“Unhappiness is my lot”. Or is it? Depression in students who perform or study music as compared to students who do not.

Michaela Korte1, Deniz Cerci2, Roman Wehry2, Renee Timmers1, Victoria J. Williamson1

1Department of Music, The University of Sheffield, United Kingdom, 2Vivantes Wenckebach-Klinikum, Berlin, Germany, 3Helios Klinikum, Klinik für Schmerzmedizin, Hildesheim, Germany

Abstract Professional musicians are more at risk of developing mental health problems, such as depression, compared to the general population. The level of mental health problems experienced by musicians is comparable to that found in professions with high stress levels such as physicians or air-craft engineers. Despite this fact, it remains unknown whether contributory and standardized factors have the potential to identify whether early career musicians are at risk of depression. This study examines known depression predictors such as anxiety, pain processing, depersonalization and coping strategies, and compares how they affect a music and non-music student population. Moreover, it investigates the extent to which professional identification has an effect on musicians’ levels of depression, and if this factor could be used as a depression predictor for student musicians. 102 under- and postgraduate students (75% UK, 16% other EU countries, 9% USA; age mean = 23.6 years) from various institutions and with different primary subjects (62% music, 38% medicine, psychology, biology) participated in an online survey featuring standardized scales for the above- mentioned depression predictors. Music college students had a significantly higher depression prevalence (31.2%) compared to musicians at university (9.3%), but not compared to university non-musicians (19.4%). The depression factors of anxiety and pain processing were lower in both musicians’ groups compared to non-music students, despite the latter group perceiving pain for shorter periods of time than all musicians. A hierarchical multiple linear regression and a regression tree analysis indicated that professional identification was not a significant predictor for depression. While anxiety accounted for the highest variance (58%), a combination of factors best predicted depression: high anxiety combined with a low-level of burnout with teaching staff. Taken together these results suggest that the culture of teaching institutions may play a vital role in the experience of depression in student musicians. Furthermore, standardized instruments such as anxiety scales, rather than profiling within the profession, offer a differentiated and therefore more promising depression risk analysis for training musicians.

Introduction While research into depression and musicians is a relatively recent phenomenon, descriptions of mental health problems within the field have been frequent across the centuries (Cordingly, 2001; Kluge, 1818; Lorusso & Porro, 2020; Rahm, 1994) , as exemplified by Mozart’s heroine from “The Abduction from the Seraglio”, who is quoted in the title of this article. In fact, descriptions of mental health are so realistic within music, that medical papers have often discussed them as if they were case studies (Chest, 2018). These include, for instance, vision madness (e.g. apparitions) and schizophrenia (Dura-Vila & Bentley, 2009; Erfurth & Hoff, 2000; Ropert, 2003), suicide (Pridmore, Auchincloss, Soh, & Walter, 2013) or (bipolar) depression (Boutolleau, 2017; Lorusso, Franchini, & Porro, 2015). This familiarity with, and realism, in portraying mental health problems within

1 compositions have led many to speculate that composers themselves might have been affected by these conditions. Indeed, people have argued that successful musicians with no history of mental health difficulties were the exception (la Motte-Haber, 2006). Reflecting on this issue, Sigmund Freud wrote that a happy individual imagines or creates nothing (Freud, 1918). But is this historical stereotype linking musicianship and mental health reflected by evidence relating to depression nowadays?

On the one hand the answer appears to be ‘yes’. Research has established that musicians – composers and performers alike – suffer more from mental health issues generally, and depression specifically, compared to the general public (Constant, 2011; Johnson et al., 2012; Kenny & Ackermann, 2015a; Lederman, 2015; Steinberg, 2016; Vitale, 2009). On this basis, musical, and thus creative, accomplishment has been predicted (Burch, Pavelis, Hemsley, & Corr, 2010; Carson, 2014; Dietrich, 2014; Janka, 2004) . The idea that musicians, creativity and depression are intertwined is so entrenched that studies showing a deviation from this statement find the need to point this out specifically (Chávez-Eakle, 2006). However, the proportion of depression in musicians needs to be seen within context (Ellis, 2010), so at-risk populations can be identified correctly and the true scale of the problem laid bare.

Depression prevalence in general has increased within the last decade and is the fourth cause for disability worldwide (WHO, 2012). In Europe, depression prevalence is the second overall highest burden of disease after substance abuse disorders. To put this into further context, cumulative weighted prevalence rates combining substance abuse, dependence and depression scores, account for 58% of the disease burden within the EU (Angst et al., 2005; Wittchen & Jacobi, 2005; Wittchen et al., 2011). The prevalence rate in an at-risk population, such as musicians’, will be higher. Another important at-risk subpopulation is younger people, since depression shows its first peak within the adolescent/young adult population (Bertha & Balázs, 2013). The distinctive mental challenges attached to being a student at this age are factors that contribute to the high depression prevalence in this group (Bacchi & Licinio, 2015; Kessler et al., 2009). A logical question is whether these strands of evidence lead to the conclusion that musicians in training, or students who play music in their spare time, are at particular risk of depression?

Despite the apparent inevitability of such a conclusion, various reasons have been offered to explain an elevated depression prevalence in musicians. Some consider a higher-than-normal depression rate sufficient to indicate that the entire musical profession is more at risk for depression (Vaag, Bjørngaard, & Bjerkeset, 2016), while others have found that the depression rate among musicians, while higher than the norm, is comparable to that of other professions with high stress levels, such as aircraft engineers or medical students (Voltmer, Kötter, & Spahn, 2012; Voltmer, Schauer, Schroder, & Spahn, 2008; Woodward, Lipari, & Eaton, 2017). This latter hypothesis assumes that depression predictors, such as (chronic) stress and (high) anxiety could cause any population to be more at risk, no matter their age or profession.

Following this thought, studies have investigated whether particular depression predictors are more commonly found within groups of musicians. Musicians were found to have higher scores for depression predictors such as (chronic) pain {Spahn:2004tr, Gasenzer:2011ki, Maric:2019dx} 2 , psycho-social stress (Hildebrandt, 2002; Holst, Paarup, & Baelum, 2012; Langendörfer, Hodapp, Kreutz, & Bongard, 2006), performance anxiety on its own or in context with other (clinical) conditions (Jabusch & Altenmüller, 2004; Kenny & Ackermann, 2015b; Spahn, Hildebrandt, & Seidenglanz, 2001) and dysfunctional attitude to discomfort (Spahn, Burger, Hildebrandt, & Seidenglanz, 2005; Zander, Voltmer, & Spahn, 2010). Since these predictors are specific to the musicians’ environments, there has been a tendency to assume that they carry more predictive weight compared to more general factors such as depersonalization (Michal et al., 2015), burnout (Frank, Nixdorf, & Beckmann, 2017) and dysfunctional coping strategies (Prinz, 2012). Furthermore, to date all depression predictors, whether specific to musicians or not, have only been analyzed on their own and not in larger models to consider their accumulative weight in predicting depression in a musicians’ environment.

The inconsistencies in musicians’ depression research identified above are mirrored in sports psychology research, particularly in elite college athletes (Nixdorf, Frank, & Hautzinger, 2013). While some studies see college athletes as more at risk of depression due to their intensive investment into sports, others did not find such a significant difference in depression levels between college athletes and non-athletes (Armstrong & Oomen-Early, 2010; Proctor & Boan- Lenzo, 2010). These latter studies identified financial stress and change in lifestyle from high school to university/college life to be sufficient to explain higher than norm depression prevalence in both groups. While the added pressure for athletes to perform/compete at the highest level, and stress caused by their relationships with coaches/trainers, have been cited as reasons for this increase, their stress levels were not consistently found to be significantly different from those experienced by non-athletes. It is conceivable that the same reasoning could apply to music students, but research into music students’ wellbeing is still catching up with sports psychology in this regard.

A prominent differentiating factor between athletes and musicians is a strand that appears to be exclusively discussed in reference to musicians and other creative professions, namely the hypothesis that depression is innate to creative artists. This factor is often referred to as the ‘the myth of creativity’ (i.e. the more creative the artist, the more mental health illness ensues) (Jamison, 1995; Lauronen et al., 2004; Ludwig, 1992). It is worth noting that whilst depression prevalence within other professional groups compares to or exceeds that amongst musicians, there is no literature claiming an innate depression vulnerability for other groups (Fan et al., 2012; Wulsin, Alterman, Timothy Bushnell, Li, & Shen, 2014). In a way, the hypothesis that musicians are more at risk compared to the norm might be seen as a logical building block for the hypothesis that depression is innate to musicians. This line of argument has some support within depression literature (Wray et al., 2018), and particularly in genetic depression research, where a ‘depression vulnerability gene’, or a variation of the DNA in a specific location on a chromosome1, was shown to make carriers of the variation more vulnerable to depression (Caspi et al., 2003). However, such a line of argument overlooks the factors that contribute to such a vulnerability within a group, besides being a part of the group itself. There is no such thing as a ‘genuine depression gene’ compared to, for instance, the one

1 Vulnerability factors were discovered in the promotor-region of the serotonin-transporter-gene, the 5-HTTLPR (5HT Transporter Lengh Polymorohic Region), on the chromosome 17q11.1-q12, for more details, please see (Keers & Pluess, 2017) 3 identified for Huntington’s (Myers, 2004), as carriers of the genetic depression vulnerability will only be subject to significantly more depression episodes, compared to those who do not carry this mutation, if they accumulate stress (Mayberg et al., 1997; Monroe & Reid, 2008; Short & Baram, 2019).

In summary, depression prevalence research is indicative when it comes to understanding who may be at risk. However, identifying an at-risk population without identifying depression predictors/ stressors that promote risk, limits both the scope of investigation and nature of the subsequent conclusions. The present study therefore looked at both prevalence and predictors within an already identified at-risk population of young training musicians in order to answer the following four questions: (1) how does depression prevalence compare between student musicians and student non-musicians? (2) How does depression prevalence compare between students of music in different higher education institutions (university and music college)? (3) What is the role, if any, of more general depression predictors in predicting depression across student groups (university students of music versus music college students)? (4) What role, if any, does the factor of ‘professional identification’ play in predicting depression prevalence in a musicians’ environment?

In answering these questions, this study will determine which known factors influence reported depression in music and non-music students, how or if they differ between groups, and if the level of professional identification can be considered a predictor variable for depression. In order to inform our choice of factors within the study, we concentrated on previously identified and well evidenced depression predictors – general and music-specific ones – such as anxiety, pain perception and pain processing (in this case, pain catastrophizing), professional identification, depersonalization, burnout and coping styles. The outcomes of this study will provide timely evidence on how we may better identify music students who are at risk of depression during their training years.

Materials and Methods

Procedure An important issue within the present study is anonymity. It has been established that students worry about being identified as having a mental disorder in case they experience (academic) disadvantages due to the stigma attached to such a condition. This fear alone, whether perceived or based on experience, can dissuade individuals from participating in studies of mental health (Gebauer, 2014; Hall, 2018; Storrie, Ahern, & Tuckett, 2010). Therefore, this study was designed to respect anonymity at the highest level, most notably in the choice of survey provider2. We prioritized automatic privacy settings to prevent the collection of personal information during the process of filling out the questionnaire. This included preventing the consolidation of participants’ email addresses with their survey answers. Participants were informed about these measures in the introduction text. Furthermore, the question of gender was omitted to prevent the involuntary tracking of participants. This was particularly important in cases where there were few students pursuing studies of a particular instrument or vocal ‘Fach’3, who might otherwise have become recognizable.

2 SoSci survey (Leiner, 2014) 3 Fach differentiates voice categories, classifying them according to their size, range, weight or color, etc. For a tenor this means, for instance, ’lyric tenor’ or ‘dramatic tenor’. Both repertoires will rarely overlap, if at all, even though both singers are tenors. For more details, please see (Legge, 1988) 4 Questions were not mandatory, but where individual questions were not answered total scores were calculated using the individual guidelines of the questionnaires and/or general depression literature (WHO, 1992). Recruitment was limited to online contact, and links to the questionnaire were emailed to students of the University of Sheffield via the university server (Jan. 2016), and to various music colleges via their respective student presidents (Jan. 2017).

This study was carried out in accordance with the recommendations of the University of Sheffield, and the protocol was approved by the Ethics Committee of the Department of Music. All subjects gave written informed consent, which was collected by the survey’s software, in accordance with the Declaration of Helsinki, prior to being able to access the questionnaire. It took participants on average 15 minutes to complete the survey.

Material: At the beginning of the online survey participants were asked to share demographic details such as age, relationship status, primary course of study and preferred methods of relaxation, choosing from a list of options (e.g. various high/low impact sports see table 1). In the following section detail is provided on each of the selected standardized scales to measure depression predictors.

The Hospital Anxiety and Depression Scale (HADS) (Zigmond & Snaith, 1983) is a self-reported questionnaire to assess depression and anxiety symptoms that uses two separate scales (seven items per scale) based on a 4- point Likert response. As a primary care screening tool, it is well-validated and routinely used and, despite its brevity, is considered a comprehensive instrument (Bjelland, Dahl, Haug, & Neckelmann, 2002). The HADS asks participants to rate their mood based on the previous week. This instruction helps to avoid classifying participants as depressed who may be experiencing a mood disturbance only on the day they complete the scales. The HADS discriminates well between anxiety and depression, and is a good fit to the Rasch Model (Bjelland et al., 2002; Pallant & Tennant, 2010), meaning it is less subject to bias (cultural or across professions). The cut-off for significant depression was set at ≥ 9 for both scales in accordance with the literature (Sadek & Bona, 2000). Increased rating on the anxiety scale suggests a more generalized anxiety disorder according to the International Statistical Classification of Diseases and Related Health Problems (ICD), while higher scores on the depression scale indicate symptoms leading to a depressive episode (ICD-10, WHO, 1992). For participants who completed all questionnaires except the HADS, anxiety and depression scores were calculated separately: in accordance with the literature high scores in pain catastrophising, sleep dysfunction, dysfunctional coping and negative identification (MIMS), combined with low scores of active and cognitive coping, projected a depression score of ≥ 9. Consequently, the reverse predicted a depression score of ≤9 (WorldHealthOrganization:1990tv}.

The GOLD-MSI (Müllensiefen, Gingras, Musil, & Stewart, 2014) is a self-reported test that assesses an individual’s propensity to engage with music. It is modelled on a multidimensional construct of musical sophistication and assumes that musical skills and habits are not only acquired through instrumental lessons, but also by actively engaging with music in all its facets. We used two of the test’s subscales: active engagement and musical training.

The Athletic Identity Measurement Scale (Brewer, Van Raalte, & Linder, 1993) is a 10-item scale 5 that assesses the strength and exclusivity of an athlete’s professional identity. The higher the score, the more a candidate identifies with being an athlete. The stronger the identification with the role of an athlete, the more the athlete is at risk of developing a depression if an obstacle is keeping him from training (e.g. injury, recovery) (Lally, 2007; Park & Tod, 2015; Wiechman & Williams, 1997). This scale has been previously used to assess identity in musicians (Vitale, 2009) and dancers (Langdon & Petracca, 2010) by changing the word athlete to musician or dancer. We followed this example. For the purpose of this study, we called the questionnaire MIMS, standing for Musicians Identification Measurement Scale.

The Cambridge Depersonalization Scale (CD-9) (Sierra & Berrios, 2000) measures depersonalization, a medical disorder that is marked by emotional regulation difficulties: a feeling of incompleteness, a sense of detachment from oneself, and the experience of one’s own actions as ‘not-feeling-quite-right’. Correlations of depersonalization with anxiety, depression, and burnout symptoms suggest a comorbidity or, at least, common neurobiological denominators between these symptoms (Hunter, Sierra, & David, 2004). The cut- off point for significance was set at the level requested by the scale authors at ≥ 19 for short, transient versions and ≥ 90 for depersonalization as a unique condition. Transient depersonalisation phases, short in frequency and duration are characterized by the main symptoms: the absence of emotions, a feeling of ‘not being quite oneself ‘or ‘feeling a like robot’. There is similarity between this description and a main symptom of depression, namely an absence of emotion (anhedonia). While transient depersonalization phases are common under stress or trauma, a permanent and unique state of depersonalization has rarely been observed (i.e. 1-2% prevalence worldwide (Dell & O'Neil, 2010)).

The Örebro Musculoskeletal Pain Screening (Linton & Boersma, 2003) is a primary care tool to assess individuals with sub-acute and chronic musculoskeletal pain. It measures how their experience of pain affects their performance at work or, in our case, their performance at university or music college. In 21 questions, it addresses beliefs and expectations about pain (yellow flags). The higher the final score, the less likely the individual is to return to work while remaining disabled by pain. Total scores of over 105 show a moderate risk of disability, and a total score of over 130 a high risk of being disabled by pain. Linton and Boersma specified that, when using a six-month prediction, 71% of patients were correctly classified (sensitivity, 72%; specificity, 70%), with a high reliability (훼 = .97, p ≤ .05), a high criterion validity (Spearman’s r =.97) and a high internal consistency (훼 = .87).

The Brief COPE (Carver, Scheier, & Weintraub, 1989) is a short version of the COPE questionnaire by the same authors. For the Brief COPE, the 14 scales from the original questionnaire were distilled into three scales via a factor analysis with varimax rotation. These three scales are: active functional coping (e.g. ‘I actively did something'), functional cognitive coping (e.g. 'I tried to find something positive in what happened.') and dysfunctional coping strategies (e.g. 'I used alcohol/other substances to help me through this situation'). This reduction to three scales allows for diverse testing of stress coping and correlation of findings. While this questionnaire overlaps in part with the Burnout Scale (to follow), it also focuses on important information about coping strategies and is the only scale to do so in the battery.

6 The Copenhagen Burnout Inventory, (CBI; (Kristensen, Borritz, Villadsen, & Christensen, 2005)) The CBI consists of three main scales: personal burnout, work burnout, and client-related burnout. The word 'client' should generally be replaced with a word suited to the environment in which the study is conducted. Borritz and Kristensen (2015) attested all three scales to have very high internal reliability (훼 = .85 - 87). This study's design was modelled on the study by (Campos, Carlotto, & Marôco, 2013) who used it to assess burnout in Brazilian and Portuguese university students . To reflect the dual ‘client’ burnout problem of students, the ‘client’ questions in this questionnaire were doubled up, exchanging the word ‘client’ with ‘fellow student’ in one set, and ‘teacher/professor’ in the other.

Participants: 102 participants took part in the study: 67 students from The University of Sheffield (TUOS) and 36 students from music colleges. The group of TUOS students was made up of two sub-groups: 31 musicians and 36 non-musicians. The musicians’ group comprised music students and students who self-identified as musicians. Irrespective of whether the University music students were part of the Music Department or not, there was no difference regarding level of practice (years practiced and daily practice), instrumental/vocal and music theory lessons taken, or engagement with music (listening to music, going to live concerts, etc.). Participants in the non-musician group had never played an instrument, had taken no music theory lessons and did not engage with music in the same way as the musicians (e.g. did not attend live concerts as much, if at all, or listen to music on a regular basis; see table 1). For ease of reference, the three different groups will from now on be referred to as university musicians, university non- musicians and college musicians (see all tables below).

The average age of participants was 23.26 years (SD:6.81). University musicians were the youngest with an age mean of 21.6 years, followed by university non- musicians with a mean of 23.3 years and then music college musicians with a mean of 27.9 years (see table 1 for further group demographics). All three groups were almost equally made up of undergraduate and postgraduate students, and students were to an equal percentage single or in a stable relationship (see table 1). All three groups similarly preferred active relaxation techniques over meditative ones or talking therapies.

Statistical analysis: Evaluation of the data was performed using the software RStudio 1.2.1335 (R Core Team, 2013) . An a priori calculation was performed with G*Power (Faul, Erdfelder, Lang, & Buchner, 2007). Power of  = .8 was considered as appropriate (Rasch, Friese, Hofmann, & Naumann, 2009). Pairwise comparisons, regressions and a tree model were performed to analyze group differences and predictability. Bonferroni corrections were applied to safeguard against multiple testing. Spearman’s correlation was applied for correlations.

There are numerous statistical techniques that could be used to model dependent variables and deal with a larger set of predictors, such as a canonical correlation (Thompson, 1984) or logistic regression (Hoerl & Kennard, 1970). Considering the structure of the data and the research question of whether ‘being a musician’ could be used as a depression predictor, we decided on two complementary models: (a) a hierarchical multiple linear regression and (b) a tree model. While depression factors seldom present a linear pattern, a

7 hypothetical approach to a linear regression modelling can still be advantageous. The focus here is to compare different regression models using statistics in order to understand the relationship between factors, and specifically, if adding more factors over and above the strongest factor significantly improves predictability (Raudenbush & Bryk, 2002). Regression tree models and forest analysis have been used in larger statistical analyses, in artificial intelligence, and, more recently, in psychology (Anglada-Tort & Müllensiefen, 2017; Jakubowski, Finkel, Stewart, & Müllensiefen, 2017; Pawley & Müllensiefen, 2012)4. While these models are typically used for larger data sets, a growing number of studies supports their use for smaller collections, such as the present data sample (Liu et al., 2007; Song & Lu, 2015).

Tree methodology is used to classify a dataset based on multiple covariates or develop predictors for a target variable. This approach is robust since the algorithm does not impose a parametric structure and can deal with complicated data sets. The advantage of using a tree model is firstly that they do not presume a functional linear relationship between variables. Secondly, they are ideal for high-order interaction effects between predictor variables, which do not need to be specified as is the case in a linear regression. Thirdly, tree models can deal with large sets of heterogeneous variables, which includes continuous, ordinal, categorical or binary ones. The heterogeneity of the variables and the assumption that depression requires certain conditions (i.e. interaction of certain variables) were the reason to adopt a tree-based approach as an additional control to the regression.

Results This results section is divided into two parts: part one reports the results from the individual depression factors – known henceforth as ‘variables’ as according to standard reporting in statistics (demographics followed by each standardized questionnaire) and part two explores the prediction of depression based on a hierarchical multiple linear regression and a tree model. Following part one, two of the predictors, depersonalization and coping, were dropped from the regression and tree models conducted in part two, due to insignificant findings. These outcomes are detailed at the end of each relevant predictor results section.

Part 1 – Scale Outcomes

Musical sophistication and musicians’ identity (GOLD-MSI and MIMS) According to the GOLD-MSI, Music college musicians spent more time on daily practice and formal instrumental/vocal lessons than uiversity musicians however, they did not accumulate more years of practice or music theory overall. Music college musicians spent more time listening attentively to music and attended more live concerts (as audience members) than university musicians. As expected, non-musicians did not play any instrument or sing, and consequently did not take any music lessons: instrumental, vocal or music theory. They also listened less to music and attended fewer concerts than both musicians’ groups (see table Gold MSI).

4 For a detailed introduction to tree methods see (Breiman, Friedman, Olshen, & Stone, 1984), and for an overview in music research see (Müllensiefen, 2010)

8 Moving on to the MIMS, a Mann-Whitney U-test determined a significant difference in the full score between music college musicians and university musicians, with a large effect size (U = 856.0, p = .001, rank-biserial correlation (effect size) = .57). This difference in identity score is comparative to the one measured in college athletes. College athletes’ identity scores dropped when they changed their career path from professional sports to a different profession. It was stable or increased when they aimed to pursue a career in sports (Lally & Kerr, 2013). Table 1 Demographics including age, ways of dealing with stress, practice time, formal musical training, formal music theory lessons, and the importance of music in daily life represented by daily attentive listening to music and attendance of concerts as member of the audience (GOLD-MSI) University College University Non- Musicians Musicians Musicians Age (mean, SD) 21.6; 3.44 27.9; 8.74 23.3; 7.85 Family/relationship status: stable 49% 57% 48% Family/relationship status: single 51% 43% 52% Stress relief, active (running, 15% 63% 87% yoga, etc.)1 Stress relief, other 67% 9% 7% (meditation, therapies, etc)1 Regular practice time (years)2 4.25 4.76 0 Regular practice time (hours per 2.45 6.14 0 day) Music theory lessons (years) 4.82 4.82 0 Formal instrumental/vocal training 4.74 6.11 0 (years) Attending live concerts (member 3.22 5.35 1.05 of audience, during past year) Attentively listening to music 30-60min 30-60min 0-15min (average per day) NOTE: 1Activities had to be carried out regularly: once or twice a week over a minimum period of two months 2 Please note that the GOLD-MSI gives ranges and not an exact number, e.g. 4-6 years, which is here noted as 4

Similar to studies comparing elite college athletes with those college athletes who were not pursuing a professional career in athletics, only the subscales self- identity (U = 853.0, p < .004) and social identity (U = 588.0, p < .004) returned significantly higher results for college musicians than for university musicians. Negative affectivity (p = .4) and exclusivity (p = .5) did not differ significantly.

Table 2 Musician Identity Measurement Scale (MIMS) results with subscales for all groups including mean and standard deviation (SD) Test Groups Music College University University Non-Musicians Musicians Musicians mean SD mean SD mean SD MIMS total 49.11 10.42 32.50 16.63 - - score Self-identity 11.71 2.88 7.22 4.20 - - Social 9.88 3.17 5.90 3.37 - - identity Negative 8.31 2.37 6.80 4.59 - - affectivity Exclusivity 10.00 3.12 9.08 4.97 - - NOTE: Higher means indicate a stronger identification with the role of a musician.

Depression and anxiety There was a significant difference in depression prevalence between the college musicians’ group and both the university musicians (z= -3.67, p = .0002), and also

9 when compared to the university non-musicians (z = 2.16, p = .003). In other words, college musicians showed significantly more depression compared to university musicians and non-musicians. The results can be found in table 3.

The comparison of depression prevalence on its own seems to support the hypothesis that ‘being a musician’ is a variable that contributes to depression. However, this argument loses its force when looking at anxiety prevalence, one of the most significant depression predictors identified by literature (WHO, 1992). Here, the highest prevalence was found in university non-musicians. When compared to music college musicians, the difference was significant (z= -2.01, p= .04). While university musicians also had a 14.9% lower anxiety prevalence compared to university non- musicians, there was no statistically significant difference between these two groups (p= .2).

TABLE 3 Results from all standardized tests for all groups: HADS, CD-9, MIMS, Brief Cope, Örebro (pain catastrophizing) and CBI (burnout) including mean, standard deviation (SD) and prevalence (if applicable). Groups Test Music College University Musicians University Non- Musicians Musicians mea SD Prevalenc mea SD Prevalenc mea SD Prevalenc n e n e n e HADS Anxiety 8.68 3.58 43.7% 4.3 4.05 40.6% 4.5 3.30 55.5% HADS 7.80 1.4 31.2% 4.41 3.26 9.3% 5.2 3.38 19.4% Depression MIMS (total 49.1 10.4 - 32.5 16.6 - - - - score) 1 2 3 CD-9 (total 28.6 6.39 ≥19 = 32.0 16.4 ≥19 = 32.9 17.6 ≥19 = 90% score) 93.3% 6 2 100% 7 Brief Cope 21.6 4.60 - 21.5 4.32 - 20.5 7.11 - active 5 5 4 functional Brief Cope 19.1 5.69 - 17.2 5.60 - 15.9 4.88 - cognitive 0 4 6 functional Brief Cope 9.65 2.64 8.65 2.53 - 8.53 2.72 dysfunctional Örebro 82.2 23.3 ≥105=13.3 73.9 25.4 ≥105=13.7 80.5 30.2 ≥105=24.1 2 2 0% 7 1 9% 4 7 3% ≥130 = 0% ≥130 = ≥130=10.3 3.4% 4% CBI (Burnout) 2.4 0.29 - 2.89 0.89 - 2.57 0.88 - Personal 15 CBI 3.06 0.88 - 3.03 0.74 - 3.05 1.28 - Work CBI 3.18 0.37 - 3.56 1.05 - 3.42 1.14 - Student CBI 3.88 .01 - 3.76 0.98 - 3.81 0.98 - Teacher/ Professor NOTE: with the exception of the brief cope active functional and cognitive functional scales, higher values indicate more depression/anxiety/depersonalization symptoms (HADS & CD-9), a less favorable way of dealing with stress (CBI), higher/worse pain catastrophising and more dysfunctional coping strategies. For the brief cope active functional and cognitive functional coping scales, higher values indicate a better way of dealing with stress.

Pain catastrophizing The first questions from the Örebro provides an overview of pain duration: 39% of music college musicians had perceived pain for 12 months or longer, compared to 32% of university musicians and 22% of university non-musicians. While music

5 The CBI was not part of the second set of questionnaires. This score was therefore estimated with the help of a tree model to predict a high/low score, using all pertinent questionnaires referring to emotional exhaustion, depersonalization, fear, being prone to illness (leave of work/study time), etc. 10 college musicians reported the most chronic pain, university non-musicians exhibited the highest rate of dysfunctional pain processing: there was a significant difference in high pain catastrophising (≥130) between music college musicians and university non-musicians (z = 1.94, p = .05). Although university musicians showed a lower prevalence of high catastrophizing, compared to university non-musicians (6.94%), this difference was not statistically significant (p = .3). These results underscore current medical literature on chronic pain, more specifically they re-confirm that pain perception does not necessarily predict how pain is processed. Musicians, no matter at which institution, perceived longer periods of pain compared to non-musicians, but did not process this pain in a negative way as non-musicians did. Put another way, non-musicians suffer more from the pain they experience and hence are more subject to depression and anxiety problems compared to musicians.

Depersonalization Depersonalization, as measured by the CD-9, is one way of looking into emotion processing. There was almost no difference in depersonalization prevalence between college musicians (93.3%), university musicians (100%) and university non-musicians (90%).

Coping Scales measuring coping skills help shed light on how individuals cope with stressful situations. Music college students, like their university counterparts (musicians or not), tend to favor active coping methods over cognitive functional ones, while dysfunctional coping strategies were comparatively low (see table 3). An ANOVA found no significant difference between groups for this variable (active functional cope: p = .6; cognitive functional cope: p = .2; dysfunctional cope: p = .7). This means that all groups had a (relatively) healthy approach to coping with stress and/or problems, favoring an active approach (e.g. actively looking to overcome a problem, asking for help) over a cognitive one (e.g. finding the positive aspect in what happened), while not giving into less favorable strategies (e.g. drinking, retreating, etc.). In terms of depression, this should mean that all groups are equally likely to cope relatively well with depression when they are affected. In order to avoid multicollinearity with MIMS, this variable will not be further investigated in the regression to follow in the next section.

Burnout Burnout sheds light onto stress specificity, or, in other words, the areas of life which are perceived as most stressful for individuals: personal interactions, workload and interactions with people at work (in this case with their peers and with teaching staff/professors). There was no statistically significant difference between groups. The highest level of burnout was experienced based on interactions with university teaching staff, followed by interactions with fellow students, work burnout and then personal burnout. Burnout and depression intersect in both directions, meaning they might cause each other to a certain degree.

Part 2 – modelling

Hierarchical linear multiple regression The a priori power calculation for a linear multiple regression fixed model, setting R2 to a medium effect size, found that the overall number of participants required for this test were 826, hence the present sample was sufficient.

6 .09; with an 훼error = .05, power 1- error probability = .08, number of tested predictors = 1 and total number of predictors = 4. 11 The overall multiple linear regression model using (HADS) anxiety, pain catastrophizing, teacher/professor burnout and the MIMS full score was significant F(4,86) = 34.05 , p <.001, R2 = .595. In this context the MIMS score was added as the variable representing ‘being a musician’. The first individual multiple linear regression model, comprising only three independent variables (HADS) anxiety, pain catastrophizing and professor/teacher burnout, was significant: F(3,87) = 33.33, p <.001, R2 = .535. All individual variables were significant: anxiety (b = .58, t(87) = 10.7, p < .001, pr2 = .568), pain catastrophising (b = .015, t(87) = 1.67, p < 001, pr2 = .031) and professor/teacher burnout (b = -.25, t(87) = - 1.72, p < 001, pr2 = -.035). Anxiety emerged as the strongest depression predictor. With every .58 unit increase in anxiety the model predicted one unit increase in depression. Both pain catastrophizing and professor/teacher burnout predicted depression with approximately similar strength, but both were weaker predictors compared to anxiety. Burnout returned an inverse score, suggesting that the higher the burnout score, the lower the depression score. This is in line with the most recent publications suggesting that, while burnout is a separate condition to depression, they can be considered related (WHO, 2019).

The second multiple linear regression model added the full professional identification score to the other three variables included in the previous model. This model was equally significant (p <.001) however, this only demonstrates that R2 was greater than zero. The important question for this calculation was if adding the variable ‘being a musician’, or in other words the MIMS full score, significantly improved the model. Put differently, does adding the variable ‘being a musician’ help to better predict depression? The comparison of both models with an ANOVA showed that this was not the case: ∆F(1,86) = 2.50, p = .11. Furthermore, the ‘being a musician’ variable was not significant within the model (p =.1). Moreover, the levels of predictability of the individual variables within the model also changed with the addition. While anxiety and professor/teacher burnout increased in importance (anxiety: b = .58, t(86) = 10.8, p < .001, pr2 = .575; teacher/professor burnout: b = -.43, t(86) = -2.34, p = .02, pr2 = -.06), pain catastrophizing became an insignificant predictor (p =.06). This was plausible, since musicians also showed the lowest levels of pain catastrophizing despite a high level of perceived pain. This then suggests that pain processing happens differently in musicians compared to non-musicians, as indicated above. In summary, the hierarchical linear regression model does not support the hypothesis that ‘being a musician’ is a valid predictor for depression.

Tree model For the tree model we calculated a decision tree to predict depression, using all variables: HADS anxiety, all four sub-scales of the professional identification scale (not the full score), duration and frequency of depersonalization, all three coping scales, pain catastrophising, all burnout scales, practice time (total of years and hours per day), study courses (undergraduate, postgraduate at university/music college), age, family status and considering oneself a musician or not. The results can be seen in fig.1.

12 Fig. 1 Decision tree predicting depression, including probability of variables and the number of participants belonging to each terminal node (bottom grey panel) NOTE: the tree model starts at the top panel (1) and offers 4 different possible outcomes from which only option 4, a combination of anxiety (>7) and teacher/professor burnout (≤ 2.8), predicts a depression score of 8.4, which for this study is below the cut-off (9) for significant depression.

The tree model can be interpreted by starting at the top of the figure, with the first predictor for depression being anxiety (HADS anxiety). Each branch can then be followed down to the next node until the final node is reached, which shows the mean depression score for the branch. The most promising combination of variables can be seen in panel 4 at the bottom of the tree: the combination of anxiety (>7) and teacher/professor burnout (≤ 2.8) predicts a depression score of 8.4, a result that approaches significance (cut-off is 9). This outcome confirms the findings from the hierarchical regression model: low levels of professor/teacher burnout predicted higher levels of depression. The tree model did not find the variable ‘being a musician’ to be a predictor for depression, regardless of the institution of higher education or chosen course of study.

In summary of both results sections, depression prevalence did not differ between student musicians and non-musicians when both musicians’ groups were considered together. However, when participants were separated into three groups there was a significant difference between all groups, with college musicians having the highest depression prevalence and university musicians the lowest depression prevalence. At first glance, this result suggests that professional identification could play a role in predicting depression. However, neither subsequent model found this variable to be a significant predictor. Rather, non-musician-specific variables, such as (general) anxiety and burnout with teaching staff, were reliable depression predictors. The variable teacher/professor burnout was significant in all groups. Despite reporting long durations of perceived pain, scores in pain catastrophizing in both musicians’ groups were low and this factor was insignificant in the model.

Discussion

The aims of the present study were to add to our knowledge of how we may better predict depression in your training musicians by increasing understanding of how known depression predictors (such as anxiety, coping, and pain) relate to depression experiences in populations of young people who are engaging with music at different levels and in different kinds of higher educational institutions (University, Music College). In this study we examined six depression predictors 13 (anxiety, depersonalization, coping strategies, professional identification, pain catastrophizing and burnout). Following group testing and regression modelling of the data, only two of these factors were found to be significant in predicting depression in our population of young people; general anxiety and burnout with teaching staff. These factors are discussed first before considering the reasons for our findings in relation to the other four predictors.

The first significant depression predictor identified by both models was anxiety. It was surprising to find that the musicians’ groups reported (significantly) less anxiety compared to the non-musicians’ group, as this goes against our expectations from the literature (Jabusch & Altenmüller, 2004; Kenny & Ackermann, 2015b; Spahn et al., 2001). However, the fact that despite lower anxiety prevalence, college musicians showed higher depression scores, does not decrease the importance of anxiety as a depression predictor. Instead, this result shows that depression is multifactorial and thus requires a combination of several predictors alongside a simple assessment of anxiety. The high depression scores for the college musicians’ group might have been influenced by stress from one or several other predictor variables that were not accounted for, such as for instance economic status (Mauz & Jacobi, 2008; Miech & Shanahan, 2000), the quality of relationships (Daniel J Buysse, 2008), or institution-based environmental stress factors (discussed below) While the anxiety prevalence of musicians was lower than that of non-musicians, all anxiety scores in this study of young people were high. To put the scores from this study into perspective, our college musicians showed a similar anxiety level to their EU counterparts (Spahn, Strukely, & Lehmann, 2004). However, non- music students showed about 28% - 30% higher anxiety levels compared to their EU counterparts with the same or similar subject (predominantly medicine/psychology) (Prinz, 2012) (Spahn et al., 2004). Indeed, medical students have been found to show a higher prevalence of anxiety and depression compared to students from other faculties, and also compared to non-students (Bacchi & Licinio, 2015; Ibrahim, Kelly, Adams, & Glazebrook, 2013).

There could be several reasons why both groups of musicians in our study reported lower anxiety levels compared to non-musicians. Firstly, there is a possibility of habituation. Musicians could have grown accustomed to stressful situations and developed a better coping routine. Anxiety may only peak in high- stress situations, such as right before a performance, and then descend to a lower anxiety base level after such a situation, a working hypothesis that could be tested in future. This theory is supported by our finding that, musicians, on the whole, showed better active and functional coping strategies compared to non- musicians. Based on this result, it is possible that due to the years of playing successfully in concerts and auditions, musicians evaluate anxiety-inducing situations differently than others, and dispose of greater self-efficacy compared to non-musicians (WHO, 2012), another hypothesis that is suited to further testing. We should also not discount the possibility that making music could also have a long term therapeutic effect on active musicians. Making music, similarly to listening to music, might offer an outlet and the possibility to channel anxiety.

Moving on from anxiety now, to burnout, the second depression predictor variable in our study. Burnout has been found to change previously engaged and motivated individuals into persons who are increasingly distant and depersonalized, who suffer progressively from physical and psychological complaints, and who may ultimately fall into substance abuse and (thoughts of) suicide (Campos et al., 2013; Dahlin & Runeson, 2007; Maslach & Jackson, 2007; 14 Willcock, Daly, Tennant, & Allard, 2004). On the surface, it is easy to confuse burnout with depression as burnout is a syndrome that results from chronic workplace stress that has not been successfully managed. However, in our study, burnout only became a depression predictor when combined with anxiety, and only a low burnout level (≤ 2.8), and not a higher one, predicted depression. This contrasts the data found in reference to depression predictors anxiety and pain catastrophizing, where a higher level predicted higher depression (Gatchel, Peng, Peters, Fuchs, & Turk, 2007; Vlaeyen & Linton, 2000; Wooley, Blackwell, & Winget, 1978). How should this disparity be explained?

Before this can be answered it is vital to note that the WHO is in the process of revising its approach to look at burnout as a condition in its own right (WHO 2019). Burnout is currently not defined as an individual condition in the ICD, since it is described in several codes (i.a. Z73). This circumstance might have contributed to burnout being either confused with depression or seen as a ‘fashionable’ diagnosis, a person might be more comfortable to reveal suffering from due to the stigma attached to depression (Bahlmann, Angermeyer, & Schomerus, 2013). From 2022 the condition will be entered separately into the new ICD-11. This step will distinguish more clearly between burnout and depression, even though, as shown in our study’s results, both share similar predictors, and might mutually cause or reinforce one another. If both conditions, depression and burnout, coincide but do not fully overlap or explain one another, this could explain why a lower level of burnout predicted a higher level of depression.

Our results of high depression and low burnout levels reflected the findings of recent studies with elite college athletes (Frank et al., 2017; Nixdorf, 2018). Based on their findings, the authors caution to not confuse both conditions or use them interchangeably. Instead, they advocate that burnout should be viewed as multifactorial, similarly to depression. Moreover, studies with elite college athletes found stress to be the significant predictor for burnout (Frank et al., 2017) and, more specifically, sports-related stressors were positively linked to burnout (Cresswell & Eklund, 2004). Adolescent college elite athletes’ perceptions of conflicts with coaches have been identified as significant predictors for burnout (DeFreese & Smith, 2013; Smith, Gustafsson, & Hassmén, 2010), and coaches with a disempowering or controlling motivational style increased the likelihood that their athletes experienced burnout (Appleton & Duda, 2016; González, Tomás, Castillo, Duda, & Balaguer, 2017). For musicians, such a disempowering or controlling motivational style, also referred to by some as ‘an imperative or Russian teaching style’, has been found to be used especially for string players (Parncutt & Grazenser, 2006) and could have a similarly negative impact. This might be one explanation why in our study teacher/professor burnout had a greater impact as depression predictor than personal burnout, work burnout or burnout with fellow students. It is also a further indication for the importance of the environment in both burnout and depression. Drawing from the parallels between elite college athletes and musicians in training, these findings and this line of thinking will help as we advance research into musicians experiencing burnout and depression. Taken together these results have two implications on future research in this area: (1) the variability of depression between musician groups, the differentiability of identification with the profession and the high depression scores in non-music students suggest that further interdisciplinary research is crucial. Contrasting student musicians and non-music students not only allows for a better 15 differentiation of depression and depression predicating variables, but also enables to weigh depression factors and isolate music specific factors. (2) The variability between higher education institutions demonstrated their vital role in (music) students experiencing depression. Further research is needed to understand the different characteristics within the various institutions that impact student musicians in terms of depression.

While only two factors, anxiety and burnout, were found to be significant in predicting depression, a third the factor, namely pain catastrophizing, merits separate consideration. Previous studies found that high pain perception inevitably led to an increase in dysfunctional coping with pain, and consequently to an increase of depression (Kenny & Ackermann, 2015b). Following this line of thought, and in keeping with the perceived pain levels and durations of our participants, our study should have found raised levels of dysfunctional coping strategies – and consequently raised depression levels – in both musicians’ groups compared to the levels in the non-musicians’ group. However, this is not what we found – neither for the individual variables (pain catastrophizing, dysfunctional coping strategies, MIMS subscale negativity), nor in the predictive models. The reason for this discrepancy can be found in general, that is non- musician specific, pain and depression research. Depression has indeed been found to increase pain perception (Bär et al., 2005; Miller & Cano, 2009). However, it is crucial to note that literature has only shown this effect when a mild depression coincided with acute pain or at the very onset of chronic pain. If pain is perceived for a longer period time, or in other words is chronic, the variable ‘pain perception’ loses in significance as depression predictor to the same extent that pain processing gains in significance (Boer et al., 2005). As (Steinmetz, 2015) noted, pain research for musicians, when compared to general pain research, is only at the beginning. The choice of overly sensitive instruments for researching pain and depression in musicians have portrayed a picture that general depression research does not support (Altenmüller, 2013). In our study, non-musicians, who mostly reported to have perceived constant pain for less than six months, did indeed show the expected behavior of dysfunctional coping with pain. Moreover, in accordance with literature, their high pain catastrophizing scores also coincided with a higher vulnerability to a further increase in both pain perception and depression (Demyttenaere et al., 2007; Vaccarino, Sills, Evans, & Kalali, 2009). These results show that depression investigations in musicians need to differentiate between perceived pain and pain processing. If, as our results suggest, the perceived pain is not processed in a dysfunctional manner, pain perception also loses its predictive function for depression.

This was the first study to explore whether another important factor, ‘professional identification’, could be considered as a valid predictor for depression in young students. Contrary to expectation our results showed that identification with being a musician was not a significant predictor for depression for our sample, despite the fact that college musicians showed the highest depression prevalence of all groups investigated. Given both musicians’ groups identification scores, and in view of the parallel findings in (elite) college athletes, it is reasonable to assume as a future hypothesis that the subscales self and social identification from the professional identification questionnaire (MIMS) have a neutral relationship with depression in young people (i.e. no predictive power), while exclusivity and negative affectivity had a positive relationship with depression (i.e. higher score in one predicts higher score in the other) (Grove, Lavallee, & Gordon, 2008). Based on this result, we can conclude firstly, that individuals belonging to a certain profession should not automatically be 16 associated with a high (or low) level of depression. In other words, rather than conclude that individuals are attracted towards musicianship because of a possible predisposition, depression research in student musicians should be approached from a different angle, that of understanding the more generalized stressors which may impact on them in the same way as people in many different professions. One such consideration is the type of study environment. The higher depression scores for college musicians compared to university musicians in our study might be explained by a different career focus or institution- specific environmental stressors. Extrapolating from studies with college athletes and professional identification, their identification level was not only based on how much they were immersed in their sports, but also fed by the environment (Grove, FISH, & EKLUND, 2010). However, as (Perkins, Reid, Araújo, Clark, & Williamon, 2017) showed, documentation of mental health in UK music colleges is still in its infancy, which makes it difficult to identify particular stressors and/or depression predictors for this environment. Without speculating, we can only say that stressors for university musicians seem to be fewer compared to those encountered by college musicians and university non-musicians.

It is important to note the limitations of this study, which need to be considered in order to allow a meaningful interpretation of the data. Firstly, despite providing more than the number of participants determined by the a priori power calculation for a minimum sample size required to allow for meaningful results, the overall sample size is modest. This could be addressed in future studies with larger numbers of participants, including a representative sample of music students across all higher education institutions. Secondly, many participants did not want to disclose information about their mental health status. While this is in keeping with the literature (Clement et al., 2015), and while the design of our study endeavored to assure participant anonymity, some students chose not to disclose some information (i.e. HADS). This pattern was most noticeable among music college musicians. It is possible that the estimated depression scores for those who did not complete the HADS, despite being carefully done in keeping with current medical knowledge, might have erred slightly. Thirdly, for anonymity reasons, variables such as gender, socio-economic background or being an international/exchange student, were excluded. The inclusion of these variables might have provided some additional information about particular depression stressors, but on balance anonymity was deemed more important in order to collect data at all. Fourthly, this study was designed based on the empirical findings from previous studies in the area of musicians’ pain perception and depression. We did not anticipate that our findings would contrast most of the current literature for musicians’ depression in these particular aspects. In hindsight, it would have further elucidated our findings, had we included additional instruments for pain processing in the exploratory research design, to help interpret the data. We would suggest that future studies take this into account as they explore this area in more detail.

Still, our findings provide several implications for the field of depression research in musicians. Our study did not find identification with the profession of a musician, or ‘being a musician’ to be a significant variable in predicting depression. Rather than being considered innate to the profession of a classical musician, depression should thus be understood on the basis of its multifactorial model, and the corresponding predictors such as anxiety, burnout and dysfunctional coping strategies, including pain processing. Identifying depression predictors can not only change the self-stigma of musicians (Pryal, 2011), but also has implications for primary depression prevention strategies in higher education. The lower depression scores in university musicians suggest that the 17 environment, the prevention strategies used here or a combination of both, are more effective in preventing depression, than those applying to university non- musicians or in the music college environment. In music college, as shown by (Spahn, Walther, & Nusseck, 2016), multimodal prevention models can be implemented successfully. Moreover, by replicating results from general pain and depression research (Bair, Robinson, Katon, & Kroenke, 2003), this study shows that no causality should be implied between ‘pain perception’ and ‘pain processing’. Instead, these should be investigated separately in depression research. While high pain catastrophizing has been shown to have an impact on depression scores, high pain perception on its own does not (Pfingsten, 2016). Furthermore, the high scores in pain catastrophizing of non-musicians suggest that dysfunctional pain processing is a greater issue for the non-musician population than for musicians. Finally, considering burnout, and drawing from previous research with elite athletes (Nixdorf, 2018), this study suggests that depression and burnout only explain each other to a certain degree. Therefore, they should not be used interchangeably. In view of the place of burnout within the statistical models, it is more likely that burnout acts as a mediator/moderator rather than as a single or individual predictor for depression.

Coming back to the statement from the title of this article, ‘unhappiness is my lot’, it would seem that rather than unhappiness, or depression, being a musician’s’ lot, any experience of this kind is based on general depression predictors and shaped by the environment. From a musical point of view, our results might be less dramatic than those of previous studies, and subsequently challenge some entrenched concepts and myths, such as those surrounding creativity and mental health. However, they provide an insight into depression in musicians in higher education institutions, endeavor to increase knowledge within this field with a view to supporting studies that aim to optimize depression prevention and mental health support.

Author contribution MK led the conception and design of the study, data acquisition, analysis and interpretation. She also drafted the article. DC and RW contributed to the design and conception and to the interpretation of the data, and critically reviewed the article. RT assisted with the data acquisition, data analysis and interpretation, and critically reviewed the article. VJW helped with the conception, design, data acquisition and analysis, as well as the interpretation of the data, and critically reviewed the article. All authors approved the submitted version.

18 References

Altenmüller, E. (2013). Kommentar zur Studie von D. Kenny und B. Ackermann. [A commentary to the study by D. Kenny and B Ackermann] Musikphysiologie Und Musikermedizin, 2(20), 67. Anglada-Tort, M., & Müllensiefen, D. (2017). The Repeated Recording Illusion: The Effects of Extrinsic and Individual Difference Factors on Musical Judgments. Music Perception: an Interdisciplinary Journal, 35(1), 94–117. http://doi.org/10.1525/mp.2017.35.1.94 Angst, J., Gamma, A., Neuenschwander, M., Ajdacic-Gross, V., Eich, D., Rössler, W., & Merikangas, K. R. (2005). Prevalence of mental disorders in the Zurich Cohort Study: a twenty year prospective study. Epidemiologia e Psichiatria Sociale, 14(02), 68–76. Appleton, P. R., & Duda, J. L. (2016). Examining the interactive effects of coach- created empowering and disempowering climate dimensions on athletes’ health and functioning. Psychology of Sport and Exercise, 26, 61–70. http://doi.org/10.1016/j.psychsport.2016.06.007 Armstrong, S., & Oomen-Early, J. (2010). Social Connectedness, Self-Esteem, and Depression Symptomatology Among Collegiate Athletes Versus Nonathletes. Journal of American College Health, 57(5), 521–526. http://doi.org/10.3200/JACH.57.5.521-526 Bacchi, S., & Licinio, J. (2015). Qualitative Literature Review of the Prevalence of Depression in Medical Students Compared to Students in Non-medical Degrees. Academic Psychiatry: the Journal of the American Association of Directors of Psychiatric Residency Training and the Association for Academic Psychiatry, 39(3), 293–299. http://doi.org/10.1007/s40596-014-0241-5 Bahlmann, J., Angermeyer, M., & Schomerus, G. (2013). „Burnout“ statt „Depression“ - eine Strategie zur Vermeidung von Stigma? [Calling it “Burnout” Instead of ‘Depression’ – A Strategy to Avoid Stigma?] Psychiatrische Praxis, 40(02), 78–82. http://doi.org/10.1055/s-0032-1332891 Bair, M. J., Robinson, R. L., Katon, W., & Kroenke, K. (2003). Depression and Pain Comorbidity: A Literature Review. Archives of Internal Medicine, 163(20), 2433–2445. http://doi.org/10.1001/archinte.163.20.2433 Bär, K.-J., Brehm, S., Boettger, M. K., Boettger, S., Wagner, G., & Sauer, H. (2005). Pain perception in major depression depends on pain modality. Pain, 117(1- 2), 97–103. http://doi.org/10.1016/j.pain.2005.05.016 Bertha, E. A., & Balázs, J. (2013). Subthreshold depression in adolescence: a systematic review. European Child & Adolescent Psychiatry. http://doi.org/10.1007/s00787-013-0411-0 Bjelland, I., Dahl, A. A., Haug, T. T., & Neckelmann, D. (2002). The validity of the Hospital Anxiety and Depression Scale: an updated literature review. Journal of Psychosomatic Research, 52(2), 69–77. Boer, den, J. J., Oostendorp, R. A. B., Beems, T., Munneke, M., Oerlemans, M., & Evers, A. W. M. (2005). A systematic review of bio-psychosocial risk factors for an unfavourable outcome after lumbar disc surgery. European Spine Journal, 15(5), 527–536. http://doi.org/10.1007/s00586-005-0910-x Boutolleau, D. (2017). Médecine & Opéra. [Medicine & Opera] DIU Pédagogie Médicale, 1–11. Breiman, L., Friedman, J. H., Olshen, R. H., & Stone, C. J. (1984). Classification and Regression Trees. Boca Raton, Florida: Taylor & Francis. Brewer, B. W., Van Raalte, J. L., & Linder, D. E. (1993). Athletic identity: Hercules’ muscles or Achilles heel? International Journal of Sport Psychology, 24(2), 327–254.

19 Burch, G. S. J., Pavelis, C., Hemsley, D. R., & Corr, P. J. (2010). Schizotypy and creativity in visual artists. British Journal of Psychology, 97(2), 177–190. http://doi.org/10.1348/000712605X60030 Campos, J., Carlotto, M. S., & Marôco, J. (2013). Copenhagen Burnout Inventory- student version: adaptation and transcultural validation for Portugal and Brazil. Psicologia: Reflexão E Crítica. Carson, S. (2014). Leveraging the “mad genius” debate: why we need a neuroscience of creativity and psychopathology. Frontiers in Human Neuroscience, 8(26), 1. http://doi.org/10.3389/fnhum.2014.00771 Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). You want to measure coping but your protocol’ too long: Consider the brief cope. Journal of Personality and Social Psychology, 56(2), 92–100. Caspi, A., Sugden, K., Moffitt, T. E., Taylor, A., Craig, I. W., Harrington, H., et al. (2003). Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science Magazine, 301(5631), 386–389. http://doi.org/10.1126/science.1083968 Chávez-Eakle, R. A. (2006). Creativity and personality. Taylor & Francis. Chest, J. S. (2018). On Doctors and Their Operas: A Critical (and Lyrical) Analysis of Medicine in Opera. Chest, 154(2), 409–415. Clement, S., Schauman, O., Graham, T., Maggioni, F., Evans-Lacko, S., Bezborodovs, N., et al. (2015). What is the impact of mental health-related stigma on help-seeking? A systematic review of quantitative and qualitative studies. Psychological Medicine, 45(1), 11–27. http://doi.org/10.1017/S0033291714000129 Constant, E. (2011). Music, composers and psychopathology: the psychiatrist's view. Bulletin Et Memoires De LAcademie Royale De Medecine De Belgique, 166(1-2), 88–96; discussion 97–8. Cordingly, J. (2001). Disordered Heroes in Opera. (C. Seymore, Ed.). Martlesham, United Kingdom: Boydell and Brewer. http://doi.org/10.7722/j.ctt1gr7ctj Cresswell, S. L., & Eklund, R. C. (2004). The athlete burnout syndrome: possible early signs. Journal of Science and Medicine in Sport, 7(4), 481–487. http://doi.org/10.1016/S1440-2440(04)80267-6 Dahlin, M. E., & Runeson, B. (2007). Burnout and psychiatric morbidity among medical students entering clinical training: a three year prospective questionnaire and interview-based study. BMC Medical Education, 7, 6. http:// doi.org/10.1186/1472-6920-7-6 Daniel J Buysse, J. A. A. G. V. A. D. E. W. R. (2008). Prevalence, Course, and Comorbidity of Insomnia and Depression in Young Adults. Sleep, 31(4), 473. DeFreese, J. D., & Smith, A. L. (2013). Teammate social support, burnout, and self-determined motivation in collegiate athletes. Psychology of Sport and Exercise, 14(2), 258–265. http://doi.org/10.1016/j.psychsport.2012.10.009 Dell, P. F., & O'Neil, J. A. (2010). Dissociation and the Dissociative Disorders. Routledge. Demyttenaere, K., Bruffaerts, R., Lee, S., Posada-Villa, J., Kovess, V., Angermeyer, M. C., et al. (2007). Mental disorders among persons with chronic back or neck pain: Results from the world mental health surveys. Pain, 129(3), 332– 342. http://doi.org/10.1016/j.pain.2007.01.022 Dietrich, A. (2014). The mythconception of the mad genius. Frontiers in Psychology, 5. Dura-Vila, G., & Bentley, D. (2009). Opera and madness: Britten's Peter Grimes--a case study. Medical Humanities, 35(2), 106–109. http://doi.org/10.1136/jmh.2009.002030 Ellis, P. D. (2010). The Essential Guide to Effect Sizes. Cambridge: Cambridge University Press. http://doi.org/10.1017/CBO9780511761676

20 Erfurth, A., & Hoff, P. (2000). Mad scenes in early 19th-century opera. Acta Psychiatrica Scandinavica, 102(4), 310–313. Fan, Z. J., Bonauto, D. K., Foley, M. P., Anderson, N. J., Yragui, N. L., & Silverstein, B. A. (2012). Occupation and the prevalence of current depression and frequent mental distress, WA BRFSS 2006 and 2008. American Journal of Industrial Medicine, 55(10), 893–903. http://doi.org/10.1002/ajim.22094 Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. http://doi.org/10.3758/ BF03193146 Frank, R., Nixdorf, I., & Beckmann, J. (2017). Analyzing the relationship between burnout and depression in junior elite athletes. Journal of Clinical Sport Psychology, 11(4), 287–303. http://doi.org/10.1123/JCSP.2017-0008 Freud, S. (1918). Der Dichter und das Phantasieren.[Creative writers and day- dreaming] (O. Jahraus, Ed.). Reclam Verlag. Gatchel, R. J., Peng, Y. B., Peters, M. L., Fuchs, P. N., & Turk, D. C. (2007). The biopsychosocial approach to chronic pain: scientific advances and future directions. Psychological Bulletin, 133(4), 581–624. http://doi.org/10.1037/0033-2909.133.4.581 Gebauer, L. (2014). Aktuelle nationale und internationale Angebote zur Stressbewältigung für Medizinstudierende. [Domestic and international stress coping management offers for medical students]. PhD thesis in medicine, Justus-Liebig University, Giessen González, L., Tomás, I., Castillo, I., Duda, J. L., & Balaguer, I. (2017). A test of basic psychological needs theory in young soccer players: time‐lagged design at the individual and team levels. Scandinavian Journal of Medicine & Science in Sports, 27(11), 1511–1522. http://doi.org/10.1111/sms.12778 Grove, J. R., FISH, M., & EKLUND, R. C. (2010). Changes in Athletic Identity Following Team Selection: Self-Protection versus Self-Enhancement. Journal of Applied Sport Psychology, 16(1), 75–81. http://doi.org/10.1080/10413200490260062 Grove, J. R., Lavallee, D., & Gordon, S. (2008). Coping with retirement from sport: The influence of athletic identity. Journal of Applied Sport Psychology, 9(2), 191–203. http://doi.org/10.1080/10413209708406481 Hall, J. (2018). Perceptions of Confidentiality and Stigma Associated with use of Counseling Services. Hildebrandt, H. (2002). Musikstudium und Gesundheit. Aufbau und Wirksamkeit eines präventiven Lehrangebotes [Music study and health. Development and efficacy of a preventative study course ](2nd ed., Vol. 1). Bern: Europäischer Verlag der Wissenschaften. Hoerl, A. E., & Kennard, R. W. (1970). Ridge Regression. Technometrics, 12, 55– 67. Holst, G. J., Paarup, H. M., & Baelum, J. (2012). A cross-sectional study of psychosocial work environment and stress in the Danish symphony orchestras. International Archives of Occupational and Environmental Health, 85(6), 639–649. http://doi.org/10.1007/s00420-011-0710-z Hunter, E., Sierra, M., & David, A. S. (2004). The epidemiology of depersonalisation and derealisation. Social Psychiatry and Psychiatric Epidemiology, 39(1), 9–18. Ibrahim, A. K., Kelly, S. J., Adams, C. E., & Glazebrook, C. (2013). A systematic review of studies of depression prevalence in university students. Journal of Psychiatric Research, 47(3), 391–400. http://doi.org/10.1016/j.jpsychires.2012.11.015

21 Jabusch, H. C., & Altenmüller, E. (2004). Anxiety as an aggravating factor during onset of focal dystonia in musicians. Medical Problems of Performing Artists, 19(2), 75–81. Jakubowski, K., Finkel, S., Stewart, L., & Müllensiefen, D. (2017). Dissecting an earworm: Melodic features and song popularity predict involuntary musical imagery. Psychology of Aesthetics, Creativity, and the Arts, 11(2), 122–135. http://doi.org/10.1037/aca0000090 Jamison, K. R. (1995). Manic-Depressive Illness and Creativity, 282(2), 62–67. Janka, Z. (2004). [Artistic creativity and bipolar mood disorder]. Orv Hetil, 145(33), 1709–1718. Johnson, S. L., Murray, G., Fredrickson, B., Youngstrom, E. A., Hinshaw, S., Bass, J. M., et al. (2012). Creativity and bipolar disorder: Touched by fire or burning with questions? Clinical Psychology Review, 32(1), 1–12. http://doi.org/10.1016/j.cpr.2011.10.001 Keers, R., & Pluess, M. (2017). Childhood quality influences genetic sensitivity to environmental influences across adulthood: A life-course Gene × Environment interaction study. Development and Psychopathology, 29(5), 1921–1933. http://doi.org/10.1017/S0954579417001493 Kenny, D., & Ackermann, B. (2015a). Performance-related musculoskeletal pain, depression and music performance anxiety in professional orchestral musicians: a population study. Psychology of Music, 43(1), 43–60. http://doi.org/10.1177/0305735613493953 Kenny, D., & Ackermann, B. (2015b). Performance-related musculoskeletal pain, depression and music performance anxiety in professional orchestral musicians: a population study. Psychology of Music, 43(1), 43–60. Kessler, R. C., Birnbaum, H., Bromet, E., Hwang, I., Sampson, N., & Shahly, V. (2009). Age differences in major depression: results from the National Comorbidity Survey Replication (NCS-R). Psychological Medicine, 40(02), 225– 237. http://doi.org/10.1017/S0033291709990213 Kluge, C. A. F. (1818). Versuch einer Darstellung des animalischen Magnetismus, als Heilmittel [Attempt to portray animal magnetism as a cure] (3rd ed.). Berlin: Realschulbuchhandlung. Kristensen, T. S., Borritz, M., Villadsen, E., & Christensen, K. B. (2005). The Copenhagen Burnout Inventory: A new tool for the assessment of burnout. Work & Stress, 19(3), 192–207. http://doi.org/10.1080/02678370500297720 la Motte-Haber, De, H. (2006). «der glückliche phantasiert nicht»: Melancholie und Musik [A happy individual does not fantasize: melancholia and music]. Neue Zeitschrift Für Musik, 167(6), 12–15. http://doi.org/10.2307/23991615 Lally, P. (2007). Identity and athletic retirement: A prospective study. Psychology of Sport and Exercise, 8(1), 85–99. http://doi.org/10.1016/j.psychsport.2006.03.003 Lally, P. S., & Kerr, G. A. (2013). The Career Planning, Athletic Identity, and Student Role Identity of Intercollegiate Student Athletes. Research Quarterly for Exercise and Sport, 76(3), 275–285. http://doi.org/10.1080/02701367.2005.10599299 Langdon, S. W., & Petracca, G. (2010). Tiny dancer: Body image and dancer identity in female modern dancers. Body Image, 7(4), 360–363. http://doi.org/ 10.1016/j.bodyim.2010.06.005 Langendörfer, F., Hodapp, V., Kreutz, G., & Bongard, S. (2006). Personality and Performance Anxiety Among Professional Orchestra Musicians. Journal of Individual Differences, 27(3), 162–171. http://doi.org/10.1027/1614- 0001.27.3.162 Lauronen, E., Veijola, J., Isohanni, I., Jones, P. B., Nieminen, P., & Isohanni, M. (2004). Links between creativity and mental disorder. Psychiatry, 67(1), 81– 98.

22 Lederman, R. J. (2015). Frederick Delius: controversies regarding his neurological disorder and its impact on his compositional output. Progress in brain research (Vol. 216, pp. 217–232). http://doi.org/10.1016/bs.pbr.2014.11.008 Legge, A. (1988). The Art of Auditioning (Leibzig, London, New York). Editions Peters. Leiner, D. J. (2014). Sosci Survey. Retrieved from http://www.soscisurvey.com Linton, S. J., & Boersma, K. (2003). Early identification of patients at risk of developing a persistent back problem: the predictive validity of the Örebro Musculoskeletal Pain Questionnaire. Clinical Jouranal of Pain, 19(2), 80–86. Liu, X., Buysse, D. J., Gentzler, A. L., Kiss, E., Sleep, L. M., 2007. (2007). Insomnia and hypersomnia associated with depressive phenomenology and comorbidity in childhood depression. Sleep, 30(1), 83–90. Lorusso, L., & Porro, A. (2020). Opera and Neuroscience: A Historical Approach and Its Relevance Today. In B. Colombo (Ed.), Brain and Art (Vol. 259, pp. 81– 94). Mannheim: Springer, Cham. http://doi.org/10.1007/978-3-030-23580-2_8 Lorusso, L., Franchini, A. F., & Porro, A. (2015). Opera and neuroscience. Progress in Brain Research, 216, 389–409. http://doi.org/10.1016/bs.pbr.2014.11.016 Ludwig, A. M. (1992). Creative achievement and psychopathology: Comparison among professions. American Journal of Psychotherapy, 46(3), 330–356. Maslach, C., & Jackson, S. E. (2007). The measurement of experienced burnout. Journal of Organizational Behavior, 2(2), 99–113. Mauz, E., & Jacobi, F. (2008). Mental Disorders and Social Inequalities in Three Birth Cohorts - Psychische Störungen und soziale Ungleichheit im Geburtskohortenvergleich. Psychiatrische Praxis, 35(07), 343–352. http://doi.org/10.1055/s-2008-1067557 Mayberg, H. S., Brannan, S. K., RK, M., Jerabeck, P. A., Brickmann, J. S., Tekell, J. L., et al. (1997). Cingulate function in depression: A potential predictor of treatment response. Clinical Neuroscience and Neuropathology, 8(4), 1057– 1067. Michal, M., Duven, E., Giralt, S., Dreier, M., Müller, K. W., Adler, J., et al. (2015). Prevalence and correlates of depersonalization in students aged 12-18 years in Germany. Social Psychiatry and Psychiatric Epidemiology, 50(6), 995–1003. http://doi.org/10.1007/s00127-014-0957-2 Miech, R. A., & Shanahan, M. J. (2000). Socioeconomic Status and Depression over the Life Course. Journal of Health and Social Behavior, 41(2), (162—176) Miller, L. R., & Cano, A. (2009). Comorbid Chronic Pain and Depression: Who Is at Risk? The Journal of Pain, 10(6), 619–627. Monroe, S. M., & Reid, M. W. (2008). Gene-Environment Interactions in Depression Research: Genetic Polymorphisms and Life-Stress Polyprocedures. Psychological Science, 19(10), 947–956. http://doi.org/10.1111/j.1467- 9280.2008.02181.x Müllensiefen, D. (2010). Statistical techniques in music psychology: An update. In R. Bader, C. Neuhaus, & U. Morgenstern (Eds.), Concepts, Experiments, and Fieldwork Studies in Systematic Musicology and Ethnomusicology (pp. 193– 215). Bern: Citeseer Müllensiefen, D., Gingras, B., Musil, J., & Stewart, L. (2014). The Musicality of Non- Musicians: An Index for Assessing Musical Sophistication in the General Population. PLoS ONE, 9(2), e89642–23. http://doi.org/10.1371/journal.pone.0089642 Myers, R. H. (2004). Huntington’s disease genetics. NeuroRX, 1(2), 255–262. Nixdorf, I. (2018). From a first insight to a deeper knowledge of depression in elite athletes. Nixdorf, I., Frank, R., & Hautzinger, M. (2013). Prevalence of depressive symptoms and correlating variables among German elite athletes. Journal of Clinical Sport Psychology, 7(4), 313–326. http://doi.org/10.1123/jcsp.7.4.313

23 Pallant, J. F., & Tennant, A. (2010). An introduction to the Rasch measurement model: An example using the Hospital Anxiety and Depression Scale (HADS). British Journal of Clinical Psychology, 46(1), 1–18. http://doi.org/10.1348/014466506X96931 Park, S., & Tod, D. (2015). Athletes’ Career Transition Out of Sport. International Review of Sport and Exercise Psychology, 1–49. Parncutt, R., & Grazenser, E. (2006). How do Musicians deal with their Medical problems? Presented at the International conference on Music Percepetion and Cognition. Pawley, A., & Müllensiefen, D. (2012). The Science of Singing Along: A Quantitative Field Study on Sing-along Behavior in the North of England. Music Perception: an Interdisciplinary Journal, 30(2), 129–146. http://doi.org/10.1525/mp.2012.30.2.129 Perkins, R., Reid, H., Araújo, L. S., Clark, T., & Williamon, A. (2017). Perceived Enablers and Barriers to Optimal Health among Music Students: A Qualitative Study in the Music Conservatoire Setting. Frontiers in Psychology, 8, 968. http://doi.org/10.3389/fpsyg.2017.00968 Pfingsten, M. (2016). Risikofaktoren und psycho biologische Mechanismen der Chronifizierung. In Rückenschmerzen und Nackenschmerzen [Risk factors and psycho-biological mechanisms of chronification] (7 ed., Vol. 15, pp. 57–66). Berlin, Heidelberg: Springer, Berlin, Heidelberg. Pridmore, S. A., Auchincloss, S., Soh, N. L., & Walter, G. J. (2013). Four centuries of suicide in opera. The Medical Journal of Australia, 199(11), 783–786. http:// doi.org/10.5694/mja13.10724 Prinz, P. (2012). Burnout, depression and depersonalisation. GMS Zeitschrift Für Medizinische Ausbildung (German Medical Science), 29(1). Proctor, S. L., & Boan-Lenzo, C. (2010). Prevalence of Depressive Symptoms in Male Intercollegiate Student-Athletes and Nonathletes. Journal of Clinical Sport Psychology, 4(3), 204–220. http://doi.org/10.1123/jcsp.4.3.204 Pryal, K. R. G. (2011). The Creativity Mystique and the Rhetoric of Mood Disorders. Disability Studies Quarterly, 31(3), 1–16. R Core Team. (2013). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from http://www.R-project.org/ Rahm, A. (1994). "Un dottor della mia sorte". [A doctor of my standing]} In U. Müller, F. Hundsnurseher, & O. Panagl (Eds.), (pp. 227–234). Presented at the Wort und Musik, Salzburger Akademische Beiträge, Salzburg, Austria. Rasch, B., Friese, M., Hofmann, W. J., & Naumann, E. (2009). Quantitative Methoden 1. Einführung in die Statistik für Psychologen und Sozialwissenschaftler. [Quantitative methods: 1. Introduction to statistics for psychologists and social scientists] Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical Linear Models. SAGE. Ropert, R. (2003). A propos de la folie et l’opéra. [About madness and opera]. Annales Médico-Psychologiques, 16, 213–214. http://doi.org/10.1016/S0003- 4487(03)00047-7 Sadek, N., & Bona, J. R. (2000). Subsyndromal symptomatic depression: a new concept. Depression and Anxiety, 12(1), 30–39. Short, A. K., & Baram, T. Z. (2019). Early-life adversity and neurological disease: age-old questions and novel answers. Nature Reviews Neurology, S1552- 5260, 33491–13. http://doi.org/10.1038/s41582-019-0246-5 Sierra, M., & Berrios, G. E. (2000). The Cambridge Depersonalisation Scale: a new instrument for the measurement of depersonalisation. Psychiatry Research, 93(2), 153–164.

24 Smith, A. L., Gustafsson, H., & Hassmén, P. (2010). Peer motivational climate and burnout perceptions of adolescent athletes. Psychology of Sport and Exercise, 11(6), 453–460. http://doi.org/10.1016/j.psychsport.2010.05.007 Song, Y.-Y., & Lu, Y. (2015). Decision tree methods: applications for classification and prediction. Shanghai Archives of Psychiatry, 27(2), 130–135. http://doi.org/10.11919/j.issn.1002-0829.215044 Spahn, C., Burger, T., Hildebrandt, H., & Seidenglanz, K. (2005). Health locus of control and preventive behaviour among students of music, 33(3), 256–268. Spahn, C., Hildebrandt, H., & Seidenglanz, K. (2001). Effectiveness of a prophylactic course to prevent playing-related health problems of music students. Med Probl Perform Art, 16(1), 24–31. Spahn, C., Strukely, S., & Lehmann, A. (2004). Health conditions, attitudes toward study, and attitudes toward health at the beginning of university study: music students in comparison with other student populations. Med Probl Perform Art, 19(1), 26–33. Spahn, C., Walther, J. C., & Nusseck, M. (2016). The effectiveness of a multimodal concept of audition training for music students in coping with music performance anxiety. Psychology of Music, 44(4), 893–909. http://doi.org/10.1177/0305735615597484 Steinberg, R. (2016). Robert Schumann in the psychiatric hospital at Endenich. Progress in Brain Research, 216, 233–275. Steinmetz, A. (2015). Muskuloskeletale Funktionsstörungen bei professionellen Musikern. [Musculoskeletal dysfunctions in professional musicians].Manuelle Medizin, 53(1), 13–16. Storrie, K., Ahern, K., & Tuckett, A. (2010). A systematic review: Students with mental health problems--a growing problem. International Journal of Nursing Practice, 16(1), 1–6. http://doi.org/10.1111/j.1440-172X.2009.01813.x Thompson, B. (1984). Canonical Correlation Analysis. Newbury Park: SAGE Publications. Vaag, J., Bjørngaard, J. H., & Bjerkeset, O. (2016). Symptoms of anxiety and depression among Norwegian musicians compared to the general workforce. Psychology of Music, 44(2), 234–248. http://doi.org/10.1177/0305735614564910 Vaccarino, A. L., Sills, T. L., Evans, K. R., & Kalali, A. H. (2009). Multiple Pain Complaints in Patients With Major Depressive Disorder. Psychosomatic Medicine, 71(2), 159–162. http://doi.org/10.1097/PSY.0b013e3181906572 Vitale, C. R. (2009). Post-performance mood in classical musicians: Where elation meets depression, PhD dissertation. Boston University, Boston. Vlaeyen, J. W., & Linton, S. J. (2000). Fear-avoidance and its consequences in chronic musculoskeletal pain: a state of the art. Pain, 85(3), 317–332. http://doi.org/10.1016/s0304-3959(99)00242-0 World Health Organization (1992). International Classification of Diseases, 10th Revision. World Health Organization, Geneva World Health Organization. (2012). The global burden of mental disorders and the need for a comprehensive, coordinated response from health and social sectors at the country level. World Health Organization, Geneva. World Health Organization (2019). Burn-out an “occupational phenomenon”: International Classification of Diseases. Retrieved June 4, 2019, from http://www.who.int/mental_health/evidence/burn-out/en/ Wiechman, S. A., & Williams, J. (1997). Relation of athletic identity to injury and mood disturbance. Journal of Sport Behaviour. Willcock, S. M., Daly, M. G., Tennant, C. C., & Allard, B. J. (2004). Burnout and psychiatric morbidity in new medical graduates. The Medical Journal of Australia, 181(7), 357–360.

25 Wittchen, H.-U. U., & Jacobi, F. (2005). Size and burden of mental disorders in Europe---a critical review and appraisal of 27 studies. European Neuropsychopharmacology, 15(4), 357–376. http://doi.org/10.1016/j.euroneuro.2005.04.012 Wittchen, H.-U. U., Jacobi, F., Rehm, J., Gustavsson, A., Svensson, M., Jönsson, B., et al. (2011). The size and burden of mental disorders and other disorders of the brain in Europe 2010. European Neuropsychopharmacology, 21(9), 655– 679. http://doi.org/10.1016/j.euroneuro.2011.07.018 Wooley, S. C., Blackwell, B., & Winget, C. (1978). A learning theory model of chronic illness behavior: theory, treatment, and research. Psychosomatic Medicine. Wray, N. R., Ripke, S., Mattheisen, M., Trzaskowski, M., Byrne, E. M., Abdellaoui, A., et al. (2018). Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nature Genetics, 50(5), 668–681. http://doi.org/10.1038/s41588-018-0090-3 Wulsin, L., Alterman, T., Timothy Bushnell, P., Li, J., & Shen, R. (2014). Prevalence rates for depression by industry: a claims database analysis. Social Psychiatry and Psychiatric Epidemiology, 49(11), 1805–1821. http://doi.org/10.1007/s00127-014-0891-3 Zander, M. F., Voltmer, E., & Spahn, C. (2010). Health promotion and prevention in higher music education. Med Probl Perform Art, 25, 45–65. Zigmond, A. S., & Snaith, R. P. (1983). The hospital anxiety and depression scale. Acta Psychiatrica Scandinavica, 67(6), 361–370.

26