I am noticing that many of my subjects have motion artifact that is strongly affecting my data. Motion correction using 3dregistration has been done in the preprocessing, and censor tr has been used to eliminate tr's with motion greater than 2 mm, but my problem persists. I am ending up with negative (blue) horizontal slices that are very clearly related to motion. I am thinking of reducing the criteria in censor tr to 1 mm to catch the motion that I am not catching at 2 mm but I am worried that I will loose too many images. Is there a different method of motion correction that I should try? One that deals with small but frequent head motion throughout the time course of the data collection. It seems to be more of a problem in our older adult population.