It seems like you would probably want to do it as a single analysis. Averaging the FAs from the separate analyses might have unintended consequences as it is a non-linear operation to compute the FA from the tensor. On the other hand, you could potentially average the tensors then compute the FA from the average tensor. Another consideration is that when you concatenate the raw data, you effectively concatenate the design matrices, which can lead to more efficient tensor estimates than doing the analyses separately. I don't have a feel for whether this would have a big impact or not. If you have a lot of direction, probably not. doug Tim Behrens wrote: > Hi Both - I agree that if there is no motion it will make no > difference, but I don't know what the possible advantage is of doing > it separately? > Clearly if there is some motion, the all in one approach will do better. > > T > > On 14 Mar 2007, at 20:54, Andreas Bartsch wrote: > >> it may depend on the number of runs. Occasionally, you may see eddy >> currents becoming less compensated the more DWI runs you acquire >> without interruption. I guess I would favor to correct for eddy >> currents and motion and then average but it's probably not much of an >> effort to simple try both ways and compare the results. >> > -- Douglas N. Greve, Ph.D. MGH-NMR Center [log in to unmask] Phone Number: 617-724-2358 Fax: 617-726-7422 In order to help us help you, please follow the steps in: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting