Original Contribution: The “off line learning approximation” in continuous time neural networks: An adiabatic theorem