Music and Biomarkers of Stress: A Systematic Review

Objective: The purpose of this paper is to review literature on music and biomarkers of stress in order to (1) Identify music interventions and (2) Detail the biomarkers of stress associated with music. Methods: PRISMA guidelines were followed in performing this systematic review. Studies published from January 1995 to January 2020 that pertain to biomarkers of stress and music were identified through the use of the PubMed database, using the keywords: „music‟ AND „biomarker‟ OR „marker‟ OR „hormone‟. Two authors independently conducted a focused analysis and reached a final consensus on 16 studies that met the specific selection criteria and passed the study quality checks. Results: The reviewed studies were all randomized controlled trials. Reviewed music interventions included Music Listening (ML), Meditational Music (MM), „Guided Imagery and Music‟ (GIM), and Singing. The studies showed that music is associated with a decreasing trend in cortisol, salivary α-amylase, heart rate, and blood pressure, as well as an increasing trend in Immunoglobulin A (IgA), oxytocin, and EEG theta wave, while testosterone was associated with sex-related differences. Conclusion: Music is associated with significant changes in biomarkers of stress, suggesting that it could be utilized for the development of stress reduction tools.


Introduction
Music is the composition of vocal or instrumental patterns to express artists" emotions [1]. Music, fostering a universal celebration of emotion and beauty, is proven to impact approximately 90% of all of the world"s population at an average of over thirty-two hours per week [2]. Music therapy is the practice of treating patients with music (usually at the patient"s discretion); although it is casually understood that music may have a positive effect on one"s mood, there is no standing measure of improvement, nor is there a standard treatment as the norm grants patients to pick their musical treatment [3]. In addition, musicologists understand the melodical aspect of feelings associated with certain progressions, while psychologists understand the scientific aspect of chemical reactions involved in emotion, but the two fields have never actually seen eye-to-eye on music for therapeutic purposes [4]. An important question, addressed by Gerra, et al. [5], focused on whether all genres of music had the same effect. They found that Techno-music was associated with a significant increase in heart rate, systolic blood pressure and in self-rated emotional states. Contrarily, classical music induced an improvement in a patient"s emotional state. Another important theory explored by Fukui and Yamashita [6] was whether the impact music had on hormones was the same between both sexes.
The most common emotional response associated with music, frisson (colloquially known as the "chills"), is widely known yet not quite scientifically understood; over the course of extensive scientific research, this widely known emotional response was recorded as "electrodermal activity and subjective arousal" [7]. Psychological frisson as a result of music has biological implications on heart rate and breathing and mimics the human stress response and the return to homeostasis as a result of stress [7]. The purpose of this systematic review is to examine the scientific literature on music and biomarkers of stress in order to (1) Describe music interventions and (2) Detail the biomarkers of stress associated with music.

Search Strategy
We performed this systematic review in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [8]. A systematic literature search was conducted on articles in the PubMed database that were published within the past 25 years, from January 1995 to January 2020 using the following keywords: "music" AND "biomarker" OR "marker" OR "hormone" and following the inclusion and exclusion criteria detailed below. More studies were added from the reference lists for identified research studies and reviews of music and biological markers.

Study Selection Criteria and Methodology
The following inclusion criteria were used: (a) articles published in English or had a published English translation; (b) articles published in a peer reviewed journal; (c) original studies in human adults (no reviews, no animal studies, no studies age<18). Exclusion criteria included studies using music in the medical illness or settings, reviews, editorials, opinion pieces, and case reports. Two authors independently conducted a focused analysis then together reached a consensus on studies that meet the specific selection criteria. The quality of each study was examined by identifying its strengths and limitations using the criteria adapted from Lohr and Carey by the Agency for Healthcare Research and Quality [9], [10]. Quality aspects assessed include sample size, patient selection methods, bias, study groups comparison, blinding, intervention details, outcome measures, and statistical analysis plans. The search method is displayed in a flow diagram in Figure 1.

Search Results
Our search strategy identified 647 articles. After elimination of the duplicates and irrelevant abstracts, 35 studies were identified to meet the pre-defined selection criteria. Two authors independently conducted a focused analysis of the gathered 35 full-text articles. The two authors then reached a consensus on what studies to include in this review, which yielded 19 studies. The findings from the study quality check method [8,9] eventually led to the exclusion of three studies due to inadequate sample size and statistical reporting [5,11,12] resulting in a final selection of 16 studies.

Overview
All study designs were randomized controlled trials (RCTs). Study samples sizes ranged from 21 to 143 subjects with intervention duration ranging from 5 minutes to 2 hours and an outcome follow-up spanning from 1 day to 12 weeks.

Music Interventions
Reviewed studies included 4 main music interventions: A) Music Listening (ML, 10 studies) is the active or passive action of listening to vocal and/or instrumental sounds combined and organized by using rhythm, melody or harmony [6,13,[19][20][21][22][23][24][25]27]. B) Meditational Music (MM, 3 studies) combines elements of mindfulness meditation, yoga movements and relaxation with music, using repetitive rhythmic sequences [15,18,26]. C) Guided Imagery and Music (GIM, 2 studies) includes relaxation, music listening and imagery where music serves as a catalyst to evoke spontaneous images to come to conscious awareness, while therapist through dialogue provides grounding and focus to the client [14,17]. D) Singing (one study) is initiated by a warm-up phase, which include breathing, stretching, and vocalization exercises, then, sections and pieces are rehearsed by an amateur choir [16].
Cortisol: Three studies showed a statistically significant decrease in cortisol (salivary cortisol) in the Music Listening (ML) intervention when compared to Meditational Music (MM) p=0.022 [19], p=0.003, [24], and p=0.025 [27]. Four additional controlled studies involving the ML intervention also showed a significant decrease in Cortisol levels, with p<0.02, p=0.0029, p=0.0035, and p<0.05 respectively [6,20,21,23]. One study comparing singing and ML [16] found that ML, led to a decrease in Cortisol (p<0.001). Two studies involving comparing Guided Imagery and Music (GIM) to wait-list control group [14,17] revealed a significant decrease in cortisol (p=0.04 and p<0.025 respectively). Lastly, one study involving intervention of MM [18] demonstrated a significant effect of MM on salivary cortisol concentration, with lower post-exposure cortisol concentrations (p=0.023).
Heart Rate: A decrease in heart rate was shown in four studies where participants experienced a lower heart rate while receiving ML compared to control groups (p<0.001, p=0.017, p<0.05, p=0.023, and p<0.05, respectively) [13,19,23,24]. Trappe et al. showed that different types of music all lead to a decrease in heart rate [13], whereas Karageorghis et al. found that faster tempo music actually had the opposite effect with fast, stimulative music inhibiting the return of heart rate toward resting levels [24]. One study found no significant difference in heart rate between the ML group and the control silence group [25]. Thoma et al. showed that heart rate measures did not significantly differ between the ML group, and the MM groups [27].
Blood Pressure: Three ML intervention studies demonstrated a significant decrease in systolic blood pressure [13,23,25] (p<0.001, p<0.05, and p<0.05, respectively). Lai et al. showed a significant decrease in self-perceived stress, which correlated to a decrease in systolic and diastolic blood pressure, (p<0.05) [23]. Trappe et al. also found a decrease in diastolic blood pressure in subjects receiving ML in the form of instrumental Romantic-era music (p=0.003) [13]. One study found no significant effects on systolic or diastolic blood pressure in participants during the recovery process, immediately after strenuous exercise [24].
Salivary -Amylase: One study involving mindfulness-based intervention (a form of MM) showed a significant decrease in salivary α-amylase in the intervention group compared to the waiting list control group (p=0.026) [26]. One study showed that salivary α-amylase reached baseline values faster in the ML intervention group compared to the rest control group (p=0.026) [27].
Testosterone: Three studies investigated how listening to music affected Testosterone plasma levels [6,14,21]. Sex-related differences in Testosterone secretion when listening to music were shown in two studies [6,21]. In 2003, Fukui et al. found that ML intervention decreased testosterone in males (p=.0460) and increased its levels in females (p=.0012) [6]. In 2013, Fukui et al. examined the difference in testosterone levels in men and women during ML. Testosterone levels in women increased with chill-inducing music and decreased with music they disliked, whereas testosterone levels decreased in men with both types [21]. However, one study examining the effects of Guided Imagery and Music (GIM) on biopsychosocial measures of work-related stress, found no significant difference when comparing testosterone levels in the GIM group and the wait list control group (p=0.93) [14].
Oxytocin: Ooshi et al. found that listening to slow-tempo music is accompanied by an increase in oxytocin [19] with a significant Tempo x Time interaction for the oxytocin level (p=0.0014). After the participants listened to the slow-tempo sequence, the study showed that oxytocin level was significantly higher compared to baseline (p=0.0007) [19].
Additional Biomarkers: Studies of other biomarkers, such as androgen receptor (AR) [6], melatonin [14], dopamine, norepinephrine, and epinephrine levels, T-lymphocyte CD4+, CD8+, CD16+ [22] showed inconclusive results, and therefore were not included in this review. Other biomarkers were not covered due to the fact that they were examined in the context of medical illness.

Discussion
In this review of 16 randomized controlled trials that met pre-defined selection criteria and quality check, we identified four music interventions: Music Listening (ML), Meditational Music (MM), "Guided Imagery and Music" (GIM), and singing. The reviewed studies showed music was associated with the following specific biomarkers: cortisol, heart rate, blood pressure, salivary α-amylase, testosterone, IgA, oxytocin, and EEG theta wave. We observed a decreasing trend in cortisol, salivary α-amylase, heart rate, and blood pressure, as well as an increasing trend in Immunoglobulin A (IgA), oxytocin, and EEG theta wave, while testosterone revealed sex-related differences. Music interventions are correlated with changes in biomarkers of stress, suggesting that it could be utilized for the development of stress reduction tools.
The human body"s response to stress involves a significant increase in cortisol levels, heart rate, and blood pressure. Interestingly, the body"s response to stress demonstrates opposite biological trends to those witnessed under musical interventions. In addition to cortisol, biomarkers of chronic stress include high levels of α-amylase, which showed a decrease in response to music. The biological response to music, then, mimic"s the body"s return to homeostasis after undergoing stress. The increases in Oxytocin, IgA, and EEG theta wave come into play as an antistress markers, consistent with prior data. In particular, the increasing trend in IgA levels, strengthens the hypothesis that music could be used as a potential immune system booster. Thus, music serves as a catalyst to stress relief and could lead to potential uses in treating immediate and chronic stress in healthy adults.
Chronic stress many times leads to a sustained increase in the biomarkers aforementioned, resulting in chronic medical conditions, such as anxiety disorders. Acknowledging the inner workings of music in healthy adults and the changes in biomarker"s responses could guide us in applying music, not only as a therapeutic tool but also as a preventative and health-promoting intervention.
The results of the current review were compared to previous reviews. A recent review by Chanda and Levitin [28] showed similar results pertaining to the biomarkers cortisol and oxytocin using similar interventions; however, it did not include information on salivary α-amylase, heart rate, blood pressure, or testosterone. The review included medical patients (surgical and lung infection patients), creating a significant challenge in confidently attributing biomarker changes to music intervention alone, without the confounding effect of medical illness. Finn and Fancourt [29] also demonstrated similar results, specifically in cortisol, among other stress hormones [29]; however, the review did not focus on specific musical interventions and included patients with medical illness [29]. Another recent review by de Witte, et al. [30], although focusing on stress response in relation to music interventions, lacked data on specific biomarkers released in response to those interventions and included clinical settings [30]. Fu, et al. [31], reviewed the effect of perioperative music on stress response to surgery. Despite including a sample of medically ill subjects, the study showed similar results, finding that music attenuated the neuroendocrine cortisol stress response [31]. Mojtabavi,et al. [32] demonstrated similar results in music"s effect on cardiac autonomic function but focused almost exclusively on heart rate [32]. A review by Fancourt, et al. [33] cited Cortisol as the most common biomarker investigated. The review did not utilize strict inclusion criteria for the inclusion of healthy adults [33].
The strengths and limitations of this review are highlighted here. A notable strength is the fact that this review focused primarily on randomized controlled trials (RCTs). Excluding patients with medical illness added the advantage of preventing misinterpretation of biomarkers which are usually influenced by medical illness. Excluding small sample and poorly designed studies ensured the integrity of this systematic review. Thus, our review adds to the literature by covering biomarkers of stress specifically shown in healthy adults and describing each biomarker influenced by music interventions. Limitations include that the majority of the reviewed studies report stressreducing effect with slow tempo or instrumental music, yet the reverse effect with fast tempo music is not consistently observed. In addition to the differences in experimental approaches, individuals" subjective preference for different genres of music might be a complicating factor. In many studies, the type of music selected are slowtempo, instrumental music, which are considered "relaxing" or "comfortable" by the experimenters, whereas in a reallife context, people utilize a variety of genres (some with faster tempos) when they seek stress-reducing or relaxing effects from music. As such, the extent of pleasure derived from listening to the "relaxing" music selected by the researchers may greatly vary. An interesting question to address in this context is whether there is a correlation between the extent of pleasure derived from music intervention and the extent of stress mediation.

Conclusions and Future Directions
Music is associated with significant changes in biomarkers of stress, suggesting that it could be utilized for the development of stress reduction tools. Future work could benefit from distinguishing music"s neurochemical effects from its common use as a simple means of distraction. The Human Computer Interface (HCI) research community has been looking into several potential means including visual, auditory and olfactory stimuli. Current wearable devices that promote well-being, such as Fitbit and Apple Watch, which are mostly designed for monitoring, could present exciting opportunities for the HCI community to implement music interventions. Recently, there has been a surge of research and development on hassle-free wearable stress monitoring platforms which continuously detect users" vital signs and/or stress-related biomarkers. New technology to measure cortisol levels using a wireless sweat sensor would expand the use of monitoring biomarkers of stress in response to music in everyday settings [34]. The delivery of music through digital gadgets could be conveniently coupled with these stress monitoring devices without the introduction of additional burdensome tests. The synergistic effect of stress-monitoring devices and automated music intervention in the foreseeable future will also allow the investigation of music interventions on stress management.