Christian Hirsch (RUG)

Modeling synaptic plasticity through dynamically reinforced random networks

Graph-based Pólya urns are a promising approach to model random processes on networks that exhibit self-reinforcing phenomena such as the ones inspired by synaptic plasticity in neuroscience. As in classic Pólya urns, the network-based analogues behave very differently in regimes of weak and strong reinforcement. At weak reinforcement, the effect of reinforcement disappears in the long-term limit. At strong reinforcement, we observe clear localization phenomena. Moreover, we present a hierarchical model with asymmetrical reinforcement, which has a small-world characteristic. The talk is based on joint work with Y. Couzinié, M. Heydenreich, M. Holmes and V. Kleptsyn.

VIDEOSTREAM