A general higher order neural model

We present a new methodology for describing the functioning of artificial neurons, including new, as yet untested, types of behaviour. It also provides the possibility of defining artificial neurons of any order, and a wide range of functions from which to choose. As an illustration of the new formulation, a practical realization is analyzed, consisting of a multilayered neural network applied to the image processing of black and white scenes, making manifest the possibilities of this new type of neuron in the field of cellular logic but with new types of processing. This is just an early stage in the development of the new neurons, so that many of their possible applications have yet to be initiated. Among them, one can already foresee those related to fuzzy models, analogue models, and many others. For these applications, it will no longer be necessary to make any change in the network design, just to make a choice from the proposed library of functions