question about adding motion as covariate in group level analysis

AFNI Message Board
1

Hello,

I am conducting a ReHo analysis on resting-state data and would like to add motion as a covariate in the group-level analysis (specifically, paired t-test using 3dttest++). My question is, what index do people usually use to control for differences in motion across groups and/or conditions? And how do I extract that from each scan using AFNI?

For example, once I finish running preprocessing pipeline using afni_proc.py, I saw that there are numbers for “average motion (per TR)” and “average censored motion”, can I record one of the numbers and then use it as a covariate in my later steps?

Thank you!
Jiaxu

2

Jiaxu,

My understanding is that motion occurs at the TR level, and thus is handled in the subject-level, not group-level, analysis.

I saw that there are numbers for “average motion (per TR)” and “average censored motion”, can I record one of the
numbers and then use it as a covariate in my later steps?

If you believe that differential head motion plays a role at the group level, you could try to use those indices as covariates at the group level and see how it pans out.

3

Thanks Gang!