|
G. Becq, S. Roth, S. Zozor, J. L. Mocellin, S. Gory-Faure, P. O. Amblard, C. Villard In order to study the behavior of small biological neural networks, we have developed a project including both experiments and simulations. In this study we present observations and modelisations of three neurons networks and the influence of electrochemical noise on some parts of the system. For the experimental part, neurons from mice hippocampus are grown in vitro on multi-electrodes arrays (MEA) for monitoring, recording and stimulation. Triangular adhesive patterns are applied on each MEA to induce the growth of several three neurons networks. For the modelisation, intracellular potentials are simulated using networks of three neurons based on quadratic integrate and fire models. Extracellular potentials and simulated measures are computed using simple models of attenuation and electrodes impedances. Simulated or recorded noise measures are added to simulated extracellular potentials. Common methods of spike detection are evaluated using receiver operating characteristics curves and are used to evaluate the minimal signal to noise ratio necessary to detect spikes. Detected spikes will serve as a basis for the development of more complicated signal processing tools for the study of communication between neurons. Even if simple models are used, it allows us to study the influence of the different important points for the observation and understanding of the dynamic of three neurons networks. Several parameters need to be fitted to obtain relevant observations and good simulations. This will allow us to study more accurately the influence of noise on the different parts of the system experimentally and theoretically. |
|