We review a technique for creating Self-0rganising Maps (SOMs) in a Feature space which is nonlinearly related to the original data space. We show that convergence is remarkably fast for this method. By considering the linear feature space, we show that it is the interaction between the overcomplete basis in which learning takes place and the mixture of one-shot and incremental learning which comprises the method that gives the method its power. We illustrate the method on real and artificial data sets.