I'm sorry if this question has been addressed before, but I couldn't find anything which would answer my question.

I have three conditions for each subject (e.g. 1,2,3), and I want to compare two pairs only (e.g., 1 vs 2 and 2 vs 3). Should I do within-subject ANOVA (with two contrasts [1 - 1 0] and [0 1 -1]), or two paired t-tests? And if I do pared t-tests, how do I control for multiple comparisons (not voxel-wise, but condition-wise)?

I performed both tests, and outcomes are similar although not exactly the same.

Thank you!

Regards,

Natalia