Comprehensive Summary
This study by Shuaiqi Liu et al. proposes the implementation of the Hebei University Emotional EEG Dataset (HBUED). This dataset incorporates a computation method to better capture EEG signals and deep learning technology to improve performance of emotion detection. The study to support these findings was conducted by providing image, audio, or video stimuli to 75 students from Hebei University. EEG data was continuously recorded for the duration of the experiment. Several metrics of evaluation were used to determine the effectiveness of the proposed emotion recognition algorithms, and the performance when applied to the conceived dataset was analyzed. The result was that the proposed model was able to correctly classify the emotional states provided in the HBUED with over 90% accuracy.
Outcomes and Implications
One limitation to the model is the sensitivity of the emotional cortexes of the brain in individuals between the ages of 18 and 35, which was the primary group studied in this article. The amygdala and prefrontal cortex may display varied characteristics, decreasing neuronal synchronization, and thus sending weaker signals to the database. Issues related to these causes for the delay of clinical implementation of emotion-detecting algorithms. However, the proposal of HBUED can be a great step forward to learning more about emotional variability in the human population as the database expands to include more diverse participants.