Hi James,

I have a very similar trial design and was wondering if you ever figured out if this could be done in randomise?

Best wishes

Kathryn

From: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of SUBSCRIBE FSL James <[log in to unmask]>
Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
Date: Friday, 17 March 2017 5:10 am
To: "[log in to unmask]" <[log in to unmask]>
Subject: [FSL] Randomise 2x3 Repeated Measures Design for TBSS and FSLVBM with Covariate of Non-Interest

Dear Experts, I would like to perform a repeated measures analyses on white matter data processed through TBSS and grey matter data processed in FSLVBM using randomise. I have two groups each scanned three times (1) baseline, 2) after a three month intervention, and 3) four months later as a followup). The data set is complete, without missing data. Furthermore, I want to use demeaned age as a covariate of non-interest and test if there is a change within and between group over time. My first issue is that I get an error that my model is rank deficient because at least one EV is a linear combination of the others. I am assuming this is due to the age covariate. Are there ways around this, possibly using a more separable metric of age? Secondly, is the aforementioned design feasible in randomise? Thank you in advance for the help, James