Starting for my master thesis, I explored the effects of network topology on dynamic properties of neural networks. I focused on two main dynamical features, synchronization properties and self-sustained activity.

First, in collaboration with researchers at the IFISC, we explored the role of network topology and delayed connections on the synchronization properties of neural networks, more details can be found here, and in my master thesis.

Local and global synchronization properties of neural networks in the coupling strength-delay time phase space.

Afterward, I investigated how network architecture shapes activation patterns on graphs, and studied the contribution of short and long cycles to self-sustained activity. These projects were done in collaboration with Annick Lesne, Marc-Thorsten Hütt and Claus C. Hilgetag, with funding from the DFG.

Left. Graphical representation of a network, its adjacency matrix, and co-activation patterns. Right. Cycle usage and average activity as a function of time steps.

All my publications on this topic: