Brain Effective Connectivity Analysis from EEG for Positive and Negative Emotion

Recently, there have been increasing evidence which supports that multiple brain regions are involved in emotion processing. Therefore, research on emotion from the perspective of brain network is becoming popular. In this study, based on the Granger causal analysis method, we constructed brain effective connectivity network from DEAP emotional EEG data to investigate how emotion affects the patterns of effective connectivity. According to our results, prefrontal region plays the most important role in emotion processing with interactions to almost all other regions. More interactions are found under negative emotion than positive one. Parietal region in charge of human's alert mechanism is more active under negative emotions. These results are consistent with the previous findings obtained in neuroscience, which illustrate the effectiveness of our methods. Furthermore, the brain effective connectivity network shows significant differences to different emotional states, so it can be used to recognize different emotional states with EEG.