Decoding emotions in expressive music performances: A multi-lab replication and extension study

Abstract

With over 560 citations reported on Google Scholar by April 2018, a publication by Juslin and Gabrielsson (1996) presented evidence supporting performers' abilities to communicate, with high accuracy, their intended emotional expressions in music to listeners. Though there have been related studies published on this topic, there has yet to be a direct replication of this paper. A replication is warranted given the paper's influence in the field and the implications of its results. The present experiment joins the recent replication effort by producing a five-lab replication using the original methodology. Expressive performances of seven emotions (e.g., happy, sad, angry, etc.) by professional musicians were recorded using the same three melodies from the original study. Participants (N = 319) were presented with recordings and rated how well each emotion matched the emotional quality using a 0-10 scale. The same instruments from the original study (i.e., violin, voice, and flute) were used, with the addition of piano. In an effort to increase the accessibility of the experiment and allow for a more ecologically-valid environment, the recordings were presented using an internet-based survey platform. As an extension to the original study, this experiment investigated how musicality, emotional intelligence, and emotional contagion might explain individual differences in the decoding process. Results found overall high decoding accuracy (57%) when using emotion ratings aggregated for the sample of participants, similar to the method of analysis from the original study. However, when decoding accuracy was scored for each participant individually the average accuracy was much lower (31%). Unlike in the original study, the voice was found to be the most expressive instrument. Generalized Linear Mixed Effects Regression modelling revealed that musical training and emotional engagement with music positively influences emotion decoding accuracy.

The Musical Emotion Discrimination Task: A New Measure for Assessing the Ability to Discriminate Emotions in Music

“…As it is possible that the effects of training may only manifest at a measurable level as a result of extensive training, Lima and Castro (2011) argue that this could have played a role in the lack of a discernable effect. More recent studies investigating musical emotion decoding have uncovered a positive association between decoding performance and self-reported musical expertise, providing additional support for the influence of musical training (Taruffi et al, 2017; Akkermans et al, 2018). In spite of this, further investigation is required to delineate the relationship between musical training and recognition of non-verbal emotional expression.…”

“…EI is typically separated into two constructs for the purpose of measurement; ability EI, measured using cognitive ability tests, and trait EI, assessed via self-report methods (Petrides et al, 2004). In keeping with a recent study of emotion decoding in music (Akkermans et al, 2018), a self-report measure of trait EI was used within the current research. Differences in recognition of emotion within speech prosody have been linked to EI (Trimmer and Cuddy, 2008), signifying the potential influence of EI on musical emotion decoding, especially when considering the strong evidence for a link between the communication of emotions in speech and music (Juslin and Laukka, 2003).…”

“…Emotional music skills can be measured using the Goldsmiths Musical Sophistication Index (GOLD-MSI) (Müllensiefen et al, 2013), a self-report tool that allows for the assessment of a wide range of musical skills and behaviors. The Emotion subscale of the Gold-MSI subscale was used in a recent study, which found that self-reported level of emotional engagement with music predicted accuracy on a musical emotion decoding task (Akkermans et al, 2018). A high level of emotional engagement with music, as measured using the Gold-MSI, may therefore be a good indicator of ability to discriminate musical emotional expression.…”

“…The idea that musical expertise may enhance emotional skills seems plausible when taking into account other cognitive advantages found to result from training (Schellenberg, 2005). Accordingly, it has been suggested that enhanced musical and acoustic processing acquired through training (Kraus and Chandrasekaran, 2010) may contribute to an enhanced sensitivity to non-verbal emotions (Taruffi et al, 2017; Akkermans et al, 2018). Empirical evidence for this claim has been provided by studies conducted by Thompson et al (2004) and Lima and Castro (2011), both of which demonstrated that a group of musicians were better able to decode emotions in speech prosody than untrained controls.…”

“…The current research was inspired by a recent replication and extension (Akkermans et al, 2018) of a study carried out by Gabrielsson and Juslin (1996). The original study investigated communication of emotion in music using a production-recognition paradigm.…”