Correct, theoretically it is possible, I am asking what is the practical difficulties, say border monitoring system where nodes can face difficulties which we may not guess.
In any system there are potential unknowns. In a system with bio inspired nodes performing a function (e.g. border monitoring), I think it is always important to gracefully deal with the unknown. Understand the bounds of your bio inspired algorithms and the range within which they can adapt to a situation. Beyond those bounds one can get discrete and effectively "catch" and "report". For a border monitoring example, its better to catch and report than to ignore. So your practical difficulties are understanding your bio inspire algorithms and defining the bounds within which they can operate/adapt and outside of which you need to catch/report.
Bio inspired models are an emerging research area in the field of networking. It is quite exciting field and the benefits possible depends on what one wants to use as we still know so less about bio inspired models. I highly appreciate your idea and think that certainly one may think in this line. However, be aware about the complexity of the algorithms to be implemented as WSN nodes have less computing power and other constraints present in WSN. However, I think that there should not be much problem in implementation. Your idea is quite through provoking so all the best for your research.