Hi Greg,

I am going to assume here that the three conditions that you mention are all within-subject effects (let me know if these represent instead three different subject groups). Assuming you have two between-subject effects (treatment and placebo, dummy coded regressors identifying the subjects in each group), and 6 within-subject conditions (named preA to preC, and postA to postC, for the pre- and post- session scans for each of your 3 conditions A,B, and C), then:

To begin with, if you just want to look at the interaction between treatment/placebo and pre/post for condition A you woud select 'treatment' and 'placebo' in the between-subject effects, and enter a between-subject contrast of [1 -1], and then select 'preA', and 'postA' in the conditions list and enter a between-conditions contrast of [-1 1]. This will highlight those areas where the connectivity differences for condition A (comparing post- vs. pre- treatment) are different between the treatment and placebo groups.

Then, if you want to do the same, but simultaneously across the three conditions (perform an F-test of the interaction between treatment/placebo and pre/post across any of your three conditions) you would select 'treatment' and 'placebo' in the between-subject effects, and enter a between-subject contrast of [1 -1], and then select 'preA', 'postA', 'preB', 'postB', and 'preC','postC' in the conditions list and enter a between-conditions contrast of [-1 1 0 0 0 0; 0 0 -1 1 0 0; 0 0 0 0 -1 1]. This will highlight those areas where the connectivity differences (comparing post- vs. pre- treatment) are different between the treatment and placebo groups, in any of the three conditions tested. 

Now, if you want to compare the above effects across conditions (e.g. you want to see whether the treatment effect is different between condition A and condition B), you would select 'treatment' and 'placebo' in the between-subject effects, and enter a between-subject contrast of [1 -1], and then select 'preA', 'postA', 'preB', 'postB' in the conditions list and enter a between-conditions contrast of [-1 1 1 -1].

Last if you want to perform the above comparison simultaneously across all conditions (to see whether the treatment effect is the same across the three conditions), you would again select 'treatment' and 'placebo' in the between-subject effects, and enter a between-subject contrast of [1 -1], and then select 'preA', 'postA', 'preB', 'postB', and 'preC','postC' in the conditions list and enter a between-conditions contrast of [-1 1 1 -1 0 0; 0 0 -1 1 1 -1].

Let me know if this clarifies.

Best
Alfonso


Originally posted by Greg Book: