Laboratoire Matière et Systèmes Complexes

Neuronal networks growth in vitro is studied from two different points of view :

From an empirical point of view. Beyond the axonal growth through microchannels, the topology of in vitro neural networks can be controlled by obstacle arrays. First of all, the growth of a single axon encountering a simple obstacle, like for instance a wall with a given incidence angle or a plot with a given diameter needs to be systematically characterized. Preliminary experiments revealed the affinity of axons for edges of microstructures and the existence of an angle at which they are released from topological guides. Empirical laws drawn from these measures will enable to design in a rational way neuronal devices able to grow networks where the connectivity will be much more controlled.

Demonstration of axons affinity for edges, and existence of a non-trivial release angle. The corticalaxons in green (immunostaining of betatubulin) are only released from the edges of PDMS structures at vertices whose angles are smaller than a critical value of about 130±10°.

From a theoretical point of view. As a preliminary task, we elaborated a stochastic model using physically relevant rules for neurites growth rate, growth direction and generation of synapses able to construct two-dimensional neuronal networks in good agreement with those experimentally observed. Moreover, we are now able to draw dynamical information from time-lapse films produced during the growth in vitro and to perform a modelling work aiming at a fundamental understanding of axonal growth ; besides axon-substrate interactions, axon-axon interactions, axon branching and the role of stress, the growth involves the cytoskeleton, suggesting that actin waves already observed experimentally should be incorporated in a dynamical model.

Top : Image of a two-population device at Weizmann Institute of Science.(R. Renault). Bottom : Simulation of a culture from our stochastic model (M. Agamennone).