Hi Shan,

I think the option "B" (full design) in this thread has is similar to the design you need: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=FSL;c56ad405.1509

The main EVs already code the interactions of what in your case would be drink vs condition, such that the 3-way interactions are simply these further compared between patients and controls.

Feel free to post again if something isn't clear.

All the best,

Anderson

On 30 March 2016 at 23:54, Shan(Hannah) Luo <[log in to unmask]> wrote:

Hi FSL experts,

I have 39 people (18 patient and 21 control) each of whom completed two sessions of a simple food cue task. At one session they had drink A, and at another session they had drink B. There are three food cue conditions in the task, neutral, high calorie and low calorie. I am interested in looking at drink (A/B) by food cue (high calorie vs. neutral, low calorie vs. neutral) by group (patient vs. control), 3-way interaction. I am wondering whether there's a way I can do it in FSL.

P.S., I've already done in the first-level analysis getting contrast maps for high calorie vs. neutral and low-calorie vs. neutral.

Your help is greatly appreciated. Thanks very much.