Hi,
It might help to supply some prior probability maps if you have any. Otherwise I'm not sure that I can help, as I don't have a lot of experience myself with this. Maybe someone else can help, although if you cannot see any intensity differences between WM and the thalamus or basal ganglia then it might not be possible to separate these with segmentation.
All the best,
Mark
On 28 Jan 2014, at 21:15, Mojdeh Zamyadi <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Hi again,
after a more careful look it seems that some gray matter structures such as thalamus and basal ganglia are being segmented as WM. I've tried 4 and 5 classes as well and still the same problem. Do you know what else I might try to get a better segmentation?
Thanks,
-Mojdeh
On Tue, Jan 28, 2014 at 3:16 AM, Mark Jenkinson <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Hi,
If you think the segmentations look fine then you have a good segmentation!
I am no expert in paediatric brains so there would be no point in me looking at the images.
If you are still unsure yourself then show it to someone with some training in paediatric brain anatomy.
As for the hard segmentation outputs, there is normally not a lot of difference and if all you want is a WM segmentation for tractography then it won't matter too much and you should be OK with the _seg version.
All the best,
Mark
On 24 Jan 2014, at 19:39, Mojdeh Zamyadi <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Thanks for the reply Mark. My subjects range from 2-16 yrs old! I need a GM/WM segmentation fro tractography (to start my tracks in WM seed only). I've used the default parameters as an initial run and the WM/GM boundary seems reasonable! Would it be OK to upload a couple of subjects so that you can also take a look and let me know if the segmentation is good.
Also, I've notices there are two hard segmentation output files: *_pveseg and *_seg ... could you tell me what's the difference? For the purpose of seeding the tractography do you think I should use the hard segmentation or the partial volume estimate thresholded at a certain threshold? if using the PVEs do you know what threshold ppl usually use for tractography?
Thanks again,
-Mojdeh
On Fri, Jan 24, 2014 at 4:02 AM, Mark Jenkinson <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Hi,
I think this will depend on the age of your subjects and the nature of the acquisition that you are taking.
Because there are often many differences in contrast due to absent and partial myelination then the settings for adult brains often don't work that well. One thing that others have tried is to increase the number of classes that you use (from the original 3 to something higher, even up to 8) in order to capture the differently myelinated tissues. So try this and see if it helps.
All the best,
Mark
On 23 Jan 2014, at 17:31, Mojdeh Zamyadi <[log in to unmask]<mailto:[log in to unmask]>> wrote:
> Hi all,
>
> I was wondering if anyone has used FAST for Grey/White matter segmentation of paediatric brains? Is there a set of parameters that works best with young brains?
>
> Thanks,
> -Mojdeh
