How do conduct separate analyses for negative and positive effects using nistats

In SPM, it seems to be standard to conduct separate analyses for negative and positive effects. For example, you would conduct first a two-sample t-test for Patients > Controls and then a second one for Patients < Controls. You would also treat those tests separately when you correct for multiple comparisons. Is it possible to do that using nistats?

Take the nistats tutorial on the OASIS dataset as an example. Is there an option to compare gender effects separately for Male > Female and Male < Female? If yes, how would you do this?

P.S.: I donā€™t really get this approach methodologically, but I need to do it anyways to replicate a former SPM-analysis by a colleague. So any methodological explanation is greatly appreciated.

This aspect only comes into play when you threshold the maps.
The function used in nistats has a parameter two_sided, that is true bue default, but you can set it to false.
This should address your concern.

1 Like