Self-organization of neuronal networks

State-dependent neural dynamics

In this project we are interested in investigating the dynamical state of neural activity during different behavioral states (e.g. attention state) or cognitive processes (e.g. memory consolidation) and across different modalities (spikes, LFP, BOLD). We analyze recorded neural activity and use computational models to suggest possible mechanistic descriptions for the state transitions in cortical dynamics.

People involved


Sub-project. Modulation of neural activity timescales during spatial attention

We develop simple network models that explain how timescales of neural activity can arise from the underlying spatial network structure. Moreover, by analyzing electrophysiological recordings from non-human primates during an attention task we show that top down attention can modulate these temporal dynamics.

Selected presentations and publications:

  • Zeraati, R.; Shi, Y.; Gieselmann, M.; Steinmetz, N.; Moore, T.; Thiele, A.; Engel, T.; Levina, A.: Timescales of ongoing activity reflect local connectivity and are modulated during attention. Computational and Systems Neuroscience Meeting (COSYNE 2020) , Denver, CO, USA (2020)
  • Zeraati, R.; Steinmetz, N.; Moore, T.; Engel, T.; Levina, A.: Signatures of network structure in timescales of spontaneous activity. 28th Annual Computational Neuroscience Meeting (CNS*2019), Barcelona, Spain. Featured oral