Effects of Autism on Emotional Sharing and Learning in the Anterior Cingulate

Background:

Impaired imitation in autism is well recognised but the developmental relationship between imitation and aspects of autism psychopathology, such as impaired empathy and shared emotion, is poorly understood. This is possibly because research has focussed on either manual imitation or emotional mimicry. We have created a task that blends facial stimuli expressing basic emotions, in which we systematically vary the relative quantity of each emotion being expressed. The task measures an individual’s ability to evidence accurate discrimination between emotionally expressive stimuli of varying similarity by expression through imitation. Task performance shows good correlations with empathic traits.

Objectives:

To determine the neural correlates of facial imitation ability and to determine whether these are disrupted in autism.

Methods:

27 participants with autism (aged 9-20 yrs, 2 females, IQ 93-133; diagnosis confirmed using ADI-R; current severity assessed with Social Responsiveness Scale [SRS]), matched on age, sex and IQ to 25 control participants undertook the facial imitation task. Of the participants who took part in the behavioural tasks, 17 participants with autism and 19 controls were able to complete the task during functional neuroimaging, in which they also completed a control task, viewing the same stimuli but instead enacting a previously learnt facial action in response to an instruction. In addition we utilised data from a previous study in which 20 adults undertook the same tasks during fMRI.

Results:

The ASD group should much lower levels of imitation accuracy on the behavioural task (ASD mean facial error=41.82, SD=5.97; control group error=36.07, SD=3.38; df=50; t=4.229; p< .001). Imitation accuracy across the whole group correlated with self-reported and parent-reported empathy quotient, as well as SRS score. In the fMRI study, using the whole group and conservative statistical thresholds (p<0.05 FWE corrected for cluster), facial imitation ability only correlated with activity in the genu of the anterior cingulate (ACCg) where activity was greater in the control task compared to imitation. Using this cluster as a region of interest we found significantly decreased activity in this area among the group with autism

Conclusions:

The ACCg serves to represent the value of actions to others and also signals differences between predicted and experienced action outcomes. We found that higher activity in this region when perception-action differences were greater predicted imitation accuracy and empathic traits. Low activity was associated with autism. We suggest that the ACCg plays a key role in the intentional control and learning of appropriate emotional responses to others’ emotionally communicative actions. This is required for development of empathy, shared emotion and the cultural transmission of behaviour according to its emotional value. A number of studies have shown abnormal anatomy and neurochemistry in this region, suggesting that the reduced prediction error signal in ACCg (and easily measured facial imitation ability) may provide a practical target for therapeutic intervention.