It is not necessary to move all the bvalues to the front and average them
(unless you plan to use some software that requires it). You should do any
averaging after some form of motion correction (which eddy_correct does).
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Gary Turner
Sent: Thursday, August 27, 2009 10:53 AM
To: [log in to unmask]
Subject: [FSL] eddy correct & series averaging follow-up
Just a quick query following up on exchange posted a wile back about the
order of EC correct and series averaging ...
Mike Cohen: "I have a 60-direction DTI scan that is repeated 3x for
averaging ... Is it better to run eddy_correct on each of three acquisitions
then average them, or average them first and then run eddy_correct?"
T. Behrens: We have exactly the same data - what we do is merge all data
sets into a single big 4D file, run eddy_correct on this big file, then
split it up into 3 averages and average these.
I have the same protocol (only 30 dirs due to time constraints
unfortunately). I have interleaved b0s throughout the 30 diffusion weighted
scans which I average and move to the beginning of the diffusion weighted
scans for subsequent processing. Should I do this before eddy_current
correction? Or should I run EC on the whole dataset as acquired,
referencing the first b0 acquisition, and then do the averaging for both
diffusion and non-diffusion-weighted volumes as per the response above?
Gary Turner, Ph.D.
Helen Wills Neuroscience Institute
University of California, Berkeley
132 Barker Hall