Total connectivity: a marker of dynamical functional connectivity applied to consciousness

In the last years functional connectivity (FC) has become one of the most popular tools to explore and characterize information contained in fMRI =me series. The classical hypothesis on FC consists of considering it as constant (or static) over the whole fMRI time series. However, it has been emphasized recently that FC should be treated as a dynamical quantity, for example by using sliding windows of the fMRI time courses in order to compute a dynamical FC.
We propose a comprehensive marker of FC based on an auto-regressive (AR) model of fMRI time series capturing its static and dynamic properties. We call it total connectivity and we illustrate the benefits of our approach on data of patients undergoing four different states of consciousness.