Electronic Thesis and Dissertation Repository


Master of Science




Dr. Stefan Köhler


Probing the functions of human hippocampal subfields is a promising area of research in cognitive neuroscience. However, defining subfield borders in Magnetic Resonance Imaging (MRI) is challenging. Here, we present a user-guided, semi-automated protocol for segmenting hippocampal subfields on T2-weighted images obtained with 7-Tesla MRI. The protocol takes advantage of extant knowledge about regularities in hippocampal morphology and ontogeny that have not been systematically considered in prior related work. An image feature known as the hippocampal ‘dark band’ facilitates tracking of subfield continuities, allowing for unfolding and segmentation of convoluted hippocampal tissue. Initial results suggest that this protocol offers sufficient precision and flexibility to accommodate inter-individual differences in morphology and produces segmentations that have improved accuracy and detail compared to other prominent protocols, with similar inter-rater reliability. We anticipate that this protocol will allow for improved anatomical precision in future research on hippocampal subfields in health and neurological disease.