Building brain atlases and segmentation tools with ex vivo MRI and histology: application to hippocampus, amygdala and beyond

Abstract: Ex vivo modalities make it possible to obtain manual delineations of brain anatomy with very high precision. These delineations can be used to build accurate atlases at very high resolution, which can in turn be applied to the segmentation of brain substructures in in vivo MRI scans. While analyzing the brain at the substructure level holds promise to increase our understanding of such a complex organ (both in health and in disease), it also poses a number of modeling challenges. In this talk, I will first describe the construction of a probabilistic atlas of the hippocampal subfields using a combination of in vivo and ex vivo MRI. Then, I will present some results on in vivo MRI segmented with the atlas. Next, I will present an extension of the atlas to the amygdaloid nuclei. And finally, I will discuss the extension of the framework to the thalamus and other brain structures using histology – a modality that yields excellent contrast and resolution, but also introduces additional difficulties.

Bio: Juan Eugenio Iglesias holds an MSc in Telecom Engineering from the University of Seville (Spain), an MSc in Electrical Engineering from the Royal Institute of Technology (Stockholm, Sweden), and a PhD in Biomedical Engineering from the University of California, Los Angeles (USA). He has also been a postdoctoral researcher at the Martinos Center for Biomedical Imaging (Massachusetts General Hospital / Harvard Medical School), and is currently a Marie Curie postdoctoral fellow at the Basque Center on Cognition, Brain and Language (San Sebastian, Spain). His main research interest is exploring generative models for brain MRI segmentation.