Emotional Bias And Decision Analysis Based On EEG And Cognitive Psychology

The illogical design of the task display leads to risk-approaching and psychological load-raising behaviors,and these negative effects reduce the performance level of the subjects.Therefore,in this paper,we analyze subjects’ decision-making behaviors when performing specific tasks from two perspectives:emotional bias and self-regulatory mechanisms,and develop mathematical models for these two types of data.Based on this,the paper reads as follows.1.The effect of emotional bias on subjects’ risky decision-making behavior during task performance is analyzed using the gambling mechanism.Negative and positive class scenario pictures are used as target stimuli,respectively,to induce subjects’ emotional bias in a specific environment.ERP analysis is conducted on subjects’ behavioral data and decision motivation and outcome assessment stages,respectively.The results show that negative bias prompts subjects to focus their attention,as evidenced by a decrease in P100 amplitude.Positive bias,on the other hand,broadens attention and triggers an increase in P100 amplitude.Both emotional biases increase attention to emotional material,which leads to an increase in P200 amplitude.Positive bias increases the sensitivity of the subject to the outcome,resulting in a greater FRN amplitude.Negative bias induces risk-seeking behavior.2.Self-regulatory behaviors are used to control the subjects’ psychological load level in the appropriate range and to enhance their performance level.A controlled trial is conducted in which a self-regulatory mechanism is introduced in the experimental group and no additional intervention is made in the control group,and both groups complete the same mental arithmetic problems.Combining behavioral data with physiological data,extracting the EEG signals of the subjects when completing the task and conducting ERP,power spectral density,and microstate analysis can effectively enhance the performance level of the subjects under the specific task difficulty and control the mental load in the appropriate range.The results indicate that self-regulatory behaviors cause additional psychological load,showing smaller P300 and P600 wave amplitudes in the central parietal area.When task difficulty is moderate,self-regulatory behaviors motivate subjects to perform better.3.Convolutional neural networks are constructed separately for two types of data characteristics.For the EEG signal part of the emotion class,a confusion-free convolutional neural network model is built.The introduction of the adversarial component in the model eliminates the confounding effect in feature extraction,effectively improving the accuracy and robustness of the model in cross-subject detection.For the EEG signal part of the mental load class,a twin convolutional neural network model based on the EEG signal is built.The model effectively identifies the psychological load level of the subjects by taking advantage of the obvious differences in the frequency domain features of such data.