Hi! I would like to include motion estimates as covariates in my statistical analyses (not first-level analyses, I already included them here, but I need to include them one more time when comparing patients vs. controls). Each subject underwent multiple fmri tasks and for each subject and task I have the usual 6 motion parameters (‘trans_x’,‘trans_y’,‘trans_z’,‘rot_x’,‘rot_y’,‘rot_z’).
My problem(s):
1.) I need one value per subject
2.) This value needs to include all tasks (so I somehow have to aggregate over the tasks). I already had the idea to simply concatenate the motion arrays here?
3.) Subjects come from different studies and I fear that some motion measures are to sensitive to spatial resolution and/or scanner setttings (for example frame-wise displacement)?
In other words: Is there a motion measure that can sum up / represent the motion for each subject over all tasks in a single value and is also insensitive to different scanner resolutions?