Artificial Neural Networks (ANNs) can be used to provide intuitive ways to design mapping strategies in novel DMIs, i.e., how sensor data are used to control sound generation. Research supports the idea that complex mappings are interesting and rewarding for expert performers, although there is currently no clear indication of how such mappings should be designed. Furthermore, it is often difficult to explicitly define mapping functions in a way that relates to a specific mathematical function. ANNs offer a means of implicitly generating complex, nonlinear mappings through the internal adaptations of the algorithm. This project uses various types of ANNs as tools for the design of DMIs that allow instrument designers and performers to link gestures and sound without having to explicitly describe the mapping relationship.