So when using dual-regression/randomise I know that you have to look at t-stat values for associated, non-interaction contrasts to interpret significant interaction effects.

However, I'm a bit unsure as to the best way to do so with FSLnets results. What output should I be looking at to interpret significant interactions -  particularly if the non-interaction contrasts are not significant?

Would I simply look at output from nets_boxplots?

Thanks