In this study, we show the specific and complementary attributes of Artificial Intelligence (AI) and of Connectionism (C). AI seems to be more adapted to modeling upper levels of data and knowledge processing performed by the brain, whereas C is more generally linked to sensory perception, reflexes or pattern recognition processes. A certain number of medical diagnosis aiding systems, combining these two paradigms, document the thesis that hybrid symbolic-connectionist architectures offer a very promising opening for the realization of complex, high level decision making systems in the years to come.