[MRIStudio-users] [Mristudio-users] MRI - Head Motion issue

[Susumu Mori]

Hi all,

I want to add few more things.

Motion artifacts are indeed clear issues for all kind of MRI studies,
especially for those based on multiple images such as DTI, fMRI, ASL, etc.
For DTI, there were active research studies on motion artifacts and many
tools were developed in late '90s and early '00s.
DtiStudio and MriCloud perform usual image registration and eddy current
corrections and provide reports about the amount of motions, eddy currents,
and the number of rejected voxels during the corrections.

However, motion correction is an ill-posed scheme. No matter which tool you
use, it can correct only a portion of problems.
Suppose you register a 3D image A and B and the registration tool says,
"there was x translation and y rotation", this statement is accurate only
if the registration is accurate.
Then, how do we know if the registration is accurate? The software judges
the accuracy based on a cost function it monitors; the minimum (or maximum)
of the cost function is the best registration. The problem is, there is no
guarantee that the minimum cost function is indeed the metric of accurate
registration. Unfortunately, we don't have a magic cost function that is
truly accurate.

Also, there are types of motion we can't correct. For example, DTI is based
on multiple 2D slices, usually axial slices. Often, the odd-number slices
are acquired first followed by even-number slices. If motion happens
between the first odd-number scan and second even-number scan, you can't
correct them by usual 3D registration scheme (the motion problems reside
within the 3D image). Imagine how things get complicated if sagittal
rotation happens during multi-slice axial 2D scans.

In addition, motion often causes signal changes. Even if voxel locations
are corrected accurately, we don't have means to correct the intensity.

So, as Barbara pointed out, at some point, we need to take some level of
motion artifacts as noise in our data. We only hope there are similar
amounts of motion artifacts across groups. If there is a possibility that
patient groups have more motion artifacts, this causes typical co-variate
issues.

It is always a good practice to monitor the registration reports (assuming
the numbers in the reports reflect the truth about the extent of motion)
and check the absence of significant differences among groups. I must admit
that I haven't fully utilized the reports, though...

On Mon, Jun 3, 2019 at 11:04 AM Barbara Kreilkamp <bakk.hbg at googlemail.com>
wrote:

> Dear Dr Mohammad,
>
> This paper you are referring to (Yendiki et al. 2013 neuroimage) reports
> on the methods used to correct for motion and the spurious group
> differences present even when accounting for motion with special
> correction tools.
>
> If there are differences in motion between groups that you are
> comparing, then yes, motion can be have confounding effects on your
> results even if they have been 'corrected'. This is because motion
> always also means loss of information (e.g. slice dropout etc.) which
> cannot be easily mitigated.
>
> FSL and Freesurfer (Tracula) offer tools to extract motion parameters
> for individual subjects so you can compare gross motion across groups
> and account for it in your analysis. It's always a great idea to
> visually check your unprocessed DTI data as well for any signal loss etc.
>
> Best wishes,
>
> Barbara
>
>
>
> On 27/03/2017 14:56, Michael McLane wrote:
> > FYI -  ?
> > ________________________________________
> > From: mristudio-users <mristudio-users-bounces at mristudio.org> on behalf
> of Shaimaa Abdelsattar <shaimaa96 at hotmail.com>
> > Sent: Sunday, March 26, 2017 2:05 PM
> > To: MRI Studio
> > Subject: [Mristudio-users] Head Motion issue
> >
> > Hello,
> > There are papers in the literature discussing the effect of head motion
> on DTI analysis as it can induce spurious differences among groups. I would
> like to ask if this motion is still a concern even after eddy current
> correction and registration into b0 image using AIR tool in DTI studio for
> tractography.
> > Thank you in advance
> > Dr.Shaimaa Mohammad, MD
> >
> > Sent from my iPad
> > _______________________________________________
> > mristudio-users mailing list
> > mristudio-users at mristudio.org
> > http://lists.mristudio.org/mailman/listinfo/
> > Unsubscribe, send a blank email to:
> mristudio-users-unsubscribe at mristudio.org
> >
> > _______________________________________________
> > mristudio-users mailing list
> > mristudio-users at mristudio.org
> > http://lists.mristudio.org/mailman/listinfo/
> > Unsubscribe, send a blank email to:
> mristudio-users-unsubscribe at mristudio.org
>
> _______________________________________________
> mristudio-users mailing list
> mristudio-users at mristudio.org
> http://lists.mristudio.org/mailman/listinfo/
> Unsubscribe, send a blank email to:
> mristudio-users-unsubscribe at mristudio.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mristudio.org/pipermail/mristudio-users/attachments/20190604/effd1133/attachment.html>