Abstract
Over the past few years, two generations of cochlear implants have been designed and tested by our group. Since the beginning, our goal was to design a system that is completely transparent to sound processing algorithms and stimulation strategies. The first generation microstimulator has eight completely independent channels and the second generation one has sixteen. Each of them is mounted on a thick film hybrid circuit together with necessary interfacing components. The electronic circuits are housed in a hermetically sealed titanium case. They are able to deliver monopolar, bipolar, simultaneous or non-simultaneous stimuli on all of their channels. Their electrodes are addressable with all possible combinations. The eight-channel stimulator has already been successfully implanted on volunteer profoundly deaf patients, validating our concepts and our approaches. The second one is endowed with more flexibility and transparency to stimulation algorithms. It offers extended possibilities and more electrodes combinations. In a second version, this one has also been mounted on a thin film hybrid circuit making its size considerably reduced and allowing implanting it in other sites rather than on the temporal bone. The sound processor supporting these systems communicates with internal parts through an inductive link. It is built around a powerful DSP and can be programmed with more than one algorithm at a time. It has been already programmed with all the known strategies and some original ones such as the technique of vector quantization. The latter uses a dictionary of pre-stored stimulation patterns mapped to sound segments. As for programming the systems, this is accomplished by mean of specific user-friendly application software using a window driven completely graphical interface.