Analysis of neural electrical activities during elicitation of human emotion based on EEG

Abstract

This work aims to analyze the neural electrical activity inside the brain during elicitation of human emotional states using Electroencephalographic (EEG) signal. Such analysis offers to understand the human emotional process in great details as earlier most of the works focused on either emotion recognition or its application only using EEG signal. To do so, firstly, the emotional states are categorized in dimensional space based on valence-arousal scores into four classes: high valence high arousal (HVHA), high valence low arousal (HVLA), low valence high arousal (LVHA), and low valence low arousal (LVLA) from the DEAP emotional dataset. The sLORETA tool has been used in this purpose, which can perform the source localization in 3D with current density estimation from the EEG signal, and functional connectivity based on coherence. The results show that the current density during HVHA, HVLA, LVHA, and LVLA are 1.8±0.51 µA/mm2, 3.7±1.1 µA/mm2, 2.5±1.2 µA/mm2, and 4.4±2.8 µA/mm2 respectively. Besides, the main active regions in cortical level during elicitation of different classes of emotions are: for HVHA, Brodmann areas 10 of superior frontal gyrus region of frontal lobe; for HVLA, Brodmann areas 6 of superior frontal gyrus region of frontal lobe; for LVHA, Brodmann areas 19 of cuneus region of occipital lobe; and for LVLA, Brodmann areas 40 of supramarginal gyrus region of parietal lobe. The results show that, for higher valence, the brain active regions in cortical level is same but the Brodmann areas are different. Such results suggest that there are different sets of neural active regions with different current densities for different classes of emotional states. Therefore, in future, more subjects including more trials and others parameters will be analyzed which could guide to understand intra-brain electrical neural activity with great details.