Visualization of Information Flows in a Brain Stemming from Alpha Waves - Using Multidimensional Directed Information

摘要

This article proposes a method to visualize information flows in the brain stemming from specific components making up electroencephalograms (EEGs). This method combines multidimensional directed information analysis, which is a method used in causality analysis of multidimensional time series, with independent component analysis to extract only directed information related to specific EEG components. This method will enable us to find problems in brain information communication that are caused by specific diseases or organic disorders in the brain, and observe several information flows occurring simultaneously but completely independent of information communication in the brain. In this paper, as an example of applying the proposed method, an attempt was made to extract only information flows related to alpha waves from information flows in the multi-channel EEGs of a healthy individual.