Psychological Stress Assessment Methods And System Study On Physiological Parameters Fusion

Mental health problems are becoming increasingly the focus of attention. It is a hot research question to assess the psychological stress state effectively. In the field of psychology, the main way to assess psychological stress is questionnaire, which is a kind of subjective method, however, lacking of accuracy and objectivity. But assessing the psychological stress based on physiological parameters can much more objectively determine the subjects’ psychological pressure. In this paper, aiming at assessment of psychological stress, make full use of EEG, ECG, EMG three physiological parameters which contain a wealth of psychological characteristics information and combining SVM and DS evidence theory method of classification decision-making to achieve the psychological pressure assessment analysis. A system to evaluate the psychological pressure based on the fusion of physiological parameters has been designed and finished.In the design of experiment, the psychological pressure is divided into two states including under-pressure and zero-pressure. The laboratory equipment multiple physiological parameters MP150 recorder simultaneously measured EEG, ECG, EMG. Then using the scale SRI determined the effectiveness of the self-assessment data. In processing signal, we used the de-noising method based on db4 wavelet in 5 layers for EEG, db5 wavelet in 7 layers for ECG, db5 wavelet in 4 layers for EMG respectively. In feature extraction of EEG, we took 3 features into account, such as the C0 complexity. In feature extraction of ECG, we took 21 features into account, such as waveform amplitude of the mean. In feature extraction of EMG, we took 29 time-domain and frequency domain features into account, such as mean power frequency,. In the classification, support vector machine for three kinds of physiological signals were characteristic layer classification to evaluate psychological stress, to get the posterior probability, confusion matrix. Then combining with DS evidence theory which has simple reasoning form and can represent the information of uncertain and unknown, achieves decision-making classification. Finally, Matlab and C # are mixed to implement a psychological stress assessment system.Contrasting the classification accuracy rate of test to single physiological parameter, for EEG classification, maximum classification accuracy was 88.00%, the average rate of 88.04%; in ECG classification, maximum classification accuracy was 84 % and average accuracy rate of 80.43%; SVM classification of EMG signal, the maximum classification accuracy rate of 80.00%, with an average accuracy rate of 77.82%. The use of decision-level fusion on DS evidence theory, the fusion of the three largest physiological signals correct classification rate was 96.00%, the average accuracy rate was 88.62%. Research results suggested that EEG, ECG, EMG three physiological parameters can effectively assess the psychological pressure, including the best classification results of EEG. To decision-classification with SVM and DS theory, classification accuracy had been significantly improved, this paper proves SVM and DS evidence theory to assess the psychological pressure of the effectiveness and feasibility.