I have found some motion anomalies by subtracting a time-shifted EPI from itself; TR5-TR4, TR6-TR5, TR7-TR6 etc...
If there is no motion between TRs, the difference images should be near zero values. Some parts of my difference images show movement in one part of the brain, while the remainder of the image values are near zero.
Is volreg picking up on this? Perhaps not; in plane motion?
Has anyone looked for a similar type of motion, and have you tried to censor these TRs?
Thanks,
Peter