Synaptic plasticity: from single cell to microcircuit

Neurons are the computational building blocks of our brains. They form complicated net- works that process the incoming sensory information, eventually leading to our actions. The connections between neurons can be altered, which is called synaptic plasticity. Synaptic plasticity is thought to underly learning and memory in the brain. In order to fully under- stand how our brain works, it is therefore crucial to understand these plasticity mechanisms. Moreover, neurons have unique morphologies with dendrites that can extend hundreds of micrometers, connecting different brain areas. In this thesis, we explored how dendrites influ- ence plasticity and how this translates to networks. In biophysically realistic neuron models, we investigated how spike-timing-dependent plasticity and dendritic-spike-dependent plas- ticity depend on the location of synapses across the dendritic tree. We then analysed the consequences of these dependencies on the connectivity between two neurons and showed how dendritic spikes can gate plasticity at other synapses. Furthermore, we explored how synaptic clusters can prolong memory retention. We then investigated this property in a net- work of simplified neurons, where we demonstrated how the learning of different associations easily depresses previously strengthened proximal synapses, but not distal synapses. Finally, we analysed how homeostasis can influence the connectivity in networks with excitatory and inhibitory neurons without dendrites. We showed how networks with similar connectivity to the one observed in primary visual cortex are more stable and can better discriminate between different activation patterns, compared to networks with a modified connectivity.Open Acces