the same. #2 showed
one significant cluster in each of the three individual comparisons,
totaling 3 clusters. However, #1 above showed only one significant cluster
corresponding to one of the three individual cluster in #2. Can anyone
explain why results are different?

Thanks,
Raghavan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20140213/8894f1d4/attachment-0001.html>