Facial mimicry is not modulated by dopamine D2/3 and opioid receptor antagonism

Abstract

Rationale: According to theories of embodied cognition, facial mimicry - the spontaneous, low-intensity imitation of a perceived emotional facial expression - is first an automatic motor response, whose accompanying proprioceptive feedback contributes to emotion recognition. Alternative theoretical accounts, however, view facial mimicry as an emotional response to a rewarding stimulus, and/or an affiliative signal, and thus reject the view of an automatic motor copy.

Objectives: To contribute to this debate and further investigate the neural basis of facial mimicry, as well as its relation to reward processing, we measured facial reactions to dynamic happy and angry faces after pharmacologically manipulating the opioid and dopamine systems - respectively, thought to subserve 'liking' and 'wanting' of rewards.

Methods: In a placebo-controlled, double-blind experiment, 130 volunteers received in a between-subjects design 50 mg of the opioidergic antagonist naltrexone, 400 mg of the dopaminergic antagonist amisulpride, or placebo.

Results: Clear occurrence of facial mimicry, measured 4 h after drug intake with electromyography (EMG) of the zygomaticus major and corrugator supercilii muscles, was found. However, facial mimicry was not affected by either compound, as shown with both frequentist statistics, and a Bayesian asymptotic regression model.

Conclusions: This null finding does not support the hypothesis that facial mimicry (of happiness) reflects an emotional response to a rewarding stimulus, leaving open the possibility of facial mimicry being an automatic motor copy. The results are relevant to the discussion about the psychological nature and the neural basis of facial mimicry, although they should be considered preliminary, given the challenges of interpreting null findings when targeting a novel effect of unknown size.

Keywords: Amisulpride; Dopamine; Electromyography; Embodiment; Facial mimicry; Naltrexone; Opioid.

Fig. 1

Facial EMG by muscle, time, and drug. During AngryToHappy trials (left), facial mimicry of smiles, as shown by ZM activation and CS deactivation, was somewhat greater in the placebo condition, but no significant differences between drug groups were found. Similarly, during HappyToAngry stimuli, facial mimicry of anger was present, as shown by CS activation and ZM deactivation, but was not modulated by either drug. Error bars represent SEM. CS, corrugator supercilii muscle; ZM, zygomaticus major muscle

Fig. 2

A An asymptotic regression line with a = 0.1, b = 0 and c = 0.25. The dashed line indicates the asymptote. B A prior prediction plot. Each line shows a sample from our prior. C Boxplots summarise each participant’s mean log EMG signal in each condition. The shaded region illustrates the 95% HPDI (highest posterior density interval) from an asymptotic regression model that ignores drug group information. Note: there are a number of outlier points that fall outside of this figure’s axis in both directions, and were omitted for clarity. CS, corrugator supercilii muscle; ZM, zygomaticus major muscle

Fig. 3

Posterior estimates for the fixed effects in our model. Remember that in our model rate of increase can only be positive, thus a greater rate for the ZM to HappyToAngry in the placebo group (lower right quadrant) should be interpreted as a greater relaxation of the muscle (towards the negative asymptote)