Mark,
Thanks for the reply.
I have previously tried to FNIRT masks from FIRST FLIRT subcortical registration to a modal mask of all the subjects. This yielded terrible results, probably because there were no features besides the image boundary to align to. I then tried to FNIRT the hippocampus itself and this still yielded terrible results. I am sure I am probably not using the optimal parameters.
I wasn't sure how to FNIRT to the standard image. Would I just segment that the standard image and then use the FIRST FLIRT results to guide the FNIRT segmentation?
I thought I had mentioned every thing I did, but no, you are right, I was using a 1mm structural produced by FIRST. I guess this WOULD make the analysis take longer, I just didn't think it would take 10 times as long.
I will try your suggestions.
Your help is greatly appreciated,
Carlos Faraco
|