Katharina Wilmes, BCCN Berlin / GRK 1589 / ITB / HU Berlin

(Dis-)inhibitory gating of excitatory synaptic plasticity: a cellular modelling approach

The neuronal correlate of learning is thought to be the experience-dependent adjustment of neuronal connections - synaptic plasticity. However, cellular processes mediating these changes are highly regulated, and can e.g. be influenced by the state of the organism. Limiting learning to behaviorally relevant episodes is for instance useful if new experiences can overwrite old memories. In this thesis, we use computational modeling to explore a mechanism by which cellular processes for learning (happening in the principal neurons of the brain) can be modulated by another cell type: local inhibitory neurons. Although these cells are known to play a role for learning, the cellular mechanisms by which they could influence synaptic plasticity are not completely understood. The aim of this thesis is hence to shed light onto the cellular mechanisms underlying the regulation of synaptic plasticity.

In the first part of this thesis, we show that inhibitory neurons can modulate signals for plasticity in the dendrites (input structures) of principal neurons in an all-or-none manner. Thereby, inhibition can provide a binary switch for plasticity, which - as we further demonstrate - can be specific for inputs arriving via different neural pathways.

An important dendritic signal for synaptic plasticity is the backpropagating action potential; the neuron fires an action potential that travels along the axon to the next neuron and additionally travels backwards into the dendrite (hence backpropagating action potential). This backward-directed signal informs synapses about the activity of the neuron and can be modulated by inhibition. We show that the timing requirement for inhibition of backpropagating action potentials is tight; especially if modulation of plasticity via this mechanism ought to preserve forward-directed stimulus processing from the synapses to the axon in the same neuron. Yet, we demonstrate that the desired timing can be accomplished if inhibition is embedded in a common neuronal network motif: an inhibitory feedforward circuit.

In the second part of this thesis, we address the question whether and how appropriately timed inhibitory feedforward circuits can be established in the brain. We propose that spike-timing dependent plasticity (plasticity that depends on the relative timing of activity in the connected neurons) at inhibitory synapses can shape microcircuits to become properly adjusted to the individual timing requirements of the modulated principal neuron. For this purpose, we propose particular inhibitory plasticity rules and demonstrate their functioning.

Organized by

Susanne Schreiber / Robert Martin