Facial expressions are fundamental emotional stimuli. They convey important information in social interaction. Most previous studies focused on the processing of isolated facial expressions. However, in everyday life, faces always appear within complex scenes. The emotional meaning of the scenes plays an important role in judging facial expressions. Additionally, facial expressions change constantly from appearance to disappearance. Visual scenes may have different effects on the processing of faces with different emotional intensities. Individual personality traits, such as trait anxiety, also affect the processing of facial expressions. For example, individuals with high trait anxiety have processing bias on negative emotional faces. The present study explored whether previously presented visual scenes affected the identification of emotions in morphed facial expressions, and whether the influences of visual scenes on the identification of facial expressions showed differences between individuals with high and low trait anxiety. Using the Spielberger State-Trait Anxiety Inventory (STAI), we placed 29 participants who scored in the top 27% in the high trait anxiety group (9 men and 20 women; mean age 19.76 +/- 1.3 years) and 28 participants who scored in the bottom 27% in the low trait anxiety group (11 males and 17 females, mean age 19.71 +/- 1.2 years). The images of faces (4 models, half male and half female) used in this study were selected from the NimStim Set of Facial Expressions. The face stimuli showed typical happy, neutral, and fearful expressions. Facial expressions were morphed to create a series of gradually varied images of facial expressions. Specifically, fearful face (100%) versus neutral face (0%) and happy face (100%) versus neutral face (0%) were morphed in 20% increments. In addition, 40 surrounding scene images were used, with 20 positive scenes and 20 negative scenes. In the face-emotion detection task, participants were asked to determine whether the emotion from the faces presented after the scenes were fearful, happy, or neutral. For the repeated measure ANOVA of the accuracy for facial expression detection, the results showed scene effects on the identification of emotions in facial expressions. The scene effects were varied between the different intensity of face emotion: for the emotionally vague faces, the detection of happy and fearful expression showed significant scene effects; for the faces with moderate emotional intensity, only the detection of the fearful faces showed significant scene effects; for the intense emotions on faces, there was a significant effect on happy and neutral faces but not on fearful faces. Trait anxiety as an individual factor was found to play a moderating role in the identification of facial expressions. For the high trait anxiety group, there were no significant differences in the accuracy of emotional detection between congruent and incongruent conditions. This means that the high trait anxiety group did not show significant scene effects. The low trait anxiety group showed a significant difference in the accuracy of identification of emotions in facial expressions between congruent and incongruent conditions, i.e., significant scene effects. In summary, the present study demonstrated that, for facial expressions with low emotional intensity, the identification of happy and fearful faces was more likely to be affected by visual scenes than the identification of neutral faces. Visual scenes were more likely to affect the identification of moderately fearful faces than moderately happy faces. Trait anxiety played a moderating role in the influence of visual scenes on emotional detection of facial expressions. Specifically, individuals with high trait anxiety were less affected by surrounding visual scenes and paid more attention to facial expressions.
