Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
BP: Fachverband Biologische Physik
BP 14: Posters: Neurophysics
BP 14.2: Poster
Montag, 16. März 2015, 17:30–19:30, Poster A
Activity propagation in feed-forward neuronal networks described by reaction-diffusion like equations — •Dmytro Grytskyy1, Markus Diesmann1,2, and Moritz Helias1 — 1INM6&IAS6, Juelich Research Centre and JARA, Jülich, Germany — 2Medical Faculty, RWTH Aachen, Germany
We investigate feed-forward neuronal networks of linear-nonlinear rate neurons with synaptic plasticity. For systems without plasticity or near to criticality, layers can be mapped onto states in time. We treat the K nearest neighbor coupling in diffusion approximation to obtain equations for the activity evolution from layer to layer. The equations are similar but not identical to those describing reaction-diffusion systems. We develop an appropriate solution scheme also applicable to generalized systems with several synapse or neuron types. We obtain the critical border separating ultimately decaying activity from possible activity explosion. On the border (requiring exact parameter tuning) and in its subcritical vicinity we analytically find long-living dissipative solitons with a plateau of arbitrary width with the decay velocity proportional to the distance to criticality. Two bumps can unite or remain disjoint within their lifetime depending on the distance in-between. For the united bump the same scenarios exist, so a kind of association tree can appear in this way.
Partially supported by the Helmholtz Association: HASB and portfolio theme SMHB, the Jülich Aachen Research Alliance (JARA), HGF Nachwuchsgruppe; VH-NG-1028, the Next-Generation Supercomputer Project of MEXT, and EU Grant 269921 (BrainScaleS).