Dear AFNI experts,
I think this question is best directed at Gang, as it concerns whole-brain interaction modeling using 3dMVM.
I have data from two groups (healthy and diseased), and they each completed three tasks during fMRI acquisition (task_1, task_2, task_3). I am interested in testing for a group interaction on the linear contrast that beta values are greatest for task_1, then task_2, and lowest for task_3, i.e.: task_1 > task_2 > task_3. My question: How can I model this using 3dMVM?
Here’s what I currently have to run the model without the “linear” contrast across three levels of task:
3dMVM -prefix 3dmvm_output -jobs 12 -bsVars "group+age" -wsVars "task" -qVars "age" \
-num_glt 3 \
-gltLabel 1 grp_task_1 -gltCode 1 "group : -1*healthy 1*diseased strategy : 1*task_1 " \
-gltLabel 2 grp_task_2 -gltCode 2 "group : -1* healthy 1* diseased strategy : 1*task_2 " \
-gltLabel 3 grp_task_3 -gltCode 3 "group : -1* healthy 1* diseased strategy : 1*task_3 " \
-dataTable @mvm_table.txt > 3dmvm_output.txt
What would the gltCode look like for task_1 > task_2 > task_3?
Thank you!
Matthew