You could edit the eddy_correct script to use spline interpolation instead of trilinear and that would reduce this effect.

Peace,

Matt.

From: Anna Greenwald <[log in to unmask]>

Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>

Date: Tuesday, April 2, 2013 9:10 AM
To: <[log in to unmask]>

Subject: Re: [FSL] eddycorrect introduces motion

Thanks, Jesper!

Follow-up question: eddy_correct introduces a good bit of implicit smoothing, so I'm hesitant to combine non-eddy-corrected datasets (from subjects who didn't move much) with eddy-corrected datasets (from subjects where eddy_correct helps reduce motion). Of course I could "mimic" the implicit smoothing by explicitly smoothing my uncorrected datasets by a small amount. Is that something you'd recommend?

Thanks,

Anna

On Mon, Apr 1, 2013 at 5:42 PM, Jesper Andersson <[log in to unmask]> wrote:

Dear Anna,

> I'm running the default eddycorrect function on a DTI dataset (5 b0 to begin with, followed by 30 different direction). I eddycorrect to the first b0. The problem is that when I look at the non-corrected and the corrected 4D file in movie mode, the latter actually has much more motion than the former. Have other people experienced this problem, and maybe found settings that ameliorated it? From looking at the movies, I'm really tempted to just continue my processing on the non-eddycorrected files, but I understand that that's generally thought to be a bad idea. Any input?

this is something that is sometimes seen. Especially if the data have little movement and/or distortions to begin with. If your visual assessment is that data is better before the "correction" then I think that is what you should use.

There is a recently released application called eddy that we have found to work better, but that also has higher demands on the data. Have a look at the FSL web-pages to see if your data is suitable.

Good Luck Jesper

>
> Thanks!
> Anna
>