Hi,

I have a question about the correct approach for applying an ROI mask to a group analysis (e.g. a between-groups t-test).

I have two groups that I am comparing during a certain contrast (e.g. emotional-neutral), and I am interested in the left and right amygdala.

Using an anatomically-derived amygdala mask (as opposed to a whole brain mask) during the 3dttest++ command with two sets would achieve this, correct?

What about the use of 3dcalc as shown below, in which "a" is the anatomically derived amygdala mask and "b" is the .BRIK output of a ttest (run via 3dttest++) previously run with a whole brain mask? The number in the parentheses being subtracted from b would be the t threshold associated with the desired p value from the whole brain ttest (i.e. 2.983 is the t value associated with a .005 p value when viewing results in the whole brain ttest).

3dcalc \
-a mask_r_amygdala+tlrc.BRIK \
-b ttest_groupA-groupB_emo-neutral+tlrc.BRIK'[1]' \
-expr "a*step(b-2.983)" \
-prefix ROI_r_amygdala_emo-neutral \

Would this use of 3dcalc resample the results of the whole brain ttest to the volume of the amygdala mask? Or does it just spit out the contents of the whole brain analysis but within the bounds of the mask (not resampled for the smaller volume, and thus a result equivalent to that seen in the whole brain ttest)?

Thanks for your help!