Lack of Emotion-Specific Facial Mimicry Responses among High-Functioning Individuals with An Autism Spectrum Disorder

Background: Previous research has shown that presentation of facial expressions of emotion automatically evokes subtle changes in the observer’s facial muscle activity as measured with electromyography (EMG), with differential muscle activity evoked depending on the affective valence of the stimulus. This phenomenon, termed facial mimicry, is hypothesized to represent a psychophysiological mechanism of emotion perception and empathy, and as such, may be a good biological marker of a disruption to the emotion perception system among individuals with Autism Spectrum Disorders (ASD).

Objectives: To examine automatic facial mimicry to happy and angry expressions of emotion among high-functioning individuals with ASD and a typically developing control group utilizing dynamic audiovisual stimuli.

Methods: Participants included nine high-functioning individuals with a diagnosis of ASD (mean CA = 20.2 yrs) and 15 typically developing controls (mean CA = 14.6). There were no significant differences between the groups with respect to chronological age, verbal and nonverbal IQ, and face recognition. Participants underwent a dynamic emotion perception task during which they judged the emotion expressed in 2-second video clips containing facial expressions and affective prosody while EMG sensors recorded facial movements on zygomatic (cheek) and corrugator (brow) muscles. Analyses focused on within-group comparisons of average levels of left zygomatic and corrugator muscle activity, 500ms to 1000ms post stimulus onset, between happy and angry stimuli.

Results: As expected, for the control group, happy stimuli elicited significantly more zygomatic activity than did the angry stimuli [t(14) = 2.98, p < .01, partial η² = .39], and angry stimuli elicited a trend toward more corrugator activity than the happy stimuli [t(14) = 2.06, p = .06, partial η² = .23]. In contrast, there were no differences in corrugator and zygomatic activity in response to either angry or happy stimuli among the ASD participants (all ps > .1).

Conclusions: Preliminary results of our analysis of EMG response to dynamic facial expressions and affective prosody suggest that unlike typically developing individuals, individuals with ASD do not show an emotion-specific differential EMG response. In this regard, our findings tentatively echo those of McIntosh and colleagues (2006), who found that participants with ASD failed to demonstrate significantly more congruent than incongruent facial EMG responses to static photographs of happy and angry facial expressions.  These findings raise the question of whether disrupted facial mimicry may be a psychophysiological substrate underlying previously documented impairments in emotion recognition and emotional responsiveness among individuals with ASD.  Consequently, in future analyses with a larger sample, we will examine group differences in patterns of EMG activity and latency to EMG response, and we will explore whether disruptions in facial mimicry have implications for social-emotional functioning among individuals with ASD.