Modelling Immune Memory for Prediction and Computation

Abstract

This paper investigates the concept of immune memory, and the potential for an Artificial Immune System (AIS), in which memory is an emergent property of the antibody population and its dynamics. Inspiration for the implementation of this concept of memory is taken from current biological theories. However, a difficulty lies in the fact that as yet these theories remain unconfirmed; no conclusive explanation has been put forward to explain how immune memory is created and sustained over time. The approach taken here is to investigate and build on two of the basic models of immune memory : (i) the memory cell model, (ii) the residual antigen model, and iii) provide some initial thoughts as to the influence of the immune network model on our theory of immune memory. We show that each model can partially explain how the immune system can remember infections, and respond quickly to re-infections, but we begin to demonstrate that none of them in isolation result in an effective immune memory response. Our initial appraisal is that all three models should be combined, and updated to reflect new biological concepts, we make some suggestions about how this might be achieved, and we illustrate our discussion with some tentative results. © Springer-Verlag 2004.