The early warning signal of complex diseases based on the network transition entropy

Abstract

Many evidences suggested that during the progression of complex diseases, the deteriorations are generally not smooth but abrupt, which may cause a critical transition from one state to another at a tipping point, corresponding to a bifurcation of the dynamical system for the underlying organism. A pre-disease state is assumed to exist before reaching the tipping point between a normal state and a disease state. Since the predisease state is defined as a limit of the normal state, which represents an early-warning signal of the disease, it is crucial to identify such a state so that remedial actions can be executed to avoid the abrupt transition to the disease state. Although most complex diseases are model free, and usually only small samples are available due to clinical limitations, we propose that an index called the network transition entropy (NTE) may serving as an early-warning indicator for predicting the critical transition. Although the theoretical deviation is based on the dynamical network biomarker (DNB), the application of NTE is DNB free.