Preveously, I had questions about Braph. Now I have another question.
In between-group differences in nodal measures, we can visually find the results more/less than 95% CI. However, in statistical analysis, it doesn’t seem to be difference significantly even in the period more/less than 95% CI, because of the fdr correction value of 0.
So, is it correct when the between-group differences are more/less than 95% CI, the results are significantly different, regardless of the fdr correction value of 0?
Braph will output for you the raw p-values. This means that Braph will consider the values you have for each brain region and each subject, and perform a permutation test for them. The permutation test is performed for each region individually and the p-values that are outputted by Braph are related to this permutation test. This tests whether the difference between groups, for the particular brain region, is statistically different than zero.
However, due to the simultaneous testing for many brain regions, one runs into the problem of multiple comparisons, i.e. controlling for the type I error. In Braph this is done by the fdr correction (http://braph.org/forums/topic/confidence-intervals-and-p-statistic/). Therefore, the fdr correction tells you whether a comparison is statistically significant after controlling for the effect of multiple comparisons.
Hope this helps. If I can do anything more to help, please let me know.