EEG responses to emotional videos can quantitatively predict big-five personality traits

Recent advances in information technology have suggested the potential possibility to assess an individual's personality automatically. The present study proposed and implemented an EEG-based personality assessment method for quantitative evaluation of people's Big Five personality. EEG data were collected from 66 participants, while they watched a total number of 28 video clips covering 9 typical emotion categories of amusement, joy, inspiration, tenderness, anger, disgust, fear, sadness and neutral. Regression analyses were performed to predict the participants' Big Five personality trait scores using the EEG responses to these emotional video clips. A nested leave-one-out cross-validation procedure was employed with a sparse feature selection strategy to evaluate the out-of-sample personality assessment performance. The established EEG-based regression models could effectively predict the participants' self-reported personality trait scores. The prediction accuracies, measured as the correlations between the EEG-predicted personality trait scores and the self-reported scores, were 0.71, 0.72, 0.86, 0.71, and 0.82 for agreeableness, conscientiousness, neuroticism, openness and extraversion, respectively. A series of tests from both the internal and external validity perspective further showed that a good reliability of the obtained results. These results suggest the proposed method as a promising alternative to conventional personality questionnaires.