Meeting Banner
Abstract #4439

Reliable Amygdala Segmentation Using Clustering of Multimodal Data at 7 Tesla.

Eugenia Solano-Castiella1, Gabriele Lohmann1, Andreas Schfer1, Robert Trampel1, Robert Turner1

1Neurophysics, Max Planck Institute, Leipzig, Sachsen, Germany


The amygdala is a relatively small brain structure. Its location next to the sphenoidal sinus results in a variety of MR image artifacts, which increase with field strength. However, brain scanning at 7T provides considerable improvement of SNR and CNR having the potential for in vivo parcellation of amygdala images into neurologically significant subdivisions that may improve interpretation of fMRI-based neuropsychological studies. To improve clustering we used the strategy of combining images (GRE and TSE) having different contrast mechanisms. The analysis and clustering techniques we have developed for this purpose may assist parcellation of other deep brain structures.