Hello,
I need help with how to analyse fMRI data using nonlinear regression. The
data I’m dealing with are the difference in activation btn emotional faces
(eg fear) vs neutral faces (ie Fear – Neutral face activation). I’m
interested in studying differences btn groups & gender (3 indept groups of
*unequal* sample size, males & females in each group). Preliminary
assessment of the data have indicated genderXgroup interactions, and
curvilinear/quadratic trends btn groups. Ideally I’d like to analyse using
ANCOVA, with age as covariate, group & gender as indept variables, with a
genderXgroup interaction - unfortunately, I don’t think SPM2 can handle such
complex ANCOVA analyses & interactions.
Also, due to greatly unequal sample sizes in the 3 groups, I thought an
alternative was to analyze by nonlinear regression, since I could re-code
group as a continuous variable (eg 1, 2, 3) & the sample size issue becomes
irrelevant. Analysing within males & females separately, I would therefore
have age as covariate, & group as indept variable. Is this appropriate? If
so, how can I analyse my data using nonlinear regression to capture such
quadratic patterns across groups? Any guidance would be greatly appreciated.
Cheers,
Stacey
|