Modeling calcium-dependent synaptic plasticity and its role in sleep-dependent memory consolidation

Abstract

It has been shown that a single neuron can encounter different firing rates during the sleep and the awake states. Those rhythms directly have an impact on the synaptic weight between the neurons. Moreover, recent evidence shows that spindle oscillations encountered during sleep influence the calcium levels in the post-synaptic spine that trigger synaptic plasticity changes. There exists a large number of synaptic plasticity rules. In particular, this thesis focuses on calcium-induced synaptic plasticity. However, the little number of calcium-based models do not take into account the calcium dynamics in much detail. Indeed, to reproduce protocols and obtain results that are consistent with experimental data, a great number of simplifications are often considered. A review of the existing calcium-based models is made in order to categorize those models in a systematic way: ‘How do they implement the calcium flow into the neuron?’, ‘What is the equation governing synaptic plasticity depending on the calcium concentration?’, etc. The thesis focuses on the calcium-dependent synaptic plasticity model developed by Graupner et al. (2016). This model has made simplifications to implement the calcium dynamics while being consistent with data obtained experimentally. The contribution of this thesis is first to integrate this abstract model into a conductance-based model which allows switching from a tonic pattern to a bursting pattern, encountered during the switch to the sleep state. This allows observing what are the consequences of this switch on the calcium-dependent synaptic plasticity. The second main contribution of the thesis is to integrate a more detailed calcium dynamics into the abstract calcium dynamics model from Graupner et al. (2016). The key message is the fact that integrating a detailed calcium dynamics into an abstract one represents a major challenge to tackle because of the large number of assumptions that have been made to construct this abstract model. This leads to the prospect that starting from a more physiological calcium dynamics then integrating a calcium-dependent synaptic plasticity rule to this model may be a more suitable way of doing.