Benjamin Lindner (Humboldt University Berlin and BCCN Berlin) will speak about
"Challenges in the study of neural networks: slow fluctuations, detection of single-cell stimuli, and signal transmission in the presence of up/down fluctuations."
To understand the spontaneous activity and signal-transmission properties of brain networks, theoreticians often study simplified networks of randomly coupled integrate-and-fire neurons. In my talk I discuss the surprisingly rich behavior of these networks and some applications for signal transmission and detection problems. Specifically, I explain how networks of deterministic neurons with fast dynamics can generate very slow asynchronous spontaneous fluctuations; I show under which conditions the strong stimulation of a single neuron can lead to weak but detectable changes at the population level; I demonstrate that up/down transitions may be beneficial for the signal flow between cortical populations.