Approche électrique de la transmission d'influx nerveux à partir des systèmes de couplage Thermo-photosensible

Abstract

The human brain is an organ made up of thousands of neurons interconnected via synapses. It is the source of our thoughts, emotions, perceptions, actions and memories. Advances in neuroscience over the past few years have given us a comprehensive understanding of brain activity. Thus, recent work has shown that biophysical effects such as light, temperature and sound, to name a few, can influence neuronal dynamics. In this thesis, we propose and analyze models of thermo-photosensitive FitzHugh- Nagumo neurons that take into account the effect of light and temperature. Considering these effects acting simultaneously on neuronal dynamics, the main properties of this model are carefully analyzed through analytical, numerical and experimental techniques. It emerges that the temperature variation modifies the number and the nature of the equilibrium points of the neuronal system and induces a bifurcation of Hopf showing that the neuron can pass from a state of rest to an oscillatory state and vice versa. Moreover, several important electrical activities of the neuron such as rest, doping, burst, chaos and other complex phenomena can be observed when there is variation of light and temperature. In addition, the effect of synaptic couplings between two thermo-photosensitive neurons have also been the sub ject of several analyzes and important results have been noted. By considering a synapse at the Josephson Junction to couple two models of thermo-photosensitive neurons, we were able to demonstrate that these neurons can undergo complete synchronization and phase synchronization when the parame- ters of the coupling pathway are well defined. Subsequently, by coupling a photosensitive neuron and a heat-sensitive neuron through a memristor synapse, we were able to identify fascinating phenomena such as the extremes of homogeneous and heterogeneous multistabilities when the internal state variable of the memristor changes periodically. In addition to these exciting results, we noticed that it is very difficult for two different coupled neurons to achieve full synchronization. On the other hand, phase syn- chronization can occur when the channel parameters are activated. Therefore, increasing the coupling strength in this last model achieves the phase synchronization regime between the two neurons. Then, in order to control the phenomenon of multistability, we introduced into this neural model a feedback term depending on one of the dynamic state variables of the neurons. We show that it is possible for a neuron model to go from a multistable state to a monostable state. The analog and experimental results obtained provide proof of the reliability and accuracy of the proposed models.