Dear Fernanda,
The arrows in the top panel point to areas where the brain
extraction in the highres image is imperfect. However, it has
not affected the alignment of any of the real brain structures
and so you can ignore it.
The arrows in the bottom panel point to parts of the brain
that are not covered in your functional images (they must
have limited FOV) and so these too can be ignored as
the parts of the brain that are imaged are aligned well.
You will find that in any situation where the functional
images do not cover the entire brain that you will have
these areas of blank areas. There is nothing else that
can be done as there is no image data to cover these
areas.
So, in summary, I still think you have great registration results! :)
All the best,
Mark
On 1 Sep 2011, at 22:21, Fernanda Palhano wrote:
> Hi Mark,
> I'm very happy that you think my results are good. =)
> But I'm still worried about the red lines. I'm sending the same image but now
> with arrows pointing to where I think that there is a problem.
> http://imago.neuro.ufrn.br/imagens/
>
> Maybe it's just a visualization problem. Can you please look again this image? Sorry to bother you with this, but I want
> to be sure that my registration is good enough.
>
> Thank you very very much!
> Fernanda
>
> On Thu, Sep 1, 2011 at 5:37 PM, Mark Jenkinson <[log in to unmask]> wrote:
> Dear Fernanda,
>
> Thank you for sending the links to the images.
> The results to me look very, very close to *perfect*!!
>
> What are you unhappy about?
>
> Your functional images have a limited Field Of View (FOV)
> and so do not cover some parts of the superior and
> inferior parts of the brain. This is why you see blank
> patches in the overlays and sharp, angled lines.
> However, the structures within the brain look like they
> align very, very well.
>
> So I don't think you need to do anything else.
> I think the results are great.
>
> All the best,
> Mark
>
>
>
> On 1 Sep 2011, at 18:39, Fernanda Palhano wrote:
>
> > Hi Mark,
> >
> > Thank you for your quick reply. I just realize that my attachment didn't go.
> > You can see the images accessing the follow address: http://imago.neuro.ufrn.br/imagens/
> >
> > I don't know if this have some relation with my problem but I used fslswapdim (with the option RL PA IS) to reorientate all my structural T1 images before doing the registration.
> > I never tried the Full Search option, but if I want to try it, should I use it in both registrations (Main Structural Image and Standard Space)?
> > And, unfortunatelly I don't have fieldmaps.
> >
> > I'm now running a new data set (from another scanner) and I'm having the same registrations problems. Although this, my statistical results seem fine. I was wondering if it's not only a misplace of the red lines...
> > Do you think this is possible?
> >
> > Thanks again,
> > Fernanda
> >
> > On Thu, Sep 1, 2011 at 2:08 PM, Mark Jenkinson <[log in to unmask]> wrote:
> > Dear Fernanda,
> >
> > There was no attachment (as the list does not allow anything except very,
> > very small attachments) so I couldn't really be sure of what you are seeing.
> > Do your images start in very different alignments (i.e. 90 or 180 degrees
> > different)? If so then you might find that running fslreorient2std on your
> > data might help - do this on either the functional or the structural or both -
> > depending on which are in different orientations compared to the template
> > MNI152 images. You can also try "Full Search" if you have not tried that.
> >
> > If you have fieldmaps then we strongly advise using those as they can
> > really help the functional to structural (example_func2highres) registration
> > hugely. Often these "failures" of registration relate to uncorrected distortion
> > in the functional EPI images.
> >
> > All the best,
> > Mark
> >
> >
> >
> >
> > On 1 Sep 2011, at 14:46, Fernanda Palhano wrote:
> >
> > > Hi,
> > >
> > > I used FEAT GUI, first-level analysis to run the data, but I noticed a registration problem when I checked the registration report, the alignment (in red) didn't fit, please refer to the photos attached, the problem is quite clear according to the photos. The alignment for the images on the top seems fit the image on the bottom, and the alignment lines (red lines) for the images on the bottom seems fit the images on the top.
> > > I'm registering my functional image (EPI) to the MNI template. For this, I used a T1 structural image as Main structural image, 6 DOF, and MNI152_T1_2mm_brain, linear 12 DOF, as standard space.
> > >
> > > Thanks in advance.
> > >
> > > Fernanda
> >
>
|