Effect of Variable Propagation Window Parameter on Multidimensional Directed Information

抄録

There are many fluctuations in life phenomena, and they are always occurring in a human brain. Although we developed a method called multidimensional directed information (MDI) analysis to visualize information flow in a human brain, it was based on the presumption that the information propagation time is constant. This paper reports how the method has been improved so that it may have applicability even when information propagation time fluctuates. First, we conducted a numerical simulation to investigate characteristics of the new method, and we confirmed that the new MDI analysis is beneficial in calculating the amount of information flow. Next, we adapted the new method to an electroencephalogram (EEG) analysis. As a result, we were able to estimate the amount of information flow between EEGs of a normal subject.