Embodied simulation as part of affective evaluation processes: Task dependence of valence concordant EMG activity

Abstract

Drawing on recent findings, this study examines whether valence concordant electromyography (EMG) responses can be explained as an unconditional effect of mere stimulus processing or as somatosensory simulation driven by task-dependent processing strategies. While facial EMG over the Corrugator supercilii and the Zygomaticus major was measured, each participant performed two tasks with pictures of album covers. One task was an affective evaluation task and the other was to attribute the album covers to one of five decades. The Embodied Emotion Account predicts that valence concordant EMG is more likely to occur if the task necessitates a somatosensory simulation of the evaluative meaning of stimuli. Results support this prediction with regard to Corrugator supercilii in that valence concordant EMG activity was only present in the affective evaluation task but not in the non-evaluative task. Results for the Zygomaticus major were ambiguous. Our findings are in line with the view that EMG activity is an embodied part of the evaluation process and not a mere physical outcome.

Keywords:

• EMG
• Embodied emotion
• Affective evaluation
• Simulation
• Valence

André Weinreich was supported by a fellowship of the International Max Planck Research School “The Life Course: Evolutionary and Ontogenetic Dynamics” (LIFE; www.imprs-life.mpg.de). We thank Peter Frensch, Torsten Schubert, Robert Gaschler and Sabine Schwager for valuable comments concerning earlier versions of the article, Werner Sommer, Thomas Pinkpank and Rainer Kniesche for providing technical support, and Kate Koennecke who helped in manuscript editing.

André Weinreich was supported by a fellowship of the International Max Planck Research School “The Life Course: Evolutionary and Ontogenetic Dynamics” (LIFE; www.imprs-life.mpg.de). We thank Peter Frensch, Torsten Schubert, Robert Gaschler and Sabine Schwager for valuable comments concerning earlier versions of the article, Werner Sommer, Thomas Pinkpank and Rainer Kniesche for providing technical support, and Kate Koennecke who helped in manuscript editing.

Notes

¹ Even though automatic evaluative processes have been shown to occur with mildly emotional stimuli (Giner-Sorolla, Garcia, & Bargh, 1999), we were highly interested in assuring that our stimuli were affectively efficient. In terms of a manipulation check, therefore, we applied the Affect Misattribution Procedure (AMP; Payne, Cheng, Govorun, & Stewart, 2005), which is widely used as an implicit measure of automatic stimulus valence (but see Bar-Anan & Nosek, 2012). Stimuli from the main experiment served as primes, while Asian ideographs were used as targets. One hundred and sixty trials were completed so that each prime was used four times while targets differed for each trial. Please see Payne et al. (2005) for a detailed account of the procedure.

² Furthermore, results of the Affect Misattribution Procedure are in line with the assumption that the stimuli used in the main part of the experiment were evaluated automatically. For the Affect Misattribution Procedure an ANOVA with individual valence (negative, neutral and positive) as repeated measures factor showed a significant effect of prime valence on the frequency of a “rather pleasant” response to the target, F _(2,40) = 12.541, p < .001, ηp² = .385. The direction of the effect was as expected. Pairwise LSD tests with Bonferroni correction revealed significant differences between negative and neutral primes, p = .033, negative and positive primes, p = .004 and neutral and positive primes, p = .008. This finding indicates that our stimuli were affectively efficient on the individual level.