Advancing the Characterization of Neuronal Cyto-Architecture by X-ray Phase-Contrast Tomography

To bring physiology and pathology of the human brain into better micro-anatomical and histological context, studies with different methodologies are required. Established techniques such as electron microscopy or histology show limitations in view of invasiveness, labor-intense and artifact-prone sample preparation, as well as an adequate ratio between resolution and volume throughput. For this reason, X-ray phase-contrast tomography (PC-CT) has been proposed as a three-dimensional non-destructive imaging technique, which requires less effort in sample preparation and can assess larger volumes. Furthermore, it offers quantitative electron density based contrast even for unstained tissue. Up to now, however, PC-CT studies fell short in number of samples, so that structural alterations caused by neurodegenerative diseases cannot be distinguished from physiological inter-subject variations. In this thesis, the scalability of PC-CT with respect to the required number of samples and resolution-to-volume-throughput is demonstrated, and the methodology is advanced with respect to data acquisition, processing and segmentation. In addition to the human cerebellum, cortex and hippocampus are studied. Concerning quantification and analysis of PC-CT data, this work introduces optimal transport analysis to obtain quantitative metrics of the cyto-architecture and to identify changes due to neurodegenerative diseases. For the case of Alzheimer’s disease, this workflow reveals a yet undescribed compactification of granular cells in the human hippocampus. This thesis also provides optimized configurations to study neural tissues with laboratory instrumentation, and – finally – provides new correlative imaging approaches, in particular with scanning electron microscopy.