Hi all, 

I would like some advice on whether my approach is valid.
I am working on 3 contrasts of interest from a task whose stimuli are relatively long blocks (10-13s) and I would like to explore several things but despite reading several papers, I have doubts about my analysis program.

1) Independently of the activations obtained in the comparisons (on SPM), I would like to explore the connectivity for each contrast of interest. So I was thinking of conducting a wRRC analysis including the atlas included in CONN.The idea would then be to see if this correlation matrix (if I observe one that passes the statistical threshold) is affected by an external variable (a score on a clinical scale). But I have seen quite few papers doing this type of analysis from a task (many do it on resting-state) and all restrict the RCC analyses to a pool of ROIs. So I guess this approach is not valid? This analysis is exploratory in nature and I won't be surprised if you tell me that wRRC analyses are not appropriate for this. 

2) I also have some ROIs defined a priori or from the activations of my task and I would like to know if their activity is correlated to those of other brain regions depending on the contrasts of interest and external variables. I think that for this type of purpose, a gPPI or weighted-GLM analysis should be performed. If I understand correctly, both approaches are valid given my stimuli (block >10s)? However, isn't there a statistical problem with conducting multiple gPPI analyses in this case (1 analysis per contrast per seed)?

If anyone has any advice, I'm interested!

Thanks, 

Alex