How much neuronal information is needed to control robots or prostheses?

The amount and quality of necessary information is a basic criterion for the environment parameter choice of robotic artefacts and neural prostheses. We start with a summary of different controllable artificial artefacts and determine how much data is needed to drive them. Latest developments of different human machine interfaces will be introduced with respect to the amount of data individually available and extractable in the present stage of their development. Founded on the present stage of interface options we will give an outlook into the near future, weighing advantages and disadvantages of each device and best field of use. Summarizing we will present some examples from the clinical practice, research laboratory and prospected devices. Closing with some examples of our own work divided in basic research in vitro applications and in vivo interfaces