Servicenavigation

Detailseite

Zurück

Projekt Druckansicht

Digital neuroanatomy, data visualisation, and modelling

Fachliche Zuordnung Kognitive, systemische und Verhaltensneurobiologie
Förderung Förderung von 2005 bis 2011
Projektkennung Deutsche Forschungsgemeinschaft (DFG) - Projektnummer 18222611
 
Erstellungsjahr 2010

Zusammenfassung der Projektergebnisse

The goal of this project was to develop and improve tools that support the creation, management, and visualization of hierarchical brain atlases. The usability of the tools should be evaluated using a standard atlas of the honeybee brain. All methods should be integrated into the visualization and data analysis framework Amira that was (and is) used by our research partners in neurobiology. First, we developed tools and methods that generalize and improve the steps in the processing pipeline for creation of surface-based atlases. These steps consist of imaging and preprocessing, segmentation, registration, surface reconstruction, and assignment of biological semantics to geometrical objects. Secondly, techniques supporting the usage of hierarchical brain atlases were developed. To improve the segmentation and registration steps, a statistical shape model of the central structures of the honeybee brain was created. Based on this shape model, an automatic segmentation method for coarse structures was developed. It computes spatial displacements and deformations of statistical shape models such that they fit to individual microscopy images. The resulting information can be used also for spatial registration of the image data. Besides this method for deriving registration information, we integrated alternative registration techniques into Amira. To enable the integration of semantic information like, e.g., biological identifiers of anatomical structures or relations between such structures, in collaboration with neurobiologists we developed a first exemplary ontology of the honeybee brain. The ontology includes concepts ranging from coarse anatomical structures such as neuropils down to individual neurons of certain types. It further embodies morphological and functional relationships between structures. The structures contained in the ontology are linked to geometric entities of the standard atlas. Using this hierarchical classification scheme, we implemented methods for query-based browsing and automated creation of visualization. In order to enhance the presentation of neurons within their anatomical context, we also designed and evaluated novel visualization methods for line-like structures in complex surface geometries. Here, knowledge from perception theory was used to improve depth perception and visual differentiation of structures. A quantitative user study substantiated these developments. We also provided tools to our cooperation partners in neurobiology for analyzing atlas structures and substructures by calculating, e.g., distance measures or other important parameters. The atlas creation pipeline was used for the generation of a standard moth brain, a standard desert locust brain, and a fly thoracic ganglion. Furthermore, the synaptic distribution of several beebrain projection neurons within mushroom body parts were evaluated with our tools. The browsing and visualization methods developed in the project improved atlas exploration, data retrieval, and the detection of segmentation and registration errors. They also supported the meaningful representation of atlas data for education, presentation, and communication.

Projektbezogene Publikationen (Auswahl)

 
 

Zusatzinformationen

© 2024 DFG Kontakt / Impressum / Datenschutz

Textvergrößerung und Kontrastanpassung