Tatjana Tchumatchenko, Max-Planck-Institute for Brain Research, Frankfurt

Shaping Network Dynamics via Individual Neurons

A defining feature of cortical circuits is their versatility. They perform multiple fundamental computations such as contrast normalization, memory storage and rhythm generation. It is currently an open question how this impressive versatility is achieved in a single circuit given that specialized models are often needed to implement each of these computation. Here, we propose that a single computational motif, the stabilized supralinear network (SSN), can support working memory, persistent activity and rhythm generation in addition to other experimentally reported features such as contrast invariance and normalization. We solved the SSN model analytically and show that it can provide a substrate for working memory by supporting two stable firing states. Furthermore, the SSN framework can sustain finite firing rates following input withdrawal and can thus serve in decision-making tasks. In addition, we show that the SSN motif can exhibit a Hopf bifurcation and lead to global oscillations. We outline the synaptic and neuronal mechanisms underlying this versatility and pave the way for a unified theory of cortical function.

Additional Information

Workshop talk; registration is necessary

Organized by

GRK 1589