 |
 |
Dear Arman, > > I am running a study of inter-subject variability of FA data due to some behavioral measurements and for motion correction I use eddy. > > I understood that each run of eddy generates a different rotated bvecs for the same subject. I wonder how these differences can affect the final statistical results. it is true that eddy will generate slightly different results on each run. You can remove that run-to-run variability by setting --initrand, but just remember that your results will still only represent one instance of all possible results. In general the run-to-run differences are _very_ small. Also, note that any effects on the end results would be dominated by the difference in corrected movement rather than the rotated bvecs. A one degree rotation of the brain can easily lead to a 50% intensity difference for some voxels, but a one degree rotation in Q-space will have very marginal effects. > > I also need to estimate the head motion in order to control the results for this artifact. What would be your suggestion for calculating the head motions? I found in the eddy_movement_rms file the displacements but what would be an efficient way to calculate the global head motion from the single displacements? As I suggested in an earlier email SNIP I would recommend using the “eddy_restricted_movement_rms” info. If you average the second column it would give you a measure of “short term movement” which would cause variance due to things like slice-to-volume movement and spin history effects. If you average the second column it would give you a measure of “long term movement” which would cause variance due to for example susceptibility-by-movement interaction and “movement within a stationary bias field”. You could then use those two as covariates in your model. SNAP I don’t think there is a right or wrong answer as to how to condense movement to a single (or small number of) parameter. So you just have to use common sense. Jesper > > Thanks so much in advance. > > Best Regards > > Arman