Emotion response and regulation to “happy” and “sad” music stimuli: Partial synchronization of subjective and physiological responses

Abstract

Music has been demonstrated to induce strong emotional responses in listeners, but listeners’ capacity to regulate this response has not yet been experimentally examined. In this study, 32 participants (M_age = 20.22 years, SD = 2.45; 72% female) listened to short music excerpts (four “happy” and four “sad” pieces) with instructions to “feel” or to “regulate” (specifically, “reduce” sadness or “strengthen” happiness) the emotional response to the music. Self-reported “sadness” and “happiness” ratings and physiological responses (skin conductance and heart rate) were recorded at baseline and continuously across all conditions. Under the “feel” instruction, both “happy” and “sad” music resulted in an increase in self-reported emotion ratings, and a decline in skin conductance and heart rate. These responses were effectively regulated for both “happy” and “sad” music across all measures except heart rate response. These findings partially support the prediction that music would induce coherent changes across self-reported and physiological measures during both emotion induction and emotion regulation.

Keywords

affect regulation conductance emotional response heart rate multicomponent skin subjective synchronized

Music is one of the most commonly reported means of inducing and regulating emotions (Schäfer, Sedlmeier, Städler, & Huron, 2013). Within experimental methods, it has stimulated reliable, intense and authentic changes in mood states (Eich, Ng, Macaulay, Percy, & Grebneva, 2007). Music plays a prominent role in everyday life, whether intentionally via purposeful music listening (e.g., via personalized listening devices), or unintentionally via passive exposure (e.g., in shopping centres) (Greasley & Lamont, 2011; Saarikallio, 2011). The debate as to whether music induces “real emotions” has, however, been extensive and is ongoing (Davies, 2010; Juslin, Liljeström, Västfall, & Lundqvist, 2010; Scherer & Zentner, 2001). Scherer and others have argued that “real” emotions involve multiple cognitive, physiological and behavioural components (Scherer & Coutinho, 2013). A synchronized response, consisting of coherent changes in self-reports of emotional “feeling” and objective measures of physiological response, motor expression (e.g., facial feedback) or action tendencies, provides robust evidence that a meaningful emotional response has occurred (Hamm, Schupp, & Weike, 2003; Levenson, 2003; Mauss, Levenson, Carter, Wilhelm, & Gross, 2005; Scherer, 2009). Regulation is also considered a component of emotion (e.g., Fontaine, Scherer, Roesch, & Ellsworth, 2007; Scherer, 2005). Demonstration of multimodal changes in response to induction of emotion as well as demonstration that emotion can be regulated by music, are therefore important concerns in this field of music psychology. Few studies have demonstrated a synchronized change in both objective and self-report measures of both induction and regulation of the emotional response to music, as predicted by current models of emotion (Scherer, 2009).

Music’s capacity to induce emotional responses has been demonstrated across a range of self-report and physiological measures (for reviews, see Gabrielsson, 2001–2002; Hodges, 2010). Happy and arousing music is generally associated with greater increases in autonomic nervous system activity (indexed by increased heart rate and skin conductance) than is sad or relaxing music (Hodges, 2010; Khalfa, Roy, Rainville, Dalla Bella, & Peretz, 1998; Krumhansl, 1997; Rickard, 2004; Witvliet & Vrana, 2007). Skin conductance changes, however, may be less differentiated by whether the self-reported emotional experience is positive or negative (i.e., valence) than by level of arousal (Lang, Greenwald, Bradley, & Hamm, 1993), and also tend to show a characteristic monotonic decline over time in experimental settings (e.g., Lundqvist, Carlsson, Hilmersson, & Juslin, 2009). Some musical stimuli have been found sufficiently potent to induce chills and activate rewards systems in the brain (Blood & Zatorre, 2001; Rickard, 2004; Salimpoor et al., 2013). Considerable variability is nonetheless observed across studies, with a substantial number of null effects also observed, particularly in studies measuring respiration rate (Hodges, 2010).

Whether music that is capable of inducing sad or happy emotional responses within listeners is a useful stimulus for studies of the regulation of emotion is also relevant to this field. To date, the types of emotion-inducing stimuli that have been employed for such purposes have been visual images and film. For example, functional magnetic resonance imaging (fMRI) has revealed that healthy individuals showed down-regulation of amygdala and insula activation (regions associated with an emotional response) when instructed to use cognitive reappraisal to regulate their response to negative images, such as a picture of a ferocious dog (Johnstone, van Reekum, Urry, Kalin, & Davidson, 2007). When viewing positive images, such as a picture of a stunning nature scene, fMRI showed that healthy individuals were able to sustain activity in related regions associated with reward and motivation (nucleus accumbens and fronto-striatal activity) when instructed to “attend” or “enhance” their emotional response to positive images (Heller et al., 2009). To date, relatively few studies of emotion regulation appear to have supplemented self-reported emotion measurement with an autonomic response measure. Troy, Wilhelm, Shallcross, and Mauss (2010) found that both self-reported sadness and skin conductance levels decreased less when participants were asked to reduce the emotional impact of sad film stimuli, than when they watched the film without instruction. Other studies have included autonomic response measures to examine the regulation of an emotional response to film stimuli that evoked disgust (Demaree, Schmeichel, Robinson, & Everhart, 2004; Gross, 1998), but not sadness. With regard to “up-regulation” of emotional responses to positive stimuli, Demaree and colleagues (2004) detected decreases in interbeat interval, a heart rate measure, when participants were instructed to exaggerate their emotional response to amusing film stimuli, but no influence on skin conductance. Given that regulation of emotion underlies effective psychological functioning and well-being, it is also often one of the key skills that is dysfunctional in psychological disorders, such as depression (e.g., Gross & John, 2003). Using multimodal assessment of the emotional response to select the musical stimuli that are most effective in evoking happiness and sadness in an individual may therefore help refine music therapeutic interventions.

One consistent caveat in this body of research, however, is that the emotional response to music, and the capacity to regulate that response, varies across contexts and individuals. For instance, the processing of emotions in response to music may be moderated by individual differences such as gender (Nater, Abbruzzese, Krebs, & Ehlert, 2006) and music background (Iwanaga & Moroki, 1999; Lehmann, 1997), while the capacity to regulate one’s responses to music may be moderated by emotion regulation style (Heller et al., 2009; Johnstone et al., 2007). Potential confounds should also therefore be taken into consideration in studies that aim to assess whether various aspects of emotion might also synchronize in response to music stimuli.

The aim of the current study was to examine the utility of music in the investigation of emotion induction and regulation, using normative music stimuli known to elicit “sad” and “happy” emotional responses, controlling for the potential confounds of gender, music background, and emotion regulation style. The study design consisted of two within-subjects factors: Instruction – “Feel” or “Regulate” (i.e., “Reduce” a response to “sad” music or “Strengthen” a response to “happy” music) and Time – Baseline or Condition, and three dependent variables – self-reported emotion ratings (self-reported sadness ratings for the “sad” music condition, and self-reported happiness ratings for the “happy” music condition), skin conductance means and heart rate means. It was hypothesized that music stimuli would elicit a synchronized emotional response in both self-reported emotion ratings and objective indices (skin conductance and heart rate) of emotion when participants were instructed to “feel” the emotions in the music. Specifically, it was anticipated that the emotional response to “happy” music would be associated with an increase in heart rate and skin conductance, and the response to “sad” music would be associated with a decrease in heart rate and skin conductance. Second, with regard to emotion regulation, it was hypothesized that emotional responses would be attenuated when participants were instructed to “reduce” their response to “sad” music, and enhanced when participants were instructed to “strengthen” their response to “happy” music. Finally, it was hypothesized that the majority (more than 50%) of participants would self-report being able to both feel and regulate their responses to music.

Method

Participants

Participants (N = 32) were recruited from an undergraduate psychology student cohort from Monash University, Melbourne, and were given course credit for participation. Age ranged from 18 to 28 years (M = 20.22, SD = 2.45) and 72% of the sample was female. Exclusion criteria applied to people: under 18 years of age; with uncorrected vision or hearing problems; taking medication that might impact their mood or concentration; and/or with a history of severe head injuries or seizure-related disorders. All participants had completed secondary school; one had completed a further diploma level certificate.

Materials

Emotional stimuli

Music excerpts (four “happy” and four “sad”) were selected from the normative music database of emotionally expressive music samples developed by Eerola and Vuoskoski (2011). The excerpts were approximately 15 seconds each in duration, and taken from film soundtracks that were believed not to be widely recognizable. They were selected on the basis of ratings for each of the target emotions (“happy” and “sad”; see Table 1). Normative ratings on a 9-point Likert-type scale for “energy” indicate that the “happy” pieces were all rated as moderately arousing (5.83–6.83), while the “sad” pieces were low in arousal (2.17–2.20). A manipulation check confirmed that the selected stimuli were generally effective in inducing the intended emotional response in this study (see Table 1), although it is notable that in the current study the “sad” pieces elicited lower sadness and higher happiness ratings than participants in the study by Eerola and Vuoskoski. All pieces were also generally liked but only of moderate familiarity (see Table 1). Eight sequences consisting of different presentation orders of the four “happy” and four “sad” pieces, paired with either “feel” or “reduce/strengthen” instructions (see Procedure section for further detail about instructions), were created (each sequence lasted a total of 136 s). This counterbalanced both the order of the music excerpts and the order of the emotion regulation instructions to control any potential carryover-emotion effects. A set of Sennheiser, HD570, headphones was used to deliver the music to the participant.

Table 1.

Mean ratings of induced (felt) emotion for the four “happy” and four “sad” music excerpts from standardized database (Eerola & Vuoskoski, 2011) and in current study.

Target emotion and piece number	Normative Database Ratings			Current study’s ratings
Target emotion and piece number	Code & Soundtrack^a	Happy	Sad	Total n	Happy	Sad	Liking	Familiarity
Sad Music 1	032 The English Patient, Track 18	1.17	7.50	32	5.35	6.21	7.09	3.99
Sad Music 2	034 Band of Brothers, Track 12	1.00	7.33	32	5.11	6.30	6.75	4.61
Sad Music 3	036 Running Scared, Track 15	1.17	7.00	32	4.79	5.90	6.64	3.88
Sad Music 4	044 Gladiator, Track 7	1.00	1.00	32	5.67	6.25	6.41	4.39
Happy Music 1	001 Shallow Grave, Track 6	7.33	1.00	31	8.32	2.72	7.53	4.31
Happy Music 2	003 The Rainmaker, Track 3	7.17	1.00	32	8.46	2.65	7.58	3.33
Happy Music 3	007 Man of Galilee CD1, Track 11	7.00	1.00	32	8.73	2.54	8.02	4.75
Happy Music 4	013 Shallow Grave, Track 3	6.20	1.00	32	8.89	2.65	8.73	5.56

Note. 7-point rating scale standardized to 9-point scale to allow comparison with database ratings; liking and familiarity ratings included also for current study. ^aCode and soundtrack refer to that used in Eerola and Vuoskoski (2011) to identify these music excerpts.

Physiological measures

Physiological data were collected using a 10 channel bioamplifer (Nexus-10; Mind Media B.V., Netherlands, 2004–2009) providing 24 Bit A/D conversion. Data were sent to the BioTrace+ computer software via Bluetooth connection. Heart rate was sampled at a rate of 256 Hz using three disposable adhesive pre-gelled ECG electrodes (Kendall Arbo*, tyco/Healthcare) connected to carbon-coated cables with active shielding. Sensors were placed on the chest and abdomen according to standard ECG Lead II configuration. Skin conductance data were sampled at a rate of 32 Hz, with resolution of up to 0.0001 microsiemens. Skin cleansing swabs were used to clean the skin site prior to electrode placement. Sensors were connected to the medial phalanx of the index and middle fingers of the non-dominant hand with adjustable Velcro straps. A stabilized voltage is derived from the Nexus-10 system power source, and creates a constant current by applying a known resistance (13MΩ) to the two flat Ag-AgCl electrodes (10 mm diameter). A voltage measurement is made across the electrodes to determine the resistance, and the reciprocal value calculated using the formula (1)

S C (t) = \frac{1}{(V s k i n (t) / ((V b a t - V s k i n (t)) / 13 M Ω))}

(1)

where SC(t) = Output value (Siemens), Vskin(t) = voltage derived from SC-electrodes Vbat = stabilized voltage from power source of encoder, and 13 MΩ = internal resistance of the SC-sensor (required to meet safety requirements).

Self-reported emotion measures

Self-reported emotional response data were collected within the experiment using self-reported emotion rating forms specifically developed for the study. The forms comprised 7-point Likert-type “happiness” and “sadness” rating scales, ranging from 1 = “Not much or not at all” to 7 = “Extremely”. Scales were provided for “Before music” and “After music” time conditions. Two additional 7-point Likert-type rating scales were provided for the participant to rate the music with regard to: “I liked it” and “I was familiar with it”. Reliability for self-reported baseline ratings for both sad and happy music were high, with Cronbach’s α = .95 (sad music), and α = .95 (happy music).

Music Use questionnaire (MUSE; Chin & Rickard, 2012)

The 24-item MUSE questionnaire assesses the extent to which the respondent listens to music, plays (or has played) a musical instrument, and has been formally trained in music. The respondents rate, on a 6-point scale (1 = “Strongly disagree”; 6 = “Strongly agree”), the extent to which they identify with statements about how they relate to and use music in their everyday lives. The indices of interest in the present study were the music training index, and two music engagement subscales (Cognitive and Emotional Regulation; Engaged Production). An example item from the Cognitive and Emotional Regulation subscale is: “Music often takes away tension at the end of the day.” An example item from the Engaged Production subscale is: “I often play challenging pieces.” Internal reliabilities for the Cognitive and Emotional Regulation subscale have been reported as Cronbach’s α = .95 (and in the current sample, α = .74); and for the Engaged Production subscale, .94 (here, α = .90).

Emotion Regulation Questionnaire (ERQ; Gross & John, 2003)

The 10-item ERQ provides an indication of the extent to which the respondent habitually uses the emotion regulation strategies of cognitive reappraisal and emotion suppression. Respondents rate, on a 7-point scale (1 = “Strongly disagree”; 7 = “Strongly agree”), the extent to which they disagree or agree with statements about how they typically regulate or manage their emotional experience and emotional expression. For example, “When I want to feel less negative emotion (such as sadness or anger), I change what I’m thinking about.” The validity of ERQ constructs has been supported and acceptable Cronbach alpha reliabilities reported in the literature (α = .79 for reappraisal; α = .73 for suppression; Gross & John, 2003) and in the current study (α = .82 for reappraisal; α = .78 for suppression).

Procedure

Prior to attending the laboratory session participants completed the online questionnaire, comprising a demographic questionnaire, as well as the MUSE and ERQ. The laboratory session was held in a furnished and temperature-controlled (21°C) laboratory. The participant was seated in a comfortable chair in front of a computer screen, which displayed instructions in large black font on a white background within the semi-automated experiment. Physiological sensors were attached to the participant, and a 2-minute resting period obtained for normalization of physiological measures.

The experimenter then explained the procedure and the instructions. For the “Feel” instruction, the participant was advised to “feel whatever it is you feel from the music naturally.” The emotion regulation instruction was adapted as relevant to “happy” and “sad” music stimuli presented. This means that for the “sad” music, the regulation instruction was always “Reduce”; while for the “happy” music, the regulation instruction was always “Strengthen”. For the “Reduce” instruction, participants were advised that, “some people might distract themselves from their feelings or from the music; or they might think differently about how they are feeling; or they might try to control or suppress their feelings.” For the “Strengthen” instruction, participants were advised that, “some strategies you could use to strengthen your emotional response to the music are to try to savour the feelings you initially have to make them last longer, or to focus on the parts that you find positive to increase your overall feeling.”

After setting the headphones in place, the participant was given a 10-min practice trial with two music excerpts, one happy and one sad (which were different from those included in the main experiment). After addressing any participant queries, the experiment proceeded with one of eight music sequences selected in a quasi-randomized order (without replacement). Each sequence commenced with a 15-s pre-music self-reported emotion rating period, during which the participant made a self-reported rating of their emotion (few seconds) using their dominant hand (avoiding movement artefacts) and then remained still. The emotion “feel” or “regulate” instruction was then given, and the music presented. Following each music excerpt, participants then made another self-reported rating of their felt emotion. An 8-s rest period was given before the next pre-music self-reported emotion rating was completed, and so on until all eight music excerpts had been presented.

The BioTrace+ software was used to collect continuous physiological measures and to automate the experiment.

Ethics statement

All procedures were approved by Monash University Human Research Ethics Committee (Project number: CF11/0627 – 2011000240). Informed consent was also obtained from all participants.

Results

Data treatment and screening

Missing data occurred for one participant who made self-reported emotion ratings on only one scale (either the happy or sad scale): these missing values were replaced with the mean value for that scale for the sample (Tabachnick & Fidell, 2007). In the analysis of the skin conductance variable, two participants’ data sets were excluded list-wise due to inadequate detection attributable to poor contact with skin conductance sensors. No other physiological data or key questionnaire data were missing from the data set.

For self-reported and physiological measures of emotion (i.e., dependent variables), the Baseline and Music Condition means were determined for the four “happy” and four “sad” music excerpts. Physiological data were first normalized by subtracting resting period raw scores, and dividing by the resting period standard deviation. Normalized skin conductance levels (µS) and heart rate (bpm) were then averaged across the 15-s duration of each baseline or condition recording period. Normality was confirmed via histograms and normal Q-Q plots. The self-reported emotion ratings were slightly skewed, although not sufficiently to warrant transformation, particularly as Analysis of Variance (ANOVA) is robust to small violations (Howell, 2002).

Examination of potential confounding variables

The relationship between potential confounds identified in the literature and each of the dependent variables was examined using Pearson’s correlations (2-tailed; p < .05). Neither self-reported emotion ratings nor skin conductance change were found to be significantly correlated with music training, music engagement subscales (assessed by the MUSE), or ERQ (Reappraisal or Suppression) style. Heart rate was, however, found to be positively correlated with ERQ Suppression scores during regulation to both the sad (r = .405, p = .021, n = 32) and happy (r = .483, p = .005, n = 32) music. ANCOVAs were therefore performed for the heart rate univariate tests, with ERQ Suppression scores entered as a covariate.

Proportion of sample reporting an emotional response, and capacity to regulate the response, to music

To assess whether the manipulations were effective, the frequency of participants reporting a change in self-reported emotion ratings before and after the “feel” instruction, for each sample of music, was calculated. To assess the proportion of the sample which felt capable of regulating this response, the frequency of participants with a different change score under the regulation instruction (i.e., “reduce” or “strengthen”) than in the “feel” instruction was also calculated for each music sample. While participants who did not “feel” a response to “happy” music could still feasibly become more positive under the “strengthen” instruction, those who did not “feel” a response to “sad” music were excluded from the tally for the “regulate” instruction as it was ambiguous as to whether participants could reduce a feeling they reported not feeling. Furthermore, one participant produced complete “feel” data, but incomplete “regulate” data for the sad music condition. Sample sizes therefore varied accordingly across the various conditions.

The proportion of the sample self-reporting successful response and regulation to happy and sad music pieces is summarized in Table 1. The proportion of participants who responded to the “sad” music stimuli (72–78%) was generally higher than the proportion of participants who responded to the “happy” music stimuli (50–81%; see Table 2). The proportion of the sample that self-reported sadness responses to at least two of the four “sad” music stimuli was the same (i.e., 88%) as the proportion of the sample that self-reported happiness responses to at least two of the “happy” music. These participants were categorized as “responders”. Of the four non-responders to sad music and the four non-responders to happy music, only two non-responders were non-responders to both sad music and happy music.

Table 2.

Number and proportion (%) of successful responses and regulations across the four “sad” and four “happy” music excerpts, based on self-reported ratings of emotion felt and regulated (total N = 32).

Music stimulus	Successful “feel” responses	Sample size	Successful “regulations”	Sample size
At least 2 of 4 sad pieces	28 (88%)	32	26 (96%)	27
Sad Music #1	25 (78%)	32	18 (75%)	24
Sad Music #2	24 (75%)	32	19 (83%)	23
Sad Music #3	23 (72%)	32	20 (91%)	22
Sad Music #4	23 (72%)	32	21 (95%)	22
At least 2 of 4 happy pieces	28 (88%)	32	10 (31%)	32
Happy Music #1	16 (50%)	32	22 (69%)	32
Happy Music #2	21 (66%)	32	10 (31%)	32
Happy Music #3	26 (81%)	32	11 (34%)	32
Happy Music #4	17 (53%)	32	18 (56%)	32

Across the four “sad” music stimuli, successful regulation occurred in 75–95% of participants. Across the four “happy” music stimuli, a large difference was seen in the proportion of participants who exhibited successful regulation for Happy Music #1 (69%), when compared to Happy Music #2 (31%) and #3 (34%). At least 96% of responders within the sample were capable of regulating their responses to at least two of the “sad” pieces, while only 31% of the sample was capable of regulating their responses to at least two of the “happy pieces”.

Emotional response to music

Data only from responders (i.e., those that successfully produced an emotional response to at least two of the four music samples within the music category) were included in all subsequent analyses. A within-subjects MANOVA (Pillai’s trace) was performed for the “sad” and “happy” music to assess whether responses were synchronized across self-reported emotion ratings and physiological measures. In the first MANOVA, self-report ratings, skin conductance and heart rate changes in response to “sad” music were the multiple dependent variables, while time and instruction were the independent variables. A significant time main effect was observed for “sad” music (V = 0.81, F(3,23) = 23.00, p < .001, $η_{p}^{2}$ = .80). In the second MANOVA, self-report ratings, skin conductance and heart rate change in response to “happy” music were the multiple dependent variables, while time and instruction were the independent variables. A significant time main effect was observed for “happy” music (V = 0.90, F(3,23) = 71.01, p < .001, $η_{p}^{2}$ = .90). The means for each dependent variable in the Baseline and Condition time periods (for both the “sad” and “happy” music conditions) are presented for comparison in Figure 1.

Figure 1.

Self-reported emotion ratings and physiological responses, and attempted regulation of responses, to music. These panels show partially synchronized changes in (A) self-reported emotion ratings, (B) skin conductance and (C) heart rate in response to sad and happy music, under “Feel” (left side of each graph) and “Regulate” (i.e., “strengthen” or “reduce”, as relevant; right side of each graph) instructions. Small bars represent the differences between before (black bars) and during (grey bars) music conditions (physiological measures were obtained continuously during the condition, whilst self-reported emotion ratings were obtained immediately “after” the condition finished). Large bars represent the time by instruction interactions. Error bars represent the standard error of the means.

Paired t-tests were used to assess whether responses to music within each instruction condition, for each measure, were significant over time. In each t-test, the independent (paired) variable was time, and the dependent variable was the self-report or physiological measure. Under the “feel” instruction, “happy” music resulted in a significant increase in self-reported happiness ratings (t(28) = -8.32, p < .001) and significant declines in both skin conductance (t(25) = 3.06, p = .005) and heart rate (t(27) = 3.11, p = .004). “Sad” music resulted in a significant increase in self-reported sadness ratings (t(27) = -10.40, p < .001), and significant declines in both skin conductance (t(25) = 4.29, p <.001) and heart rate (t(27) = 5.48, p < .001).

Regulation of the emotional response to music

When instructed to “reduce” the emotional response to “sad” music, all responses were attenuated. Paired t-tests (with time as the independent (paired) variable, and self-report or physiological measures as the dependent variable) confirmed that skin conductance was no longer significantly decreased by “sad” music (t(25) = 1.44, p = .161). While self-reported sadness ratings were still significantly increased (t(27) = -2.35, p = .027), and heart rate (t(27) = 2.55, p = .017) significantly decreased, instructions to “reduce” this response did reduce the effect size. Univariate repeated measures ANOVAs were conducted to determine whether the emotion regulation instruction modulated each measure’s response over time, in response to “sad” music. Time and instruction were independent variables in all ANOVAs, and the self-report or physiological measure was the dependent variable. As ERQ-Suppression correlated with heart rate, it was also entered as a covariate in the heart rate ANOVAs. The Time (baseline, during music) × Instruction (“feel”, “regulation”) interaction effect was significant for self-reported sadness ratings (F(1,27) = 72.24, p < .001, $η_{p}^{2}$ = .73) but not for skin conductance changes (F(1,25) = 2.05, p = .165, $η_{p}^{2}$ = .08), nor heart rate changes (with ERQ Suppression entered as a covariate; F(1,26) = 1.00, p = .328, $η_{p}^{2}$ =. 04).

When instructed to “strengthen” the emotional response to happy music, self-reported happiness ratings were again significantly enhanced (t(27) = -8.38, p < .001), and skin conductance (t(25) = 2.20, p = .037) and heart rate (t(27) = 27.42, p < .001) were significantly reduced by the music. Furthermore, this instruction to “strengthen” the response increased the effect size for both self-reported ratings and heart rate when compared to the “feel” response. A univariate repeated measures ANOVA confirmed a significant Time (baseline, during music) x Instruction (“feel”, “regulation”) interaction effect for the increased self-reported happiness ratings (F(1,27) = 4.90, p = .036, $η_{p}^{2}$ = .15), and for the attenuated skin conductance response (F(1,25) = 5.83, p = .023, $η_{p}^{2}$ =.19). Despite a significant reduction in heart rate from baseline to during the “strengthen” condition, the Univariate ANOVA (with ERQ Suppression entered as a covariate) revealed no significant Time (baseline, during music) × Instruction (“feel”, “regulation”) interaction effect (F(1,26) = 0.29, p =.596, $η_{p}^{2}$ = .01).

Discussion

This study used a multimodal assessment method to examine the utility of music as a stimulus to explore emotion responsiveness and regulation. As hypothesized, music was found to be an effective stimulus for eliciting a synchronized emotional response, with significant changes observed across both self-reported emotion ratings and physiological measures. The majority of participants reported feeling an emotional response to both the “sad” and “happy” music samples, and the majority were also reported being able to “regulate” their response to at least two of the “sad” pieces and one of the “happy” pieces. In support of the second hypothesis, self-reported emotional responses were associated with significant changes in autonomic nervous system activity (heart rate and skin conductance). Finally, our data reveal the first evidence to our knowledge to show partially synchronized changes in a self-reported and physiological measure of emotion response as a result of regulation of emotions experienced in response to music.

Participants’ self-reported ratings of their emotional response were concordant with that predicted from the normative perceived emotion data of the musical excerpt source (Eerola & Vuoskoski, 2011) with regard to the direction of the emotional change (i.e., increased sadness to “sad” music and increased happiness to “happy” music). In the current study, the “sad” music pieces induced these congruent self-reported emotional responses in a greater proportion of the sample than did the “happy” music. However, the “sad” music also induced a more mixed response than reported by Eerola and Vuoskoski, with moderate happiness induced also. Whether this difference in responsiveness to “sad” and “happy” music reflects a real difference in human capacity to be more responsive to emotionally sad stimuli than to happy stimuli warrants further investigation in a larger sample using a greater variety of stimuli of positive emotions. Furthermore, such individual variation in happiness responsiveness may be explained by individual factors, such as differences in trait resilience (for a review, see Tugade & Fredrickson, 2007). Participants can also have some difficulty differentiating the instruction to “feel” an emotion expressed in the music, from the tendency to describe the emotion they “perceive” is intended by the music (Gabrielsson, 2001–2002). It remains possible that differences in adherence to this instruction accounted for some of the variability in reported effectiveness of the music to induce self-reported happiness and sadness across individuals.

The self-reported emotional responses to both “happy” and “sad” music in this study were accompanied by significant changes in averaged skin conductance and heart rate. The direction of physiological changes for the “sad” music was consistent with previous research, with decreases in skin conductance and heart rate (Khalfa et al., 2008; Krumhansl, 1997; Lundqvist et al., 2009; Nykliček, Thayer, & Doornen, 1997). This represents a synchronized emotional response to music stimuli in line with current theories of emotion (Levenson, 1994; Scherer, 2009). The decrease in heart rate and skin conductance in response to “happy” music was, however, inconsistent with previous research, which has demonstrated increased autonomic nervous system activity to arousing and pleasant music (Hodges, 2010; Khalfa et al., 2008; Lundqvist et al., 2009). This inconsistent physiological response may reflect error in measurement or it may be a result of mixed emotional responses to the pieces of music used in this condition, which when averaged over the 15-s period, appear contradictory. For instance, while participants reported liking the “happy” music, it is possible that some participants may have associated this music with sad or relaxing visual imagery or episodic memories (see Juslin & Västjfäll, 2008), resulting in a decline in autonomic arousal associated with an emotional experience mixed with peace, nostalgia or feeling moved. Furthermore, the experimental demand (in particular, cognitive load) involved here may have had its own additive influence on the resulting autonomic response. Given the dynamic nature of music, future research may attain greater sensitivity if a time series analytical approach is taken instead, to explore whether changes in responsiveness across measures are correlated more closely at this more temporally defined level. Nevertheless, the current findings are consistent with recent demonstration that skin conductance appears to be more closely synchronized with a self-reported negative than positive emotional response (Balconi, Grippa, & Vanutelli, 2015).

The current study also demonstrated that individuals were capable of regulating their emotional response to music. Regulation of emotional responses to pictures has been previously demonstrated using fMRI (Heller et al., 2009; Johnstone et al., 2007; Kanske, Heissler, Schönfelder, & Wessa, 2012), but multimodal demonstration of regulation of the emotional response to music has been lacking. The emotional response to “sad” music observed in both self-reported sadness ratings and skin conductance changes was weakly attenuated by the instruction to “reduce” the emotional response. This is consistent with similar findings by Troy et al. (2010), who demonstrated that individuals who had recently experienced a stressful life event were able to down-regulate their sadness response when viewing a sad film, as measured using self-reported emotion ratings and skin conductance responses.

When participants were instructed to strengthen their response to “happy” music, as expected, self-reported happiness ratings were significantly potentiated when compared to those for the feel instruction. The instruction to “strengthen” the emotional response to “happy” music was, however, apparently more difficult than the instruction to “reduce” responses to sad music, with only around half the sample self-reporting an intensification of happiness. This indicates that there was substantial individual variability in the capacity to up-regulate already positive emotions. It may also be possible that people are more accustomed to down-regulation of negative emotions, and find up-regulation of positive emotions (e.g., savouring) more challenging, although research suggests this is not so (e.g., Driscoll, Tranel, & Anderson, 2009). A substantial proportion of the sample reported a reduction in happiness, indicating that there may have been some confusion about this regulation instruction. Confusion with instructions was occasionally raised during practice trials, but readily addressed with further explanation. Nevertheless, this is another possible experimental artefact, potentially adding to the cognitive demand, that may account for the observed attenuation of the skin conductance response, despite self-reports that happiness had indeed increased in response to the “strengthen” instruction with “happy” music.

Alternatively, regulation of emotions may be subject to some desynchronization of emotional elements. Ryff and Singer (2003) suggest that self-reported and physiological indices of emotion may become discrepant under certain conditions. For instance, when self-reported affect is positive, but physiological indices are consistent with negative affect, an individual may be in denial of their emotional state, which can have long-term negative health consequences. Similarly, if emotion regulation is not entirely effective, there may be only partial attenuation, or inconsistent regulation, of the multimodal emotional response. For instance, suppressing emotion expressive behaviours has been reported to enhance autonomic activation even as subjective affect is reduced (Gross, 1998; Gross & Levenson, 1997). Given the absence of heart rate modulation under regulation conditions, and the inconsistency in skin conductance responses with happy music regulation, the current findings suggest that emotion regulation in this instance may have been associated with some desynchronization of emotional indices. Further exploration of this possibility with more intentional focus on type of emotion regulation strategy (e.g., antecedent versus response-focused) would be beneficial in this regard. Methodologically, however, this variability in regulation success across the four pieces highlights the importance of selecting a range of music pieces in studies of this type.

These experiments were intended as an initial exploration of the capacity of people to regulate their emotional response using music as the emotional stimulus. As an initial exploration, the sample size was relatively small, and the study therefore requires replication. Nonetheless, music was demonstrated to be a powerful tool for examining emotion and its regulation, despite the excerpts used in this study being very brief and experimenter- (rather than participant-) selected. While emotions are understood to be elicited rapidly in response to a stimulus (in milliseconds; Ekman, 1992), future studies would benefit from using musical excerpts that are of longer duration (Gomez & Danuser, 2004; Hodges, 2010) in an attempt to elicit a stronger, and more ecologically valid, emotional response. Music could also be personally selected by the individual (idiographic) in an attempt to enhance personal relevance of the stimuli strength (Ellard, Farchione, & Barlow, 2012), although this design comes at the expense of reproducibility and generalizability. In addition to examining responses to “happy” and “sad” music, it would be interesting to examine both emotion response and regulation of music characterized as of high positive energy or high tension (Eerola & Vuoskoski, 2011), or perhaps even as “beautiful” or “awe-inspiring” (Juslin, 2013; Zentner, Grandjean, & Scherer, 2008). Such music might induce emotions that are more strongly associated with action tendencies (approach or avoidance) or aesthetic emotions.

In conclusion, this study provides evidence that short samples of experimenter-selected music can be effective stimuli for the multimodal assessment of emotional responsiveness and regulation. Both “happy” and “sad” music samples were capable of inducing changes in self-reported emotion ratings and in skin conductance and heart rate. When participants were instructed to regulate their emotional responses to “sad” and “happy” music, the self-reported sadness ratings and skin conductance responses were generally modulated. These findings contribute to research supporting the component process model of emotion (Scherer, 2009), and further validate the utility of music as a tool in emotion research.

Footnotes

Acknowledgements

We thank Dr Will Randall and Ms Chantal Roddy for their technical and experimental assistance.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Balconi

Grippa

Vanutelli

M.E.

(2015). What hemodynamic (fNIRS), electrophysiological (EEG) and autonomic integrated measures can tell us about emotional processing. Brain and Cognition, 95, 67–76.

Blood

A. J.

Zatorre

R. J.

(2001). Intensely pleasurable responses to music correlate with activity in brain regions implicated in reward and emotion. Proceedings of the National Academy of Sciences, 98, 11818–11823.

Chin

T. C.

Rickard

N. S.

(2012). The Music USE (MUSE) Questionnaire: An instrument to measure engagement in music. Music Perception, 29, 429–446.

Davies

(2010). Emotions expressed and aroused by music: Philosophical perspectives. In Juslin

P. N.

Sloboda

J. A.

(Eds.), Handbook of music and emotion: Theory, research, applications (pp. 15–44). Oxford, UK: Oxford University Press.

Demaree

Schmeichel

Robinson

Everhart

D. E.

(2004). Behavioural, affective, and physiological effects of negative and positive emotional exaggeration. Cognition and Emotion, 18, 1079–1097.

Driscoll

Tranel

Anderson

(2009). The effects of voluntary regulation of positive and negative emotion on psychophysiological responsiveness. International Journal of Psychophysiology, 72, 61–66.

Eerola

Vuoskoski

J. K.

(2011). A comparison of the discrete and dimensional models of emotion in music. Psychology of Music, 39(1), 18–49.

Eich

J. T. W.

Macaulay

Percy

A. D.

Grebneva

(2007). Combining music with thought to change mood. In Coan

J. A.

Allen

J. J. B.

(Eds.), Handbook of emotion elicitation and assessment (pp. 124–136). New York, NY: Oxford University Press.

Ekman

(1992). An argument for basic emotions. Cognition and Emotion, 6(3/4), 169–200.

10.

Ellard

K. K.

Farchione

T. J.

Barlow

D. H.

(2012). Relative effectiveness of emotion induction procedures and the role of personal relevance in a clinical sample: A comparison of film, images, and music. Journal of Psychopathology and Behavioural Assessment, 34, 232–243.

11.

Fontaine

J. R. J.

Scherer

K. R.

Roesch

E. B.

Ellsworth

P. C.

(2007). The world of emotions is not two-dimensional. Psychological Science, 18, 1050–1057.

12.

Gabrielsson

(2001–2002). Emotion perceived and emotion felt: Same or different? Musicae Scientiae, 5(1 Suppl.), 123–147.

13.

Gomez

Danuser

(2004). Affective and physiological responses to environmental noises and music. International Journal of Psychophysiology, 53, 91–103.

14.

Greasley

A. E.

Lamont

(2011). Exploring engagement with music in everyday life using experience sampling methodology. Musicae Scientiae, 15(1), 45–71.

15.

Gross

J. J.

(1998). Antecedent- and response-focused emotion regulation: Divergent consequences for experience, expression and physiology. Journal of Personality and Social Psychology, 74, 224–237.

16.

Gross

J. J.

John

O. P.

(2003). Individual differences in two emotion regulation processes: Implications for affect, relationships, and well-being. Journal of Personality and Social Psychology, 85, 348–362.

17.

Gross

J. J.

Levenson

R.W.

(1997). Hiding feelings: The acute effects of inhibiting positive and negative emotions. Journal of Abnormal Psychology, 106, 95–103.

18.

Hamm

A. O.

Schupp

H. T.

Weike

A. I.

(2003). Motivational organization of emotions: Autonomic changes, cortical responses, and reflex modulation. In Davidson

R. J.

Scherer

K. R.

Goldsmith

H. H.

(Eds.), Handbook of affective sciences (pp. 187–211). New York, NY: Oxford University Press.

19.

Heller

A. S.

Johnstone

Shackman

A. J.

Light

S. N.

Peterson

M. J.

Kolden

G. G.

…Davidson

R. J.

(2009). Reduced capacity to sustain positive emotion in major depression reflects diminished maintenance of fronto-striatal brain activation. Proceedings of the National Academy of Sciences, 106(52), 22445–22450.

20.

Hodges

D. A.

(2010). Psycho-physiological measures. In Juslin

P. N.

Sloboda

J. A.

(Eds.), Handbook of music and emotion: Theory, research, applications (pp. 279–311). Oxford, UK: Oxford University Press.

21.

Howell

D. C.

(2002). Statistical methods for psychology. Belmont, CA: Duxbury.

22.

Iwanaga

Moroki

(1999). Subjective and physiological responses to music stimuli controlled over activity and preference. Journal of Music Therapy, 36, 26–38.

23.

Johnstone

van Reekum

C. M.

Urry

H. L.

Kalin

N. H.

Davidson

R. J.

(2007). Failure to regulate: Counterproductive recruitment of top-down prefrontal-subcortical circuitry in major depression. Journal of Neuroscience, 37(33), 8877–8884.

24.

Juslin

P. N.

(2013). From everyday emotions to aesthetic emotions: Towards a unified theory of musical emotions. Physics of Life Reviews, 10, 235–266.

25.

Juslin

P. N.

Liljeström

Västfjäll

Lundqvist

L-O.

(2010). How does music evoke emotions? Exploring the underlying mechanisms. In Juslin

P. N.

Sloboda

J. A.

(Eds.), Handbook of music and emotion: Theory, research, applications (pp. 605–642). Oxford, UK: Oxford University Press.

26.

Juslin

P. N.

Västfjäll

(2008). Emotional responses to music: The need to consider underlying mechanisms. Behavioral and Brain Sciences, 31, 559–621.

27.

Kanske

Heissler

Schönfelder

Wessa

(2012). Neural correlates of emotion regulation deficits in remitted depression: The influence of regulation strategy, habitual regulation use, and emotional valence. NeuroImage, 61, 686–693.

28.

Khalfa

Roy

Rainville

Dalla Bella

Peretz

(2008). Role of tempo entrainment in psychophysiological differentiation of happy and sad music? International Journal of Psychophysiology, 68, 17–26.

29.

Krumhansl

(1997). An exploratory study of musical emotions and psychophysiology. Canadian Journal of Experimental Psychology, 51, 336–352.

30.

Lang

P. J.

Greenwald

Bradley

M. M.

Hamm

A. O.

(1993). Looking at pictures: Evaluative, facial, visceral, and behavioural responses. Psychophysiology, 30, 261–273.

31.

Lehmann

A. C.

(1997). Affective responses to everyday life and music listening. Psychology of Music, 25, 84–90.

32.

Levenson

R. W.

(1994). Human emotions: A functional view. In Ekman

Davidson

R. J.

(Eds.), The nature of emotion: Fundamental questions (pp. 123–126). New York, NY: Oxford University Press.

33.

Levenson

R. W.

(2003). Autonomic specificity and emotion. In Davidson

R. J.

Scherer

K. R.

Goldsmith

H. H.

(Eds.), Handbook of affective sciences (pp. 212–224). New York, NY: Oxford University Press.

34.

Lundqvist

L.-O.

Carlsson

Hilmersson

Juslin

P. N.

(2009). Emotional responses to music: Experience, expression, and physiology. Psychology of Music, 37(1), 61–90.

35.

Mauss

I. B.

Levenson

R. W.

Carter

Wilhelm

F. H.

Gross

J. L.

(2005). The tie that binds? Coherence among emotion experience, behaviour and physiology. Emotion, 5, 175–190.

36.

Nater

U. M.

Abbruzzese

Krebs

Ehlert

(2006). Sex differences in emotional and psychophysiological responses to musical stimuli. International Journal of Psychophysiology, 62, 300–308.

37.

Nykliček

Thayer

J. F.

Van Doornen

L. J. P.

(1997). Cardiorespiratory differentiation of musically-induced emotions. Journal of Psychophysiology, 11, 304–321.

38.

Rickard

N. S.

(2004). Intense emotional responses to music: A test of the physiological arousal hypothesis. Psychology of Music, 32(4), 371–388.

39.

Ryff

C. D.

Singer

B. H.

(2003). The role of emotion on pathways to positive health. In Davidson

R. J.

Scherer

K. R.

Goldsmith

H. H.

(Eds.), Handbook of affective sciences (pp. 1083–1104). New York, NY: Oxford University Press.

40.

Saarikallio

(2011). Music as emotional self-regulation throughout adulthood. Psychology of Music, 39(3), 307–327.

41.

Salimpoor

V. N.

van den Bosch

Kovacevic

McIntosh

A. R.

Dagher

Zatorre

R. J.

(2013). Interactions between the nucleus accumbens and auditory cortices predict music reward value. Science, 340, 216–219.

42.

Schäfer

Sedlmeier

Städtler

Huron

(2013). The psychological functions of music listening. Frontiers in Psychology, 4, 511. doi:10.3389/fpsyg.2013.00511

43.

Scherer

K. R.

(2005). What are emotions? And how can they be measured? Social Science Information, 44, 693–727.

44.

Scherer

K. R.

(2009). The dynamic architecture of emotion: Evidence for the component process model. Cognition and Emotion, 23(7), 1307–1351.

45.

Scherer

Coutinho

(2013). How music creates emotion: A multifactorial process approach. In Cochrane

Fantini

Scherer

K.R.

(Eds.), The emotional power of music (pp. 121–146). Oxford, UK: Oxford University Press.

46.

Scherer

K. R.

Zentner

M. R.

(2001). Emotional effects of music: Production rules. In Juslin

P. N.

Sloboda

J. A.

(Eds.), Music and emotion: Theory and research (pp. 361–392). Oxford, UK: Oxford University Press.

47.

Tabachnick

B. G.

Fidell

L. S.

(2007). Using multivariate statistics (5th ed.). Boston, MA: Pearson Education.

48.

Troy

A. S.

Wilhelm

F. H.

Shallcross

A. J.

Mauss

I. B.

(2010). Seeing the silver lining: Cognitive reappraisal ability moderates the relationship between stress and depressive symptoms. Emotion, 10(6), 783–795.

49.

Tugade

M. M.

Fredrickson

B. L.

(2007). Regulation of positive emotions: Emotion regulation strategies that promote resilience. Journal of Happiness Studies, 8, 311–333.

50.

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 & Emotion, 21, 3–25.

51.

Zentner

Grandjean

Scherer

K. R.

(2008). Emotions evoked by the sound of music: Characterization, classification and measurement. Emotion, 8, 494–521.