theeffectofmusicalexperienceonemotionalself-reports and ... - … · 2016. 6. 30. · norris,...
TRANSCRIPT
-
The effect ofmusical experience on emotional self-reports
and psychophysiological responses to dissonance
DELPHINE DELLACHERIE,a,b MATHIEU ROY,c LAURENT HUGUEVILLE,d,e,f
ISABELLE PERETZ,c and SÉVERINE SAMSONa,baLaboratoire deNeurosciences Fonctionnelles et Pathologies, CNRS-UMR8160, University of Lille-Nord de France, Lille, FrancebEpilepsy Unit, La Salpêtrière Hospital, Paris, FrancecBRAMS Laboratory, University of Montréal, Montréal, Québec, CanadadUniversité Pierre et Marie Curie-Paris 6, Centre de Recherche de l’Institut du Cerveau et de la Moelle Epinière, UMR-S975, Paris, FranceeInserm, U975, Paris, FrancefCNRS, UMR 7225, Paris, France
Abstract
To study the influence of musical education on emotional reactions to dissonance, we examined self-reports and
physiological responses to dissonant and consonant musical excerpts in listeners with low (LE: n5 15) and high (HE:
n5 13) musical experience. The results show that dissonance induces more unpleasant feelings and stronger phys-
iological responses in HE than in LE participants, suggesting thatmusical education reinforces aversion to dissonance.
Skin conductance (SCR) and electromyographic (EMG) signals were analyzed according to a defense cascade model,
which takes into account two successive time windows corresponding to orienting and defense responses. These
analyses suggest that musical experience can influence the defense response to dissonance and demonstrate a powerful
role of musical experience not only in autonomic but also in expressive responses to music.
Descriptors: Emotion, Music, Dissonance, Musical experience, Psychophysiology, SCR, EMG, HR
Since Helmholtz (1877), musical dissonance has been considered
by psychologists and neuroscientists as a puzzling topic at the
interface between auditory perception and emotional processing.
The simplest definition of dissonance refers to an unpleasant
sensation induced by the simultaneous presentation of two
sounds. As proposed by Pythagoras two and half millennia ago,
this affective sensation produced by a pair of tones is related to
the ratio of the fundamental frequencies of the sounds that are
played together. When the ratio is simple, consonance is heard
(e.g., octave 2/1; perfect fifth 3/2) and when the ratio is more
complex, dissonance occurs (e.g., tritone chord, 45/32) (Plomp&
Levelt, 1965; Tramo, Cariani, Delgutte, & Braida, 2001). This
physical component of dissonance results in a perception of
beating that is perceived as unpleasant and is referred to as
‘‘sensory’’ dissonance, which appears when chords are played in
isolation (Tramo et al., 2001). This phenomenon should be dis-
tinguished from ‘‘musical’’ dissonance, which refers to disso-
nance manipulated by composers in a harmonic or melodic
context in order to induce tension and expectation effects
(Meyer, 1956). In the present study, we did not consider musi-
cal dissonance but focused rather on sensory dissonance.
According to neuropsychological, electrophysiological, and
brain imaging studies, it has been suggested that perceptual and
emotional processing of dissonance depends on distinct cerebral
structures. Whereas perception of dissonance requires the func-
tion of auditory cortical areas within the superior temporal gyrus
(Fishman, Volkov, Noh, Garell, Bakken, et al., 2001; Peretz,
Blood, Penhune, & Zatorre, 2001), its emotional processing in-
volves various brain structures including mesial temporal lobe
and precuneus regions (Blood, Zatorre, Bermudez, & Evans,
1999; Gosselin, Samson, Adolphs, Noulhiane, Roy, et al., 2006;
Koelsch, Fritz, Von Cramon, Muller, & Friederici, 2006).
At the behavioral level, it is well known that dissonant chords
induce more negative valence judgments than consonant chords
(Blood et al., 1999; Brattico, Pallesen, Varyagina, Bailey, Anour-
ova, et al., 2009; Gosselin et al., 2006; Khalfa, Roy, Rainville,
Dalla Bella, & Peretz, 2008; Koelsch et al., 2006; Koelsch, Remp-
pis, Sammler, Jentschke,Mietchen, et al., 2007; Pallesen, Brattico,
Bailey, Korvenoja, Koivisto, et al., 2005; Passynkova, Neubauer,
& Scheich, 2007; Peretz et al., 2001; Sammler, Grigutsch, Fritz, &
Koelsch, 2007; Schoen, Regnault, Ystad, & Besson, 2005). Even
infants prefer consonance to dissonance, suggesting a natural
The authors are grateful to Laurence Conty, Daniela Sammler, Séve-
rine Farley, and SeanHutchins for their helpful assistance and comments
on previous versions of the manuscript. This study was supported by a
PhD scholarship from the Regional Council of Nord-Pas de Calais to
Delphine Dellacherie and by a grant from ‘‘Agence Nationale pour la
Recherche’’ of the French Ministry of Research (project no. NT05-3-
45987) to Séverine Samson and from Eisai Inc. Isabelle Peretz is
supported by grants from the Canada Research Chair in Neurocognition
of Music and Mathieu Roy by a Canadian fellowship from NSERC.Address correspondence to: Séverine Samson, Department of Psy-
chology, Université de Lille 3, BP 60 149, 59653 Villeneuve d’AscqCedex, France. E-mail: [email protected]
Psychophysiology, ]]] (2010), 1–13. Wiley Periodicals, Inc. Printed in the USA.Copyright r 2010 Society for Psychophysiological ResearchDOI: 10.1111/j.1469-8986.2010.01075.x
1
mailto:[email protected]
-
aversion to dissonance very early in life (Masataka, 2006; Trainor
& Heinmiller, 1998; Trainor, Tsang, & Cheung, 2002; Zentner &
Kagan, 1996). However, it has also been suggested that emotional
responses to dissonance can be influenced by musical experience.
As reported by Pallesen et al. (2005) and Schoen et al. (2005),
adults with musical experience considered dissonant chords as
more unpleasant than did musically untrained individuals. These
self-reported results suggest that aversion to dissonance could in
some instances be reinforced by formal musical training and pos-
sibly by learning mechanisms. Further lines of evidence suggest
that, at the neurophysiological level, musical experience could fa-
cilitate the processing of dissonance in musicians in comparison to
nonmusicians (Brattico et al., 2009; Minati, D’Incerti, Pietrocini,
Valentini, Scaioli, et al., 2009; Regnault, Bigand, & Besson, 2001;
Schoen et al., 2005). However, the effect of musical education on
emotional responses to dissonance is not a well-established phe-
nomenon (Brattico et al., 2009).
To test the effect of musical experience in emotional responses
to dissonance, we recorded psychophysiological correlates of
musical listening, which is an appropriate method to examine
objectivemeasures of emotional experiences (Bradley, Codispoti,
Cuthbert, & Lang, 2001; Bradley & Lang, 2000; Lang, Bradley,
& Cuthbert, 1998; Sanchez-Navarro, Martinez-Selva, & Ro-
man, 2006; Witvliet & Vrana, 1995). Several lines of evidence
show that physiological changes in skin conductance responses
(SCRs), heart rate (HR) and electromyographic (EMG) re-
sponses vary systematically with judgments of affective valence
or arousal to emotional stimuli, although evidence in the musical
domain is more mixed.
In general, SCR and HR responses that are indexes of au-
tonomic activity are shown to be higher for negative than for
positive emotion (Cacioppo, Bernston, Klein, & Poehlmann,
1997; Cacioppo, Bernston, Larsen, Poehlmann, & Ito, 2000;
Taylor, 1991). For example, by comparing the results obtained in
13 physiological studies, Cacioppo et al. (2000) found that heart
rate was significantly greater during negative than positive emo-
tions. The authors interpreted this effect as part of a ‘‘negativity
bias,’’ which was defined in behavioral studies as ‘‘the greater
impact of negative information on people’s evaluations than
comparably extreme positive information’’ (Peeters &Czapinski,
1990). Classical animal and behavioral studies (for review, see
Cacioppo et al., 2000; Taylor, 1991) as well as more recent phys-
iological studies (Ito, Larsen, Smith, & Cacioppo, 1998) inves-
tigated the concept of ‘‘negativity bias,’’ which can be considered
as a normalmanifestation of adaptive functioning. Cacioppo and
colleagues (Cacioppo & Bernston, 1994; Cacioppo, Gardner, &
Berntson, 1997) have then incorporated the negativity bias in a
more general model of evaluative space in which positive and
negative evaluative processes are assumed to involve separate
motivational substrates (Lang et al., 1998). In this context, neg-
ativity bias refers to a tendency for the negative motivational
system to respond more intensely than the positive motivational
system to comparable amounts of activation (Ito, Cacioppo, &
Lang, 1998; Ito et al., 1998). This is the definition that we
adopted in the present study, and we hypothesized that this neg-
ativity bias exists also with music.
In the musical domain, several studies have focused on the
effect of various musical variables on autonomic responses to
emotional stimuli (Baumgartner, Esslen, & Jancke, 2006;
Chapados & Levitin, 2008; Grewe, Nage, Kopiez, & Altenmull-
er, 2007; Khalfa, Peretz, Blondin, & Robert, 2002; Khalfa, Roy,
et al., 2008; Koelsch, Kilches, Steinbeis, & Schelinski, 2008;
Krumhansl, 1997; Steinbeis, Koelsch, & Sloboda, 2006; Witvliet
& Vrana, 2007), but only three studies specifically explored the
effect of musical valence on autonomic responses (Nater,
Abbruzzese, Krebs, & Ehlert, 2006; Roy, Mailhot, Gosselin,
Paquette, & Peretz, 2008; Sammler et al., 2007). A greater HR
deceleration in response to aversive (heavy metal) music was re-
ported by Nater et al. (2006). Sammler et al. (2007) also found a
significant decrease of HR induced by dissonant music in com-
parison to consonant music. This cardiac deceleration does not
fit with the classic cardiac defense response obtained with aver-
sive loud noises, which is mainly characterized by a late accel-
eration (Cook & Turpin, 1997). However, when unpleasant
pictures or sounds are presented, a marked cardiac deceleration
has been repeatedly reported (Bradley et al., 2001; Cook & Tur-
pin, 1997; Sanchez-Navarro, Martinez-Selva, & Roman, 2006).
Such a cardiac deceleration is characteristic of an orienting re-
sponse (Bradley et al., 2001; Cook & Turpin, 1997). Following
Cacioppo et al. (1997, 2000), this greater response to unpleasant
than to pleasant affective stimuli could be interpreted as resulting
from the negativity bias. According to these results, we could
predict a higher HR deceleration in response to dissonant stimuli
than to consonant stimuli.
Although SCR has been traditionally linked to variation of
arousal (Bradley et al., 2001; Bradley & Lang, 2000), some ev-
idence suggests that SCR duration and amplitude can also be
influenced by emotional valence (Cacioppo et al., 1997, 2000;
Norris, Larsen, & Cacioppo, 2007; Ohman & Wiens, 2003). In
the musical domain, Baumgartner et al. (2006) and Nater et al.
(2006) found that SCR was more elevated for unpleasant or
negatively valenced music than for pleasant or positively va-
lenced music. Therefore, SCR could be larger for dissonant than
for consonant music.
According to the defense cascade model (Bradley et al., 2001;
Lang, Bradley, & Cuthbert, 1997; Ohman &Wiens, 2003), emo-
tional autonomic responses to aversive stimuli display a two-step
dynamic characterized by an initial orienting response followed
by a subsequent defense response. This model describes an aver-
sive motivational circuit that triggers reactions ranging from
orienting to fight/flight. The orienting response elicits a large
SCR response as well as a HR deceleration in an early time
window and is part of an automatic lower level appraisal that
could be processed in a short period of time, without any ap-
parent implication of cortical association areas or of awareness
(Ohman & Soares, 1994; Ohman & Wiens, 2003). The defense
response defined as a controlled appraisal of the emotion is re-
flected by further increase of SCR and HR acceleration in a
subsequent time window. In addition to these two steps of re-
sponse, HR responses to emotional stimuli typically show a third
decelerative component that corresponds to the return to base-
line. Because this third component is not part of the affective
response per se, we focused here on the two first components of
the HR response. Here, we hypothesized that negative emotions
created by musical dissonance should elicit a similar pattern with
two steps of responses (an orienting response, followed by a
defense response).
EMG responses of facial expression provide another index of
somatic activity related to emotional valence (Cacioppo, Klein,
Bernston, & Hatfield, 1993). More specifically, corrugator ac-
tivity (used in frowning) is increased for unpleasant or negatively
valenced stimuli (pictures: Bradley et al., 2001; mental imagery:
Witvliet & Vrana, 1995; sounds: Bradley & Lang, 2000; films:
Ellis & Simons, 2005). Conversely, zygomatic activity increases
2 D. Dellacherie et al.
-
with pleasantness (Ellis & Simons, 2005; Lang et al., 1998;
Witvliet & Vrana, 1995). Although the corrugator activity in
response to emotional valence inmusic has been already reported
(Roy et al., 2008; Witvliet & Vrana, 2007), no study to our
knowledge has recorded corrugator and zygomatic EMGactivity
in response to dissonance. Based on previous results in non-
musical domains, we can hypothesize that corrugator muscles
known to be influenced by negative valence should be more ac-
tivated when listening to dissonance than to consonance. How-
ever, zygomatic muscles generally modulated by positive valence
should be more activated in response to consonance than to dis-
sonance. Since no study interpreted facial EMG responses in
relation to the defense cascade model, no specific prediction
about the dynamic changes of such responses was proposed.
Finally, it has also been demonstrated that automatic phys-
iological responses to aversive stimuli can be modulated by our
own experience with the material. Some evidence in a non-mu-
sical domain suggests that the relevance of the task for a given
subject could modulate SCRs. For example, parachutists exhibit
larger skin conductance orienting response to parachutist-rele-
vant words and pictures than do inexperienced controls (Epstein
& Fenz, 1962; Fenz & Epstein, 1962). Since musical experience
acquired through lessons or practice can influence the emotional
response to dissonance, it seems relevant to explore the effect of
musical experience on emotional reaction to dissonance by com-
bining subjective rating judgments with objective autonomic
(SCR and HR) and somatomotor (EMG) measurements in re-
sponse to musical listening.
For this purpose, consonant musical excerpts and their dis-
sonant counterparts of the same musical pieces matched for
arousal were presented to participants with low and highmusical
experience that were carefully controlled in terms of age, sex,
education, and mood. Subjective ratings of emotional valence
(according to pleasantness) as well as autonomic (i.e., HR, phasic
SCR) and somatomotor (i.e., facial EMG) indexes related to
valence judgments were simultaneously recorded. We predicted
that dissonance will induce more negative valence judgments
than consonance, and this effect would be larger in high than in
low musically experienced participants. For all the participants,
we also predicted higher SCR and larger HR deceleration in
response to dissonance as compared to consonance, these effects
being modified by musical experience as well. Based on the de-
fense cascade model with two-step responses, two successive
patterns of autonomic responses were expected as a function of
the time window, early and late time windows corresponding to
orienting and defense responses, respectively. Therefore, we can
suppose that SCR and HR can be modified by musical experi-
ence differently in early and late time window. We also predicted
that facial EMG would be modulated by emotional valence: co-
rrugator muscles known to be influenced by negative valence
should be more activated when listening to dissonance than to
consonance, whereas zygomatic muscles generally modulated by
positive valence should be more activated in response to conso-
nance than to dissonance.
Methods
Participants
Twenty-seven participants (13 men and 14 women), aged be-
tween 19 and 31 years (mean age5 29.41. SD � 8.53) took part
in this study. Based on their responses on a musical experience
questionnaire, participants were sorted into Low Experience
(LE, n5 15) and High Experience (HE, n5 12) groups. The
musical experience questionnaire included items related to their
music listening habits (listening subscale) as well as their level of
musical education and actual practice (practice subscale), in ac-
cordance with a multi-dimensional definition of musicianship
(Ehrle, 1998). All HE participants, except one who was an au-
todidact, received training in classical music, and none of them
reported listening to music genres which generally have a lot of
dissonance, such as contemporary music or free jazz. As dis-
played in Table 1, the mean scores on both subscales were higher
for HE (mean listening score5 6, SD5 1.65 and mean practice
score5 6.92, SD5 2.57) than LE (mean listening score5 3.87,SD5 1.25; mean practice score5 3.87, SD5 1.04) participants(t(25)5 3.829, po.005 for listening and t(25)5 9.16, po.001 forpractice). Finally, the two groups did not differ on variables
unrelated to musical experience, such as age (mean age for
LE5 31.2, (SD5 9.08); for HE5 27.2 (SD5 2.18);
t(25)5 0.94, n.s.), years of education (mean years of educationfor LE5 14.33 (SD5 3.24); for HE5 15.83 (SD5 2.25);t(25)5 0.92, n.s.), and gender (for LE: 9M, 6F; for HE: 4M,
8F; w2(1)5 1.90, n.s). The results remain the same when onlyparticipants kept in the SCR and HR analyses were included in
the analyses.
Mood Questionnaires
Twomood questionnaires were administered: the State and Trait
Anxiety Inventory (STAI; Spielberger, 1983) and the Profile of
Mood Scales (POMS;McNair et al., 1992). The STAI comprises
two subscales, one assessing the individual general level of anx-
iety (i.e., trait anxiety), which is presumed to be stable over time,
and the other one assessing the present level of anxiety (i.e., state
Psychophysiology of musical emotion 3
Table 1. Description of the Musical Experience of the Two Groups
of Participants with Low (LE) and High Experience (HE)
Lowexperience
(LE)
Highexperience
(HE)
(A) Scores on the musical experience questionnaireGlobal score 4.13 (1.25)n 12.91 (3.21)n
Listening subscale score 3.87 (1.24)n 6 (1.65)n
Practice subscale score 3.87 (1.03)n 6.92 (2.57)n
(B) % of the participants reportingListening HabitsListen music every day 80 91.60Listen music with attention 20 66.60Listen classical music 33 58.30Go to concert 46.60 75
PracticePractice actually 0 33Have received an institutional training 6.60 75Have received lessons during at least 3years
6.60 92
Self-educated 0 8.30(C) Details on practice for HE
Age at start of study – 8.54 (3.56)Duration of the practice – 8.41 (4.30)
Notes: (A) Mean scores on the musical experience questionnaire (stan-dard deviation in parentheses). (B) Percentage of participants reportingdifferent listening habits and musical practice details. (C) For HE par-ticipants, mean age at start of study andmean number of years ofmusicalpractice (standard deviation in parentheses).npo.05.
-
anxiety). The POMS is a questionnaire comprising 30 items, each
consisting of an emotional adjective for which subjects have to
rate to what extent it describes their current mood (0F‘not atall,’ 4F‘extremely’). Six emotional sub-scales can be derivedfrom the 30 items: Tension, Depression, Anger, Vigor, Fatigue,
and Confusion.
Stimuli
Tenmusical excerpts of classical music were taken from a set used
in previous studies (Gosselin et al., 2006; Khalfa, Guye, Peretz,
Chapon, Girard, et al., 2008; Peretz et al., 2001; Peretz, Gagnon,
& Bouchard, 1998). They were instrumental in that they were not
originally sung with lyrics. The excerpts were selected to evoke
happiness and were played in major mode (e.g., ‘‘Brindisi’’ from
Verdi’s ‘‘Traviata’’), with a fast tempo (the quarter note value
varies from 80 to 255, conventionally written by M.M. for Met-
ronome Marking). All stimuli were transcribed for piano, com-
puter-generated, and delivered with a piano timbre. Each excerpt
had duration of 7 s. The dissonant versions of each original
consonant excerpt were created by shifting the pitch of the tones
of the leading voice by one semitone either upward or downward,
leading to 20 dissonant versions of the 10 original consonant
excerpts. During the experiment, each consonant version was
presented twice to match the number of dissonant excerpts.
Experimental Procedure
The general time course of the experiment is depicted in Figure 1.
All participants started the experiment by filling out the musical
experience and the mood questionnaires. Then, the physiological
sensors were attached while the participants sat comfortably in a
quiet room. Before the experiment started, participants per-
formed a handgrip squeeze task in which they were asked 4 times
to squeeze repeatedly on a handgrip for 7 s. This control taskwas
performed in order to obtain an index of individual electrodermal
reactivity in each participant before the beginning of the exper-
iment. The experimental task comprised four testing blocks.
Each block consisted of 10 excerpts, including 5 consonant
and 5 dissonant excerpts presented in an alternating order. The
orders of presentation of the consonant and dissonant trials
within each block were counterbalanced within subjects follow-
ing an ABBA design, as well as between subjects with half of the
participants starting with a consonant excerpt and the other half
with a dissonant one. All blocks started with a loud and sudden
burst of white noise (50 ms, 85–90 db) intended to elicit a startle
response. These startle responses served as another index of in-
dividual electrodermal reactivity and of habituation effect on the
signal throughout the experiment.
Each trial started with a warning signal asking the subject not
to move in order to avoid contamination of physiological re-
cordings by movement artifacts. This signal was followed by a
rest period lasting between 7.1 s and 7.7 s, in which subjects were
asked to do nothing. The length of the rest period varied ran-
domly in order to reduce subject’s expectancies and response
habituation. Themusical excerpt was then played for 7 s andwas
followed by another rest period of 1 s, in which subjects were
again asked not to move. This second rest period allowed the
recording of music-related physiological activity overflows and
also prevented the activity recorded at the end of the musical
excerpt from being contaminated by response preparation arti-
facts. The end of this 1-s rest period was indicated by another
signal allowing participants to move. Then, participants were
asked to judge the excerpt they just heard using a rating scale for
emotional valence (from � 55 ‘unpleasant’ to 55 ‘pleasant’)and arousal (from � 55 ‘relaxing’ to 55 ‘stimulating’). No timeconstraints were given for the ratings. All participants signed
informed consent to participate in the study.
Data Acquisition and Transformation
SCR, HR, and EMGs were monitored continuously using an
MP150 Biopac system (Biopac Systems, Inc., Goleta, CA) at a
sampling rate of 2000 Hz and processed using AcqKnowledge
4 D. Dellacherie et al.
A musicalmusicalexpertence
+ handgripexpert
+ handgrip+mood
handgripsqueeze Experimental Experimental Experimental Experimental
+mood
handgripsqueeze Experimental Experimental Experimental Experimental
questionnaires task block 1 block 2 block 3 block 4 questionnaires task block 1 block 2 block 3 block 4
B
llstartleprobe DCDDCDCCDCstartleprobe DCDCDCprobe
+DCDDCDCCDC
Musicalt i l
orprobe
+DCDCDC
Musicalt i l
orrecoveryperiod CDDCDCCDCD
trialsrecoveryperiod CDCDCD
trialsperiodperiod
10 sec10 sec ≈ 350 sec
C
“ l “valence
and“ l “valence
and“ l “valence
and“pleasedon’t rest
“youcan
andarousal
“pleasedon’t rest
“youcan
andarousal
“pleasedon’t rest
“youcan
andarousaldon t
move” rest period musical excerptrest
periodcan
move”arousalratings
don tmove” rest period musical excerpt
restperiod
canmove”
arousalratings
don tmove” rest period musical excerpt
canmove”
arousalratings
2 sec 7 1 to 7 7 sec 7 sec 1 sec 1 5 sec7 1 to 7 7 sec 7 sec 1 sec 1 5 sec7 1 to 7 7 sec 7 sec 1 sec 1 5 sec ≈ 10 sec2 sec 7.1 to 7.7 sec 7 sec 1 sec 1.5 sec7.1 to 7.7 sec 7 sec 1 sec 1.5 sec7.1 to 7.7 sec 7 sec 1 sec 1.5 sec 10 sec
Figure 1. Experimental procedure: (A) Participants first filled out the musical experience andmood questionnaires and performed the handgrip squeeze
task. The musical excerpts were then presented in four experimental blocks. (B) Each block started with a startle probe. Five consonant and 5 dissonant
excerpts were presented in an alternate order. The blocks either started with a consonant ‘‘C’’ or dissonant ‘‘D’’ excerpt. (C) Each musical trial started
with a rest period in which participants were asked not to move. At the end of each musical excerpt, participants were asked to remain still for 1 s before
they had to rate the valence and arousal of the excerpt.
-
software. Skin conductance was recorded on the index and mid-
dle finger of the non-dominant hand using the EDA finger
transducer BSL – SS3LA filled with an isotonic conducting gel.
The signal was filtered between 0.05 and 10 Hz. Then, the SCRs
to each musical trial were visually inspected and checked for
failures of the measuring device as well as movement-related ar-
tifacts. Trials in which there was electrodermal activity prior to
the onset of the musical excerpts were excluded. Responses were
selected if they occurred in a 1–4-s latency window following
stimulus onset. Three participants from the low expertise group
that showed no or very feeble responses to grasping and startle
trials were considered as non-responders and excluded from the
analysis.
Facial EMGs were recorded over the left corrugator and
zygomatic sites as recommended by Fridlund and Cacioppo
(1986), using 8 mm Ag/AgCl shielded electrodes, and were fil-
tered between 100 and 500 Hz. After the recording, EMG was
rectified using the root mean square function of the software and
then smoothed by a factor of 200 samples. HR was recorded
using a bipolar montage with an electrode on the right carotid
artery and another below the left ribs. The signal was filtered (0.5
to 35 Hz). Instantaneous inter-beat (RR) intervals (in ms), cor-
responding to the inverse of HR, were calculated from the elec-
trocardiogram using a peak detection algorithm to detect
successive R-waves and obtain a continuous RR tachogram.
Careful examination of the electrocardiogram and the tacho-
gram ensured that the automatic R-wave detection procedure
had been performed correctly. Careful examination of each par-
ticipant’s electrocardiogram signal led to the exclusion of four
participants (3 LE and 1 HE) due to technical failures. Electro-
cardiograms of the remaining participants were then cleaned by
removing the trials in which the presence of artifacts prevented
the exact extraction of RR Intervals (RRI). The participants who
were excluded from the SCR analyses were not the same ones
who were excluded from the HR analyses.
Results
Before analyzing the data, we checked for group differences be-
tween LE and HE participants on mood questionnaires. Then,
we tested the effects of dissonance and musical experience on
valence and arousal ratings, as well as on the physiological re-
sponses to the musical excerpts. Finally, we used regression an-
alyses to explore the relationship between valence ratings and
personal (mood and musical experience) and physiological
(SCR, facial EMGs, HR) variables. For all analyses, the Fmaxstatistic was used to test if the homogeneity of variance assump-
tion was met. Following the guidelines of Tabachnick and Fidell
(2001) for similar samples sizes (i.e., within a ratio of 4 to 1), the
Fmax was in an acceptable range (i.e., below 10) for all analyses,
indicating that the variances between samples were sufficiently
homogeneous to proceed with the analyses. Partial eta-squared
(Z2) were used as the effect sizes for the analyses of variance(ANOVAs). According to Cohen’s (1988) guidelines, Z2 5 .01corresponds to a small effect, Z2 5 .09 to a medium effect, andZ2 5 .25 to a large effect.
Mood Questionnaires
The mean ratings of the mood questionnaires STAI and POMS
for the LE andHEgroups are displayed in Table 2, alongwith the
results of the t-tests to verify whether baseline mood levels dif-
fered between the LE andHEgroups. As can be seen in this table,
there were no differences on any of the mood parameters that
were assessed, confirming that both groups did not differ in this
regard.
Valence and Arousal Ratings
Separate ANOVAs with one repeated measure (consonance vs.
dissonance) were carried out on the valence and the arousal rat-
ing scores. Figure 2 shows the mean ratings of valence and
arousal for consonant and dissonant excerpts as a function of
musical experience. As expected, dissonant excerpts were rated as
more unpleasant than consonant ones (F(1,25)5 73.19, po.05,Z2 5 0.75). This effect was modulated by the degree of musicalexperience of the listener, as revealed by the significant Musical
experience by Dissonance interaction (F(1,25)5 8.32, po.05,Z2 5 0.25). HE participants rated dissonant excerpts as moreunpleasant than LE participants did (F(1,25)5 8.81, po.05,Z2 5 0.26), whereas their ratings of consonant excerpts did notdiffer across groups (F(1,25)5 0.02, p5 n.s., Z2 5 0.001). Theanalysis of arousal ratings showed no effect of Dissonance
(F(1,25)5 0.28, p5 n.s., Z2 5 0.015) or of Musical experience(F(1,25)5 0.97, p5 n.s., Z2 5 0.51) nor any interaction(F(1,25)5 0.08, p5 n.s., Z2 5 0.003).
Psychophysiology of musical emotion 5
Table 2. Inter-Group Comparisons of the Mean (� SD) Ratings on the Mood Questionnaires and Electrodermal Reactivity Tests
Musical experience
Dependent variable Low experience High experience Result of t-test
STAI subscalesTrait anxiety 37.33 (� 8.86) 34.83 (� 7.63) t(25)5 1.51, p5n.s.State anxiety 34.40 (� 7.15) 30.58 (� 5.65) t(25)5 0.77, p5n.s.
POMS subscalesAnger 1.40 (� 1.45) 0.73 (� 1.10) t(25)5 1.28, p5n.s.Anxiety 1.73 (� 1.62) 1.55 (� 2.25) t(25)5 0.25, p5n.s.Depression 1.33 (� 1.80) 0.91 (� 1.64) t(25)5 0.62, p5n.s.Confusion 4.27 (� 1.87) 4.90 (� 0.83) t(25)5 1.06, p5n.s.Vigor 9.80 (� 2.68) 11.55 (� 3.59) t(25)5 0.42, p5n.s.Fatigue 3.13 (� 2.92) 4.18 (� 2.27) t(25)5 0.99, p5n.s.
Electrodermal reactivityHandgrip squeeze 0.45 (� 0.46) 0.48 (� 0.23) t(22)5 0.28, p5n.s.Startle 0.38 (� 0.34) 0.58 (� 0.40) t(22)5 0.86, p5n.s.
Note: n.s.5not significant.
-
Physiological Recordings
Electrodermal Responses
After exclusion of three non-responder participants, there
were 12 LE and 12 HE left in the SCR analyses. To make sure
that the LE and HE groups exhibited similar levels of electrode-
rmal reactivity, the mean SCRs for the grasping and startle trials
were compared by means of t-tests (Table 2). The results showed
that there were no differences in baseline electrodermal reactivity
between the two groups (Handgrip: t(22)5 0.28, p5 n.s.; Star-tle: t(22)5 0.86, p5 n.s.). The SCRs to the consonant and dis-sonantmusical excerpts were then analyzed for the two groups of
participants. First, the mean SCR values were extracted by slices
of 500 ms and averaged by musical condition and experimental
group in order to provide average response curves for the dis-
sonant and consonant excerpts and LE and HE participants (see
Figure 3A). Visual inspection of the resulting graph led to the
identification of two response periods according to the guidelines
proposed by Dawson (2007). The first response ranged from 1.5
to 6 s and corresponded to the orientation response triggered by
the onset of the excerpt. The second response ranged from 6 to 8
seconds and appeared to reflect later responses to the musical
excerpts. The amplitude of each SCR was then computed by
extracting the maximum value of the response in each time win-
dow and subtracting it from the baseline level. For the first time
window, the baseline period was the mean skin conductance level
of the 1-second baseline preceding the onset of the excerpt. For
the second time window, the mean skin conductance level be-
tween the 5th and 6th second served as the baseline. The resulting
values were transformed in log (SCR11) and averaged for each
participant.
An ANOVA with Dissonance (consonance vs. dissonance)
and Time window (early or later) as within-subjects factors were
carried out on the SCR values for the two groups of participants
(LE vs. HE). The averaged peaks for the LE and HE groups are
displayed in Table 3. SCRs were higher when participants lis-
tened to dissonant than to consonant excerpts (F(1,22)5 9.85,po.05, Z2 5 0.31), and this effect was influenced by musical ex-perience, as revealed by the significant interaction
(F(1,22)5 4.32, po.05, Z2 5 0.16) which demonstrates that thedifference in SCRs between consonance and dissonance was
higher for HE (F(1,11)5 7.89, po.05, Z2 5 0.42) than LE(F(1,11)5 2.05, p5 n.s., Z2 5 0.16) group.
The effect of Dissonance alsomarginally interactedwith Time
window (F(1,22)5 4.04, p5 .057, Z2 5 0.001), the effect of Dis-sonance being more pronounced in the early (F(1,22)5 9.01,
po.05, Z2 5 0.29) than in the late time window (F(1,22)5 1.63,p5 n.s., Z2 5 0.066). Finally, the interaction between Disso-nance, Time window, andMusical experience was not significant
6 D. Dellacherie et al.
Figure 2. Mean valence and arousal ratings of the dissonant and consonant musical excerpts for the low and high experience groups.
Table 3.MeanValues (� SD) of the Physological Recordings for Consonant andDissonant Excerpts byTimeWindow andLevel ofMusicalExperience
First window Second window
Consonant Dissonant Consonant Dissonant
SCR amplitude Low experience 0.08 (� 0.08) 0.10 (� 0.07) 0.01 (� 0.03) 0.01 (� 0.03)Log (maximum 11), mS High experience 0.12 (� 0.08) 0.20 (� 0.17) 0.04 (� 0.05) 0.07 (� 0.03)Corrugator EMG Low experience 0.03 (� 0.28) 0.06 (� 0.24) 0.03 (� 1.19) 0.94 (� 1.53)Area under the curve, mVnsec High experience 0.36 (� 0.90) 0.43 (� 0.65) 0.10 (� 01.17) 1.92 (� 4.53)Zygomatic EMG Low experience � 0.12 (� 0.22) � 0.03 (� 0.56) � 0.05 (� 1.43) 0.10 (� 2.51)Area under the curve, mVnsec High experience � 0.26 (� 0.42) � 0.11 (� 0.26) 0.26 (� 2.44) 3.03 (� 6.70)RR intervalIn the first window : Maximum acceleration, msec Low experience 33.73 (� 13.78) 29.50 (� 11.02) � 36.82 (� 28.55) � 35.92 (� 26.96)In the second window : Minimum deceleration, msec High experience 24.41 (� 12.64) 29.27 (� 11.44) � 32.06 (� 20.57) � 32.38 (� 14.66)
Note: All displayed values are differences from baseline.
-
(F(1,22)5 0.17, p5 n.s., Z2 5 0.075). For exploratory reasons,we tested if the interaction between Dissonance and Musical ex-
perience was different in the two time windows by performing
ANOVAs in the first and second time window with Dissonance
and Musical experience as factors. The results showed that the
effect of Dissonance did not interact with Musical experience in
the early time window (F(1,22)5 1.78, p5 n.s., Z2 5 0.075), butit did interact withMusical experience in the second time window
(F(1,22)5 4.11, po.05, Z2 5 0.16), indicating that SCRs werehigher for dissonant than for consonant excerpts in the late time
window. This effect holds for HE (t(11)5 2.39, po.05,Z2 5 0.25) but not for LE (t(11)5 1.91, po.05, Z2 5 0.05) par-ticipants.
Heart Rate
There were 12 LE and 11 HE participants left in the HR
analyses. RRI was extracted by slices of 500 ms and averaged by
musical condition and experimental group in order to provide
average response curves for the dissonant and consonant ex-
cerpts and LE and HE participants (see Figure 3D). Inspection
of the mean RRI curves revealed a typical triphasic pattern
(Bradley et al., 2001) comprising an initial increase of RRI (i.e.,
deceleration of heart rate), followed by a decrease of RRI (i.e.,
acceleration of heart rate), and then by a final increase of RRI.
The amplitude of the first increase and of the following decrease
were then extracted and subtracted from a 1-s baseline preceding
the onset of the musical excerpts. An ANOVA with Dissonance
and Phase of response was carried out on the RRI for the two
groups of participants. The amplitude of the first increase and
secondary decrease in RRI for the HE and LE groups are dis-
played in Table 3. No significant effect of Dissonance
(F(1,21)5 0.001, p5 n.s., Z2 5 0.00) or of Musical experience(F(1,21)5 0.01, p5 n.s., Z2 5 0.00) was observed. Only the ex-pected effect of Phase was observed (F(1,21)5 162.53, po.05,Z2 5 0.87), confirming that RRI were higher during the first in-crease than during the following decrease.
Facial Electromyography
The corrugator and zygomatic EMG responses to musical
excerpts were calculated as the difference between the raw signal
during each musical excerpts and the mean activity observed in
an 800-ms baseline preceding the onset of the excerpts. The re-
sulting response curves were then averaged by slices of 100ms for
the dissonant and consonant excerpts and for LE andHE groups
(see Figures 3B and 3C). Visual inspection of the resulting graphs
indicated that facial muscles appeared to show two phases of
Psychophysiology of musical emotion 7
ACorrugator EMG
BSkin Conductance Responses
0.6000.150
0.1252nd
0.50
2nd
window 21st0.100 0.40window 21
window0.100
0 301st window0.075
0 .30so
lts
μvo
0.0500.20
ns
me
nsi
eμ
0.100.025
00 1 2 3 4 5 6 7
0
1 2 3 4 5 6 7 8 9 −0.10−0.025
Time in seconds Time in seconds
DCZygomatic EMG R-R interval
101.20
8
2nd2nd window1.00 6 2 window2nd window
4
nds0.80
0
2
eco
n1st0.60
0lis
e1 2 3 4 5 6 7
window0.40 −2
mil
l
1 window1 2 3 4 5 6 7o
ltsvoμ
0.40
−4 1 window0.20
−8
−60
−101 2 3 4 5 6 7−0.20−12
Time in secondsTime in seconds
windowwindownd
Figure 3. Average time-courses of the physiological responses during the presentation of consonant and dissonant excerpts for low and high musical
experience groups. Each physiologicalmeasure was divided in two phases of responding. (A) For skin conductance responses, a first orientation response
(0–6 s) was followed by a second slower response (6–8 s). Note that the averaging of slight inter-trials and inter-subjects differences in the onsets of the
second non-orientation response smoothes the shape of themean response curves, compared to individual trials, fromwhich the analyzed responses were
extracted. (B) Corrugator EMG started to diverge between consonant and dissonant excerpts after the 2nd second. (C) Zygomatic EMG started to
diverge between consonant and dissonant excerpts after the 2nd second. (D) RR Intervals first increase between 0–3.5 s (slowing of heart rate) and
decrease between 3.5–7 s (speeding of heart rate).
-
responding: a first phase ranging from 0 to 2 s, where there were
no differences between the consonant and dissonant trials, and a
second phase from 2 to 7 s, where responses to consonant and
dissonant musical excerpts started to diverge. The area under the
curve was extracted within these two time windows and was
averaged for consonant and dissonant versions for each partic-
ipant. The averaged area under the curve for the LE and HE
groups are displayed in Table 3. An ANOVA with two repeated
measures, Dissonance and Time window, was carried out on the
averaged area under the curve for the LE and HE groups.
Corrugator activity differed between consonant and disso-
nant excerpts as a function of Time window, as revealed by the
significant interaction between Dissonance and Time window
(F(1,25)5 4.09, p5 .05, Z2 5 0.14). Decomposition of the inter-action revealed that corrugator activity was higher for dissonant
than for consonant excerpts in the second window
(F(1,25)5 4.06, p5 .05, Z2 5 0.14), but not in the first window(F(1,25)5 0.40, p5 n.s., Z2 5 0.02). Zygomatic activity differedas a function of Dissonance, Time window, and Musical expe-
rience, as revealed by the significant interaction between the three
factors (F(1,25)5 4.09, po.05, Z2 5 0.14). Decomposition ofthe interaction showed that there was no difference between
consonant and dissonant excerpts in the first time window
(F(1,25)5 0.01, p5 n.s., Z2 5 0.00) whereas zygomatic activitywas higher for dissonant than for consonant excerpts in the sec-
ond window for the HE (F(1,11)5 4.84, p5 .05, Z2 5 0.31), butnot for the LE group (F(1,14)5 0.01, p5 n.s., Z2 5 0.00).
Regression Analyses
Stepwise Regressions
In order to assess the relationship between personal variables,
valence ratings, and physiological responses, we performed two
stepwise regressions with the effects of dissonance on valence
ratings as the dependent variable and physiological responses
(SCR, facial EMGs, and RRI within all time windows) and per-
sonal variables (Musical experience and STAI and POMS sub-
scales) as the independent variables. The goal of these regressions
was to identify the physiological responses andpersonal variables
that best predicted the effects of dissonance on valence ratings.
For all the variables included in the analysis, the difference be-
tween the mean value for the dissonant and consonant excerpts
(dissonant–consonant) was used as an index of the effects of
dissonance on these variables. Because the regression models in-
cluded all the physiological measurements, we only kept the
participants for which SCR and HR data were available (9 LE
and 11 HE).
The results of the first stepwise regression revealed that
zygomatic activity in the second time window was the best pre-
dictor of the decrease in valence in reaction to dissonant excerpts
(r5 0.61, po.05). None of the other variables (SCRs, corrugatoractivity, RRI, and zygomatic activity in the first time window)
significantly improved the proportion of variance explained by
the model once the variance explained by the zygomatic activity
in the second time window was taken into consideration. The
results of the second stepwise regression showed that musical
experience was the best predictor of the decrease in valence in
reaction to dissonant excerpts (r5 0.53, po.05). Again, allother personal variables (STAI and POMS subscales) did not
significantly improve the proportion of variance explained by the
model. Thus, the participants who scored higher on the musical
experience questionnaire or who had the largest increases in
zygomatic activity in reaction to dissonant excerpts were the ones
showing the largest decreases in valence ratings in response to
dissonance. A final correlational analysis indicated that musical
experience was also correlated to the amount of zygomatic
activity in response to dissonance (r(26)5 0.59, po.05),suggesting that musical experience predicted zygomatic
responses to dissonance, which in turn strongly predicted the
decreases in valence ratings in response to dissonance.
Discussion
The purpose of our study was to determine the effect of musical
experience on emotional reaction to dissonance by recording
emotional self-reports as well as their psychophysiological corre-
lates. For this purpose, we compared the emotional responses of
two categories of listeners. One consisted of nonmusicians with
low musical experience (LE). The other one was composed of
musicians with high musical experience (HE) who had received at
least 3 years of formal training or who were still practicing a mu-
sical instrument although they did not reach professional levels.
Therefore, we were able to investigate the interaction between
musical experience and emotional pleasantness judgments on au-
tonomic and somatic responses by measuring SCR, HR, and fa-
cial EMG for the zygomatic and corrugatormuscles in response to
consonant and dissonant musical excerpts. Before we discuss the
results in further detail, we should first point out that the two
groups did not differ in terms of age, sex, education, or mood.
Moreover, given that the only difference between the stimuli con-
cerned the pitch shift of the melodic line, tempo differences could
not have contaminated physiological measurements, in contrast to
most previous studies onmusical emotionswhere differentmusical
excerpts were used to test different emotions.
Behavioral Measures
As predicted, the behavioral results confirm previous find-
ings, indicating that listening to dissonance induces a more neg-
ative affect than listening to consonance (Blood et al., 1999;
Gosselin et al., 2006; Koelsch et al., 2006; Pallesen et al., 2005;
Passynkova et al., 2007; Peretz et al., 2001; Sammler et al., 2007).
Arousal ratings, which were used as a proxy for motivational
activation (Ito, Larsen, et al., 1998) did not differ between con-
sonant and dissonant excerpts. In agreement with Pallesen et al.
(2005) and Schoen et al. (2005), we also found that the effect of
dissonance on emotional valence judgments ismore salient inHE
than in LE participants. Given that arousal judgments obtained
for consonant and dissonant stimuli did not differ between the
two groups of listeners, we can therefore suggest that musical
experience modulates valence judgments. Moreover, regression
analyses showed that musical experience was a very good pre-
dictor of valence ratings in the present study, suggesting that
musical experience may have enhanced the participants’ sensi-
tivity to dissonance. However, since almost all HE participants
were trained in classical music, it remains to be determined if the
observed results can be generalized to training in other musical
genres more marked by dissonance, such as free jazz or contem-
porary music. Nevertheless, this result appears to be consistent
with data of another study (Bigand, Parncutt, & Lerdahl, 1996)
showing that dissonant chords induce stronger ratings of musical
tension than consonant chords, this effect being more pro-
nounced in musicians than in nonmusicians. We can therefore
conclude that musical experience can modulate emotional judg-
ment of music even if the role of the specific background remains
to be clarified.
8 D. Dellacherie et al.
-
Physiological Responses
The analysis of physiological measures revealed that disso-
nance elicited stronger responses than consonance in the two
groups of participants. Given that dissonant excerpts were
judged by all participants as more unpleasant than consonant
excerpts but equally arousing, we interpret this reaction to dis-
sonance as resulting from the emotional negativity bias, defined
as the tendency for the negative motivational system to respond
more intensely than the positive motivational system to compa-
rable amounts of activation (Cacioppo et al., 2000; Taylor, 1991).
The main finding of our study confirmed that autonomic and
somatomotor responses to musical dissonance are influenced by
musical experience, with HE participants presenting stronger
physiological responses to dissonance in comparison to conso-
nance than LE participants. Since HE participants judge disso-
nance as even more unpleasant than consonance as compared to
LE participants, the differential reactivity to dissonance between
the two groups seems to be related to the subjective ratings of
emotional valence.
HR Measures
HR is normally responsive to the affective valence of the
stimuli. Emotional stimuli generally prompt a triphasic HR re-
sponse characterized by an initial brief deceleration, followed by
a short acceleration and a latemoderate deceleration (Lang et al.,
1997; Sanchez-Navarro et al., 2006). Although we found the
predicted triphasic response when listening to the musical ex-
cerpts, it was not modulated by dissonance. This result does not
seem to be consistent with previous findings obtained with un-
pleasant pictures (Bradley et al., 2001). However, these authors
found an effect of valence on HR response only with high (e.g.,
erotic pictures or mutilation scenes) but not with low arousal
pictures from International Affective Picture System (Lang,
Bradley, & Cuthbert, 1999). Similarly, Sammler et al. (2007)
found an effect of dissonance on HR deceleration with real but
not synthetic excerpts that might have been more arousing than
our computer-synthesized versions of the stimuli. It might be
therefore possible that the musical stimuli used in the present
study do not induce a sufficiently high emotional response to
affect HR. Notwithstanding the fact that we did not observe any
significant difference inHR responses between the two emotional
conditions, the successive phases of decelerative and accelerative
cardiac responses clearly argue in favor of the two-step defensive
cascade model (Bradley et al., 2001; Lang et al., 1997).
SCR
The analysis of phasic SCR revealed two different peaks of
response to emotional musical excerpts in the 2–6 s and in the
6–8 s windows. As the late response could not be ascribed to the
onset of the musical excerpt, this second bump indicates a sec-
ond, superimposed skin conductance response. Indeed, SCRs are
monophasic, and a secondary increase indicates that another
SCR has been triggered after the first one (Dawson, 2007). These
two successive responses match our hypothesis based on a two-
step cascade model and are consistent with the HR response
pattern (Bradley et al., 2001). The initial response, which occurs
immediately after the onset of the stimulus (early time window)
and the following response beginning a few seconds later (late
time window) are both larger for dissonant than for consonant
excerpts. Based on the cascade defense model (Bradley et al.,
2001; Lang et al., 1997; Ohman & Wiens, 2003), these two suc-
cessive responses could correspond to the (automatic) orienting
response in the first time window and the (more controlled) de-
fense response in the second one. Curiously, such a biphasic re-
sponse is rarely seen in most psychophysiological studies of
emotion, where visual stimuli are used to induce emotions. For
instance, in Bradley and Lang’s study (2000), although the affec-
tive pictures had a 6-s duration, no second phasic response after
the initial orienting response was observed. Classically, a ‘‘more
sustained’’ electrodermal response is observed as defense re-
sponse (Norris et al., 2007; Ohman & Wiens, 2003). One expla-
nation for this original finding could be that musical emotions
unfold in time (Grewe et al., 2007) whereas for pictures, all the
relevant emotional information is available immediately. Inter-
estingly, musical experience seems to have contributed to the
enhancement of this second, more controlled response. Indeed,
after verifying that baseline electrodermal reactivity did not differ
between the two groups of participants, we showed that the SCR
difference between dissonance and consonance is more pro-
nounced in HE than in LE group, as predicted. Subsequent SCR
analysis showed that musical experience interacted with disso-
nance in the second but not in the first time window, suggesting
that these orienting and defense responses were differently mod-
ulated by musical experience.
The orienting response is known to be processed at a pre-
attentive level and without implication of cortical association
areas (Ohman &Wiens, 2003). The orienting SCR to dissonance
that we observed in the present study could be interpreted as an
automatic response reflecting an emotional unconscious pro-
cessing. Given that this orienting SCR was higher for dissonance
than for consonance in HE and LE participants, we suggest that
nonmusicians can discriminate dissonance from consonance on
the basis of emotional feeling as musicians do. We were not able
to demonstrate an interaction between musical experience and
dissonance in this first time window, although larger SCRs were
recorded for HE than for LE participants. This finding indicates
that LE as well as HE participants showed stronger orienting
responses to dissonance than to consonance. We note that this
result was observed despite the lack of formal musical training in
LE participants.
At a more evaluative level (defense response in the second
time window), we also observed larger responses to dissonance
(unpleasant) than to consonance (pleasant) but only in HE par-
ticipants. This response seems to be enhanced in musicians. The
influence of musical valence on the SCR in the second time win-
dow is compatible with previous results demonstrated by re-
cording the tonic (and not phasic) SCRs when listening to
musical stimuli for more than 1 min (Baumgartner et al., 2006;
Nater et al., 2006). The fact that a sustained and higher response
was obtained for dissonance than for consonance suggests that
dissonance affects a more elaborate processing in addition to the
automatic response. This later processing of musical stimuli
could be specific to musicians. An interpretation of this result is
that formal musical learning led musicians to consciously reject
dissonant stimuli in a controlled defense response that follows the
pre-attentive orienting one. Taken together, analysis of SCR
suggests that musical experience could modulate musical disso-
nance processing, particularly in the late SCR. This result indi-
cates that musical education and training enhance emotional
response to dissonance.
An alternative interpretation of these results could be that
stronger SCRs to dissonance than to consonance may be related
to an effect of unfamiliarity of dissonance. It is clear that con-
sonant excerpts are more frequent or familiar than dissonant
Psychophysiology of musical emotion 9
-
ones because we are immersed in a consonant musical environ-
ment. Moreover, Terhard (1984) hypothesized that prenatal ex-
posure to the overtone structure of maternal speech could serve
as the basis for developing preferences for consonance. It is also
clear that musical experience contributes to the fact that some
people are more exposed to consonance than others are, even
more so if the formation is classical as it is in the present study. It
is therefore impossible to differentiate the pleasantness induced
by consonance from the frequency or exposure effects (Zajonc,
1980) and to dissociate the unpleasantness to dissonance from a
novelty or incongruity effect. However, the aim of the study was
not to explain the origin of aversion to dissonance, which is still a
hotly debated topic (Hauser &McDermott, 2003), but rather to
explore the relationships between physiological and self-re-
ported measures of emotional responses to dissonance (Bradley
et al., 2001; Ohman & Wiens, 2003). Because subjective emo-
tional ratings and physiological responses converge, the inter-
pretation of physiological responses as reflecting emotional
reactions to dissonance appears highly probable, even if this
emotion might be at least partly explained by the unfamiliarity
of dissonance.
Several lines of evidence in non-musical domains showed that
the relevance of the task for a given subject plays a role in mod-
ulating the SCR (Dindo & Fowles, 2008; Epstein & Fenz, 1962;
Fenz & Epstein, 1962; Lang et al., 1998; Ohman & Soares, 1994;
Perpina, Leonard, Bond, Bond, & Banos, 1998; Stormark, La-
berg, Nordby, & Hugdahl, 2000). These studies, carried out with
psychiatric patients and normal volunteers, revealed that higher
SCRs specific to a given object were dependent on individual ex-
perience. For example, studies examining phobic individuals
(Lang et al., 1998; Ohman & Soares, 1994) showed hyperactivity
to phobic objects, by expressing an intense and sustained fear in
response to stimuli thatwere not necessarily dangerous.According
to Mineka and Ohman (2002), the learning of fear could arise
froma classic Pavlovian conditioning: phobic patients would learn
to be frightened by a neutral stimulus because of an association
between such a stimulus and an aversive event. Because formal
musical education trains listeners to detect and reject dissonance
when it is not integrated in an appropriate musical context, we can
hypothesize by analogy that musical experience leads to a hyper-
sensitivity to dissonance. Consequently, HE participants could
produce larger SCRs to dissonance than do LE participants in the
same way as phobic subjects respond to phobia-relevant stimuli
(Lang et al., 1998). In other terms, even if aversion for dissonance
is natural (Schellenberg & Trainor, 1996; Zentner &Kagan, 1996)
or influenced by an exposure effect (Witvliet & Vrana, 2007), HE
participants could be seen as conditioned to react to dissonance,
resulting in an enhanced autonomic reactivity.
EMG Measures
Another index of the emotional appraisal of the stimuli was
the EMG responses. Corrugator and zygomatic muscles are
known to accompany unpleasant and pleasant emotions, re-
spectively. In the present study, we found a response of these
muscles that arose late (during the second time window), which
probably reflected a controlled mechanism. We can clearly see
that there is no zygomatic activity before 2 s after the onset of the
stimulus. Thus, the zygomatic response to musical excerpts is
particularly late. This cannot be explained by the characteristics
of EMG latencies. Indeed, unlike skin conductance responses,
which generally take around 2 s to develop because of sweat
gland physiology, EMG responses can be triggered within 100
ms. Therefore, this late response cannot be interpreted as an
orienting automatic response. Following the two-step cascade
model and given the coherent HR and SCR responses in regard
to this model, we thus propose to interpret a posteriori the late
EMG reactivity (more than 2 s) as reflecting more controlled
psychological processes related to conscious emotional evalua-
tion, which is characteristic of the defense response (Ohman &
Wiens, 2003). This interpretation is confirmed by the regression
analysis, which shows a strong relation between EMG responses
and valence ratings. Indeed, the zygomatic activity in the second
window was the best predictor of valence ratings.
Moreover, 2 s after the onset of the aversive stimulus, co-
rrugator activity was higher for dissonant than for consonant
excerpts, confirming that such a measure is relevant to the study
of emotional reaction to music (Roy et al., 2008; Witvliet &
Vrana, 2007). One probable reason why this facial response
seems to take more time to be initiated in the case of music
compared to pictures (Bradley et al., 2001) is the fact that the
musical emotions unfold in time whereas for pictures, all the
relevant emotional information is immediately available. More
studies on the dynamics of the response to musical emotion are
needed to clarify the specificity of emotional responses to music
in comparison to other types of stimuli, such as faces or brief
sounds.
In addition, we discovered that the late increase of zygomatic
contraction 2 s after the stimulus onset was specific to experienced
participants when they were listening to dissonant excerpts. This
unexpected result seems to be in contradiction with the responses
observed in other studies showing such a response with pleasant
stimuli (Ellis & Simons, 2005; Lang et al., 1998; Witvliet & Vrana,
1995). This zygomatic activity might be interpreted as an ironic
smile or as a grimace related to the displeasure induced by disso-
nance. Since we witnessed that some participants paradoxically
laughed (Ansfield, 2007; Craig & Patrick, 1985; Keltner & Bon-
anno, 1997; Papa&Bonanno, 2008; Prkachin&Solomon, 2008) at
the dissonant excerpts, we tend to favor the former interpretation,
although we lack objective empirical data to confirm it. Regression
analysis revealed that more unpleasant dissonances were associated
with stronger zygomatic activities and that this smile/grimace is the
best predictor of the subjective responses to dissonance. The
zygomatic response could therefore constitute not only a controlled
but also a communicative part of the emotional response to dis-
sonance, and this reaction could be more developed in HE partic-
ipants than in LE participants. Mimics to dissonance would have a
communicative function that musical learning might have contrib-
uted to create in HE participants. We could hypothesize that ex-
perience enhances the ability to communicate musically induced
negative emotions. Taken together, these results show that musical
experience plays a role in physiological and communicative re-
sponses induced by dissonance that are linked to emotional valence.
Based on our results, it seems that dissonance produces larger
SCR and EMG responses in HE than LE listeners. The results
could be further explained by brain reorganization induced by
musical experience. Indeed, even in amateur musicians or in
children, musical training can produce functional and morpho-
logical brain changes in auditory areas (Gaser & Schlaug, 2003;
Schneider, Scherg, Dosch, Specht, Gutschalk, & Rupp, 2002) as
well as in emotional structures such as the amygdala and the
insula (James, Britz, Vuilleumier, Hauert, & Michel, 2008), the
anterior cingulate cortex (Foss, Altschuler, & James, 2007) and
the frontal-lobe areas (Koelsch, Fritz, Schulze, Alsop, & Schl-
aug, 2005; Minati et al., 2009).
10 D. Dellacherie et al.
-
Taken together, based on the obtained physiological results,
we argue for the existence of a first automatic response to dis-
sonance, which is mainly characterized by an orienting SCR.
This initial physiological reaction seems to be followed by a sec-
ond phase, which is less automatic and more controlled. This
second step is characterized by facial EMG responses and a sec-
ond non-orienting SCR. The first response suggests the existence
of an alarm system that can automatically interpret dissonant
music as an aversive stimulus and then evaluate this stimulus in
terms of displeasure. In the nonmusical domain, an alarm system
in response to fear was shown to involve the amygdala (Liddell,
Brown, Kemp, Barton, Das, et al., 2005). Given that responses
of the autonomic nervous system are linked to the amygdala and
other structures such as the cingulate cortex, the effect of dis-
sonance on SCR is compatible with neuroimaging studies sug-
gesting that affective responses to dissonance could mainly
involve mesio-temporal lobe structures (Blood et al., 1999; Ko-
elsch et al., 2006, 2007). The stronger late responses obtained in
HE than in LE participants (second phase response in SCR and
zygomatic response) seem consistent with the more important
involvement of cortical structures such as the cingulate cortex or
frontal lobe areas in musicians than in nonmusicians in response
to dissonance and consonance (Foss et al., 2007; Minati et al.,
2009). Considering that the cingulate cortex is known to con-
tribute in regulating autonomic functions as well as in attentional
control (Critchley, Mathias, Josephs, O’Doherty, Zanini, et al.,
2003; Devinsky, Morrell, & Vogt, 1995), we may hypothesize
that the SCR and the activity observed with fMRI in the cingu-
late cortex in musicians may reflect a similar mechanism, result-
ing in a hypersensitivity to dissonance. This hyperactivation may
result in a stronger negativity bias observed in autonomic reac-
tivity in HE than in LE participants in the second phase of pro-
cessing dissonance.
Conclusion
The data of the present study emphasize the effect of musical
experience on emotional responses to music and the role of emo-
tional valence in determining the autonomic and somatomotor
components of these responses. The results indicate that SCR and
EMG responses are sensitive to dissonance. These measures
therefore appear to be appropriate tools to assess emotional re-
action to unpleasant music. Stronger responses to dissonant than
to consonantmusic were observed in the orienting SCR, and there
were no differences between HE and LE participants in the am-
plitude of this response, suggesting that the first step of emotional
response to dissonance could be independent of musical experi-
ence. Late skin conductance and EMG responses (corrugator and
zygomatic) confirmed the presence of a subsequent, more con-
trolled, response to dissonance. Moreover, whereas corrugator
activity when listening to dissonancewas found in all participants,
late SCRs and zygomatic activity were found only in experienced
musicians, suggesting a specific emotional response in this group.
This specific reaction expressed by stronger physiological late re-
sponses to dissonance in HE than in LE participants was con-
firmed by self-report responses. This suggests that experience
could influence the negativity bias. An interpretation is that mu-
sical education could have reinforced the representation of dis-
sonance for musicians by a long and sustained associative
learning between dissonance and unpleasant emotions. These re-
sults add arguments in favor of both physiological and psycho-
logical origins of the feeling of dissonance. Moreover, the smile
observed in musicians represents an original finding indicating
that musical experience influences not only conscious appraisal of
emotional significance but also its means of communication. In
other terms, learned aesthetic preference might play a powerful
role in affective and expressive response to music.
REFERENCES
Ansfield, M. E. (2007). Smiling when distressed: When a smile is a frownturned upside down. Personality and Social Psychology Bulletin, 33,763–775.
Baumgartner, T., Esslen, M., & Jancke, L. (2006). From emotion per-ception to emotion experience: Emotions evoked by pictures andclassical music. International Journal of Psychophysiology, 60, 34–43.
Bigand, E., Parncutt, R., & Lerdahl, F. (1996). Perception of musicaltension in short chord sequences: The influence of harmonic function,sensory dissonance, horizontal motion, and musical training.Perception & Psychophysics, 58, 125–141.
Blood, A. J., Zatorre, R. J., Bermudez, P., & Evans, A. C. (1999). Emo-tional responses to pleasant and unpleasant music correlate withactivity in paralimbic brain regions.Nature Neuroscience, 2, 382–387.
Bradley, M. M., Codispoti, M., Cuthbert, B. N., & Lang, P. J. (2001).Emotion and motivation I: Defensive and appetitive reactions in pic-ture processing. Emotion, 1, 276–298.
Bradley, M. M., & Lang, P. J. (2000). Affective reactions to acousticstimuli. Psychophysiology, 37, 204–215.
Brattico, E., Pallesen, K. J., Varyagina, O., Bailey, C., Anourova, I.,Jarvenpaa, M., et al. (2009). Neural discrimination of nonprototyp-ical chords in music experts and laymen: An MEG study. Journal ofCognitive Neuroscience, 21, 2230–2244.
Cacioppo, J. T., & Bernston, G. G. (1994). Relationship between atti-tudes and evaluative space: A critical review, with emphasis on theseparability of positive and negative substrates. PsychologicalBulletin, 3, 401–423.
Cacioppo, J. T., Bernston, G. G., Klein, D. J., & Poehlmann, K. M.(1997). The psychophysiology of emotion across the lifespan. AnnualReview of Gerontology and Geriatrics, 17, 27–74.
Cacioppo, J. T., Bernston, G. G., Larsen, J. T., Poehlmann, K. M., &Ito, T. A. (2000). The psychophysiology of emotion. In R. Lewis & J.M.Haviland-Jones (Eds.),The handbook of emotion (2nd Edition, pp.173–191). New York: Guilford Press.
Cacioppo, J. T., Gardner, W. L., & Berntson, G. G. (1997). Beyondbipolar conceptualizations and measures: The case of attitudes andevaluative space. Personality and Social Psychology Review, 1, 3–25.
Cacioppo, J. T., Klein, D. J., Bernston, G. G., &Hatfield, E. (1993). Thepsychophysiology of emotion. In Handbook of emotions (pp. 119–142). New York: Guilford Press.
Chapados, C., & Levitin, D. J. (2008). Cross-modal interactions in theexperience of musical performances: Physiological correlates. Cogni-tion, 108, 639–651.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences(2nd Edition). Hillsdale, NJ: Lawrence Erlbaum Associates Inc.
Cook, E. I., & Turpin, G. (1997). Differentiating orienting, startle, anddefense responses: The role of affect and its implications for psycho-pathology. In Attention and orienting: Sensory and motivational pro-cesses (pp. 137–164). Mahwah, NJ: Lawrence Erlbaum Associates.
Craig, K. D., & Patrick, C. J. (1985). Facial expression during inducedpain. Journal of Personality and Social Psychology, 48, 1080–1091.
Critchley, H. D., Mathias, C. J., Josephs, O., O’Doherty, J., Zanini, S.,Dewar, B. K., et al. (2003). Human cingulate cortex and autonomiccontrol: Converging neuroimaging and clinical evidence. Brain, 126,2139–2152.
Dawson,G. J. (2007). The electrodermal system. In J. T. Cacioppo, L.G.Tassinary, & G. G. Bernston (Eds.), Handbook of psychophysiology(Third Edition, pp. 159–181). Cambridge: Cambridge UniversityPress.
Psychophysiology of musical emotion 11
-
Devinsky, O., Morrell, M. J., & Vogt, B. A. (1995). Contributions ofanterior cingulate cortex to behaviour. Brain, 118, 279–306.
Dindo, L., & Fowles, D. C. (2008). The skin conductance orienting re-sponse to semantic stimuli: Significance can be independent ofarousal. Psychophysiology, 45, 111–118.
Ehrle, N. (1998). Traitement temporel de l’information auditive et lobetemporal. Unpublished thesis, Université de Reims Champagne-Ardennes.
Ellis, E., & Simons, R. F. (2005). The impact of music on subjective andphysiological indices of emotion while viewing films. Psychomusicol-ogy, 19, 15–40.
Epstein, S., & Fenz, D. (1962). Theory and experiment on the measure-ment of approach-avoidance conflict. Journal of Abnormal and SocialPsychology, 64, 97–112.
Fenz, D., & Epstein, S. (1962). Measurement of approach-avoidanceconflict along a stimulus dimension by a thematic apperception test.Journal of Personality, 30, 613–632.
Fishman, Y. I., Volkov, I. O., Noh, M. D., Garell, P. C., Bakken, H.,Arezzo, J. C., et al. (2001). Consonance and dissonance of musicalchords: Neural correlates in auditory cortex of monkeys and humans.The Journal of Neurophysiology, 86, 2761–2788.
Foss, A. H., Altschuler, E. L., & James, K. H. (2007). Neural correlatesof the Pythagorean ratio rules. NeuroReport, 18, 1521–1525.
Fridlund, A. J., & Cacioppo, J. T. (1986). Guidelines for human elect-romyographic research. Psychophysiology, 23, 567–589.
Gaser, C., & Schlaug, G. (2003). Brain structures differ betweenmusicians and non-musicians. The Journal of Neuroscience, 23,9240–9245.
Gosselin, N., Samson, S., Adolphs, R., Noulhiane, M., Roy, M., Has-boun, D., et al. (2006). Emotional responses to unpleasant musiccorrelates with damage to the parahippocampal cortex. Brain, 129,2585–2592.
Grewe, O., Nage, F., Kopiez, R., & Altenmuller, E. (2007). Emotionsover time: Synchronicity and development of subjective, physiolog-ical, and facial affective reactions to music. Emotion, 7, 774–788.
Hauser, M. D., & McDermott, J. (2003). The evolution of the musicfaculty: A comparative perspective.Nature Neuroscience, 6, 663–668.
Helmholtz, H. (1954/1877).On the sensation of tones. New York: Dover.Ito, T. A., Cacioppo, J. T., & Lang, P. J. (1998). Eliciting affect using the
International Affective Picture System: Trajectories through evalua-tive space. Personality and Social Psychology Bulletin, 24, 855–879.
Ito, T. A., Larsen, J. T., Smith,N.K., &Cacioppo, J. T. (1998). Negativeinformation weighs more heavily on the brain: The negativity bias inevaluative categorizations. Journal of Personality and Social Psy-chology, 75, 887–900.
James, C. E., Britz, J., Vuilleumier, P., Hauert, C. A., & Michel, C. M.(2008). Early neuronal responses in right limbic structures mediateharmony incongruity processing in musical experts. Neuroimage, 42,1597–1608.
Keltner, D., & Bonanno, G. A. (1997). A study of laughter and disso-ciation: Distinct correlates of laughter and smiling during bereave-ment. Journal of Personality and Social Psychology, 73, 687–702.
Khalfa, S., Guye, M., Peretz, I., Chapon, F., Girard, N., Chauvel, P., etal. (2008). Evidence of lateralized anteromedial temporal structuresinvolvement in musical emotion processing. Neuropsychologia, 46,2485–2493.
Khalfa, S., Peretz, I., Blondin, J. P., & Robert, M. (2002). Event-relatedskin conductance responses to musical emotions in humans.Neuroscience Letters, 328, 145–149.
Khalfa, S., Roy, M., Rainville, P., Dalla Bella, S., & Peretz, I. (2008).Role of tempo entrainment in psychophysiological differentiation ofhappy and sad music? International Journal of Psychophysiology, 68,17–26.
Koelsch, S., Fritz, T., Schulze, K., Alsop, D., & Schlaug, G. (2005).Adults and children processing music: An fMRI study. Neuroimage,25, 1068–1076.
Koelsch, S., Fritz, T., Von Cramon, D. Y., Muller, K., & Friederici, A.D. (2006). Investigating emotionwithmusic: An fMRI study.HumanBrain Mapping, 27, 239–250.
Koelsch, S., Kilches, S., Steinbeis, N., & Schelinski, S. (2008). Effects ofunexpected chords and of performer’s expression on brain responsesand electrodermal activity. PLoS ONE, 3, 2631.
Koelsch, S., Remppis, A., Sammler, D., Jentschke, S., Mietchen, D.,Fritz, T., et al. (2007). A cardiac signature of emotionality. EuropeanJournal of Neuroscience, 26, 3328–3338.
Krumhansl, C. L. (1997). An exploratory study of musical emotions andpsychophysiology. Canadian Journal of Experimental Psychology, 51,336–353.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1997). Motivated at-tention: Affect, activation, and action. In P. J. Lang, R. F. Simons, &M. T. Balaban (Eds.), Attention and orienting: Sensory and moti-vational processes (pp. 97–135). Mahwah, NJ: Lawrence ErlbaumAssociates.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1998). Emotion, mo-tivation, and anxiety: Brain mechanisms and psychophysiology.Biological Psychiatry, 44, 1248–1263.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1999). InternationalAffective Picture System (IAPS): Technical manual and affective rat-ings. Gainsville, FL: The Center for Research in Psychophysiology,University of Florida.
Liddell, B. J., Brown,K. J., Kemp, A.H., Barton,M. J., Das, P., Peduto,A., et al. (2005). A direct brainstem-amygdala-cortical ‘alarm’ systemfor subliminal signals of fear. NeuroImage, 24, 235–243.
Masataka, N. (2006). Preference for consonance over dissonance byhearing newborns of deaf parents and of hearing parents. Develop-mental Science, 9, 46–50.
McNair, D. M., Lorr, M., & Droppleman, L. F. (1992). Profile of moodstates (Revised). San Diego, CA: EdITS: Educational and IndustrialTesting Service.
Meyer, L. B. (1956). Emotion and meaning in music. Chicago: Universityof Chicago Press.
Minati, L., Rosazza, C., D’Incerti, L., Pietrocini, E., Valentini, L.,Scaioli, V., et al. (2009). Functional MRI/event-related potentialstudy of sensory consonance and dissonance in musicians and non-musicians. NeuroReport, 20, 87–92.
Mineka, S., & Ohman, A. (2002). Phobias and preparedness: The selec-tive, automatic, and encapsulated nature of fear. Biological Psy-chiatry, 52, 927–937.
Nater, U. M., Abbruzzese, E., Krebs, M., & Ehlert, U. (2006). Sexdifferences in emotional and psychophysiological responses tomusical stimuli. International Journal of Psychophysiology, 62,300–308.
Norris, C. J., Larsen, J. T., & Cacioppo, J. T. (2007). Neuroticism isassociatedwith larger andmore prolonged electrodermal responses toemotionally evocative pictures. Psychophysiology, 44, 823–826.
Ohman, A., & Soares, J. J. (1994). ‘‘Unconscious anxiety’’: Phobic re-sponses to masked stimuli. Journal of Abnormal Psychology, 103,231–240.
Ohman, A., & Wiens, S. (2003). On the automaticity of autonomic re-sponses in emotion: An evolutionary perspective. In R. J. Davidson,K. R. Scherer, & H. H. Goldsmith (Eds.), Handbook of affectivesciences (pp. 256–275). Oxford: Oxford University Press.
Pallesen, K. J., Brattico, E., Bailey, C., Korvenoja, A., Koivisto, J.,Gjedde, A., et al. (2005). Emotion processing of major, minor, anddissonant chords: A functional magnetic resonance imaging study.Annals of the New York Academy of Science, 1060, 450–453.
Papa, A., & Bonanno, G. A. (2008). Smiling in the face of adversity: Theinterpersonal and intrapersonal functions of smiling. Emotion, 8,1–12.
Passynkova, N., Neubauer, H., & Scheich, H. (2007). Spatial organiza-tion of EEG coherence during listening to consonant and dissonantchords. Neuroscience Letters, 412, 6–11.
Peeters, G., & Czapinski, J. (1990). Positive-negative asymmetry in eval-uations: The distinction between affective and informational nega-tivity effects. In W. Stroebe & M. Hewstone (Eds.), European reviewof social psychology (Vol. 1, pp. 33–60). New York: Wiley.
Peretz, I., Blood, A. J., Penhune, V., & Zatorre, R. (2001). Corticaldeafness to dissonance. Brain, 124, 928–940.
Peretz, I., Gagnon, L., & Bouchard, B. (1998). Music and emotion:Perceptual determinants, immediacy, and isolation after braindamage. Cognition, 68, 111–141.
Perpina, C., Leonard, T., Bond, J. T., Bond, A., & Banos, R. (1998).Selective processing of food- and body-related information and au-tonomic arousal in patientswith eating disorders.The Spanish Journalof Psychology, 1, 3–10.
Plomp, R., & Levelt, W. J. (1965). Tonal consonance and critical band-width. The Journal of the Acoustical Society of America, 38, 548–560.
Prkachin, K. M., & Solomon, P. E. (2008). The structure, reliability andvalidity of pain expression: Evidence from patients with shoulderpain. Pain, 139, 267–274.
12 D. Dellacherie et al.
-
Regnault, P., Bigand, E., & Besson, M. (2001). Different brain mech-anisms mediate sensitivity to sensory consonance and harmonic con-text: Evidence fromauditory event-related brain potentials. Journal ofCognitive Neuroscience, 13, 241–255.
Roy, M., Mailhot, J. P., Gosselin, N., Paquette, S., & Peretz, I. (2008).Modulation of the startle reflex by pleasant and unpleasant music.International Journal of Psychophysiology, 71, 37–42.
Sammler, D., Grigutsch, M., Fritz, T., & Koelsch, S. (2007). Music andemotion: Electrophysiological correlates of the processing of pleasantand unpleasant music. Psychophysiology, 44, 293–304.
Sanchez-Navarro, J. P., Martinez-Selva, J. M., & Roman, F. (2006).Uncovering the relationship between defence and orienting in emo-tion: Cardiac reactivity to unpleasant pictures. International Journalof Psychophysiology, 61, 34–46.
Schellenberg, E. G., & Trainor, L. J. (1996). Sensory consonance and theperceptual similarity of complex-tone harmonic intervals: Tests ofadult and infant listeners. Journal of the Acoustical Society of Amer-ica, 100, 3321–3328.
Schneider, P., Scherg, M., Dosch, H. G., Specht, H. J., Gutschalk, A., &Rupp, A. (2002). Morphology of Heschl’s gyrus reflects enhancedactivation in the auditory cortex ofmusicians.NatureNeuroscience, 5,688–694.
Schoen, D., Regnault, P., Ystad, S., & Besson, M. (2005). Sensory con-sonance: An ERP Study. Music Perception, 23, 105–117.
Spielberger, C. D. (1983).Manual for the State-Trait Anxiety. Palo Alto:Consulting Psychologists Press Inc.
Steinbeis, N., Koelsch, S., & Sloboda, J. A. (2006). The role of harmonicexpectancy violations in musical emotions: Evidence from subjective,physiological, and neural responses. Journal of CognitiveNeuroscience, 18, 1380–1393.
Stormark, K. M., Laberg, J. C., Nordby, H., & Hugdahl, K. (2000).Alcoholics’ selective attention to alcohol stimuli: Automated pro-cessing? Journal of Studies on Alcohol and Drugs, 61, 18–23.
Tabachnick, B., & Fidell, L. S. (2001). Using multivariate statistics (4thEdition). Boston: Allyn & Bacon.
Taylor, S. E. (1991). Asymmetrical effects of positive and negative events:The mobilization-minimization hypothesis. Psychological Bulletin,110, 67–85.
Terhard, E. (1984). The concept of musical consonance: A link betweenmusic and psychoacoustics. Music Perception, 1, 176–195.
Trainor, L. J., &Heinmiller, B.M. (1998). The development of evaluativeresponses to music: Infants prefer to listen to consonance over dis-sonance. Infant Behavior & Development, 21, 77–88.
Trainor, L. J., Tsang, C. D., & Cheung, V. H. W. (2002). Preference forsensory consonance in 2- and 4-month-old infants.Music Perception,20, 187–194.
Tramo, M. J., Cariani, P. A., Delgutte, B., & Braida, L. D. (2001).Neurobiological foundations for the theory of harmony in Westerntonal music. Annals of the New York Academy of Science, 930,92–116.
Witvliet, C. V., & Vrana, S. R. (1995). Psychophysiological responses asindices of affective dimensions. Psychophysiology, 32, 436–443.
Witvliet, C. V., & Vrana, S. R. (2007). Play it again Sam: Repeatedexposure to emotionally evocative music polarises liking and smilingresponses, and influences other affective resports, facial EMG, andheart rate. Cognition and Emotion, 21, 3–25.
Zajonc, R. B. (1980). Feeling and thinking: Preferences need no infer-ences. American Psychologist, 35, 151–175.
Zentner, M. R., & Kagan, J. (1996). Perception of music by infants.Nature, 383, 29.
(Received April 20, 2009; Accepted April 13, 2010)
Psychophysiology of musical emotion 13