Dear experts,
We have a clinical dataset pre-post treatment (cognitive task) which failed to produce significant results using regular GLM analyses. We are hoping for better luck with a multivariate approach (e.g., Melodic). Would you recommend including ICA-AROMA and diff_pow (motion parameters) in our preprocessing for group MELODIC?
We've regressed motion spikes in our original preprocessing. If we did the ICA-AROMA, should we still regress out motion spikes or should we skip that?
Thank you in advance,
-Marina
|