Facial Expression Perception In Relatives of ASD Children: Is There a Reliable Endophenotype?

Background:  

The ability to recognize facial expressions (FE) of emotions is crucial for mediating human social interactions, a domain in which individuals with autism spectrum disorders (ASD) have great difficulties. Recent research showed that abnormalities in FE processing are present also in first-degree relatives of individuals with ASD, and might be part of a set of features associated with the disorder. (e.g. Adolphs et al., 2008; Bolte & Poutska, 2003).

Objectives:  

Our study investigated the presence of subtle abnormalities in the perception of FE of four emotions (Anger, Fear, Happiness, and Disgust) in a sample of families with a child with ASD (ASD relatives), as compared to families with typically developing children (controls).

Methods:  

Participants: The ASD families group included 40 relatives (10 fathers, 17 mothers, 13 typically developing siblings) of ASD probands between 5 and 23 years old. All probands met DSM-IV criteria for ASD. Participants were recruited through special schools in London. Controls included 33 subjects (8 fathers, 12 mothers, 13 children). Groups were globally matched for age and IQ.

Material: Four photographs portraying the FE of Anger, Fear, Happiness, and Disgust, posed by an actor were used as prototypes. The actual stimuli shown in the experiment were two sequences of morphed pictures. Within each sequence, morphed FE described a linear transition between two endpoints (Anger-Fear, [A-F] and Happiness-Disgust, [H-D], respectively.

Task and procedure: For each FE pair, participants performed a forced-choice identification task. In each trial, one stimulus was displayed for 3 s, and participants had to decide which prototype the stimulus is more similar to (e.g. “Anger” or “Fear”). We recorded the response and the RT (in ms) measured from stimulus onset. Each stimulus was presented several times, in random order.

Results:  

For each FE pair and each group, we computed the psychometric function relating the probability of giving a certain answer (e.g. P(“Fear”)) to the rank order of the morphed stimulus. The parameters (i.e. point of subjective equality [PSE] and just noticeable difference [JND]) of the psychometric function afford an assessment of the recognition ability of the observer. There was no systematic difference between groups in these parameters. The absence of differences between JNDs for the two groups suggests that ASD relatives can recognize FE as accurately as control parents and children. However, we found that ASD parents and siblings respond to FE significantly faster (p <.05) than control parents and children. The difference is bigger for the pair A-F (mean RT: ASD = 641 ms; controls: 966 ms) than for H-D (mean RT: ASD = 558 ms; controls: 691 ms).

Conclusions:  

Our findings suggest that, although ASD relatives might not exhibit obvious deficits in the recognition of emotional FE, they nonetheless differ from controls in some aspects of FE processing, as evident when RTs are taken into account. The further observation that the effect is more compelling when Anger or Fear are involved might be interpreted as reflecting an automatic avoidance of negative facial affect signalled predominantly by the eyes.