help > RE: Tricky second-level contrast
Aug 14, 2019  09:08 AM | Alfonso Nieto-Castanon - Boston University
RE: Tricky second-level contrast
Dear Fred,

Typically you would create a new second-level covariate directly encoding the "delta clinical" measures that you would like to evaluate (e.g. clinical scores after treatment minus clinical scores before treatment), and then, for your second-level analysis, simply select the subject effects 'AllSubjects' and 'deltaClinical' (and enter a between-subjects contrast [0 1]), and select the conditions 'Pre' and 'Post' (and enter a between-conditions contrast [-1 1]).

Naturally you can also extend these analyses to include additional between-subject factors beyond the simple regression framework above (e.g. A/B groups, for a repeated-measures ANCOVA framework, other covariates of interest to be used as control covariates for a multiple regression analysis, etc.)

Hope this helps
Alfonso

Originally posted by fred Sampedro:
Dear CoNN experts,

Even after checking the very good examples of the conn manual regarding second-level contrasts (attached), I've been stuck with the following contrast:

I have two groups of subjects (A and B), and each subject underwent a resting-state session Pre and Post treatment. I have some behavioral variables that had Pre-Post changes. I would like to know whether some Pre-Post connectivity changes were related to those behavioral changes in the whole sample. In other words, I am interested to perform some kind of "delta clinical - delta connectivity" regression analysis, in order to investigate whether the post-treatment clinical changes that we observed were driven/associated to connectivity changes after treatment.

Thanks a lot in advance!

Threaded View

TitleAuthorDate
fred Sampedro Aug 9, 2019
RE: Tricky second-level contrast
Alfonso Nieto-Castanon Aug 14, 2019