Recurrence Method in the Analysis of Electroen- cephalographic Signals in Healthy Adults under Light-Dark Conditions

Abstract We analyze electroencephalographic signals (EEG signals) using numerical methods. These signals were collected during an experiment where each person was exposed to light of two different colors (blue and red) in contrast to the dark. In addition to linear signal filtering methods applied to EEG signals, we used in current study a non-linear method, recurrence plot, to characterize the system that produces them. Taking into account Takens theorem, we obtain a set of vectors that were built from the value of the EEG signals. Then we analyze the distance matrix between vectors obtained from the one chosen EEG signal. Based on these matrices we perform recurrence quantification analysis, and calculate: percent determinism, Shannon entropy and average line length. They gives information about changes in the dynamics of the sources producing the EEG signal under light-dark conditions. We demonstrate that light selectively affects EEG signal sources. Their effects are also differentiated by changes in the dynamic properties of these sources.