Bereiche | Tage | Auswahl | Suche | Downloads | Hilfe

BP: Fachverband Biologische Physik

BP 14: Neuronal and Sensory Systems

BP 14.6: Vortrag

Mittwoch, 25. März 2009, 15:30–15:45, HÜL 186

Extensive Chaotic Dynamics of Spiking Neuron Networks in the Balanced State — Michael Kreissl¹, •Siegrid Löwel², and Fred Wolf¹ — ¹Max Planck Institute for Dynamics and Self-Organization and BCCN in Göttingen, Germany — ²Friedrich Schiller University and BGCN in Jena, Germany

Based on the calculation of the spectrum of Lyapunov exponents we reveal extensive, spatiotemporal chaos in deterministic neural networks of canonical type I neurons in the balanced state. In the balanced state of cortical networks, neurons are driven by strongly fluctuating inputs that result from balanced recurrent inhibition and excitation. It is the prevailing explanation of asynchronous, irregular firing patterns often observed in vivo. While its robust emergence from the collective dynamics of spiking neuron networks has been shown in several theoretical studies, the precise nature of the network dynamics remains controversial. It depends strongly on the single neuron dynamics. Initially, using binary neurons, Vreeswijk and Sompolinsky found that nearby trajectories diverge faster than exponential. Contrary, using leaky integrate and fire neurons, Zillmer et. al and Jahnke et. al recently showed that nearby trajectories converge. In our study of sparse networks of theta neurons we find conventional chaos with a fat attractor and high entropy production. Because theta neurons exhibit the same type of bifurcation from resting to spiking as real cortical neurons, we expect that this extensive chaotic dynamics is characteristic of the balanced state in biophysically realistic network models.