Thank you Martin for your quick responses. I have another question, if anyone could help. After completing the 2nd level analysis and creating my contrast, I noticed I had no significantly activated clusters compared to the series of within-subjects contrasts I created earlier. I was curious why that may be; possibly a lack of power (i.e. only 17 subjects) to detect any significant difference between the two treatments? Again, here is a review of my design for my crossover study:
Session 1 --> Treatment 1 --> Scan (Visual paradigm: Active vs. Control)
Session 2 --> Treatment 2 --> Scan (Visual paradigm: Active vs. Control)
Also, would it be reasonable to run a one-way ANOVA for each session (as opposed to a paired t-test) and compare which regions are significantly activated. With this approach I was concerned I would not be able to infer one treatment was signifcantly different than another, would I be correct? Thanks for any comments in advance.