Brain and Cognition xxx (2011) xxx–xxx

Contents lists available at ScienceDirect

Brain and Cognition
journal homepage: www.elsevier.com/locate/b&c

Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting A scientist’s tribute to Maria Callas
Felicia Rodica Baltes, Julia Avram, Mircea Miclea, Andrei C. Miu ⇑ ß
Emotion and Cognition Neuroscience Laboratory, Department of Psychology, Babes-Bolyai University, Cluj-Napoca, CJ 400015, Romania

a r t i c l e

i n f o

a b s t r a c t
Operatic music involves both singing and acting (as well as rich audiovisual background arising from the orchestra and elaborate scenery and costumes) that multiply the mechanisms by which emotions are induced in listeners. The present study investigated the effects of music, plot, and acting performance on emotions induced by opera. There were three experimental conditions: (1) participants listened to a musically complex and dramatically coherent excerpt from Tosca; (2) they read a summary of the plot and listened to the same musical excerpt again; and (3) they re-listened to music while they watched the subtitled film of this acting performance. In addition, a control condition was included, in which an independent sample of participants succesively listened three times to the same musical excerpt. We measured subjective changes using both dimensional, and specific music-induced emotion questionnaires. Cardiovascular, electrodermal, and respiratory responses were also recorded, and the participants kept track of their musical chills. Music listening alone elicited positive emotion and autonomic arousal, seen in faster heart rate, but slower respiration rate and reduced skin conductance. Knowing the (sad) plot while listening to the music a second time reduced positive emotions (peacefulness, joyful activation), and increased negative ones (sadness), while high autonomic arousal was maintained. Watching the acting performance increased emotional arousal and changed its valence again (from less positive/sad to transcendent), in the context of continued high autonomic arousal. The repeated exposure to music did not by itself induce this pattern of modifications. These results indicate that the multiple musical and dramatic means involved in operatic performance specifically contribute to the genesis of musicinduced emotions and their physiological correlates. Ó 2011 Elsevier Inc. All rights reserved.

Article history: Accepted 31 January 2011 Available online xxxx Keywords: Operatic music Music-induced emotions Physiological differentiation of emotions

‘‘Maria Callas exploded the concept of what beautiful singing means: Is it pretty sounds and pure tones? Or should beauty evolve from text, musical shape, dramatic intent and, especially, emotional truth?’’ (Anthony Tommassini in ‘‘A Voice and a Legend That Still Fascinate; Callas Is What Opera Should Be’’, The New York Times, September 15, 1997)
Abbreviations: DBP, diastolic blood pressure; ECG, electrocardiogram; GEMS, Geneva Emotional Music Scale; HF-HRV, power in the high frequency band of HRV; HR, heart rate; HRV, heart rate variability; IBI, cardiac interbeat intervals; LF-HRV, power in the low frequency band of HRV; NA, negative affect; PA, positive affect; PANAS, Positive and Negative Affect Schedule; RR, respiratory rate; RSA, respiratory sinus arrhythmia; SAM, Self-Assessment Manikin; SBP, systolic blood pressure; SCL, skin conductance level; SEM, standard error of the mean; VLF-HRV, power in the very low frequency band of HRV. ⇑ Corresponding author. Address: 37 Republicii, Cluj-Napoca, CJ 400015, Romania. Fax: +40 264 590967. E-mail address: andreimiu@gmail.com (A.C. Miu). 0278-2626/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.bandc.2011.01.012

1. Introduction We are often emotionally moved by musical performances. However, emotions induced by music have only recently drawn the attention of scholars in cognitive and affective sciences (Juslin & Vastfjall, 2008; Scherer & Zentner, 2001). Field studies have confirmed that music pervades everyday life and some of its most important functions are related to mood change and emotion regulation (DeNora, 1999; Juslin, Liljestrom, Vastfjall, Barradas, & Silva, 2008; Sloboda & O’Neil, 2001). In daily life, music generally increases positive affect, alertness, and focus in the present (Sloboda, O’Neil, & Ivaldi, 2001). In addition, it provides opportunities for venting strong emotions, increasing their intensity, or calming down (DeNora, 1999). Therefore, music has been related to the genesis and control of emotions. Despite previous debates on whether music induces emotions in listeners (i.e., the so-called ‘‘emotivist’’ position), or only expresses emotions that listeners can recognize (i.e., the ‘‘cognitivist’’

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

2

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

position) (Kivy, 1990; Scherer & Zentner, 2001), the recent literature has generally supported the former view that music induces subjective (e.g., self-reported sadness), behavioral (e.g., crying), and physiological changes (e.g., heart rate [HR – see list of acronyms] deceleration) that are characteristic of emotions (Bharucha, Curtis, & Paroo, 2006; Juslin & Vastfjall, 2008; Koelsch, 2005; Scherer & Zentner, 2001). In addition, the mechanisms by which music induces emotions (e.g., semantic associations, emotional contagion based on observation of facial and vocal expressions; see Bezdek & Gerrig, 2008; Hietanen, Surakka, & Linnankoski, 1998; Lundqvist & Dimberg, 1995) may not be specific to music, but this possibility has only recently started to be investigated (for reviews, see Juslin & Vastfjall, 2008; Scherer & Zentner, 2001). The present report stems from the emotivist approach, and will examine the effects of opera on listeners’ physiological responses and subjective ratings of their own emotions. One way to investigate these issues has been to identify physiological responses during music listening (e.g., Krumhansl, 1997; Nyklícek, Thayer, & Van Doornen, 1997). This approach has extended the studies on the physiological differentiation of emotions induced by facial expressions (e.g., Ekman, Levenson, & Friesen, 1983), images (e.g., Codispoti, Bradley, & Lang, 2001), and even natural sounds (e.g., Bradley & Lang, 2000). Previous studies indicated that only certain emotions (e.g., fear, disgust) can be distinguished based on their autonomic signatures (for review see Levenson, 1992), but the effect sizes were small or medium at best (Cacioppo, Berntsen, Klein, & Poehlmann, 1997). These findings are not surprising considering the limited emotional saliency of images and words presented in laboratory settings. Recent psychophysiological studies have used more complex stimuli such as films, and consequently induced more robust experiences of emotion and physiological responses (e.g., Frazier, Strauss, & Steinhauer, 2004; Kreibig, Wilhelm, Roth, & Gross, 2007). 1.1. Psychophysiology of music-induced emotions Like films, music has been shown to produce physiological changes that can distinguish between emotions. In two landmark studies, Krumhansl (1997), and Nyklícek et al. (1997) measured a large array of cardiovascular, respiratory, and electrodermal responses in association with self-report measures of emotions induced by music. Emotions were differentiated based on certain physiological responses such as respiratory sinus arrhythmia (RSA) and cardiac interbeat intervals (IBI) (Nyklícek et al., 1997). For instance, sadness ratings correlated positively with IBI, systolic (SBP) and diastolic blood pressure (DBP), and negatively with skin conductance level (SCL) (Khalfa, Peretz, Blondin, & Manon, 2002; Krumhansl, 1997). Emotional arousal was best explained by physiological changes, which accounted for 62.5% of the variance (Nyklícek et al., 1997). There is only one psychophysiological field study that measured emotional ratings, electrodermal and respiratory responses in a sample of spectators (i.e., 27 listeners) during several live performances of Wagner’s operas given in the festival theater of Bayreuth in 1987–1988 (Vaitl, Vehrs, & Sternagel, 1993)1. In contrast to laboratory studies, these limited field results suggested that physiological responses differed between opera leitmotivs, but there was a weak correspondence between physiological and subjective measures of emotions. Psychophysiological studies have thus focused on the coherence between subjective, behavioral, and physiological components of music-induced emotions. Lundqvist, Carlsson, and Juslin (2009) reported an association between music-induced happiness and

greater SCL, and supported the emotivist position. In contrast, another study found that increased emotional arousal occurred without changes in SCL (Grewe, Nagel, Kopiez, & Altenmuller, 2007a). The latter pattern of results was interpreted as evidence for the cognitivist position, although the participants were clearly instructed to rate the emotional arousal they felt, and not that expressed by the music. These apparently divergent results might be explained by methodological differences, considering that one study used a self-report instrument that measured changes in several basic emotions (Lundqvist et al., 2009), and the other measured changes in arousal and valence across emotions (Grewe et al., 2007a). In addition, there are emotions specifically induced by music that are not captured by basic emotion measures such as the one used by Lundqvist et al. (2009). 1.2. Specific music-induced emotions It has been argued that aesthetic emotions are deeper and more significant (Sloboda, 1992), nuanced and subtle (Scherer & Zentner, 2001) than other more general emotions. Indeed, the range of music-induced emotions goes beyond the emotions captured by the basic emotion models. A recent field study showed that a nine-factor model best fitted the emotion descriptors that were chosen by music listeners who attended a classical music festival (Zentner, Grandjean, & Scherer, 2008). It included emotion categories (e.g., wonder, transcendence) that are not part of any current model of emotion. The Geneva Emotional Music Scale (GEMS) is the first questionnaire designed to measure music-induced emotions (Zentner et al., 2008). To our knowledge, no study has investigated the correlation between physiological responses and music-induced emotions measured by GEMS. 1.3. Music-induced chills Music-induced emotions are often accompanied by physical sensations such as chills (i.e., tremor or tingling sensations passing through the body as the result of sudden keen emotion or excitement). Two landmark studies indicated that the great majority of people were susceptible to chills (Sloboda, 1991), and these bodily phenomena were associated with music-induced emotions, especially sadness and melancholy (Panksepp, 1995). Musical events such as crescendos or a solo instrument (e.g., a soprano’s voice) emerging from a softer orchestral background induced chills (Grewe, Nagel, Kopiez, & Altenmuller, 2007b; Panksepp, 1995). Psychophysiological studies have shown that music-induced chills correlated with increases in SCL and HR (Grewe et al., 2007b; Rickard, 2004). The present study aims to integrate the measurement of chills, music-induced emotions reflected by GEMS, and a wider range of physiological changes. 1.4. The duration of musical stimuli One important aspect that differentiates studies of music-induced emotions is the duration of stimuli. For instance, many studies used short (i.e., several seconds), monotonic musical stimuli. It has been suggested that even less than one second of music is sufficient to prime an emotional meaning (e.g., Bigand, Vieillard, Madurell, Marozeau, & Dacquet, 2005; Peretz, Blood, Penhune, & Zatorre, 2001; Watt & Ash, 1998). However, this approach has at least two limitations. First, it usually involves forced-choice responses that increase the difficulty of emotional valence processing (Bigand et al., 2005; Peretz et al., 2001). Second, the correct categorization of the emotional content of music may only reflect the emotions that listeners perceive in music. One second may not be enough time to develop an emotional response. At any rate, longer durations of musical stimuli increase the magnitude of

1 A recent laboratory study on psychophysiological changes induced by opera came to our attention while this article was under review. See Bernardi et al. (2009).

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

3

psychophysiological responses in music-induced emotions (Witvliet & Vrana, 2007). Psychophysiological studies generally used longer stimuli (i.e., ranging from 6 to 600 s), and it has been argued that the use of full music pieces has greater external validity when investigating emotional responses to music (Grewe et al., 2007a; Nater, Abbruzzese, Krebs, & Ehlert, 2006; Rickard, 2004). 1.5. Multiple sources of emotion in operatic music The duration of musical stimuli, as well as the integration of music with congruent visual and verbal cues are important contributors to emotional responses that people develop to musical performance (Bezdek & Gerrig, 2008; Scherer & Zentner, 2001). Operatic music performance involves both singing and acting, which multiplies the mechanisms by which emotions are induced in listeners. Opera adds the power of the dramatic plot and the personality of the performer to the affective message of the musical score and the emotional expressivity of voice (Scherer, 1995). The rich audiovisual background arising from the orchestra and elaborate scenery and costumes are also important. The objective of the present study was to investigate for the first time the cumulative contributions of music listening, learning the context of the events it portrays (i.e., plot), and watching the acting performance to emotions induced by opera. These sources may support the genesis of emotion either independently or in concert. Research on film music supports the latter possibility. For instance, music presented during the opening scene of a film influenced the emotional valence of words that participants used in their continuations of the narratives (Vitouch, 2001). In addition, judgments of characters displaying neutral emotions were significantly affected by the emotional content of the music that accompanied the film (Tan, Spackman, & Bezdek, 2007). Lyrics are also important in emotional responses to music. For instance, the emotional effects of music and lyrics were investigated by combining musical excerpts with lyrics that conveyed the same emotion or another emotion (Ali & Peynircioglu, 2006; Stratton & Zalanowski, 1994). These studies indicated that lyrics enhanced emotion in sad and angry music. Furthermore, these emotions readily transferred to images that were arbitrarily associated with songs (Ali & Peynircioglu, 2006). In addition, visual cues such as facial expressions are preattentively integrated with vocal cues and influence the emotional judgment of the latter (de Gelder, Bocker, Tuomainen, Hensen, & Vroomen, 1999). Therefore, it seems likely that facial expressions of singers influence the emotional processing of music. Overall, music, lyrics, and visual cues seem to significantly contribute to the genesis of music-induced emotions, and their concerted contribution may explain why operatic music is so effective in inducing emotions. However, this complex issue has not been investigated to date. 1.6. Objectives of the present study

The present study had three experimental conditions that investigated the contributions of music, plot, and acting performance to emotional responses. First, participants listened to the musical excerpt. Then, they read a summary of the plot and listened to the same musical excerpt again. In the third condition, they re-listened to music while they watched the subtitled film of this acting performance. In between conditions, we measured music-induced emotions using both dimensional, and specific music-induced emotion questionnaires. During the experimental conditions, cardiovascular, electrodermal and respiratory responses were continuously recorded, and the participants kept track of their musical chills. Since there are very few psychophysiological studies of emotions in operatic music (and operatic music is so diverse), the present study was consequently exploratory. Based on the musical and dramatic content of this musical excerpt, we expected that it would induce a pattern of emotions characterized by increased unpleasant emotions (e.g., sadness) and decreased pleasant emotions (e.g., joyful activation, peacefulness). In addition, based on the literature in related areas (e.g., sadness induced by films), we expected a change in the sympathovagal balance, with vagal withdrawal and sympathetic activation, as well as decreases of SCL and respiratory rate (RR). We were specifically interested in the way each successive layer of complexity influenced music-induced emotions and their physiological correlates. 2. Methods 2.1. Participants N = 37 healthy, right-handed Romanian volunteers (25 women; mean age = 21.4 years, ranging between 19 and 24 years), with good hearing, were selected for this study (out of an initial pool of 45 volunteers). The sample size was determined by using a priori statistical power analysis (power = 0.95; alpha = 0.05; effect size f = 0.25) run in the G-Power 3.1 software (Faul, Erdfelder, Lang, & Buchner, 2007). The participants had no significant musical education, but they reported that music was an important part of their lives. None of the participants reported having listened to Tosca before, a preference for classic or operatic music, or understood Italian. These inclusion criteria were important in order to control for the degree of familiarity with the selected musical piece, and understanding of the lyrics. None of the participants reported cardiovascular or neurological problems, or any kind of medical treatment that would interfere with cardiovascular and autonomic functions. Participants were asked to refrain from alcohol, caffeine and smoking at least four hours before the experiment. All the participants signed an informed consent to participate to the experiment and the procedures complied with the recommendations of the Declaration of Helsinki for human studies. 2.2. Materials

We investigated subjective and physiological emotional responses to operatic music. In order to maximize external validity, we chose a dramatically coherent and musically complex excerpt from Tosca by Giacomo Puccini. The soprano Maria Callas and the baritone Tito Gobbi gave a legendary interpretation of the main characters in Tosca, and their 1964 live performance at Covent Garden was fortunately recorded on film. In this performance, both artists impress by their emotional identification with the characters, and the way they deliver the mixture of lust and hate, fear, emotional vulnerability and indignation through their voice (Huck, 1984). Studying the psychophysiology of emotion during this performance offers us an opportunity to catch a scientific glimpse of the emotional force that artists such as Maria Callas have inspired.

We used an excerpt from Giacomo Puccini’s Tosca (Act II), filmed at Covent Garden in 1964, starring Maria Callas as Floria Tosca, Tito Gobbi as Scarpia, and Renato Cioni as Mario Cavaradossi (Zeffirelli, 2002). We selected and juxtaposed two excerpts (i.e., excerpt 1 from 110 :0000 [Scarpia: Ed or fra noi parliam da buoni amici] to 220 :3100 [Scarpia: Io? Voi!], and excerpt 2 from 230 :3600 [Tosca: Quanto?] to 310 :3500 [Tosca: Perché me ne rimuneri cosi?]) for the following reasons. First, these excerpts contain the plot (see Supplementary materials) involving all the three main characters (i.e., Tosca, Scarpia, and Cavaradossi). Second, these excerpts are musically and dramatically heterogenous, with a variety of rhythmical dynamics, ascending and descending scales, large vocal range and emotional tension. In

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

4

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

addition, our approach to inducing music-related emotions explicitly relied on using longer excerpts (e.g., 190 :3000 in the present study) from popular operatic compositions in order to credibly replicate the musical context that induces emotions in the real world (Grewe et al., 2007a; Juslin & Vastfjall, 2008; Rickard, 2004). Music was presented using Technics RP-F600 high-quality noise canceling closed headphones. Before the start of the experiment, a test tone was played, giving participants the opportunity to adjust the loudness to an individually comfortable level. After the participants read the plot before the second experimental condition, the experimenters checked how well the plot was understood by asking the participants the following questions: (1) who are the main characters; (2) what happens in this opera; and (3) what happens in this excerpt of the opera? The great majority of the participants answered correctly to these questions, but those who omitted or were not sure of certain details were allowed to read the summary of the plot again and assisted with supplementary explanations by the experimenters. This experimental condition started only after each participant correctly answered all the questions regarding the plot. The video was displayed on a Samsung SyncMaster 205BW monitor (50.8 cm), located 1.5 m in front of the participant’s chair. The experimental room was small and dimly lit, and was maintained at a comfortable ambient temperature. 2.3. Procedure There were three conditions of musical experience: (1) music listening; (2) music re-listening after learning the plot; and (3) music re-listening while watching the acting performance. Previous studies revealed that the psychophysiological responses induced by music are not significantly affected by repeated exposure (Grewe et al., 2007a, 2007b). However, we also included a control condition in which an independent sample of N = 9 participants (five women) successively listened three times to the same musical excerpt, in order to check whether the repeated exposure to music influenced the subjective and physiological measures. The same questionnaires and physiological recordings were used in the main experiment and the supplementary control condition, except SBP and DBP that were not measured in the latter condition. The participants in this control experiment met all the inclusion criteria that applied to the main experiment. At the arrival to the laboratory, each participant completed the general scales of the Positive and Negative Affect Schedule (PANASI) (Watson & Clark, 1994), in order to control for differences in affective mood before the start of the experiment. After a habituation period during which participants were explained that several non-invasive recordings will be taken during music listening, the physiological electrodes for SCL and electrocardiogram (ECG), as well as the respiration transducer and an arm cuff coupled to an automatic blood pressure monitor were attached. Participants were instructed to sit comfortably and relax, and carefully listen to the music while monitoring the music-related emotions they felt without trying to control them in any way. They were instructed to identify emotions they felt during music listening, and not emotions that the music expressed. They were also requested to keep a count on a scratch sheet of the number of chills they experienced during each condition. Each condition was preceded by a 5 min interval during which baseline physiological recordings were made. Participants completed each condition and unless they wanted a break, they moved onto the following condition. First, they listened to the musical excerpt. In the second condition, they were given a summary of the plot (see Supplementary materials). Using a brief questionnaire, the experimenters first made sure that participants understood the plot and knew the characters, and then music was played again. In the third condition, the participants listened to music

while also watching the acting performance. In order to facilitate the complete understanding of the plot and acting performance, the movie was subtitled in Romanian. After each condition, participants were required to rate the emotional arousal (1 – non-arousing to 5 – arousing) and valence (1 – unpleasant to 5 – pleasant) induced by music; and completed GEMS (Zentner et al., 2008) for music-induced emotions. 2.4. Self-report measures The positive (PA) and negative affect (NA) scales of PANAS-I (Watson & Clark, 1994) include 20 items each, which measure the affective mood in the past few weeks until present. Emotional arousal and valence were measured using the Self-Assessment Manikin (SAM) (Bradley & Lang, 1994). SAM is a non-verbal pictorial assessment technique that directly measures the pleasure and arousal (as well as dominance, which was not used in the present study) associated with a person’s affective reaction to a wide variety of stimuli. For the measurement of emotions induced by music (e.g., wonder, transcendence, tenderness, peacefulness), we used the long (i.e., 45 items) variant of GEMS (Zentner et al., 2008). GEMS scores are grouped on nine factors: wonder; transcendence; tenderness; nostalgia; peacefulness; power; joyful activation; tension; and sadness. Whereas the dimensional rating allowed us to document general changes of emotional arousal and valence, GEMS offered us the possibility of actually identifying the specific emotions that were induced by each experimental condition. Self-reports of chills were also collected. 2.5. Physiological measures ECG, SCL, and respiration were continuously recorded during the baseline and experimental conditions, using a BIOPAC MP150 system and specific electrodes and transducers. Blood pressure was intermittently measured at fixed intervals during the experimental condition. 2.5.1. Cardiovascular measures ECG was recorded using disposable pregelled Ag/AgCl electrodes placed in a modified lead II configuration, at a sample rate of 500 samples/s, and amplified using an ECG100C module. After visual inspection of the recordings and editing to exclude artifacts in AcqKnowledge 3.9.0.17, all the recordings were analyzed using Nevrokard 7.0.1 (Intellectual Services, Ljubljana, Slovenia). We calculated HR, and HR variability (HRV) indices in the time and frequency domains: mean IBI between successive R waves (HR and IBI are negatively correlated); power in the high frequency (HF-HRV) band ($0.15–0.4 Hz in adults) of HRV, also known as RSA; power in the low (LF-HRV) ($0.05–0.15 Hz), and very low frequency (VLF-HRV) ($0–0.05 Hz) bands of HRV, as well as LF/HF ratios. The latter three measures, obtained by spectral analysis, are reported in normalized units (see Task Force Report, 1996). RSA reflects vagal modulation of the heart, whereas LF-HRV reflects a complex interplay between sympathetic and vagal influences (see Eckberg, 1997; Kingwell et al., 1994; Miu, Heilman, & Miclea, 2009; Task Force of the European Society of Cardiology and Electrophysiology, 1996). These measures were derived from each baseline and experimental conditions. The statistical analyses of RSA included respiration frequency as covariate in order to control for the influence of respiration on this measure. Therefore, the results reported here controlled for the influence of respiration on RSA. 2.5.2. Skin conductance After cleaning and abrading the skin of the palms, TSD203 electrodermal response electrodes filled with isotonic gel were attached to the volar surfaces of the index and medius fingers. SCL

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

5

recordings were amplified using a GSR100C module. We estimated SCL by extracting the area under the curve (lS/s) from each baseline and experimental condition, after the downdrift in the SCL waves was eliminated using the ‘‘difference’’ function of AcqKnowledge, as described in (Bechara, Damasio, Damasio, & Lee, 1999; Miu, Heilman, & Houser, 2008). 2.5.3. Respiration One channel of respiration was measured using a top respiration band placed on the chest, below the breast. The data were recorded with the RSP100C module and the TSD201 Transducer of the Biopac system. TSD201 can arbitrarily measure slow to very fast thoracic and abdominal respiration patterns with no loss in signal amplitude, optimal linearity and minimal hystheresis. RR (in cycles per minute) was calculated breath by breath using AcqKnowledge software. 2.5.4. Blood pressure SBP and DBP (in millimeters of mercury) were measured intermittently with an automatic blood pressure monitor (Digital Blood Pressure monitor, Vital System) through an arm cuff at the participant’s right upper arm. Inflation was initiated at the end of the baseline, at minutes 5, 10, 15, and at the end of the musical condition. 2.6. Data reduction For the continous physiological measurements (i.e., all except SBP and DBP), we calculated difference scores by subtracting each baseline measure (i.e., the quiet sitting period immediately preceding each musical experience condition) from the corresponding experimental condition measure (see Kreibig et al., 2007). In the case of SBP and DBP that were intermittently measured, we first calculated the arithmetic mean of the physiological data from baseline and experimental conditions, and then derived the same difference score. The raw scores were transformed to T scores for normalization. 2.7. Statistical analysis Data were inspected for outliers (Stevens, 2002, pp. 14–17) – only 0.8% of the data were excluded. We used repeated measure ANOVA and ANCOVA, followed by post hoc tests, in order to determine whether there were differences in emotion experience and physiological responses between the musical experience conditions. Effect sizes for t-tests and AN/COVA are reported as Cohen’s d and g2 , and interpreted as follows: d = 0.2 or g2 = 0.01 – small efp p fect size; d = 0.5 or g2 = 0.059 – medium effect size; and d = 0.8 or p 2 gp = 0.138 – large effect size (Cohen, 1988). We also used the Friedmann non-parametric test to analyze potential differences between the frequency of chills in the experimental conditions. In addition, correlation analyses allowed us to test the association between emotion experience, physiological responses, and chills. Simple regressions were used to test whether affective mood (i.e., PA and NA) predicted affect (i.e., dimensional and specific emotion ratings) and physiological responses. The data are reported in the graphs as means ± one standard error of the means (SEM). 3. Results 3.1. General affect A 3 (musical experience: music listening vs. learning the plot vs. watching the acting performance) Â 2 (sex: women vs. men)

Fig. 1. Changes in emotional arousal and valence (SAM) induced by music listening (1), learning the plot (2), and watching the acting performance (3).

ANCOVA indicated that musical experience had a significant main effect on self-reported emotional arousal (F[4, 32] = 6.19, p = 0.002, g2 = 0.12). NA and PA were included as covariates in these analyses p in order to control for the affective mood of participants before the experiments. The analyses of the data from the supplementary control sample indicated that the repeated exposure to music had no significant effects on emotional arousal and valence (p = 0.3 for both) (see Supplementary Fig. 1). In addition, we compared the first music listening condition in the control experiment to the music listening condition from the main experiment, in order to verify their similarity. Indeed, there were no significant differences between the arousal (t[45] = 1.29, ns) and valence scores (t[45] = 1.23, ns) in the first conditions of the main and control experiments, respectively. Although emotional arousal and valence were not measured before the first condition because it would have been hard to find an equally complex, but emotionally neutral stimulus to which to compare the first experimental condition, we explored the affective experience that music listening induced by one sample Student ttests. The expected mean was the mid-value of the SAM rating scale. These analyses indicated that music listening was associated with increased emotional arousal (t[35] = 2.42, p = 0.02, Cohen’s d = 0.3) and valence scores (t[35] = 8.57, p < 0.0001, Cohen’s d = 1). Next, by comparing between the three experimental condition, we found that watching the acting performance significantly increased emotional arousal compared to learning the plot, and music listening (see Fig. 1). Neither the main effect of sex, nor the interaction of sex  musical experience on emotional arousal and valence were statistically significant.

3.2. Music-induced emotions The effects of musical experience and sex on music-induced emotions measured by GEMS were also investigated. A 3 (musical experience: music listening vs. learning the plot vs. watching the acting performance) Â 2 (sex: women vs. men) ANCOVA indicated that musical experience induced specific emotions. NA and PA were again included as covariates in these analyses. The analyses of the data from the supplementary control sample indicated that the repeated exposure to music had no significant effects on any of the GEMS measures (p > 0.1 for all) (see

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

6

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

decreased IBI (F[4, 32] = 2.98, p = 0.05, g2 = 0.08), and SCL p (F[4, 32] = 3.2, p = 0.04, g2 = 0.09) in comparison to music listening. p

3.4. Experienced chills The repeated exposure of the independent control sample to music had no significant effect on self-reported chills (p = 0.3). There were no differences between the frequency of chills in the control and main experiments, respectively. A Friedman non-parametric test compared between the three experimental conditions in the main experiment and indicated that the exposure to the acting performance significantly increased the number of reported chills (v2 = 8.92, p = 0.01) in comparison to learning the plot and music listening.

3.5. Relationships between music-induced affect, chills, and physiological responses We analyzed the correlations between emotions, chills, and physiological responses within each musical experience condition. The following paragraph reports the main patterns of correlations for which we had a priori hypotheses (for detailed results, see Tables 1–3). These analyses indicated that LF-HRV positively, and RSA negatively correlated with emotional arousal after learning the plot. In the same condition, the frequency of chills also correlated with arousal. In contrast, RR positively correlated with emotional valence (i.e., increased RR for positive valence) during music listening. The analyses of music-induced emotions showed that LF-HRV positively, and RSA negatively correlated with the level of wonder, power, and joyful activation after learning the plot. Also, LF-HRV positively and RSA negatively correlated with the frequency of chills both after learning the plot, and while watching the acting performance. Chills consistently correlated positively with the levels of wonder and transcendence in all three musical experience conditions. We also checked if this correlation was replicated in the control experiment and we confirmed that chills correlated significantly with wonder (r = 0.65, p = 0.05) and marginally with transcendence (r = 0.6, p = 0.08). Another consistent pattern of positive correlations was that between RR, wonder (during all three musical experience conditions), and transcendence (during music listening, and watching the acting performance).

Fig. 2. Changes in GEMS scores induced by music listening (1), learning the plot (2), and watching the acting performance (3).

Supplementary Fig. 2). We also compared the pattern of GEMS scores between the first conditions of the main and control experiments. There was only one significant difference on tenderness (t[45] = 2.25, p = 0.02), with higher scores in the music listening condition of the main experiment. By comparing between the three experimental condition, we found that learning the plot and watching the acting performance had significant effects on distinct music-induced emotions (see Fig. 2). On the one hand, learning the plot reduced the scores of peacefulness (F[4, 32] = 7.84, p = 0.0009, g2 = 0.23) and joyful actip vation (F[4, 32] = 5.85, p = 0.004, g2 = 0.17), and increased sadness p (F[4, 32] = 10.98, p = 0.0001, g2 = 0.32). On the other hand, watchp ing the acting performance increased the scores of wonder (F[4, 32] = 8.13, p = 0.0007, g2 = 0.23) and transcendence p (F[4, 32] = 4.02, p = 0.02, g2 = 0.11). Neither the main effect of sex, p nor the interaction of sex  musical experience on specific emotions were statistically significant.

3.3. Physiological responses The analyses of the data from the supplementary control sample indicated that the repeated exposure to music had no significant effects on any of the physiological measures (p > 0.39 for all) (see Supplementary Fig. 3). However, a couple of physiological measures were significantly different between the first conditions of the main and control experiments: IBI (t[45] = 4.77, p < 0.0001) and RR (t[45] = 2.09, p = 0.04), with lower values in the first condition of the control experiment. There were significant main effects of musical experience on physiological responses. In comparison to baseline measures, music listening (i.e., the first condition) significantly decreased RR (F[4, 32] = 9.12, p = 0.005, g2 = 0.29), IBI (F[4, 32] = 3.11, p = 0.02, p g2 = 0.09), and SCL (F[4, 32] = 29.76, p < 0.0001, g2 = 0.75). In the p p following experimental conditions, both learning the plot, and watching the acting performance specifically influenced physiological measures (Fig. 3). On the one hand, learning the plot significantly decreased RSA (F[4, 32] = 3.05, p = 0.05, g2 = 0.08) and p increased LF-HRV (F[4, 32] = 3.49, p = 0.03, g2 = 0.09) and LF/HF p (F[4, 32] = 3.77, p = 0.02, g2 = 0.1) in comparison to music listening. p On the other hand, watching the acting performance significantly 3.6. Previous mood and music-induced affect PA and NA significantly correlated (r = À0.45, p < 0.01), but the low correlation allowed us to use both as predictors (i.e., negligible multicollinearity). Our hypotheses were that NA would positively predict unpleasant emotions measured by GEMS (i.e., nostalgia, sadness, tension), and negatively predict pleasant emotions (i.e., wonder, transcendence, power, tenderness, peacefulness, joyful activation). We also expected that PA would negatively predict unpleasant emotions and positively predict pleasant emotions. In addition, based on the work of Panksepp (1995), we also hypothesized that NA would negatively predict chills and RSA. On the assumption that only the first condition (i.e., music listening) would be directly affected by previous mood, regression analyses were run on music-induced emotions and chills recorded during the first condition. The results indicated that power (R = 0.53, p = 0.0009, g2 = 0.28) and joyful activation (R = 0.45, p = 0.05, p g2 = 0.21) were negatively predicted by NA. In contrast, PA posip tively predicted power (R = 0.51, p < 0.001, g2 = 0.26) and joyful p activation (R = 0.38, p = 0.02, g2 = 0.15). p

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

7

Fig. 3. Changes in interbeat intervals (IBI), heart rate (HR), power in the very low frequency (VLF), and low frequency (LF) bands of heart rate variability, respiratory sinus arrhythmia (RSA), sympathovagal balance (LF/HF), skin conductance level (SCL), systolic blood pressure (SBP), diastolic blood pressure (DBP), and respiratory rate (RR) induced by music listening (1), learning the plot (2), and watching the acting performance (3).

4. Discussion The results of this study confirmed that music listening, learning the plot, and watching the acting performance had specific effects on emotional responses measured at the subjective and physiological levels. 4.1. Effects of music, plot and acting In comparison to expected mean scores, music listening increased, as one would expect, emotional arousal and valence. In addition, music listening decreased RR, IBI and SCL, in comparison to baseline physiology. These results seem to extend previous observations that sad music is associated with decreased SCL, and sadness induced by music is well discriminated by respiratory changes (Krumhansl, 1997; Nyklícek et al., 1997). Moreover, our observation of decreased SCL associated with this music excerpt is also in line with studies that induced sadness by directed facial action tasks (Ekman et al., 1983; Levenson, 1992). It may seem that the pattern of reduced RR and SCL, and increased HR (i.e., decreased IBI) in the music listening condition is contradictory. Early observations indicated that the minor tonalities of music increased HR (Hyde & Scalapino, 1918), whereas the tempo of music influenced RR (Diserens, 1920). Bernardi and colleagues (2009) have recently reported that music crescendos or emphases (e.g., in Nessun dorma from Puccini’s Turandot) induced skin vasoconstriction along with increases in blood pressures and HR. There was also increased breath depth during music crescendos, but these modulations of respiratory power were independent of cardiovascular modulations. The present study also shows that music listening independently modulated RR and HR, and the former correlated with negative valence, wonder and transcendence. Also in line with the present results, Nakahara, Furuya, Francis, & Kinoshita, (2010) found that playing Bach’s No. 1 Prelude with emotional expression increased HR and decreased RR in pianists, in comparison to playing the same piece

without emotional expression. Therefore, these studies suggest that music-induced emotions can independently modulate cardiovascular and respiratory activities, and this pattern of physiological changes may contribute to the receptiveness or arousal to music (Bernardi et al., 2009; the present study) and the capacity of performers to incorporate expressiveness in their performance (Nakahara, Furuya, Francis, & Kinoshita, 2010). Our control analyses on the data from an independent sample indicated that re-listening to the musical excerpt for three times did not increase emotional arousal and valence, or induced additional physiological changes by itself. Whereas there were no differences between the conditions of the control experiment, which argued that repeated music listening alone did not affected the subjective and physiological measurements, the relevance of the physiological measurements from the control experiment is limited. There were differences in IBI and RR between the sample used in the main and control experiments, respectively. This was probably due to the reduced sample size in the control experiment (N = 9, in comparison to N = 37 in the main experiment). Overall, the control data supported the view that the changes observed in the main experiment were not due to repeated music listening alone, although this inference should be taken with caution in regard to some of the physiological results. Replicating the control findings with a sample size that is similar to that of the main experiment would be necessary in order to unequivocally confirm that the repeated music listening alone does not change physiological activity. Learning the plot before listening to the musical excerpt the second time (in the main experiment) induced a pattern of emotional changes that included reduced peacefulness, joyful activation, and increased sadness. At the physiological level, learning the plot decreased RSA and increased LF-HRV. The change in RSA reflects vagal suppression that has been associated with negative emotional states and traits, such as anxiety and depression (Bleil, Gianaros, Jennings, Flory, & Manuck, 2008; Miu et al., 2009). The summary of the plot that the participants read before they

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

8

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

re-listened to the musical excerpt described negative emotional events (e.g., Scarpia tortures Cavaradossi and harasses Tosca; see Supplementary materials). Therefore, we argue that the sadness induced by learning the plot triggered vagal suppression that was neither explained by concomitant respiratory changes (i.e., RR was controlled for in the analyses of RSA), nor by re-listening to the musical excerpt by itself. The increase in LF-HRV suggests that learning the plot also facilitated sympathetic activity. However, LF probably reflects a complex interplay between sympathetic and vagal influences on the heart (Eckberg, 1997; Miu et al., 2009), so the effect of learning the plot on sympathetic activity should be taken with caution. Overall, learning the plot significantly influenced music-induced emotions and changed sympathovagal balance in the direction of greater preparedness for action. Watching the acting performance increased emotional arousal and valence (SAM) compared to the first two experimental conditions. Furthermore, it increased wonder and transcendence (GEMS). Notably, wonder and transcendence are emotions that are specifically induced by music (Zentner et al., 2008). In comparison to music listening and learning the plot, watching the acting performance added social-emotional and visuospatial cues to the musical experience: facial expressions, gestures and postures, translated lines, and scenery. These factors probably contributed to the semantic processing of music and vocal expressions, and we argue that this experimental condition best approximated the full musical experience of listeners attending a live opera performance. Watching the acting performance decreased IBI and SCL in comparison to music listening. Previous studies reported that music-induced sadness ratings correlated positively with IBI and negatively with SCL (Krumhansl, 1997; Nyklícek et al., 1997). In addition, watching the acting performance was also related to significantly more music-induced chills. Another recent study showed that music-induced chills correlated with increased SCL and HR (Guhn, Hamm, & Zentner, 2007). The apparent divergence between these previous results and the present findings of increased wonder and transcendence associated with decreased IBI and SCL, and increased music-induced chills may be explained by differences in experimental design and measures. First, previous studies used short excerpts from classical orchestral music, whereas we focused on opera. Second, the previous studies investigated music listening alone, whereas our observations are based on a condition that involved music listening while watching the acting performance. Third, their conclusions are based on comparisons between music expressing negative and positive emotions, identified using basic emotions questionnaires. In the present experiment, watching the acting performance induced wonder and transcendence measured using GEMS. Overall, our results show for the first time that watching the acting performance contributes to music-induced wonder and transcendence that are associated with decreased IBI and SCL, and increased chills. In summary, both music listening (compared to baseline), and watching the acting performance (compared to music listening) decreased IBI and SCL. As shown in Fig. 3, IBI followed the same decreasing trend, whereas SCL remained at the same level after learning the plot compared to music listening. This means that learning the plot did not significantly influence these physiological variables, but they nonetheless remained at the level induced by music listening (i.e., they did not return to baseline). Therefore, music listening decreased RR, IBI, and SCL, learning the plot had no effect on these measures, and watching the acting performance significantly decreased IBI and SCL again. This indicates that IBI and SCL are the main physiological variables that are influenced by music listening and watching the acting performance. The only variables that were specifically influenced by learning the plot were RSA and LF-HRV, which indicates that they are sensitive to the addition of meaning in this context.

Chills

Tension

À0.09 À0.11 0.11 0.27 À0.21 0.18 À0.03 0.04 À0.15 0.05 Sadness À0.3 0.15 0.15 À0.04 À0.1 0.11 À0.1 0.11 À0.13 0.00 À0.17 À0.15 À0.02 0.12 0.02 À0.04 0.07 À0.21 0.25 0.1 0.03 À0.11 0.01 À0.08 0.01 0.02 0.45** À0.13 0.16 À0.15 À0.29 0.02 0.14 0.32 0.15 À0.12 0.11 À0.29 0.27 À0.22 À0.13 À0.06 À0.12 0.23 À0.01 0.03 0.36* À0.03 0.07 0.09 À0.34* À0.15 À0.14 À0.7 À0.21 À0.18 À0.6 À0.7 0.01 0.16 À0.36* À0.16 0.15 0.35* 0.01 À0.01 0.02 À0.09 0.05 0.07 0.12 À0.17 0.12 0.4* À0.11 0.08 0.31 À0.05 0.01 0.15 À0.23 À0.06 0.08 À0.01 À0.05 0.04 0.21 0.1 À0.02 0.15 À0.23 À0.13 0.05 0.35* À0.17 0.14 0.07 À0.003 À0.07 0.19 0.07 0.03 À0.07 0.00 0.02 À0.06 À0.08 0.09 À0.04 0.04 Joyful activation Peacefulness Nostalgia Tenderness Power Transcendence Wonder Valence Arousal Systolic blood pressure (SBP) Diastolic blood pressure (DBP) Skin conductance level (SCL) Respiratory rate (RR) Cardiac interbeat intervals (IBI) Heart rate (HR) Power in the very low frequency of heart rate variability (VLF-HRV) Power in the low frequency of heart rate variability (LF-HRV) Respiratory sinus arrhythmia (RSA or HF-HRV) Ratio between low and high frequency powers of heart rate variability (LF/HF) Chills 0.22 p 6 0.05. ** p 6 0.01.
*

Geneva Emotional Music Scale

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

Table 1 Correlations between physiological responses, chills, and affect during music listening.

Self-Assessment Manikin

0.08

0.44**

0.46**

0.11

À0.00

0.19

À0.12

0.12

0.1

0.26

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

9

4.2. Coherence between subjective and physiological changes
Chills À0.1 À0.05 0.18 0.45** À0.09 0.07 0.13 0.35* À0.35* 0.31

Nostalgia

0.05 0.23 0.18 0.1 À0.16 0.13 À0.2 0.43** À0.43** 0.07

0.27

There has been an active emotivist vs. cognitivist debate between scholars who argue that music listeners really experience emotions, or only identify emotions that music expresses (Kivy, 1990; Scherer & Zentner, 2001). This study integrated subjective and physiological measures of emotional responses, thus adding to the developing literature on the psychophysiology of music. In this line, a novel and important contribution of the present study is that we correlated music-induced emotions measured with a domain-specific instrument (i.e., GEMS), with an extensive array of emotion-related physiological changes. For instance, we found that music-induced wonder positively correlated with RR and chills across conditions. Moreover, by comparing the correlations of subjective and physiological changes between the three experimental conditions, one would observe that the psychophysiological coherence increases the most after learning the plot. This might suggest that the addition of meaning may be more closely related to the coherence between subjective and physiological changes induced by music, than the provision of additional sensory information (e.g., watching the acting performance). 4.3. Affective mood and sex The present findings that affective mood predicted emotions induced by music listening (e.g., power, joyful activation) suggests that future studies of music-induced emotions should control for this potential confound. Specifically, NA negatively predicted, and PA positively predicted power and joyful activation induced by music listening. This argues for the role of affective mood in the genesis of music-induced emotions, which is also in line with other studies (see Kreutz, Ott, Teichmann, Osawa, & Vaitl, 2008). In a recent field study (F.R. Baltes, M. Miclea, & A.C. Miu, unpublished observations), we have confirmed and extended the relationship between the affective mood that the participants reported before the beginning of a live opera performance, and the music-induced sadness and transcendence (GEMS). This indicates that the influence of affective mood is not limited to wonder and transcendence. However, the specificity of this association in relation to the musical stimuli, and the physical setting (i.e., laboratory vs. field studies) might be investigated by future studies. We also controlled for sex differences in the present analyses. A previous study showed that in comparison to men, women rated the chill-producing songs as being sadder (Panksepp, 1995). Another study reported that women showed elevated SCL to heavy metal compared to Renaissance music (Nater et al., 2006). In light of these results, the present study tested the effects of sex, and the interaction of sex and musical experience. We expected that after learning the plot, and especially during watching the acting performance, women would be more reactive due to increased emotional empathy with the female character in the musical excerpt. However, we found no significant main effect, or interaction of sex with musical experience, on subjective or physiological responses. 4.4. Potential limitations and implications One potential limit is that the mere repeated exposure may have influenced the present pattern of results. However, this possibility was controlled by measuring the same subjective and physiological responses while an independent control sample re-listened to the same musical excerpt for three times. The results from this sample indicated that the emotional arousal and valence, the music-induced emotions, or the physiological measures did not change with mere re-listening. This is also in line with the studies of Grewe et al. (2007a, 2007b). However, we acknowledge that a real limitation of the present study comes from the small size of

Tension

À0.01 À0.08 0.47** À0.01 À0.18 0.22 0.13 0.29 À0.29 À0.07 À0.04 À0.00 0.02 0.01 0.1 À0.09 À0.02 0.44** À0.42** 0.00 À0.05 0.11 0.08 0.32* À0.33* 0.29 0.16 0.47** À0.47** 0.11 0.01 0.19 À0.06 À0.04 À0.04 0.02 À0.15 À0.01 À0.00 0.03 0.08 0.39** À0.04 0.13 À0.11 0.05 À0.07 0.39* À0.41** 0.08 À0.24 À0.13 À0.06 0.26 À0.34* 0.35* 0.02 0.46** À0.46** 0.17 À0.09 0.16 0.09 0.42 À0.17 0.13 À0.07 0.56** À0.56** 0.14 À0.23 0.00 0.13 0.33* À0.27 0.25 À0.08 0.56** À0.56** 0.09 À0.1 0.15 À0.07 0.29 0.21 0.16 À0.23 0.02 À0.01 0.15 À0.12 0.02 À0.02 0.08 À0.09 0.05 0.02 0.36* À0.34* 0.16

Sadness

Joyful activation

Peacefulness

Tenderness

Power

Geneva Emotional Music Scale

Transcendence

Table 2 Correlations between physiological responses, chills, and affect during music listening after learning the plot.

Wonder

Valence

Self-Assessment Manikin

Arousal

Systolic blood pressure (SBP) Diastolic blood pressure (DBP) Skin conductance level (SCL) Respiratory rate (RR) Cardiac interbeat intervals (IBI) Heart rate (HR) Power in the very low frequency of heart rate variability (VLF-HRV) Power in the low frequency of heart rate variability (LF-HRV) Respiratory sinus arrhythmia (RSA or HF-HRV) Ratio between low and high frequency powers of heart rate variability (LF/HF) Chills

0.34* p 6 0.05. p 6 0.01.
* **

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

0.03

0.36*

0.39*

0.28

0.02

À0.1

0.4*

0.12

0.36*

10

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

the control sample in comparison to the sample from the main experiment. In light of the previous literature, musical expertise and (not) understanding the original language performance are also unlikely to have confounded our results. For instance, Bigand et al. (2005) showed that the classification of musical excerpts according to the emotional content did not differ between music graduates and nonmusicians. Another study found that emotional responses are not affected by song translation of non-native original language performance (Chiaschi, 2007), such as we did in our third experimental condition. It is also unlikely that listening to music with eyes open influenced music-induced emotions in the present study (Kallinen, 2004). However, future studies might control for personality variables (e.g., absorption) that are known to affect emotional arousal induced by music (Kreutz et al., 2008). These results have theoretical and methodological implications. First, they contribute to the literature supporting the emotivist position in the psychology of music. Second, they also add evidence in favor of the physiological differentiation of emotions. Third, considering that psychophysiological measures tended to correlate more highly with GEMS scores, and wonder and transcendence played a particularly prominent role, the present results emphasize the utility of domain-specific instruments to assess music-induced emotions. Fourth, many previous studies have paid a high price for experimental control, by using sound clips lasting a few seconds and crude measures of emotion (Peretz et al., 2001; Vieillard et al., 2008). Although these studies contributed to the understanding of the minimal conditions that are necessary to express an emotional meaning, it remains often unclear whether findings from such studies have any bearing on the experience of music in real life. Consequently, we chose to use a 19 min excerpt from Tosca, edited to contain a coherent plot, in order to realistically simulate the real life conditions in which opera induces emotions. The rich and complex pattern of psychophysiological results in the present study underscores the importance of external validity in laboratory studies of music-induced emotions. Each experimental condition in the present study manipulated an additional variable in relation to the previous conditions (i.e., the plot for the second condition, and the visual context for the third condition). The rationale behind this type of within-subject design is that any change that develops in one condition relative to the previous one is determined by the additional variable that was manipulated in that condition. However, it is possible that rather than being specifically induced by each new variable that was manipulated in a certain experimental condition, the changes could be due to simply increasing the sensory and semantic complexity of the musical experience. For instance, the visual context that was added in the third experimental condition might have clarified the meaning of the music, or allowed increased depth of processing in relation to the first two conditions. Other studies have used similar approaches by juxtaposing music and images, or lyrics and music, and claimed that emotional changes were specifically induced by the variable that differed between conditions (e.g., Ali & Peynircioglu, 2006). One may wonder whether this pattern of findings might generalize to all opera, or is unique to this style of operatic music performance (i.e., pertaining to verismo), composer, composition, excerpt, or interpretation. Scherer and Zentner (2001) have emphasized that music-induced emotions depend on several factors, such as structural features of music (i.e., pitch, melody, tempo, rhythm, harmony), performance features (e.g., physical appearance, expression, reputation, technical and interpretative skills of the performer), listener features (e.g., musical expertise, personality, affective mood), and contextual features (e.g., location of the performance, social framing of the event). The present study investigated the influence of affective mood, and controlled for

Chills

Tension

0.39** 0.21 À0.1 0.22 0.07 À0.02 0.13 À0.04 0.05 0.27 0.22 0.1 0.16 0.28 0.21 À0.17 0.00 0.00 À0.12 0.06 0.09 0.13 0.02 0.27 0.03 0.00 0.2 0.03 À0.02 0.15 0.1 À0.01 0.05 À0.01 À0.12 0.16 À0.25 0.11 À0.06 0.03 0.21 À0.04 0.1 0.08 0.01 0.08 À0.12 0.11 À0.11 0.07 0.2 0.12 À0.01 0.2 0.03 0.06 À0.15 0.01 À0.02 0.08 0.09 0.1 À0.09 0.34* 0.01 0.04 0.12 À0.01 0.03 0.19 0.39* 0.32 0.05 0.49** 0.25 À0.25 0.04 À0.1 0.1 0.22 0.15 0.25 0.04 0.39** 0.18 À0.17 0.13 À0.16 0.16 0.16 À0.18 À0.23 À0.16 0.2 À0.05 0.00 0.00 À0.06 0.12 0.03 0.17 0.13 0.00 0.32 0.17 À0.19 0.02 À0.08 0.02 0.05

0.02 0.017 0.09 0.53** À0.04 À0.00 0.14 0.36* À0.38* 0.29 Sadness Joyful activation Peacefulness Nostalgia Tenderness Power Transcendence Wonder Valence Arousal Systolic blood pressure (SBP) Diastolic blood pressure (DBP) Skin conductance level (SCL) Respiratory rate (RR) Cardiac interbeat intervals (IBI) Heart rate (HR) Power in the very low frequency of heart rate variability (VLF-HRV) Power in the low frequency of heart rate variability (LF-HRV) Respiratory sinus arrhythmia (RSA or HF-HRV) Ratio between low and high frequency powers of heart rate variability (LF/HF) Chills 0.2 p 6 0.05. p 6 0.01.
** *

Table 3 Correlations between physiological responses, chills, and affect during music listening while watching the acting performance.

Geneva Emotional Music Scale

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

Self-Assessment Manikin

0.12

0.36*

0.34*

0.27

0.12

0.22

À0.12

0.28

0.24

0.023

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß

11

important listener features (i.e., musical expertise, familiarity with the selected musical piece, preference for classic or operatic music). In addition, all the participants listened to the music in the same physical setting (i.e., our laboratory). This argues for the generality of our findings. It was beyond the purpose of this study to investigate the influence of musical structure, and performance features. It is likely that the stellar performance of Maria Callas and Tito Gobbi in this Tosca performance increased the effectiveness of this excerpt in inducing emotions. However, we speculate that the pattern of emotions reported here would not have been different had we used another interpretation of this opera by artists that are vocally and dramatically comparable (or at least close) to Maria Callas and Tito Gobbi. Future studies might investigate whether these findings can be replicated with excerpts from other operas. 4.5. Conclusion In conclusion, this study found that music listening, learning the plot, and watching the acting performance had specific effects on music-induced emotions and their physiological correlates. Opera poses enormous challenges to research due to the multitude of musical and dramatic means by which it induces emotions. Although the present study only scratched the surface, it opens new perspectives for future studies on the mechanisms of musicinduced emotions in opera. Acknowledgments We are grateful to Dr. Laurel J. Trainor and two anonymous reviewers for important suggestions that helped us in improving the present article, and Dr. Marcel Zentner for permission to use the Geneva Emotional Music Scale (GEMS-45) in the study. We also thank Silviu Matu for help with data collection. This research was supported by the 2010 Arnold Bentley Award from the Society for Education, Music, and Psychology (SEMPRE) to R.F.B. and A.C.M., and grant 411/2010 from the National University Research Council to A.C.M. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bandc.2011.01.012. References
Ali, S. O., & Peynircioglu, Z. F. (2006). Songs and emotions: Are lyrics and melodies equal partners? Psychology of Music, 34, 511–534. Bechara, A., Damasio, H., Damasio, A. R., & Lee, G. P. (1999). Different contributions of the human amygdala and ventromedial prefrontal cortex to decision-making. Journal of Neuroscience, 19(13), 5473–5481. Bernardi, L., Porta, C., Casucci, G., Balsamo, R., Bernardi, N. F., Fogari, R., et al. (2009). Dynamic interactions between musical, cardiovascular, and cerebral rhythms in humans. Circulation, 119(25), 3171–3180. Bezdek, M. A., & Gerrig, R. J. (2008). Musical emotions in the context of narrative film. Behavioral and Brain Sciences, 31, 578. Bharucha, J. J., Curtis, M., & Paroo, K. (2006). Varieties of musical experience. Cognition, 100(1), 131–172. Bigand, E., Vieillard, S., Madurell, F., Marozeau, J., & Dacquet, A. (2005). Multidimensional scaling of emotional responses to music: The effect of musical expertise and of the duration of the excerpts. Cognition and Emotion, 19, 1113–1139. Bleil, M. E., Gianaros, P. J., Jennings, J. R., Flory, J. D., & Manuck, S. B. (2008). Trait negative affect: Toward an integrated model of understanding psychological risk for impairment in cardiac autonomic function. Psychosomatic Medicine, 70(3), 328–337. Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The self-assessment manikin and the semantic differential. Journal of Behavior Therapy and Experimental Psychiatry, 25(1), 49–59. Bradley, M. M., & Lang, P. J. (2000). Affective reactions to acoustic stimuli. Psychophysiology, 37(2), 204–215.

Cacioppo, J. T., Berntsen, G. B., Klein, D. J., & Poehlmann, K. M. (1997). Psychophysiology of emotion across the life span. Annual Review of Gerontology & Geriatrics, 17, 27–74. Chiaschi, M. (2007). The effect of song translation vs. non-native original language performance in Japanese on emotional response of Japanese participants. M.A. thesis, The Florida State University. <http://etd.lib.fsu.edu/theses/available/etd12062006-230735/>. Codispoti, M., Bradley, M. M., & Lang, P. J. (2001). Affective reactions to briefly presented pictures. Psychophysiology, 38(3), 474–478. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. de Gelder, B., Bocker, K. B., Tuomainen, J., Hensen, M., & Vroomen, J. (1999). The combined perception of emotion from voice and face: Early interaction revealed by human electric brain responses. Neuroscience Letters, 260(2), 133–136. DeNora, T. (1999). Music as a technology of the self. Poetics, 27, 31–36. Diserens, C. M. (1920). Reaction to musical stimuli. Psychological Bulletin, 20, 173–199. Eckberg, D. L. (1997). Sympathovagal balance: A critical appraisal. Circulation, 96(9), 3224–3232. Ekman, P., Levenson, R. W., & Friesen, W. V. (1983). Autonomic nervous system activity distinguishes among emotions. Science, 221(4616), 1208–1210. Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G⁄Power 3: Flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2). Frazier, T. W., Strauss, M. E., & Steinhauer, S. R. (2004). Respiratory sinus arrhythmia as an index of emotional response in young adults. Psychophysiology, 41(1), 75–83. Grewe, O., Nagel, F., Kopiez, R., & Altenmuller, E. (2007a). Emotions over time: Synchronicity and development of subjective, physiological, and facial affective reactions to music. Emotion, 7(4), 774–788. Grewe, O., Nagel, F., Kopiez, R., & Altenmuller, E. (2007b). Listening to music as a recreative process: Physiological, psychological, and psychoacoustical correlates of chills and strong emotions. Music Perception, 24, 297–314. Guhn, M., Hamm, A., & Zentner, M. (2007). Physiological and musico-acoustic correlates of the chill response. Music Perception, 24(5), 473–483. Hietanen, J. K., Surakka, V., & Linnankoski, I. (1998). Facial electromyographic responses to vocal affect expressions. Psychophysiology, 35(5), 530–536. Huck, W. (1984). Tosca: Four callas toscas. The Opera Quarterly, 2(3), 175–178. Hyde, I. H., & Scalapino, W. (1918). The influence of music upon electrocardiograms and blood pressure. American Journal of Physiology, 46, 35–38. Juslin, P. N., Liljestrom, S., Vastfjall, D., Barradas, G., & Silva, A. (2008). An experience sampling study of emotional reactions to music: Listener, music, and situation. Emotion, 8(5), 668–683. Juslin, P. N., & Vastfjall, D. (2008). Emotional responses to music: The need to consider underlying mechanisms. Behavioral and Brain Sciences, 31(5), 559–575 (discussion 575-621). Kallinen, K. (2004). Emotion related psychophysiological responses to listening to music with eyes-open versus eyes-closed: Electrodermal (EDA), electrocardial (ECG), and electromyographic (EMG) measures. Paper presented at the International Conference on Music Perception & Cognition. Khalfa, S., Peretz, I., Blondin, J.-P., & Manon, R. (2002). Event-related skin conductance responses to musical emotions in humans. Neuroscience Letters, 328(2), 145–149. Kingwell, B. A., Thompson, J. M., Kaye, D. M., McPherson, G. A., Jennings, G. L., & Esler, M. D. (1994). Heart rate spectral analysis, cardiac norepinephrine spillover, and muscle sympathetic nerve activity during human sympathetic nervous activation and failure. Circulation, 90(1), 234–240. Kivy, P. (1990). Music alone: Philosophical reflections on the purely musical experience. Cornell University Press. Koelsch, S. (2005). Investigating emotion with music: Neuroscientific approaches. Annals of the New York Academy of Sciences, 1060, 412–418. Kreibig, S. D., Wilhelm, F. H., Roth, W. T., & Gross, J. J. (2007). Cardiovascular, electrodermal, and respiratory response patterns to fear- and sadness-inducing films. Psychophysiology, 44(5), 787–806. Kreutz, G., Ott, U., Teichmann, D., Osawa, O., & Vaitl, D. (2008). Using music to induce emotions: Influences of musical preference and absorption. Psychology of Music, 36(1), 101–126. Krumhansl, C. L. (1997). An exploratory study of musical emotions and psychophysiology. Canadian Journal of Experimental Psychology, 51(4), 336–353. Levenson, R. W. (1992). Autonomic nervous system differences among emotions. Psychological Science, 3, 23–27. Lundqvist, L. O., Carlsson, F., & Juslin, P. N. (2009). Emotional responses to music: Experience, expression, and physiology. Psychology of Music, 37, 61–90. Lundqvist, L. O., & Dimberg, U. (1995). Facial expressions are contagious. Journal of Psychophysiology, 9(3), 203–211. Miu, A. C., Heilman, R. M., & Houser, D. (2008). Anxiety impairs decision-making: Psychophysiological evidence from an Iowa Gambling Task. Biological Psychology, 77(3), 353–358. Miu, A. C., Heilman, R. M., & Miclea, M. (2009). Reduced heart rate variability and vagal tone in anxiety: Trait versus state, and the effects of autogenic training. Autonomic Neuroscience, 145(1–2), 99–103. Nakahara, H., Furuya, S., Francis, P. R., & Kinoshita, H. (2010). Psycho-physiological responses to expressive piano performance. International Journal of Psychophysiology, 75(3), 268–276.

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

12

F.R. Baltes et al. / Brain and Cognition xxx (2011) xxx–xxx ß Stratton, V. N., & Zalanowski, A. H. (1994). Affective impact of music vs. lyrics. Psychology of Music, 12(2), 70–83. Tan, S. L., Spackman, M. P., & Bezdek, M. A. (2007). Viewers’ interpretations of film characters’ emotions: Effects of presenting film music before or after a character is shown. Music Perception, 25(2), 135–152. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology (1996). Heart rate variability: Standards of measurement, physiological interpretation and clinical use. Circulation, 93, 1043–1065. Vaitl, D., Vehrs, W., & Sternagel, S. (1993). Prompts-leitmotif-emotion: Play it again, Richard Wagner! In N. Birbaumer & A. Ohman (Eds.), The structure of emotion: Psychophysiological, cognitive, and clinical aspects (pp. 169–189). Seattle, WA: Hogrefe & Huber. Vieillard, S., Peretz, I., Gosselin, N., Khalfa, S., Gagnon, L., & Bouchard, B. (2008). Happy, sad, scary and peaceful musical excerpts for research on emotions. Cognition and Emotion, 22(4), 720–752. Vitouch, O. (2001). When your ear sets the stage: Musical context effects in film perception. Psychology of Music, 29(1), 70–83. Watson, D., & Clark, A. C. (1994). The PANAS-X. Manual for the positive and negative affect schedule – expanded form. The University of Iowa. Watt, R. J., & Ash, R. L. (1998). A psychological investigation of meaning in music. Musicae Scientiae, 2, 33–54. Witvliet, C. V. O., & Vrana, S. R. (2007). Play it again Sam: Repeated exposure to emotionally evocative music polarises liking and smiling responses, and influences other affective reports, facial EMG, and heart rate. Cognition and Emotion, 21, 3–25. Zeffirelli, F.(Director). (2002). Maria callas – At Covent Garden 1962 and 1964. US and Canada: EMI Classics. Zentner, M., Grandjean, D., & Scherer, K. R. (2008). Emotions evoked by the sound of music: Characterization, classification, and measurement. Emotion, 8(4), 494–521.

Nater, U. M., Abbruzzese, A., Krebs, M., & Ehlert, U. (2006). Sex differences in emotional and psychophysiological responses to musical stimuli. International Journal of Psychophysiology, 62(2), 300–308. Nyklícek, I., Thayer, J. F., & Van Doornen, L. J. P. (1997). Cardiorespiratory differentiation of musically-induced emotions. Journal of Psychophysiology, 11, 304–321. Panksepp, J. (1995). The emotional sources of ‘‘chills’’ induced by music. Music Perception(13), 171–207. Peretz, I., Blood, A. J., Penhune, V., & Zatorre, R. (2001). Cortical deafness to dissonance. Brain, 124(Pt 5), 928–940. Rickard, N. S. (2004). Intense emotional responses to music: A test of the physiological arousal hypothesis. Psychology of Music, 32, 371–388. Scherer, K. R. (1995). Expression of emotion in voice and music. Journal of Voice, 9(3), 235–248. Scherer, K. R., & Zentner, M. R. (2001). Emotional effects of music: Production rules. In P. N. Juslin & J. A. Sloboda (Eds.), Music and emotion: Theory and research (pp. 361–392). Oxford; New York: Oxford University Press. Sloboda, J. (1991). Music structure and emotional response: Some empirical findings. Psychology of Music, 19, 110–120. Sloboda, J. (1992). Music as a language. In F. Wilson & F. Roehmann (Eds.), Music and child development (pp. 28–43). MMB Music. Sloboda, J., & O’Neil, S. A. (2001). Emotions in everyday listening to music. In P. N. Juslin & J. A. Sloboda (Eds.), Music and emotion: Theory and research. Oxford; New York: Oxford University Press. Sloboda, J., O’Neil, S. A., & Ivaldi, A. (2001). Functions of emotions in everyday life: An exploratory study using the experience sampling method. Musicae Scientiae, 5, 9–32. Stevens, J. (2002). Applied multivariate statistics for the social sciences (Applied Multivariate STATS) (4th ed.). New Jersey: Lawrence Erlbaum Assoc.

Please cite this article in press as: Baltes, F. R., et al. Emotions induced by operatic music: Psychophysiological effects of music, plot, and acting. Brain and ß Cognition (2011), doi:10.1016/j.bandc.2011.01.012

Sign up to vote on this title
UsefulNot useful