Hi Kaitlin,

Yes, you are exactly right, you could define three subject "groups" (one for your single patient "pre" scan, another for your single patient "post" scan, and another for your multiple-subject "control" group), and then use a [1 0 -1] contrast to evaluate the differences between your patient "pre" scan and your control group, a [0 1 -1] contrast to evaluate the differences between your patient "post" scan and your control group, and a [-1 1 0] to evaluate the "post vs. pre" differences in connectivity against the between-subjects variance in your control group. All of these will be valid analyses, yet note that the latter will be a somewhat conservative test for those post vs. pre differences. A more standard and powerful test would evaluate the post-pre differences in your patient against comparable post-pre differences in your control group (that would typically have more power simply because the within subject post-pre difference contrast is expected to have considerably less between-subjects variance than each of the individual post or pre effects) but for that you would need to have repeated measures of your control group as well (in a setting that is comparable to the "pre" and "post" conditions in your patient).

In any way, to set this up you simply need to define in Setup.Basic the number of subjects as N+2 (where N is the number of controls), the number of sessions per subject as 1, then enter in Setup.Functionals the "pre" and "post" scans of your patient as if they were individual scans from different subjects, define a single "rest" condition in Setup.Conditions, and last, define your three second-level covariates in Setup.CovariatesSecondLevel as: "PatientPre" (e.g. [1 0  zeros(1,N)]), "PatientPost" (e.g. [0 1 zeros(1,N)]), and "Controls" (e.g. [0 0 ones(1,N)]). 

Last, and just for completion, if you do have repeated ("pre" and "post") measures for your control group, then you would instead define in Setup.Basic the number of subjects as N+1 (where N is the number of controls), the number of sessions per subject as 2, then enter in Setup.Functionals the "pre" and "post" scans for all your controls and patient as separate sessions, define two conditions "pre" and "post" in Setup.Conditions associating each condition with an individual session, and last, define two second-level covariates in Setup.CovariatesSecondLevel as: "Patient" (e.g. [1 zeros(1,N)]), and "Controls" (e.g. [0 ones(1,N)]). 


Hope this helps
Alfonso

Originally posted by Kaitlin Cassady: