This project is part of the Sinergia project "Neural Circuit Reconstruction".

Understanding structure-function relationships is one of the most successful strategies in biology that in the context of neuroscience requires exhaustive measurements of connectivity between neurons to extract the topology of synaptic interactions in neuronal circuits. This goal is considered "one of the holy grails of Neuroscience" since it would pave the way for a systematic analysis of neuronal circuit function based on computational modeling. The recent development of 3-D electron microscopy (3DEM) techniques allows the ultrastructural analysis of biological tissues throughout large volumes (up to a few hundred micrometers in each dimension). Although 3DEM has an enormous potential for advancing our understanding of brain function in health and disease, its biomedical application is currently impeded by the lack of accessory technologies, particularly informatics tools for the automated analysis of large and complex sets of image data.

The goal of this project is to develop methods for automatic and semiautomatic segmentation of neural structure in 3DEM data. Serial block-face scanning electron microscopy (SBFSEM) and focused ion beam scanning electron microscopy (FIB-SEM) data of the olfactory bulb of zebrafish larvae is provided by our project partner.
The semiautomatic segmentation should be interactive and be included in a tool that assists the biologist when tracing and annotating neural structures. On training data gained from (semiautomatic) segmentation a statistical model will be developed that incorporates local and global information. This model will be used for the automatic segmentation of neural structures.