Content uploaded by Joydeep Bhattacharya
Author content
All content in this area was uploaded by Joydeep Bhattacharya on Jul 26, 2019
Content may be subject to copyright.
GOLDSMITHS Research Online
Article (refereed)
Logeswaran, Nidhya and Bhattacharya, Joydeep
Crossmodal transfer of emotion by music
Originally published in Neuroscience LettersThe publisher's version is
available at: http://dx.doi.org/10.1016/j.neulet.2009.03.044
You may cite this version as: Logeswaran, Nidhya and Bhattacharya,
Joydeep, 2009. Crossmodal transfer of emotion by music. Neuroscience
Letters, 455 (2). pp. 129-133. ISSN 03043940 [Article]: Goldsmiths Research
Online.
Available at: http://eprints.gold.ac.uk/4213/
This document is the author’s final manuscript version of the journal article,
incorporating any revisions agreed during peer review. Some differences
between this version and the publisher’s version remain. You are advised to
consult the publisher’s version if you wish to cite from it.
Copyright © and Moral Rights for the papers on this site are retained by the
individual authors and/or other copyright owners.
http://eprints-gro.goldsmiths.ac.uk
Contact Goldsmiths Research Online at: lib-eprints@gold.ac.uk
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
Neuroscience Letters 455 (2009) 129–133
Contents lists available at ScienceDirect
Neuroscience Letters
journal homepage: www.elsevier.com/locate/neulet
Crossmodal transfer of emotion by music
Nidhya Logeswarana, Joydeep Bhattacharyaa,b,∗
aDepartment of Psychology, Goldsmiths College, University of London, London SE14 6NW, United Kingdom
bCommission for Scientific Visualization, Austrian Academy of Sciences, Vienna A1220, Austria
article info
Article history:
Received 31 October 2008
Received in revised form 3 March 2009
Accepted 11 March 2009
Keywords:
Music
Face
Emotion
Priming
ERP
abstract
Music is one of the most powerful elicitors of subjective emotion, yet it is not clear whether emotions
elicited by music are similar to emotions elicited by visual stimuli. This leads to an open question: can
music-elicited emotion be transferred to and/or influence subsequent vision-elicited emotional process-
ing? Here we addressed this question by investigating processing of emotional faces (neutral, happy and
sad) primed by short excerpts of musical stimuli (happy and sad). Our behavioural experiment showed
a significant effect of musical priming: prior listening to a happy (sad) music enhanced the perceived
happiness (sadness) of a face irrespective of facial emotion. Further, this musical priming-induced effect
was largest for neutral face. Our electrophysiological experiment showed that such crossmodal priming
effects were manifested by event related brain potential components at a very early (within 100 ms post-
stimulus) stages of neuronal information processing. Altogether, these results offer new insight into the
crossmodal nature of music and its ability to transfer emotion to visual modality.
© 2009 Elsevier Ireland Ltd. All rights reserved.
Music is often considered as the language of emotion and one of
the oldest held views is that music arises principally from human
communication—a performer delivers some message to a receptive
listener. This message is supposed to be an emotional one and this
emotional communication is postulated to be the principal pur-
pose of music [20]. In an extensive review of music performance
[9], the analysis of communication accuracy showed that profes-
sional music performers are able to communicate basic emotions
(e.g., happy, sad, anger) to listeners with an accuracy almost as
high as in facial and vocal expression of emotions. Further, there
is considerable empirical evidence supporting the statement that
emotion is an integral part of a musical experience (see Ref. [10] for
a review).
But are musically induced emotions similar to other emotional
experiences [23]? An early EEG study [4] demonstrated a character-
istic difference in cortical brain activationpatterns: positive musical
excerpts produced a more pronounced lateralisation towards the
left fronto-temporal cortices, whereas negative musical excerpts
produced a right fronto-temporal activation pattern. This early
result is supported by recent studies showing that left frontal areas
are involved with the processing of positive music and right frontal
areas with the negative music [1,8,16]. Similar frontal asymmetry is
well reported for the processing of affective visual stimuli [2,3].
∗Corresponding author at: Department of Psychology, Goldsmiths College, Uni-
versity of London, New Cross, London SE14 6NW, United Kingdom.
Tel.: +44 2079197334; fax: +44 2079197873.
E-mail address: j.bhattacharya@gold.ac.uk (J. Bhattacharya).
Therefore, it is reasonable to infer that there are some overlaps
between musical emotions and visual emotions.
But can these musically induced emotions arising through the
auditory channel influence our interpretation of emotions aris-
ing through other sensory channels (i.e. visual)? Research on
crossmodal integration of auditory and visual emotions [5] shows
that rating of affective information in one sensory modality can
be biased towards the direction of the emotional valence of
information in another sensory modality. Event-related-potential
(ERP) studies presenting emotionally congruent and incongru-
ent face–voice pairs reveal early ERP effects (N1, P2 components)
for congruent face–voice pairs, suggesting an early interaction
between auditory and visual emotional stimuli [15,22]. Therefore,
musical emotion can interact with visual emotion for simultaneous
music and visual processing.
But can musical emotion interact with or even influence the
visual emotion for non-simultaneous processes? In other words,
can music be used as an affective priming stimulus which could
systematically influence the emotional processes of target visual
stimuli? Music was earlier used as a priming stimulus in semantic
context [12,21]. To the best of our knowledge, the current study is
the first to address this issue in a crossmodal context by using both
behavioural and ERP experiments.
We performed two separate experiments – (i) behavioural and
(ii) electrophysiological (EEG) – on a total of 46 adult human par-
ticipants. Thirty participants (15 males and 15 females, mean age
26.1±4.31 years) took part in (i) without any cash incentive, and
sixteen participants (8 males and 8 females, mean age 27.5±5.88
years) took part in (ii) against a small cash incentive. All partici-
0304-3940/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.neulet.2009.03.044
Author's personal copy
130 N. Logeswaran, J. Bhattacharya / Neuroscience Letters 455 (2009) 129–133
Fig. 1. (a) Emotion ratings of happy, sad and neutral faces, regardless of musical primes. (b) Ratings for six individual conditions: happy, sad and neutral faces primed by
happy or sad musical excerpts. (c) Difference (happy −sad) in ratings for three facial emotions. Note that the largest effect was found for neutral facial emotion.
pants were healthy right-handed university students, had normal
hearing, normal or corrected-to-normal vision, and had no special
musical expertise or musical education. The study was conducted
in accordance with the Declaration of Helsinki, and was approved
by the Internal Ethics Committee at Goldsmiths College, University
of London. All participants gave informed written consent before
both experiments.
All musical stimuli were taken from a previous study [1]. Briefly,
there were 120 instrumental musical excerpts belonging to two
emotional categories: happy and sad. Each piece was played for
15s with both beginning and end faded in and out, respectively, to
minimize surprise. The visual stimuli were faces of 40 different indi-
viduals with each individual showing threetypes of facial emotions:
happy, sad and neutral (http://www.macbrain.org/stim/faces.htm).
There were 90 trials equally divided into six possible conditions
(2 musical emotions ×3 facial emotions). Each trial lasted for 16s,
where a 15-s musical excerpt was followed by a facial stimulus pre-
sented for 1 s. At the end of each trial, participants were required
to rate the facial emotion on a 7-point scale: 1= extremely sad,
2 = moderately sad, 3 = slightly sad, 4 = neutral, 5 = slightly happy,
6 = moderately happy, and 7 = extremely happy. Participants were
told to try and to feel the emotion of the musical stimuli and
to rely mainly on their feelings while they rated the facial stim-
uli.
The EEG study followed a similar procedure to the behavioural
study but with the following exceptions. There were 120 trials with
20 trials for each condition. Further, instead of an explicit emo-
tional evaluation, the participants were asked to press a button
whenever a female face was shown. This minimized the explicit
components of emotional processing, and the remaining differ-
ences, if any,would reflect the implicitness of emotional processing.
Trials were randomized within each block and across participants.
EEG signals were recorded from 28 Ag/AgCl scalp electrodes (Fp1,
Fp2, F3, F4, F7, F8, Fz, FC3, FC4, FCz, C5, C6, C3, C4, Cz, CP5, CP6, CP3,
CP4, CPz, P7, P8, P3, P4, Pz, O1, O2, Oz) according to the Interna-
tional 10/20-system. Horizontal and vertical eye movements were
recorded from electrodes placed around the eyes. Impedances were
kept below 5 k. All scalp electrodes were algebraically referenced
to the average of two earlobes. The sampling frequency was 500 Hz.
Perceived emotional ratings were assessed using factorial
repeated-measures analysis of variance (ANOVA) with the factors,
musical emotion (two levels: happy and sad) and facial emotion
(three levels: happy, sad and neutral). Further post hoc compar-
isons between pairs of specific conditions were carried out by using
Fig. 2. Grand average ERP profiles at 13 selected electrodes (see the electrode locations on the top) during processing of neutral facial stimuli primed by either happy music
(thick line) or sad music (thin line).
Author's personal copy
N. Logeswaran, J. Bhattacharya / Neuroscience Letters 455 (2009) 129–133 131
paired t-tests. Greenhouse–Geisser correction wasapplie d toadjust
for degrees of freedom.
The data pre-processing was carried out by EEGLAB [6]. Stim-
ulus epochs of 1500ms starting 500 ms before the face onset
were generated offline from the continuous data. Extracted epochs
were subsequently checked for artefacts by visual inspection. In
order to correct for ocular artefacts including eye blinks, Inde-
pendent Component Analysis was performed on the remaining
epochs. ERP was obtained by averaging artefact free epochs for
each of the six conditions. A series of 2 ×2 factorial repeated-
measures ANOVAs with factors, priming (happy vs. sad) and region
(as selected after scalp maps), were conducted on mean ERP
amplitudes at specific regions of interest (ROI) between any two
conditions with identical facial emotion but with different musical
primes. Temporal regions of interest were selected on the basis of
global field power (GFP) [13] which quantifies the instantaneous
global activity across the spatial potential fields. Spatial regions
of interest were selected on the basis of scalp maps of mean ERP
amplitudes at the pre-selected temporal region of interest. Across
statistical comparison, we found that spatial regions of interest
consisted of two levels—anterior and posterior. Instead of a data-
blind procedure of selecting region of interests on an ad-hoc basis,
this data-driven method selected only a few regions of interests,
thereby minimizing the error variance and maximizing the effect
size.
Fig. 1(a) shows the mean ratings of happy, neutral and sad
faces regardless of the type of musical primes. It was clear that
the facial emotions were rightly rated and classified by our par-
ticipants. Mean ratings for six conditions, separately, are shown
in Fig. 1(b). The happy faces when primed by happy music were
rated (6.13 ±0.36) higher (i.e. more happy) than when primed
by sad music (5.56 ±0.35), and the sad faces when primed by
sad music were rated (1.87±0.30) lower (more sad) than when
primed by happy music (2.44 ±0.30). Further, when the neutral
faces were primed by happy music, the rating (4.68±0.33) was
much higher than when it was primed by sad music (3.37±0.54).
Therefore, the differential effects of priming (happy−sad) were
similar for happy and sad faces (mean difference rating of 0.57)
but was almost doubled (mean difference rating of 1.31) for
neutral face (Fig. 1(c)). Repeated-measures ANOVA showed that
there were highly significant main effects for musical emotion
(F(1,29)= 103.29, p< 0.001) and facial emotion (F(2,28) = 1358.89,
p< 0.001). The music ×faces interaction effect was also found to
be highly significant (F(2,58) = 34.37, p< 0.001). These results show
that the effect of musical priming was largest for emotionally neu-
tral target stimuli.
ERPs were always compared between two conditions with same
facial emotion but with different priming, and same analysis pro-
cedure was applied for all three types of facial emotions. Since the
behaviour study showed largest effect for neutral faces, we strate-
gically emphasized the results for neutral faces in details as follows,
and the results for other two types of facial emotions will be briefly
mentioned.
Visual inspections revealed that ERP profiles for neutral faces
primed by happy music were markedlydif ferent from those primed
by sad music (Fig. 2). Enhanced N1 component was seen across all
frontal and central electrodes for happy, as opposed to sad, musi-
cal primes. The classical N170 face component was exhibited in
occipital and parietal (P7, P8, not shown) regions bilaterally for
both priming conditions. Between 180 and 250 ms, an increased
positivity (or reduced negativity) for happy primes as compared to
sad primes was noticed in frontal, central. At a later stage (300 ms
onward), posterior positivity was observed for both conditions. GFP
values were plotted in Fig. 3 (top panel) and the two profiles were
separated as early as from 50ms till 150ms, and then again for the
time period 190–210ms. Scalp maps at these time windows were
Fig. 3. Global field power values for three emotional facial stimuli: neutral face
(upper panel), happy face (middle panel), and sad face (lower panel). For each emo-
tion type, two types of musical priming, happy and sad, were shown in thick and
in thin lines, respectively. The high GFP values correspond to pronounced potential
fields with high peaks (both positive and negative)and steep gradients, whereas low
GFP values correspond to flat potential fields. Foreach facial emotion, time windows
where the two profiles showed maximal separation between two musical priming
were used for successive statistical analysis.
displayed in Fig. 4 for both conditions and their differences. As com-
pared to negative musical prime, positive musical prime showed
an enhanced negativity in frontal and central brain regions during
50–150ms and enhanced positivity during 190–210 ms in similar
anterior brain regions. For the N1 time window (50–150 ms), statis-
tical analysis showed a significant effect of priming (F(1,15)= 5.35,
p= 0.03), and a priming ×region interaction effect (F(1,15)= 8.62,
p= 0.01). For the later time window (190–210 ms), a significant
priming effect (F(1,15) = 4.66, p= 0.047) was found.
GFP plots for other two types of facial emotions were shown in
Fig. 3 (middle and lower panels). For happy facial stimuli, differ-
ences between happy and sad musical primes were found between
0–50 and 160–210ms. Statistical analysis of mean ERP ampli-
tudes showed a near significant priming effect during 160–210 ms
(F(1,15) = 4.33, p= 0.06) and an almost significant priming×region
interaction effect between 160 and 210 ms (F(1,15) = 4.16, p= 0.06).
For sad facial stimuli, the early (0–50ms) difference between
the two priming conditions was also found, but the later differ-
ences were found during 430–450 ms. Statistical analysis of mean
ERP amplitudes showed significant region effects, F(1,15)= 7.54,
p= 0.015 and F(1,15) = 32.29, p<0.001, for both time windows,
Author's personal copy
132 N. Logeswaran, J. Bhattacharya / Neuroscience Letters 455 (2009) 129–133
Fig. 4. The scalp distribution of mean ERP amplitudes at two different time windows (50–150 and 190–210ms) for target neutral faces primed by happy (left column) and
sad (middle column) musical stimuli. The right column shows the same but for the difference potentials (sad−happy). As compared to sad musical prime, happy musical
prime produced enhanced negativity at the first time window and enhanced positivity at the second time window, both over anterior brain regions.
0–50 ms and 430–450 ms, respectively, and a near significant effect
of priming (F(1,15) = 4.30, p= 0.06) was found at the later time win-
dow.
Our behavioural experiment confirmed that emotional rating
of the target facial stimuli could be biased towards the direction
of the emotional valence of the musical primes. Earlier research
[5,22] reported that emotions in auditory stimuli interactwith emo-
tions in visual stimuli for simultaneously presented auditory–visual
stimuli. But our result extends it further by showing that such
interaction could also occur for non-simultaneous processing, i.e.
when the emotional auditory stimuli precede the emotional visual
stimuli. In other words, priming musical stimuli can systematically
influence the perceived emotional contents in target visual stimuli.
Music was earlier used in semantic priming paradigms. Using chord
sequences, either consonant or dissonant, as priming stimuli, it was
shown [21] that target words are faster recognized for emotionally
congruent chord–word pairs than for incongruent ones. Further,
when target words are preceded by semantically unrelated musical
primes, N400, an ERP component reflecting contextual integration,
effect is reported [12]. Altogether, this suggests that music has an
excellent potential to be used as an emotional priming stimulus.
Our behaviour data also shows that the largest effect of musi-
cal priming was found for neutral faces with an effect size almost
twice those for happy or sad faces. It was shown earlier [14] that
as compared to emotionally distinct (i.e. happy, angry, sad) facial
stimuli, emotionally ambiguous (i.e. neutral) facial stimuli are more
likely to be influenced by auditory cues in a facial emotion detec-
tion task. The information-content of neutral faces are supposedly
lower than those of happy or sad faces, and since the brain relies on
cues from multiple sensory stimuli to create an optimal represen-
tation of the external environment, emotionally neutral stimuli is
being influenced by emotionally conspicuous stimuli, even though
being generated by different senses. Although in our paradigm,
there is no such explicit requirement of integration of informa-
tion across musical and visual stimuli, our findings suggest that
a generic mechanism of multimodal affective interaction might
exist also in a priming paradigm. Alternatively, this could also be
explained by the affect-as-information hypothesis [19] which relies
on the assumption that affective processes mainly occur outside
of awareness. In contrast to traditional assumption of judgement
and decision making which emphasizes the role of target related
features, this hypothesis states that when making evaluative judge-
ments, participant often ask themselves, “How do I feel about it?”
[18], and therefore, “they may mistake feelings due to a pre-existing
state as a reaction to the target” [17]. Since the emotionally neutral
facial stimuli contain less information than emotionally conspic-
uous (happy or sad) facial stimuli, the transient affect from the
priming musical stimuli has the maximal impact in determining
the evaluation of the neutral facial stimuli.
Our ERP data showed that for neutral facial emotion, happy
music, as compared to sad music, showed a significant effect
during the N1 time period (50–150ms). Earlier Pourtois et al.
[15] have found that simultaneous presentation of emotionally
congruent face–voice pairs produce an enhancement of auditory
N1 component as compared to incongruent pairs, suggesting an
early crossmodal binding. In contrast to this study which reported
enhancement over auditory cortex, our N1 effect was predomi-
nant over fronto-central and midfrontal regions (FC3/4, FCz, Fz, Cz).
Taken together, this suggests that happy or positive auditory emo-
tion is more likely to influence neutral visual emotion by engaging
brain regions responsible for top-down projections.
ERP results also showed priming related enhancement of P2
(190–210ms) component for happy and neutral target faces but
not for sad target faces. Similar modulation of P2 has also recently
been reported [22] for happy picture–voice pairs presented simul-
taneously but not for sad pairs. Enhanced positivity at similar time
window has also been found for processing isolated emotional faces
as compared to neutral faces [7]. However, the functional role of
P2 is not yet clear (but see Ref. [22] for some possible explana-
tions) in mediating interaction between priming musical stimuli
and emotionally selective (happy and neutral, but not sad) target
faces.
Finally, let us offer a few practical remarks. First, the current
paradigm of music-induced emotional priming is quite different
from other mood-induction procedures which are associated with a
longer lasting change of emotional states, whereas our study inves-
tigated the effect of emotional changes on a much shorter time
scale [11]. Secondly, unlike previous ERP studies of facial emotion
Author's personal copy
N. Logeswaran, J. Bhattacharya / Neuroscience Letters 455 (2009) 129–133 133
processing, we alwayscompared the same facial emotional type but
differed only in priming. Therefore, our results indicate a more sub-
tle component in early neural responses which can potentially bias
subsequent emotional evaluation occurring at later stages. This was
also manifested by our robust behavioural findings which called
for an explicit evaluation of facial emotions. Thirdly, our ERP data
primarily reflects an implicit emotional processing since the par-
ticipants were naïve with respect to the specific aims of the study.
Further, as the task of gender detection did not require the par-
ticipants to focus on the presented emotions, the results are less
likely to be attributed to differences in directed attention as a func-
tion of presented emotions. Therefore, the ERP results suggest an
early processing of emotional facial expression primed by musical
stimuli.
In summary, the results of our behavioural and ERP study
revealed some patterns of crossmodal influences by musical emo-
tion. Behavioural data clearly showed that listening to musical
excerpts, albeit short, could significantly influence the subsequent
explicit evaluation of visual emotional stimuli. ERP data showed
that such musical priming could also influence implicit visual emo-
tional processes.
Acknowledgements
The study was supported by JST.ERATO Shimojo project (JB). We
are thankful to Prof. Eckart Altenmüller for the music stimuli, to
Rob Davis for technical support, to Job Lindsen for help in data
pre-processing, and to Prof. Rolf Reber for his helpful comments
as a reviewer. Author contributions: J.B. conceived research; N.L.
collected data; J.B. and N.L. analyzed data and wrote the paper.
References
[1] E. Altenmuller, K. Schurmann, V.K. Lim, D. Parlitz, Hits to the left, flops to the
right: different emotions during listening to music are reflected in cortical
lateralisation patterns, Neuropsychologia 40 (2002) 2242–2256.
[2] T.Canli, J.E. Desmond, Z. Zhao, G. Glover, J.D.E. Gabrieli, Hemispheric asymmetry
for emotional stimuli detected with fMRI, Neuroreport 9 (1998) 3233–3239.
[3] R.J. Davidson, Anterior cerebral asymmetry and the nature of emotion, Brain
and Cognition 20 (1992) 125–151.
[4] R.J. Davidson, G.E. Schwartz, C. Saron, J. Bennett, D.J. Goleman, Frontal versus
parietal EEG asymmetry during positive and negative affect, Psychophysiology
16 (1979) 202–203.
[5] B. de Gelder, J. Vroomen, The perception of emotions by ear and by eye, Cogni-
tion & Emotion 14 (2000) 289–311.
[6] A. Delorme, S. Makeig, EEGLAB: an open source toolbox for analysis of single-
trial EEG dynamics including independent component analysis, Journal of
Neuroscience Methods 134 (2004) 9–21.
[7] M. Eimer, A. Holmes, Event-related brain potential correlates of emotional face
processing, Neuropsychologia 45 (2007) 15–31.
[8] E.O. Flores-Gutierrez, J.L. Diaz, F.A. Barrios, R. Favila-Humara, M.A. Guevara, Y.
del Rio-Portilla, M. Corsi-Cabrera, Metabolic and electric brain patterns during
pleasant and unpleasant emotions induced by music masterpieces, Interna-
tional Journal of Psychophysiology 65 (2007) 69–84.
[9] P.N.Juslin, P. Laukka, Communication of emotions in vocalexpression and music
performance: different channels, same code? Psychological Bulletin 129(2003)
770–814.
[10] P.N. Juslin, J.A. Sloboda (Eds.), Music and Emotion, Oxford University Press,
Oxford, 2001, p. 487.
[11] P.N. Juslin, D. Vastfjall, Emotional responses to music: the need to consider
underlying mechanisms, The Behavioral and Brain Sciences 31 (2008) 559–575
(discussion 575–621).
[12] S. Koelsch, E. Kasper, D. Sammler, K. Schulze, T. Gunter, A.D. Friederici, Music,
language and meaning: brain signatures of semantic processing, Nature Neu-
roscience 7 (2004) 302–307.
[13] D. Lehmann, W. Skrandies, Reference-free identification of components of
checkerboard-evoked multichannel potential fields, Electroencephalography
and Clinical Neurophysiology 48 (1980) 609–621.
[14] D.W. Massaro, P.B. Egan, Perceiving affect from the voice and the face, Psycho-
nomic Bulletin & Review 3 (1996) 215–221.
[15] G. Pourtois, B. de Gelder, J. Vroomen, B. Rossion, M. Crommelinck, The time-
course of intermodal binding between seeing and hearing affectiveinformation,
Neuroreport 11 (2000) 1329–1333.
[16] D.L.Santesso, L.A. Schmidt, L.J. Trainor, Frontal brain electrical activity(EEG) and
heart rate in response to affective infant-directed (ID) speech in 9-month-old
infants, Brain and Cognition 65 (2007) 14–21.
[17] N. Schwarz, Feelings as information: informational and motivational functions
of affective states, in: E.T. Higgins, R. Sorrentino (Eds.), Handbook of Motivation
and Cognition: Foundations of Social Behaviour, vol. 2, Guildford Press, New
York, 1990, pp. 527–561.
[18] N.Schwarz, G.L. Clore, How do I feel about it? The informative function of mood,
in: K. Fiedler,J. Forgas (Eds.), Affect, Cognition and Social Behaviour, C.J. Hogrefe,
Toronto, 1988, pp. 44–62.
[19] N. Schwarz, G.L. Clore, Mood, misattribution, and judgments of well-
being—informative and directive functions of affective states, Journal of
Personality and Social Psychology 45 (1983) 513–523.
[20] M. Serafine, Music as Cognition: The Development of Thought in Sound,
Columbia University Press, New York, 1988.
[21] B. Sollberger, R. Reber, D. Eckstein, Musical chords as affective priming context
in a word-evaluation task, Music Perception 20 (2003) 263–282.
[22] K.N. Spreckelmeyer, M. Kutas, T.P. Urbach,E. Altenmuller, T.F. Munte, Combined
perception of emotion in pictures and musical sounds, Brain Research 1070
(2006) 160–170.
[23] L.J. Trainor, L.A. Schmidt, Processing emotionsinduce d bymusic, in: I. Peretz, R.
Zatorre (Eds.), The Cognitive Neuroscience of Music, OxfordUniversity of Press,
Oxford, 2007, pp. 310–324.
