LAUSR

Abstract Background Schizophrenia patients and individuals at ultra-high risk for psychosis(UHR) have shown impaired facial emotion recognition(FER). Previous studies have reported lower accuracy and negative bias of FER in schizophrenia and UHR. These impairments have been studied with various factors such as schizotypy and paranoid level, but the results were inconsistent. This study aimed to identify the impairments of FER in UHR individuals and further to examine how these impairments relate to schizotypy and paranoid level. Methods orty-three UHR individuals and 57 normal controls (NC) were requested to perform the facial emotion recognition(FER) task that consist of 60 facial photographs selected from standardized photographs of Ekman and Friesen series. For exploratory correlation analysis, schizotypy (Revised physical anhedonia scale, Magical ideation scale) and paranoid level (Paranoia scale, Persecution/suspicious item of Positive and Negative Syndrome Scale) were also examined in UHR individuals. Results The UHR individuals showed lower accuracy rate for total FER task (70.6% vs. 75.6%, p=0.010) and more “fear” responses for neutral faces (14.5% vs. 6.0%, p=0.003) than NC. In exploratory correlation analysis for UHR individuals, the total accuracy rate of FER task showed significant correlation with both scales for schizotypy, but not with both scales for paranoid level. Among threat-related emotion response rates for neutral face, only “Disgust” response rate for neutral face was correlated with all scales for paranoid level, but not with scales for schizotypys in UHR individuals. Discussion In this study, we could identify inaccuracy and negative bias of FER in UHR individuals, Furthermore, we found that inaccuracy and negative bias were associated with schizotypy and paranoid level, respectively. These findings imply that inaccuracy and negative bias of FER in UHR individuals are of different nature. Future studies on the clinical implications of these findings would be needed.