Frontal EEG asymmetry and affective states: A multidimensional directed information approach

Abstract

Electroencephalogram (EEG)-based emotion recognition is a rapidly growing research field that examines the ability of EEG signals to discriminate emotions in humans. One of the major problems that this emotion recognition approach deals with is the reliability of the emotion induction to the subjects that participate in relevant experiments. In this paper, Multidimensional Directed Information analysis was adopted to identify frontal EEG asymmetry and, thus, evaluate the emotion elicitation procedure followed to evoke affective states from 16 healthy right-handed subjects using the IAPS database. The derived results have shown that the adopted method has the potential to become an efficient emotion elicitation evaluation criterion. © 2010 International Federation for Medical and Biological Engineering.