Memory capacity of dynamical cortical networks

Recurrent neuronal networks are thought to serve as a physical basis for learning and memory. For example, the recurrent network in the CA3 region of hippocampus, exhibits the replay of stored sequences of previously experienced events. This replay is accompanied by the fieldpotential-phenomenon of sharp wave-ripple complexes.  In collaboration with the group of Dietmar Schmitz (Charite, Berlin) we study patch-clamp recordings from hippocampal neurons during sharp wave ripple episodes in-vitro to unveil the mechanistic basis and the underlying synaptic dynamics of this phenomenon. In a further part of this project we develop data analysis methods to study the local connectivity in brain slices by photolytic uncaging of glutamate. Currently we study connectivity in the entorhinal and piriform cortex. Goal of the project is to use both the dynamical and the structural constraints to build computational models of recurrent networks and in particular to gain understanding of the computations that are performed by the recurrent connections.

Example of an afferent map of an entorhinal cortex stellate cell showing the excitatory compound signals as a function of the stimulation point. In this example, the excitatory inputs cover the whole extent of layer II of the medial entorhinal cortex.