The interface man / machine had followed of many researches in biotechnology. A part of these researches concern the interconnections brain / machine. Indeed, the brain arranges numerous intellectual connections thanks to neurons. These neurons communicate between them and propagate information thanks to a "bio-electric" signal called action potential. The objective of my doctoral thesis is to measure this signal thanks to ion sensitive field effect transistor (ISFET). The ISFET process was modified to obtain a new kind of sensor, named NeuroFET. Chips containing NeuroFETs were completely microfabricated in cleanroom. The growth of neurites must be directed to pass on the NeuroFET gate. For this to be successfully addressed, we chose to force them mechanically thanks to channels in SU-8 resin. After testing different but not decisive methods, we developed our own technique based on photolithography by projection. By modifying the parameters of focus and exposure, it was possible to obtain channels in the shape of broken bows in a single insulation. Thanks to this method named SU-8 3D, we realized a biocompatible, SU-8-based microchannel network Our partnership with the Institut de la Vision in Paris allowed us to organize the growth of neurites of neurons of rat retina using this microchannel network. Chips were packaged on a printed circuit board, isolated electrically and covered with a cone of culture allowing the neuronal culture on the scale of the individual chip. As action potential of neurons of rats retinas were barely important to be measured using our measurement interface, we used neurons of pool snails Lymnaea Stagnalis. After three days of growth, we applied to cells a cycle of various toxins allowing to alternate activity of spontaneous actions potential and state of rest. This cycle allowed us to observe a neuronal activity and so to validate the functionality of the system.
