I would like to ask you for advice on whether to use 3dMVM or 3dLME for my analyses.
My data consists of 1 between subjects factor (group (2 levels)) and 2 within subject factors: task conditions (5 levels), and time bins (4 levels; estimated with a FIR model). There’s no missing data. I would like to test (per condition) whether the shape of the activity over time differs based on group, so whether there is a groupxtime interaction.
I have initially set this up in 3dMVM (see code below), but now I have read the following in the 3dMVM manual, and I am unsure this is the right approach: “For an F-test involving a within-subject variable, use -glfCode with 3dLME instead.” Does this warning also apply to my analysis? Would 3dLME be a better approach after all?
3dMVM -prefix cond1_groupxtime
-glfLabel 1 ‘cond1_group’ -glfCode 1 'Condition : 1cond1 Group : 1group1 -1group2’
-glfLabel 2 ‘cond1_grouptime’ -glfCode 1 'Condition : 1cond1 Group : 1group1 -1group2 Time : 1t1 -1t2 & 1t2 -1t3 & 1t3 -1t4’ \