 |
 |
Hi, In my experience you should be fine with this level of accuracy for registration purposes. However, the best test is just to run the registration and check the output of the registration. If you are not sure whether the registration could get better, then it might be worth hand-editing one example brain extraction and then look at the difference in the quality of the registration with or without the hand editing. Hopefully there will be negligible difference, in which case you are fine with the current brain extraction. If the registration is noticeably different then this would indicate that you should look into improving the brain extraction (not necessarily hand editing, as there are quite a few options for tweaking brain extraction). I hope this helps. All the best, Mark > On 25 Jan 2017, at 14:54, SUBSCRIBE FSL Anonymous <[log in to unmask]> wrote: > > Hello all, > > I'd like to know how big of an issue it is to leave non-brain tissue in the brain extracted images. It seems that, if I change the parameters so as to not miss any voxels, I will always end up with some bits of skull and non-brain tissue. All messages say that registration is pretty robust to non-brain tissue, but the question is <how> robust? I've provided an example image. Is it possible that I'll have troubles for registration when using this image? Will this affect the registration in any way? > > Any help is greatly appreciated. Thanks. > > PS: I apologize if this message was posted twice as I wasn't sure if it got accepted the first time and I wasn't able to locate it. > <Non_brain.png>