Bereiche | Tage | Auswahl | Suche | Downloads | Hilfe
DY: Dynamik und Statistische Physik
DY 43: Signals and neuronal Networks
DY 43.1: Vortrag
Donnerstag, 30. März 2006, 11:30–11:45, SCH 251
Precise Timing in Strongly Heterogeneous Neural Networks with Delay — •Raoul-Martin Memmesheimer1,2 and Marc Timme1,2,3 — 1Max-Planck-Institut für Dynamik und Selbstorganisation (MPIDS) Göttingen — 2Bernstein Center for Computational Neuroscience BCCN Göttingen — 3Center for Applied Mathematics, Cornell University, Ithaca, USA
Precise timing of spikes is discussed to be a key element of neural
computation [1], but it is still an open question how patterns of precisely
timed spikes emerge in the dynamics of neural networks [2]. Here we
demonstrate that and how deterministic neural networks which simultaneously
exhibit delayed interactions [3], complex topology [4] and strong
heterogeneities can yet display periodic patterns of spikes that are precisely
timed. We develop an analytical method to design networks that display a given
non-degenerate pattern with
realistic temporal extent and complicated temporal structure. We
point out that the same pattern can exist in very different networks; its
stability depends on the particular coupling architecture. Using a nonlinear
stability analysis, we show that networks with purely inhibitory (or purely
excitatory) coupling can either store only stable or only unstable patterns.
M. Abeles, Science 304:523 (2004).
I.J. Matus Bloch, C. Romero Z., Phys. Rev. E 66:036127 (2002);
D.Z. Jin, Phys. Rev. Lett. 89:208102 (2002); M. Denker et al., Phys. Rev. Lett. 92:074103
(2004).
U. Ernst, K. Pawelzik, T. Geisel, Phys. Rev.
Lett. 74:1570 (1995).
M. Timme, F. Wolf, T. Geisel,
Phys. Rev. Lett. 89:258701 (2002).