Through optical measurements with fluorescent dyes, we can observe brain activity in larger populations of neurons or in entire brain regions. With the advances in neurobiological measurement technology, analyzing the data has become equally challenging and time-consuming as recording them. For example, individual neurons or other objects have to be detected automatically and reliably in movies of brain activity. For neighboring or overlapping objects, the optical recording technique can cause signal mixtures that need to be unmixed with the help of signal processing methods.
We develop methods for processing and analyzing optical recordings of brain activity. This is a highly diverse data source with a large number of subdomains defined by animal, brain region, microscopical technique, staining dye and experimental setting, while the sample size for each subdomain is typically small, which complicates neuron detection with learning-based algorithms. The focus of our work is thus on generally applicable methods for neuronal signal extraction, as well as on the use of artificial data that can be generated in large quantities from realistic models. After neuronal signal extraction and denoising have taken place, we perform higher-level data analysis in collaboration with neurobiologists, in particular research on brain areas that are responsible for odor sensing.
New theses are regularly advertised in the area of Neuroinformatics. In addition to the general overview, there are also numerous topics that have not yet been advertised, which will be gladly presented in a personal conversation.
- Markus Rothermel, Cell Biology, University of Veterinary Medicine Hannover
- Lindy McBride, Ecology and Evolutionary Biology, Princeton University
- DFG Research Centre, “ACTIVE– Aachen Center for Biomedical Image Analysis, Visualization and Exploration”, Projektnr. 233509121
- DFG GRK 2416, “MultiSenses-MultiScales: Neue Ansätze zur Aufklärung neuronaler multisensorischer Integration”, Projektnr. 368482240