Decoding visual network-related dynamic functional connectivity for eyes-open and eyes-closed using machine learning.