Modeling immune cognition

We believe that organism's defences have gradually evolved into a real cognitive system after cell membrane receptors underwent diversification in early fish, some 300 million years ago. The somatic repertoires grow by combinatorial power, as receptors are assembled from germ line information, by an aleatory and error prone process. There is no doubt that a cellular automaton is well suited as the vehicle for modeling these aspects of the immune response. In its simplest form, the model consists of a 15/spl times/15 bidimensional grid, containing different cell types, each expressing receptors in a form of 8-bit binary strings. Antigens and antibodies interact if they meet within a site and if the respective bits are complementary. The rules and cell programs are drawn with a keen eye to current biological consensus, and the model behaves like an immune system. We review the principle and the achievements of the IMMSIM model and present preliminary results of the combined cellular and humoral responses to a viral infection.