FWE、FDR、AlphaSim correction

Submitted by wujingxi on Sat, 03/12/2016 - 16:43

First, when do two sample t test, I need a mask.

And I want to use one sample t test's results (corrected) of two groups to make a mask.

In REST, I use FWE、FDR、AlphaSim correction respectively for one sample t test's results, and all successful.

But Now, my question is that ...

                                                 which correction method's result should I choose ?

Submitted by ZangYF on Sun, 03/13/2016 - 21:57

The purpose is to create a mask. So any of the three is fine.

If you want to use a larger mask, you can use a more liberal threshold for any of the three correction methods on the one-sample t-test. And vice versa.