Dear FreeSurfer experts

I have a question regarding the correct data analysis method for my study design and could unfortunately not find an example of a similar analysis on your mailing list or documentation on the wiki.

I have repeated structural MRI data from one group (within-subjects design) at two time points (T0 and T1). Each subject has been randomly assigned to receive treatment or placebo either at T0 or T1 (i.e., half of them had treatment at T0 and placebo at T1 and half of them had placebo at T0 and treatment and T1). I am interested in the putative effects of treatment on cortical thickness. One major challenge I face stems from the fact that the scan interval (between T0 and T1) varies between subjects and, most importantly, that I expect a differential impact of treatment as compared to placebo at T0 on thickness at T1 (long-lasting increases). Thus, the time between scans and the order of the two conditions (treatment vs placebo) should be taken into account.

Repeated measures ANOVA seems not appropriate. I have been thinking about conducting linear mixed effects models to analyse my data but also been thinking about the longitudinal two-stage model you describe on the wiki. I am unsure which method appears most appropriate and I am unsure how to model the time between scans correctly.

Any suggestion regarding the most appropriate method and how to model the time between scans or reference to a similar data analysis question would be highly appreciated!

Thank you very much for your help and all best wishes, Martina