Qualitative and quantitative evaluation of brain activity in emotional stress

This paper proposes a new emotional stress detection system using multi-modal bio-signals. Since EEG is the reflection of brain activity and is widely used in clinical diagnosis and biomedical researches, it is used as the main signal. Methods and Materials: We designed an efficient IAPS acquisition protocol to acquire the EEG and psychophysiological signals under picture induction environment (calm-neutral and negatively excited) for participants. Data such as skin conductance (SC), Blood Volume Pulse (BVP), respiratory rate (RR) and EEG were continuously recorded through bio-sensors placed on the participant. In order to choose the proper EEG channels we used the cognitive model of the brain under emotional stress. After pre-processing the bio-signals, linear features were employed to extract the psychophysiological signals and chaotic (or nonlinear) invariants like fractal dimension by Higuchi’s algorithm, correlation dimension and approximate entropy were used to extract the characteristics of the EEG signals. For emotional stress detection, Genetic Algorithm (GA) and Elman neural network are applied to design the emotional stress classifier are investigated. The results show that, classification accuracy with fusion link between EEG and psychophysiological signals was 82.6% using the Elman classifier in two classes from emotional stress space. Chaotic analysis can be representing good of human brain and behaviour in emotional stress states.�This is a good improvement in results compared to other similar published researches.