摘要：Purpose: To explore the differences in facial emotion recognition among patients with unipolar depression(UD), bipolar depression(BD), and normal controls. Methods: 30 patients with UD, 30 patients with BD and 30 normal people respectively were recruited. All patients met the diagnostic criteria of DSM-IV in Zhumadian Second People's Hospital during July 2018 to August 2019. The HAMD, HAMA, and YMRS were used to assess the depressive symptoms, anxiety symptoms of patients and the severity of mania symptoms. 15 groups of facial expression recognition pairs were performed on 6 facial expressions for all participants. Results: The differences were found in the happy-sad, happy-angry, happy-surprised, and disgust-surprised facial expression in three groups. The independent sample T-test analysis showed that compared with the normal control group, there were differences in the happy-sad and happy-angry groups in patients with BD, and the accuracy of facial expression recognition in normal people was higher than that in patients with BD. Compared with patients with UD, there were differences between the happy-sad and happy-angry groups, and the identification accuracy of patients with UD was higher than that of patients with BD.The time of facial expression recognition in the three groups was analyzed by single factor ANOVA, and the time of facial expression recognition in the normal control group was shorter than that in the patient group.Independent sample T-test analysis showed that patients with BD took longer to identify happy expressions, while patients with UD took longer to identify negative and neutral expressions, but there was no statistical difference. Using happiness-sadness to distinguish unipolar and bipolar depressions, the area under the ROC curve (AUC) is 0.933, the maximum Youden index is 0.822, the specificity is 0.889, and the sensitivity is 0.667. Using happiness-anger to distinguish unipolar and bipolar depression, the AUC was 0.733, the maximum Youden index was 0.378, the specificity was 0.778, and the sensitivity was 0.600. Conclusion: Patients with UD had lower performance in recognizing negative expressions. Those with BD had lower accuracy in recognizing positive expressions. Rapid facial expression recognition performance may be as a potential endophenotype for identification of unipolar and bipolar depression.
