Emotions play a powerful and significant role in human beings everyday life. The work presented in this paper aims at assessing human emotional stress using psychological signals like skin onductance (SC), respiration photoplethysmograph (PPG) as well as electroencephalographic (EEG) signals. Two specific areas of the valence-arousal emotional space are defined, corresponding to negatively excited and calm-neutral states. Our acquisition protocol is based on a subset of pictures which correspond to two specific areas of valance-arousal emotional space. We chose to calculate the power spectrum, correlation dimension and Entropy. Genetic Algorithm as a feature selector has been used for LDA and Elman Classifiers. The classification accuracy in two classes of emotional states is 84.12% and 78.45% corresponding to Elman and LDA classifiers. Comparison between two classifiers has shown better results of using Elman classifier and it seems the reason is due to the problem of LDA with high dimensionality of EEG features as input signals.

کلمات کلیدی