Optimal State for Computations in the Brain

The human brain is an amazing computational device. It can perceive, process, and store great amounts of information using its 80 billion neurons, each interacting with thousands of other neurons. Only recently, experimental techniques have been developed for sampling the activity of hundreds of neurons in parallel, offering unprecedented access to the collective dynamics. To make use of the data and to understand the capabilities of a system of such complexity, theoretical understanding is essential. Statistical physics provides unique tools to deal with systems of many interacting units. In my talk, I will demonstrate some ways of how to use them to characterize the collective dynamics of spiking neural networks, and investigate the interplay between dynamics, topology, and information processing capabilities of the network.

Dr Anna Martius (née Levina) studied Mathematics at St Petersburg State University. In 2004, she relocated to the University of Göttingen where she completed her doctorate in 2008. Her first postdoctoral position was at Göttingen’s Max Planck Institute for Dynamics and Self-Organisation from where she transferred to the Max Planck Institute for Mathematics in the Sciences in Leipzig in 2011. From 2015 to 2017, Anna Martius was a Fellow at the Institute of Science and Technology in Klosterneuburg, Austria. Since April 2017, she is a junior research group leader at the Tübingen University. Anna Martius is a recipient of Sofja Kovalevskaja Award 2017.