This chapter addresses optimisation of a class of biological neural networks, called Central Pattern Generators (CPGs), with a view to providing autonomous, reactive control to otherwise non-adaptive operators. CPGs are self-contained neural circuits which govern rhythmic motor activities such as locomotion, breathing and digestion. Neurons in this system interact to produce rhythmic oscillations without requiring sensory or central input. These phasic firing patterns can be adaptively adjusted, through neuromodulation, and in response to fluctuations in the environment. Thus, CPGs provide autonomous, self-modulatory control and are an ideal candidate to evolve and utilise for practical engineering solutions. An empirical study is described which generates CPG controllers with a wider range of operation than their counterparts. This work is precursory to producing controllers for marine energy devices with similar locomotive properties. Neural circuits are evolved using genetic algorithm techniques. The lamprey CPG, responsible for swimming movements, forms the basis of evolution, and is optimised to operate with a wider range of frequencies and speeds. Results demonstrate that simpler versions of the CPG network can be generated, whilst outperforming the swimming capabilities of the original network [34].