Dear List,

I posted an email some time ago about non-linear coregistration of
the brain (MPRAGE) with subdural electrode implantation.
I am hoping if FSL experts would give us some advice to solve the
problem we have.

The MRI (MPRAGE) is taken after the grid implantation to identify the
location of electrordes (as a signal void due to property of platinum
electrode) for epilepsy patients having invasive evaluations.

I would like to coregister the brain with subdural electrodes (i.e.,
partially distorted because of the implantation) to the brain before
the electrode implantation (non-distorted brain of the same patient).
Attached, please see an example (MRI taken after grid implantation).
In the majority of cases the following configuration parameters do
well, but in some it did not work well.

While FLIRT using bet'ed MRI did a relatively good job, FNIRT
ocassionally locally enlarge/warp the lateral parietal area more than
we expected (cortex enlarged out of the lateral parietal surface of
MPRAGE taken before the implantation).

For configuration files for FNIRT, we used T1_2_MNI152_2mm.conf
(default config file of FNIRT) except
--in=MPRAGEwithGrid.nii.gz
--ref=MPRAGEbeforeGrid.nii.gz
--inmask=skull stripped MPRAGEwithGrid.nii.gz
--refmask=skull stripped MPRAGEbeforeGrid

I used --inmask and --refmask because the MRI after grid implantation
usually has locall swelling outside the brain (some fuid collection
epidurally or in the scalp).

I understand that FNIRT is originally made for better coregistration
of individual brain into MNI space, and not for this sort of purpose.
So far FNIRT is doing better job than IRTK for our patients.

I would greatly appreciate it very much if you could kindly give us
any advice on the cofigration parameters.

Sicnerely,

Riki
--------------------
Riki Matsumoto M.D., Ph.D.
Department of Neurology
Kyoto University Graduate School of Medicine
Shogoin, Sakyo, Kyoto, JAPAN 606-8507
Tel: +81-(0)75-751-3772
Fax:+81-(0)75-751-9416 (for large transfer, use +81-(0)75-761-9780)
--------------------


>Delivered-To: [log in to unmask]
>X-RAL-MFrom: <[log in to unmask]>
>X-RAL-Connect: <heian.kuhp.kyoto-u.ac.jp [130.54.188.225]>
>X-Mailer: QUALCOMM Windows Eudora Version 6.2J rev4.2
>X-CCLRC-SPAM-report: 0.05 : FORGED_RCVD_HELO
>X-Scanned-By: MIMEDefang 2.57 on 130.246.193.104
>Date: Thu, 29 Jan 2009 17:45:45 +0900
>Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
>Sender: FSL - FMRIB's Software Library <[log in to unmask]>
>From: Riki Matsumoto <[log in to unmask]>
>Subject: [FSL] FNIRT failure for grid patients
>To: [log in to unmask]
>List-Help:
><http://jiscmail.ac.uk/cgi-bin/webadmin?LIST=FSL>,<mailto:[log in to unmask]>
>List-Unsubscribe: <mailto:[log in to unmask]>
>List-Subscribe: <mailto:[log in to unmask]>
>List-Owner: <mailto:[log in to unmask]>
>List-Archive: <http://jiscmail.ac.uk/cgi-bin/webadmin?LIST=FSL>
>
>Dear List,
>
>Please let me ask for your professional advices.
>
>I would like to coregistrate 3DMRI (MPRAGE) after electrode
>implantation (for invasive evaluation of epilepsy surgery) into one
>before electrode implantation (the same patient).
>
>MPRAGE was taken twice: one before the electrode implantation
>(MPRAGEbeforeGrid) and the other taken after the subdural electrodes
>(MPRAGEwithGrid) were implanted (i.e., grid is on the surface of the
>brain). Usually, the brain get somehow distored locally by the grid.
>Electrodes were made of platinum and detected as a round signal void.
>
>We used FLIRT, and then FNIRT to do so. It was successful in about
>2/3 of the patients.
>MPRAGE (both one before and with grid implantation) has 1x1x1 resolution.
>
>We used the T1_2_MNI152_2mm.conf except
>--in=MPRAGEwithGrid.nii.gz
>--ref=MPRAGEbeforeGrid.nii.gz
>--inmask=skull stripped MPRAGEwithGrid.nii.gz
>--refmask=bet'ed MPRAGEbeforeGrid
>
>As for --refmask, we used the bet'ed MPRAGEbeforeGrid.
>As for --inmask , we bet'ed MPRAGEwithGrid, and then manually
>trimmed the edge of the surface (e.g. got rid of the electrode artifact).
>Since there were some high intensities within the cortex around the
>grid (probably due to the grid), we also thresholded out these high
>intensity spots (by fslmatsh -thr (above 2SD intensity for
>example)). We did this because there was a tendency of misalignment
>with FNIRT in the area with this kind of high intensity on MPRAGE.
>
>FLIRT did actually a good work (for the starting point of FNIRT),
>but FNIRT did locally enlarge/warp the lateral parietal area more
>than we expected (cortex enlarged out of the lateral parietal
>surface of MPRAGEbeforeGrid).
>
>Any suggestions to improve coregistration?
>
>Attached, please find MPRAGEwithGrid (jpeg).
>If allowed I could upload/attach other images (FLIRTed, FNIRTed
>images, MRIbeforeGrid).
>
>Many thanks in advance.
>
>Riki
>