link to homepage

Institute of Neuroscience and Medicine

Navigation and service

Talk by Jürgen Jost

Max-Planck-Institute for Mathematics, University of Leipzig

05 Feb 2018 15:15
05 Feb 2018
Bldg. 15.9 U, Seminar Room 4001b

Information decompositions

When two or more sources influence a target, like sensory input and context information from other neurons come together at some neuron, the question emerges how to disentangle the respective contributions. Which part of the information is unique to the individual sources, what is shared between them and what emerges only from their combination? The field of information decomposition was started a few years ago. It offers more refined tools of analysis than traditional concepts like transfer entropy. I shall describe the formal concepts and the current state of the theory and hope for discussions with neuroscientists about potential applications.


Dr. Moritz Helias