CELL-DIVISION AND MIGRATION IN A GENOTYPE FOR NEURAL NETWORKS

Much research has recently been dedicated to applying genetic algorith ms to populations of neural networks. However, while in real organisms the inherited genotype maps in complex ways into the resulting phenot ype, in most of this research the development process that creates the individual phenotype is ignored. In this paper we present a model of neural development which includes cell division and cell migration in addition to axonal growth and branching. This reflects, in a very simp lified way, what happens in the ontogeny of real organisms. The develo pment process of our artificial organisms shows successive phases of f unctional differentiation and specialization. In addition, we find tha t mutations that affect different phases of development have very diff erent evolutionary consequences. A single change in the early stages o f cell division and migration can have huge effects on the phenotype, while changes in later stages usually have a less dramatic impact. Som etimes, changes that affect the first developmental stages may be reta ined, producing sudden changes in evolutionary history.