Hi Kaitlin,

Sorry linear mixed models are not implemented in CONN (to do that you could always export the BETA_Subject*_Condition*_Source*.nii files resulting from your first-level analyses to other software packages to then run your desired group-level analyses).

Within the context of GLM analyses (those implemented in CONN), it is a bit more limited what you can do, but a simple approach would be to limit the number of sessions to a fixed set (e.g. 2) and evaluate how connectivity changes over time using some explicit model (e.g. linear or higher-order) of those temporal changes. For example, you could define a between-conditions contrast to evaluate differences in connectivity between two sessions (e.g. between the first and last session for each subject) and use a linear model  across subjects with 'duration' as a covariate (or more complex models if needed) in order to obtain a fit characterizing the observed changes in connectivity over time. 

Hope this helps

Alfonso

Originally posted by Kaitlin Cassady:

Hello!

I have a somewhat complicated dataset: about 50 subjects, each containing at least 2 sessions but some have as many as 4 sessions. All sessions are at least 6 months apart (i.e., "longitudinal"). The problem is that the duration between sessions is not the same for all subjects. Some subjects have 6 months between sessions, others have years. Is there a way to look at change over time in FC in these subjects using some kind of linear mixed model or another method? Any help would be greatly appreciated!

 

Thank you!