Bad registration of Freesurfer segmentation of the hippocampus

(Olof Lindberg) #1

was wondering if anyone have experience with freesurfer segmentation of the hippocampus in SPHARM.

I tried various tricks but the registration really does not work. It is obvious that there are several irregularities of the shape from FS but I find it a bit strange that I cannot get the registration to work … as the overall shape is relatively similar between subjects.

So far i tried to increase or decreas spherical complexity 10 - 15 - 20 - did not work for any of these.

I also tried to use different subjects as reference for the registration (choosing one that seem to have a relatively regular shape).

Is there anything else that I can try?


(Beatriz Paniagua) #2

Hi Olof,

Sorry this is being a challenge. Do you have an example snapshot on how the segmentation looks like?

Thank you,

(Olof Lindberg) #3

So attached are some images … Highly irregular I admint - but I had hoped that the overall shape and curvature of the structure would make it possible to make an ok registration.