Dear all,
I used fsl_motion_outliers to detect the head motion in the data. When there are significant head-motion volumes in the middle of the sequential data, do we just discard the whole dataset due to knock-on effect or is it fine to add the confounding variables generated from fsl_motion_outliers to remove the volumes in the analysis?
When we want to remove the volumes, there are 2 ways to do:
(1) just click “Add the confound EVs” in the Stats tab of FEAT, or
(2) create EV which is 0 everywhere and 1 for the volume we want to remove and then add it into our design matrix without convolution.
Is there any difference between these two methods?
Thanks.
|