Can People with Asperger Syndrome Identify Fake Smiles?

Background: Although adults with Asperger Syndrome (AS) are usually not impaired at identifying happy faces, one previous study indicates that AS adults find it difficult to identify fake smiles (Boraston et. al., 2008). In both genuine (“Duchenne”) and fake smiles, the zygomatic major muscle contracts, pulling up the corners of the lips. However, in genuine smiles there is also contraction of the external orbicularis oculi muscle, causing “crow’s feet” to appear on the outer sides of the eyes, the cheeks to raise and the eye aperture to narrow. Whether people with AS can distinguish real vs. fake smiles is important because of its relevance to social skills, and because this could reflect a specific instance of impairment in empathy (emotion recognition and attribution of intent). However, in the Boraston study, stimuli comprised just ten repeated female faces, and smile intensity in the mouth region differed between the real and fake smiles. There is therefore a need to test if this finding replicates, using improved stimuli.

Objectives: To replicate the finding that AS adults are impaired at identifying fake smiles with improved stimuli.

Methods: 15 males with AS and 20 control males aged 18-40 years, matched for IQ, participated. In Task 1, participants had to indicate whether they thought a smile was real or fake using a simple button box. Stimuli consisted of 92 photographs of smiling real faces, half of which were genuine smiles and half were fake smiles, each presented for 2955ms in a random order. The fake smiles were created by merging the mouth region from the genuine smile with a neutral expression. There were 25 female actors and 21 male actors who each appeared twice, once with a genuine smile and once with a fake smile. In Task 2, the 46 pairs of smiles were presented side by side and participants had to indicate which of the pair was the genuine smile. Accuracy and reaction times (RTs) were recorded.

Results: On Task 1, the AS group correctly identified 72% of the real smiles but just 49% of the fake smiles. In the control group, the figures were 78% and 63% respectively. Analysis showed that both groups were more accurate with real smiles and that, overall, controls were more accurate than the AS group. Both groups were faster on the real smiles and there was no group difference in RTs. Controls were better than the AS group at judging which smile was fake when the faces were presented in pairs (92% vs. 83%). The group difference for RT approached significance, with the control group being faster than the AS group. Actor sex had no effect throughout.

Conclusions: Participants with and without AS find it difficult to identify fake smiles. Presenting a fake and real smile simultaneously improves performance in both groups, but the AS group is still measurably less accurate than the control group. Because it is challenging, the Fake Smile Test could be used as a subtle test of empathy.