Thanks for the reply. The children are 4-12, I have been using the first_flirt script when I do linear registrations, and there is not a significant amount of neck in the image. On a whole registration with either first_flirt or FNIRT seem quite close, apart from the medial border of the caudate. I have been trying to play with different mask arrangements to alter the weight of this region, do you have any suggestions for how to go about this? Also, is there a way to generate new bvars and vtk files from an edited segmentation (such as cleaning up the image using the mask tool)?
From: FSL - FMRIB's Software Library [[log in to unmask]] On Behalf Of Brian Patenaude [[log in to unmask]]
Sent: Thursday, April 01, 2010 7:03 PM
To: [log in to unmask]
Subject: Re: [FSL] Registration to MNI in Children
What are the ages of the children? A good portion of the training data are
children starting from the age of 4. When you say you've used FLIRT, have
you used the first_flirt script? If not, were you registering with or
without skull? If so, what is it the first stage that went wrong? Is there
a lot of neck in the image?
> I have had trouble getting close registration of images from child
> subjects to the MNI152 standard brain. Particularly, I am having issues
> with the caudate and lateral ventricles. It seems the ventricle in the
> standard brain is significantly larger than those in our child subjects
> leading to significant misregistration and poor segmentation during FIRST.
> I have tried to use -inweight with FLIRT to increase the weighting of this
> region and FNIRT without success. Any suggestions or advice would be
> greatly appreciated. Also, I am curious if there is capability to generate
> my own standard brain and associated FIRST models?
> Rob North