Abstract:The rapid detection of changes in facial expressions is an important social and survival skill. The detection of multiple facial emotions includes not only the information of emotional valence but also differences in emotional valence, that is, emotional valence consistency and inconsistency. Thus, we explored whether changes in multiple facial expressions could be automatically detected, as indexed by the visual mismatch negativity (vMMN) response. Participants were presented with a set of facial stimuli while performing a visual facial identity detection task; the stimulus was presented in the center of the visual field. The facial stimuli set consisted of five different facial identities and were presented in an oddball sequence, with four peripherally expressing the same positive or negative emotion and one in the center expressing congruent or incongruent emotions. We found vMMN responses to changes in positive congruent deviant emotions between 210 and 320 ms and in all deviant emotions between 480 and 560 ms over bilateral temporal-occipital sites. In addition, at 480–520 ms, the positive congruent stimulus versus the incongruent stimulus and the negative incongruent stimulus versus the congruent stimulus induced more negative vMMN amplitude in the left temporal-occipital electrodes. This shows that individuals can automatically identify the changes in multiple faces' emotional differences (emotional valence inconsistency), and that the emotional valence of the target face affects the automatic processing of multi-face emotional valence differences information. Furthermore, these results can be utilized in future research investigating automatic processing mechanisms.

Automatic change detection of multiple facial expressions: A visual mismatch negativity study