Thank you Stam! From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of Stamatios Sotiropoulos Sent: 19 February 2013 10:04 To: [log in to unmask] Subject: Re: [FSL] combining dwi images Hi 1) Yes 2) It's the rotations that matter most. If you have large rotations, it's reasonable not to average and treat each acquisition separately by using the corresponding bvecs. Cheers Stam On 19 Feb 2013, at 09:59, Patai, Zita wrote: Dear FSLUsers I have 3 acquisitions of DWI data per subject. I need to combine these and I was wondering: 1) practical question: does the eddy current correction account for subject motion as well ? 2) theoretical question: if the movement is large-ish, does it make sense to create a combined image when the gradient directions may not be equivalent in all three acquisitions given the displacement of the voxels? Thanks for your help! z _________________________ Eva Zita Patai Postdoctoral Research Associate UCL, Institute of Child Health [log in to unmask]<mailto:[log in to unmask]> 020 7905 2730