An information hypothesis for the evolution of homeostasis.

A prevailing view among physiologists is that homeostasis evolves to protect organisms from damaging variation in physiological factors. Here, we propose that homeostasis also evolves to minimize noise in physiological channels. Fluctuations in physiological factors constitute inescapable noise that corrupts the transfer of information through physiological systems. We apply information theory to homeostasis to develop two related ideas. First, homeostatic regulation creates quiet physiological backgrounds for the transmission of all kinds of physiological information. Second, the performance of any homeostatic system influences information processing in other homeostatic systems. This dependence implies that multiple homeostatic systems, embedded within individual organisms, should show strongly nonadditive effects.