Dear Alfonso,
I have a further question on how to set contrasts at the second level for a ROI-to-ROI analysis.  I have two group of subjects (patient vs. control) and three conditions (task with low working memory load, task with high working memory load and baseline). Now, I hope to examine the difference in functional connectivity between task (combined high and low load) and baseline. 
I assume to set  [1 -1] for Subject effect and [0.5 0 0; 0 0.5 0; 0 0 -1] for Conditions,   or
                        [1 -1] for subject effect and  [0.5 0.5 -1]  for Conditions
 Could you tell me which one is correct? 
Thank you in advance.
Ivan