Abstract
The objective of this study was to evaluate the functional relationships among brain sites using electroencephalography (EEG) with emotional stimuli, and to further evaluate discrimination using Artificial Neural Network (ANN). Twenty-two healthy subjects were assessed using the Cornell Medical Index (CMI) and divided into two groups: normal and pre-neurotic groups. EEG was measured with emotional tasks using audio-visual stimuli and analyzed using coherence analysis. The coherence values in each group, session, frequency band, and region were analyzed using analysis of variance (ANOVA), and discrimination was evaluated in the two groups using ANN. The coherence values of the pre-neurotic group in the pleasant and unpleasant sessions toward lateral sagittal and medial coronal were significantly larger than those of the normal group. Discrimination between the normal and pre-neurotic groups showed over 80% accuracy, sensitivity, and specificity. In addition, the coherence values using ANN differed in subjects with different neurotic states. These results suggest that cross-correlation of the information processes concerning emotional stimuli in the brain differs depending on the psychosomatic state.
