Background: Previous studies suggested that the audio-visual integration ability in autism preserved well in the non-social contingencies, but was deficit in the advanced social cognitive processing, such as emotion and language (Williams et al.2004; James et al.2006; van der Smagt et al.2007GSmith et al.2007; Mongillo et al.2008GMagnee et al
Objectives: According to the audio-visual integration model (Belin, 2007), we used the paradigm of voice to facial expression interference to explore the mechanism of emotional audio-visual integration in autism.
Methods: Sixteen pairs of teenagers with and without autism (MA=13:3 years), which were matched on Chinese norms of RSPM score, CA, gender and manual laterality, participated in the study. In Experiment 1, emotional face and voice were presented simultaneously, the participants were asked to discriminate the facial expressions (sadness or happiness) by pressing different keys. In Experiment 2, faces and voice with neutral emotion were presented simultaneously, the participants were asked to discriminate the gender of the face (male or female) by pressing different keys.
Results: Results of Experiment 1 showed a significant main effect of consistency of emotional face and emotional voice, and a significant interaction effect of group and consistency. This interaction effect disappeared when emotional voice effect was controlled. Results of Experiment 2 showed a significant main effect of consistency, but no significant interaction effect of group and consistency, and no that of modal (unimodal vs. bimodal) and group.
Conclusions: According to the audio-visual integration model, results of the present study suggested that the adolescents with autism preserved well emotional audio-visual integration ability at structure analysis stage, but abnormal emotional audio-visual integration at effect analysis stage. We further found that the cause of abnormal emotional audio-visual integration was from the difficult of emotional voice recognition which was result of voice processing deficit at voice effect analysis stage.