[Brains-users] brain mask - problems with automatic definition
Ronald Pierson
ronald-pierson at uiowa.edu
Tue Feb 20 06:44:42 PST 2007
The images are actually from a 3T scanner, but it is an average of 4 1-NEX
scans, with a resampled resolution of 0.5 mm. So, it isn't a standard scan.
I like the raw images from the 3 T scanners, but they have a fair amount of
inhomogeneity in them, and require some bias field correction to be worked
up well. At the moment we do not like how our tissue classification handles
3T scans, there are too many cases with poor classifications that need a
little assistance. I think it is primarily driven by the inhomogeneity, but
we need more time to figure that out.
Some of our researchers are using 3T structural sequences, but I don't have
the details. Does anyone want to chime in with their opinions?
Oh, and right now the ANN training is not possible with the current version
of brains2. The training software ran on the SGIs, and I am not sure it has
been upgraded. We are putting the time into a new ANN package that will be
pretty straightforward to train, as well as more accurate.
Ron
-----Original Message-----
From: Christoph Christmann [mailto:christoph.christmann at zi-mannheim.de]
Sent: Tuesday, February 20, 2007 1:28 AM
To: Ronald Pierson; eugene-zeien at uiowa.edu
Cc: 'Brains-users at www.psychiatry.uiowa.edu'
Subject: RE: [Brains-users] brain mask - problems with automatic definition
This looks pretty different from our AC-PC lines, we will try it!
Since I noticed the high quality of your pictures in respect of GM-WM
contrast I have a further question regarding the measurements. Until
recently we were gathering all data on an old 1.5T machine. Since we
now have a new 3T device we think about setting up a new protocol.
What could you recommend regarding sequences and parameters? We
actually could obtain T1 (FLASH or MPRAGE), T2, PD and a dark fluid
sequence. Moreover, we wonder if it could be preferable to use an
anisotropy map from a DTI sequence instead of a PD.
Christoph
At Monday 19.02.2007 22:39, Ronald Pierson wrote:
>Here is what I think is pretty normal for our resampling. AC is picked
just
>above the AC as it crosses the midline. This brain has the commisure going
>superiorly on the left side, but we just look at it on the midline.
>
>PC is one we argue about...I think the most anterior part is just a little
>"flip" of tissue that doesn't extend left or right. If you look carefully
>for the band of white matter that crosses the midline, I believe that most
>of the time you will find it just posterior to the flip. However, there is
>an atlas that shows it as centered in the "flip", and many people go by
>that. Bottom line, for most scans, the resolution makes it hard to see, so
>it is important to have a rule that is very consistently usable, regardless
>of image quality.
>
>As a quick visual check, I use the rule that the ACPC line, seen on the
>sagittal view, will have the corpus calosum resting on it anteriorly, and
>posteriorly it will just clip off a small piece of the occipital lobe.
>
>
> Ron
>
>
>-----Original Message-----
>From: Christoph Christmann [mailto:christoph.christmann at zi-mannheim.de]
>Sent: Monday, February 19, 2007 10:41 AM
>To: Ronald Pierson; eugene-zeien at uiowa.edu
>Cc: 'Brains-users at www.psychiatry.uiowa.edu'
>Subject: RE: [Brains-users] brain mask - problems with automatic definition
>
>Oops,
>
>this could hit the nail on the head: we are actually setting the AC
>line in the middle of the commisure while setting the PC line in the
>middle or even superior (since it is not alway quite clear where it
>ends in inferior direction). This is our appraoch coming from the
>Talaraich transformation of the BrainVoyager world, i.e. fMRI. So we
>will definitely test to set the AC and PC points according to your
>standards.
>
>Could you be so kind to provide us some significant pics, for example
>a sagittal and a related axial view of AC and PC in the usual
>Iowanian brains? ;)
>
>Second we would like to know if there was a way to train the net
>anyway since this we hope to still better the results. Correcting the
>brains mask is a very boring work for our students.
>
>Thanks a lot for your support
>
>Christoph
>
>
>
>
>At Monday 19.02.2007 17:31, Ronald Pierson wrote:
> >Just thought I'd chime in - wanted to earlier, but have been too busy.
> >
> >I think one of the main problems we run into is when the brain is in an
> >orientation that places part of the brain outside of where our
probability
> >map says it should be. Take a look at the sagittal window on the
>attachment
> >that Eugene sent last week. Then, take a look at the midsagittal slice
of
> >your typical scan. Now, I can't tell for sure from your snapshots, but
it
> >would appear to me that your brain is tilted back a fair amount compared
to
> >our standard, average orientation. Is that the case? If so, then the
> >superior frontal lobe will be higher than normal, and the cerebellum
lower
> >than normal inside the Talairach bounds, pushing them outside of the
>regions
> >where the ANN expects to find the brain.
> >
> >Is this a consistent problem? If so, maybe you are picking the AC and PC
> >different from us. We pick the AC point to be just above the AC, and the
>PC
> >point just below it. If you picked both of them in their centers, the
head
> >would be rotated back further than our "typical" orientation.
> >
> >We are currently developing new methods to remove the dependence on
> >Talairach bounds and to better handle the nonstandard brains and
> >orientations better. Hopefully in the next 6 months or so we will
release
>a
> >method for a good no-trim brains on >90% of the cases.
> >
> >Ron Pierson
> >
> >
> >
> >-----Original Message-----
> >From: brains-users-bounces at psychiatry.uiowa.edu
> >[mailto:brains-users-bounces at psychiatry.uiowa.edu] On Behalf Of Eugene
>Zeien
> >Sent: Monday, February 19, 2007 10:17 AM
> >To: Christoph Christmann
> >Cc: Brains-users at www.psychiatry.uiowa.edu
> >Subject: Re: [Brains-users] brain mask - problems with automatic
definition
> >
> >I have, in fact, had very similar problems. "standard workup"
> >is not a part of my daily routine, but I needed to generate a
> >brain mask for a project. Rather than ask a more experienced
> >person to create one, I decided "I can do this". Then I
> >proceeded to do just about everything wrong. :-) One by one,
> >I found the problems and fixed them. Eventually (about 1 day)
> >I had the neural net making pretty brains.
> >
> >The basic issue I see with your mask is:
> >The neural net is not finding the expected anatomical boundaries. So,
> >the end result is merely the probability
> >mask where more than 50% of the subjects used to train the
> >neural net had a region identified as brain.
> >
> >Why? Well, there can be several causes.
> > incorrect talairach bounds
> > over-inclusive blood traces which lead to
> > bad segmentation
> > incorrect image specified for the neural net's use
> >
> >There may be other ways to confuse the neural net, but these
> >are the things I had done. Of those problems, you've checked
> >three. The only one left is if the segmented image has CSF
> >classified as blood. So, load up the image, then
> > Viewers->Tracker
> >and set the crosshairs in brains2 to somewhere in the
> >ventricles. The values (poke around several places) should
> >be around 30. If you see the values showing up as 0 or 1, then
> >something has gone wrong in the segmentation. The values
> >outside the brain/head should be 0. You should only find 1's
> >where major blood vessels exist.
> >
> >On Mon, 2007-02-19 at 10:15 +0100, Christoph Christmann wrote:
> > > hi gene,
> > >
> > > we always used proper talairach bounds. here, we just added the image
> > > to some former work where we used the 'talairach box' lines for
> > > different orientation purposes.
> > >
> > > are there any further hints regarding the brain mask? dont you have
> > > similar problems?
> > >
> > > christoph
> > >
> > > At Friday 16.02.2007 16:52, you wrote:
> > > >Christoph,
> > > > Are the red lines on the images you sent the Talairach bounds?
> > > >If so, they are nowhere near the correct location. Try re-defining
> > > >the talairach boundaries, then run the neural net again.
> > > > I've attached a picture showing proper talairach boundaries.
> > > >The 2 points near the center are supposed to be the anterior and
> > > >posterior commissures.
> > > >gene
> > > >
> > > >On Fri, 2007-02-16 at 12:35 +0100, Christoph Christmann wrote:
> > > > > hi,
> > > > >
> > > > > the segmented image looks reasonable. I attach some pictures
> > > > > elucidating the actual situation. I am especially surprised that
the
> > > > > cerebellum border is not found correctly although it seems to be
> > > > > obvious as you can see in the brain_mask_cor_seg.jpg file. You can
> > > > > see in the other two images, that T1 and T1 is realigned quite
well
> > > > > (cost function < 0.45).
> > > > >
> > > > > If you would have any idea to use other programs to get the brain
> > > > > mask I would try this way, too. Actually, we started with
BrainSuite
> > > > > but found it to be less precise, too - although in other regions.
> > > > > Maybe we end up with an SPM or FSL segmentation for the mask.
> > > > >
> > > > > Best regards
> > > > >
> > > > > Christoph
> > > >
> > > >
> > >
> >_______________________________________________
> >Brains-users mailing list
> >Brains-users at www.psychiatry.uiowa.edu
> >http://www.psychiatry.uiowa.edu/mailman/listinfo/brains-users
>
>--
>Christoph Christmann christoph.christmann at zi-mannheim.de
>Psychologist, MEE Fon/Fax +49 621 1703 63 18/05
>Central Institute of Mental Health http://www.zi-mannheim.de
>Department of Clinical D-68159 Mannheim J5
>and Cognitive Psychology
>
>
>
>
--
Christoph Christmann christoph.christmann at zi-mannheim.de
Psychologist, MEE Fon/Fax +49 621 1703 63 18/05
Central Institute of Mental Health http://www.zi-mannheim.de
Department of Clinical D-68159 Mannheim J5
and Cognitive Psychology
More information about the Brains-users
mailing list