Operant Conditioning of Facial Displays of Pain

Objective: The operant model of chronic pain posits that nonverbal pain behavior, such as facial expressions, is sensitive to reinforcement, but experimental evidence supporting this assumption is sparse. The aim of the present study was to investigate in a healthy population a) whether facial pain behavior can indeed be operantly conditioned using a discriminative reinforcement schedule to increase and decrease facial pain behavior and b) to what extent these changes affect pain experience indexed by self-ratings. Methods: In the experimental group (n = 29), the participants were reinforced every time that they showed pain-indicative facial behavior (up-conditioning) or a neutral expression (down-conditioning) in response to painful heat stimulation. Once facial pain behavior was successfully up- or down-conditioned, respectively (which occurred in 72% of participants), facial pain displays and selfreport ratings were assessed. In addition, a control group (n = 11) was used that was yoked to the reinforcement plans of the experimental group. Results: During the conditioning phases, reinforcement led to significant changes in facial pain behavior in the majority of the experimental group (p <.001) but not in the yoked control group (p >.136). Fine-grained analyses of facial muscle movements revealed a similar picture. Furthermore, the decline in facial pain displays (as.observed during down-conditioning) strongly predicted changes in pain ratings (R(2) = 0.329). Conclusions: These results suggest that a) facial pain displays are sensitive to reinforcement and b) that changes in facial pain displays can affect self-report ratings.