afni.nimh.nih.gov

> you are saying these compare the slopes of age between emotions?

Yes, that's correct.

> I thought maybe these tests were comparing the effects of emotion at mean centered age

For those you need (which by default control each covariate at its center value)

'Emotion : 1*Afraid -1*Calm '
'Emotion : 1*Afraid Age'
'Emotion : 1*Calm Age '

> "This option only makes sense if a within-subject factor has more than 3 level."

Typo here: '3' should be '2' in the sentence above.

Gang