SKM 2023 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

DY: Fachverband Dynamik und Statistische Physik

DY 45: Poster: Nonlinear Dynamics, Pattern Formation and Networks

DY 45.16: Poster

Donnerstag, 30. März 2023, 13:00–16:00, P1

Homeostatic plasticity in a minimal model for brain criticality — •Marco Schmidt and Stefan Bornholdt — Institut für Theoretische Physik, Universität Bremen

The ’criticality hypothesis’ asserts that real-world neural networks operate near a critical phase transition. Experimental evidence exists and numerous models studying the possible underlying mechanisms accumulated during the last 20 years.

Early models based on simple threshold networks tune to a critical connectivity K=2, which is not a realistic value when compared to real-world neural networks.

However, a phase transition in high degree threshold networks using the inhibition to excitation ratio as a control parameter does exist [1], as well as a corresponding self-organized critical toy model [2]. It features an adaptive threshold network, self-tuning to the critical inhibition to excitation ratio by using an activity based rewiring process that results in a highly clustered network and reaches criticality independent of K.

Here we present a new version of the model, incorporating a simple homoestatic plasticity mechanism as it appears in biological systems.

[1] L. Baumgarten, S. Bornholdt, Critical excitation-inhibition balance in dense neural networks, Phys. Rev. E 100, 010301 (2019).

[2] L. Baumgarten, S. Bornholdt, A toy model for brain criticality: self-organized excitation/inhibition ratio and the role of network clustering, arXiv:2202.03330.