Research Article: Memory for facial expression is influenced by the background music playing during study

Date Published: August 21, 2012

Publisher: University of Finance and Management in Warsaw

Author(s): Michael R. Woloszyn, Laura Ewert.


The effect of the emotional quality of study-phase background music on subsequent
recall for happy and sad facial expressions was investigated. Undergraduates
(N = 48) viewed a series of line drawings depicting a happy
or sad child in a variety of environments that were each accompanied by happy or
sad music. Although memory for faces was very accurate, emotionally incongruent
background music biased subsequent memory for facial expressions, increasing the
likelihood that happy faces were recalled as sad when sad music was previously
heard, and that sad faces were recalled as happy when happy music was previously
heard. Overall, the results indicated that when recalling a scene, the emotional
tone is set by an integration of stimulus features from several modalities.

Partial Text

Moviemakers have long known about the effect that a stirring piece of music can have
on one’s perception of a scene. Testifying to the power of this phenomenon is
the fact that the use of music to convey information and set a mood for a film
predates the use of sound in a movie altogether (Cook, 1990). Depending on the emotional quality of the melody, music can
be used to bias a viewer’s expectations of plot development (Vitouch, 2001), to influence the perceived
emotional content of otherwise valence-neutral film clips (Eldar, Ganor, Admon, Bleich, & Hendler, 2007) and static
pictures (Spreckelmeyer, Kutas, Urbach,
Altenmüller, & Münte, 2006), or possibly to generate an
emotional response from the observer (Goldstein,
1980; Steinbeis, Koelsch, & Sloboda,
2006), although this last assertion is not without its critics. As Koneni
(2008) somewhat controversially pointed
out, at least some readers (and researchers) appear to confuse the ability of a
piece of music to represent an emotion with that same piece evoking one (e.g., Sloboda & Lehmann, 2001), and that the
available evidence in favour of short snippets of music leading to a change in the
emotional state of the participant is less prevalent than one might think (see Koneni, 2008, for a detailed review). Despite
these concerns, however, most theorists and researchers would agree that music as a
stimulus has the ability to represent and, at least under certain conditions, to
evoke measureable emotional responses in listeners in the form of changes in skin
conductance, respiration, heart rate, or self-report measures (Baumgartner, Esslen, & Jäncke, 2006; Kreutz, Ott, Teichmann, Osawa, & Vaitl,

Participant responses were transformed into proportion correct for each trial type,
and submitted to a 2 (facial expression: happy, sad) × 3 (musical
accompaniment: no music, happy music, sad music) × 2 (presentation order)
mixed-design ANOVA. Figure 2 depicts the means
and 95% confidence intervals for each condition, collapsed across order (the
presentation order variable did not enter into any significant effect or

The results of the present study support the hypothesis that the valence of music
presented during study of visual stimuli can affect subsequent recall for facial
expressions. This is apparent in the obtained pattern of errors, in that happy faces
paired with sad music resulted in fewer “happy” responses relative to
the same faces being previously paired with happy music (and vice versa for sad
faces). Additionally, it is interesting that the obtained pattern of results does
not conform to what one would expect based on a combination of the bulk of research
that indicates an advantage of happy over sad music and faces described in the
Introduction. Specifically, it appeared that contrary to the expectation of happy
music and faces in conjunction being associated with the highest accuracies, and sad
music and faces being associated with the lowest, we found that when the two
modalities were in concert with one another, the accuracies were higher than when
they were in conflict. Furthermore, such a pattern would not be expected based on a
mood-arousal mechanism (Husain et al, 2002)
which would predict an overall improvement in memory for faces previously paired
with happy but not sad music. Such a main effect did occur. However, the interaction
does not fit with Husain et al.’s explanation, in that the happy music only
produced a significant benefit when a happy face was viewed. As a result, the
question remains as to what mechanism might be responsible for this phenomenon.




Leave a Reply

Your email address will not be published.