Music and stress: what 100+ trials show

Evidence label: Well studied (music listening and stress measures)

Evidence label: Speculative (claims for specific tunings such as 432 Hz, 528 Hz and the Solfeggio tones)

The first label is for the main finding: many randomised trials exist and have been pooled. It does not mean that every trial was good, that the effects are large, or that music treats anything. The trials measured things such as heart rate and how anxious people said they felt, mostly over a single session. The second covers claims for particular tunings, which none of these trials tested; they are labelled 🟠 Speculative where we discuss them. Nothing in this article is a health claim.

Author: Raymond Hackland · Last reviewed: 21 October 2026 · Reading time: about 10 minutes

In short

What this is: A look at the large body of trials on music and stress: what they measured, how big the differences were, and how sure we can be. It sets the baseline for the rest of our Music & Tone articles, and it explains why “relaxing music” trials say nothing about particular tunings or frequencies.

What the evidence says: A 2020 meta-analysis pooled 104 randomised trials with 9,617 participants. On average, people who listened to music, or took part in music therapy, had somewhat lower heart rate, blood pressure and stress-hormone levels, and reported less anxiety and stress, than people in comparison groups [1]. The differences were small to medium, came mostly from single sessions, often in hospitals, and the self-reported results could not be blinded [1, 3, 5]. None of these trials compared tunings, so they tell us nothing about 432 Hz, 528 Hz or the Solfeggio tones.

What it means for your practice: Listening to music you find calming is a reasonable part of practice. On average, though, the measured change is modest and short-lived. Listening is not a treatment, and the choice of tuning is a musical decision, not a health decision.

What science says

Label for this section: 🟢 Well studied (music listening and stress measures).

What “stress” means in these trials

Researchers can’t measure stress directly, so trials measure things that change with it. Some are physical: heart rate, blood pressure, and hormones such as cortisol, usually measured in saliva or blood. Others are psychological: questionnaires in which people rate how anxious, tense or stressed they feel at that moment. A common one is the state part of the State-Trait Anxiety Inventory (STAI) [1, 3]. When this article says a trial “found less stress”, it means one or more of these measures differed between groups. It does not mean that anyone’s health changed.

The big picture: 104 randomised trials

The largest pooled analysis we found, covering all settings, is by Martina de Witte and colleagues, a Dutch team that included researchers at the University of Amsterdam, published in 2020 [1]. Its search ended in November 2017. The later reviews we found were narrower, mostly about surgery. The authors searched for every randomised trial that tested a music intervention against a comparison group and measured stress or state anxiety in adults. They found 104 trials with 9,617 participants and ran two separate meta-analyses:

• Physical measures (61 trials, 3,188 participants): an average effect of d = 0.38, which the authors call “small-to-medium”. Heart rate showed the clearest difference (d = 0.46), blood pressure (d = 0.34) and hormone levels (d = 0.35) somewhat less [1].
• Psychological measures (79 trials, 6,800 participants): an average effect of d = 0.55, which the authors call “medium-to-large” in their results and “medium” in their discussion. These were self-reported state anxiety and stress [1].

What “d” means. It is the difference between the music group and the comparison group, measured in standard deviations, the usual spread of scores between people. A d of 0.38 means that the two groups’ average scores were a bit more than a third of that spread apart. That is a real, noticeable difference on average, but it leaves plenty of overlap: many people in the comparison groups did as well as many people in the music groups.

The authors checked for publication bias (missing negative studies) and found no sign of it [1]. They also found that results varied between trials more than chance would explain, which is why they looked for factors that might explain the differences.

What made a difference, and what didn’t

De Witte and colleagues tested many possible explanations for why some trials found bigger effects than others [1]. Most made no clear difference:

• No significant difference between music therapy with a trained therapist and simpler music listening; between live and recorded music; between music chosen by the participant and music chosen by the researcher; between “relaxation” music and the participant’s own style; between music with and without lyrics; between medical and non-medical settings; and between studies the authors rated as strong, moderate or weak in quality (EPHPP tool, not GRADE) [1].
• Trends that were not statistically significant: for the self-reported measures, slow music of 60–80 beats per minute tended to show larger effects than music with another or unstated tempo (p = 0.064), and a single session tended to show larger effects than several (p = 0.078) [1]. The authors note that some of these groups contained only a few studies, which limits what can be concluded [1]. Tempo and breathing are the subject of a later article in this series (coming soon).
• What wasn’t measured: the authors write that musical features such as “timbre, melodic and harmonic aspects, and rhythmic accentuation” were not reported in the included trials [1]. Tuning and frequency were not among the things analysed at all.

Who took part, and where

The setting matters for what these numbers mean. Most trials were run in hospitals and clinics, before, during or after surgery or other medical procedures. Only 10 of the 61 trials of physical measures, and 12 of the 79 trials of self-reported measures, were in non-medical settings such as workplaces or universities [1]. People under 18 and people with dementia were excluded. In nearly all trials, people listened to music alone, for example through headphones [1]. And the results come “mainly” from short interventions, “most of the time on a single occasion”; the authors note that the effects of long-term music interventions “have been seldom examined” [1].

One trial up close: music before a stress test

Pooled numbers can hide how mixed single studies are. In a carefully designed trial at the University of Zurich, 60 healthy young women were randomly assigned to one of three conditions for ten minutes before a standard laboratory stress test: they listened to a calm choral piece (Allegri’s Miserere), listened to the sound of rippling water, or rested in silence [2]. The stress test was a mock job interview and mental arithmetic in front of an audience.

The results did not line up neatly [2]:

• Heart rate and self-reported stress and anxiety did not differ between the three groups.
• Salivary alpha-amylase, a marker linked to the nervous system’s stress response, was back at its starting level sooner after music: by the saliva sample taken 15 minutes after the stress test ended (25 minutes after it began), compared with 30 minutes after it ended (40 minutes after it began) after silence or water sounds. Its overall course over time did not differ significantly between the groups (p = 0.23); only this recovery measure did (p = 0.022).
• Cortisol, once the authors adjusted for differences between the women, rose most in the music group and least in the water-sound group.

The authors concluded that music “predominantly affected the autonomic nervous system (in terms of a faster recovery)” [2]. It is a good reminder that “music lowers stress” is a summary of many trials, and that a single well-run study can find some measures moving one way, some the other, and some not at all.

Before surgery: what a Cochrane review found

A Cochrane review looked at one specific situation: anxiety in people waiting for surgery [3]. It included 26 trials with 2,051 participants; in all of them, people listened to recorded music in addition to standard care, and they were compared with standard care alone. On the state anxiety questionnaire, the music groups scored on average 5.72 points lower (95% confidence interval 4.17 to 7.27 points lower). The review also found a small difference in heart rate and diastolic blood pressure, but no support for differences in systolic blood pressure, breathing rate or skin temperature. No adverse effects were identified [3].

The reviewers themselves are careful: music listening “may have a beneficial effect” on anxiety before surgery, and because most trials were at high risk of bias, mainly because of lack of blinding, “these results need to be interpreted with caution” [3]. The review’s searches of most databases ran to August 2012 (one database, AMED, only to April 2011).

Music therapy: bigger numbers, weaker designs

A second meta-analysis by the same Dutch group looked only at music therapy, delivered by a trained music therapist: 47 studies with 2,747 participants [4]. The average effect was larger, d = 0.72. But the effects were also larger in non-randomised studies than in randomised ones, and larger when the comparison was a waiting list rather than usual care or another intervention the authors classed as stress-reducing [4]. Weaker designs and weaker comparisons can make effects look bigger, which is one reason why the more modest results from randomised trials against active or usual-care comparisons are the better guide.

How certain is all this?

Several things limit how far these results go:

• Blinding is impossible for the listener. You know whether you are hearing music. A large study of 146 meta-analyses found that, for subjective outcomes such as questionnaires, unblinded trials tend to exaggerate effects; for objective outcomes there was little evidence of this bias [5]. The self-reported results above are the subjective ones. The physical measures are less exposed to this, and their effects are smaller.
• Compared with what? Most comparison groups received usual care, meaning regular medical care without music; only a minority of trials compared music with another activity [1]. It is our own reasoning, not a finding of these trials, that part of the difference may come from having something pleasant to do rather than from music as such; the trials can’t separate the two.
• Healthy people, outside the clinic. A 2022 meta-analysis looked only at healthy adults who listened to music after a laboratory stress test, outside medical settings. Across 14 experiments with 706 people, it found no significant overall effect on recovery compared with silence or another sound (g = 0.15), and the results varied widely between studies [6].
• Short term. Most trials measured the effect of one session, often minutes after it ended. They say little about whether regular listening changes anything over months [1].
• No formal certainty rating. De Witte and colleagues rated the quality of individual studies but did not grade the certainty of the evidence with GRADE [1]. The Cochrane review’s abstract gives no GRADE certainty rating either, and it points to a high risk of bias [3]. For how these ratings work, see Study types explained: from anecdote to meta-analysis.

So why “Well studied”? Because the question has been tested in more than a hundred randomised trials, they have been pooled several times, the direction of the results is consistent, and there is no sign that negative trials went missing [1, 4]. What is well studied is a modest, short-term difference in stress measures during or after listening, mostly in medical settings. That is narrower than it sounds in headlines. For healthy people recovering from a laboratory stress test, the pooled result is not significant [6].

What tradition says

Label for this section: 🔵 Tradition only.

Long before trials, Chinese thinkers linked music with the inner state. Early texts paired the five tones of the Chinese scale with the seasons, the Five Phases and the organs, as a system of correspondences rather than measured effects. Our article The Chinese five tones and the Five Phases traces these texts, and it covers the modern “five-element music therapy” trials separately, with their own, lower label. These correspondences belong to tradition. The trials in this article neither confirm nor refute them, because they did not test them.

Where popular claims go beyond the evidence

Label for this section: 🟠 Speculative.

“Science proves music cures stress.” The trials measured short-term differences in heart rate, blood pressure, hormones and self-reported anxiety, mostly after one session. Stress is not a disease that music “cures”, and the trials were not designed to show lasting change [1].

“This tuning is the most relaxing: 432 Hz, 528 Hz, the Solfeggio tones.” None of the 104 trials in the main meta-analysis compared tunings, and the authors did not analyse frequency at all [1]. The few studies that did compare tunings are small and mixed; we cover them in 432 Hz vs 440 Hz: history, myths and evidence and in Solfeggio frequencies: where they come from and what the evidence says, which also covers the 528 Hz studies. A product that cites “music reduces stress” research to sell a particular frequency is borrowing evidence that was never about frequency.

“This one study showed cortisol dropped, so music works.” Single studies point in different directions. In one careful trial, cortisol actually rose most in the music group, after the authors adjusted for differences between the women [2]. The pooled result is the better guide, and it is modest [1].

“Music works as well as medication.” One trial in the Cochrane review compared music with a sedative before surgery [3]. That is one trial, in one medical situation, and no basis for replacing any treatment. We make no such claim. If anxiety or stress affects your health, talk to a doctor.

“The more you listen, the better.” The trials mostly tested one session, and the effects of long-term listening have “seldom” been studied [1]. More is not known to be better.

How this relates to the Six Healing Sounds, Qigong and neigong

Label for this section: 🟢 Well studied (music listening and stress measures) · 🔵 Tradition only (qi and the organ pairings of the six sounds) · 🟠 Speculative (any health effect of a particular tuning).

Resonance Heritage Qi is a music-assisted practice, so it is fair to ask what the music adds on its own. Here is what this research does and doesn’t tell us.

• What it supports: listening to calm music is, on average, linked with modestly lower heart rate and less self-reported anxiety in the short term [1, 3]. That fits with using music to settle into practice. It is a description of averages, not a promise for any one session.
• What it doesn’t cover: in the Six Healing Sounds you are not only listening. You breathe slowly, voice or whisper a sound on a long out-breath, and move gently. None of the trials tested that combination, and no study has tested my practice. Slow breathing has its own research, covered in Slow breathing and heart rate variability. Making sound yourself is also different from listening: our article on humming and nitric oxide in the nose describes one measured effect of humming, which is about a gas in the nose, not about stress.
• Tuning: on the main site, each of the six sounds is paired with a Solfeggio tone, described there as “an invitation to resonance, not a health claim”. The research in this article is consistent with that wording. The tones are a musical choice. Any claim that a particular frequency acts on the body, an organ or qi is 🟠 Speculative (see the Solfeggio article).
• Qi and the organs: the traditional idea that each sound relates to an organ is 🔵 Tradition only (see Liu Zi Jue: where the Six Healing Sounds come from). Stress trials don’t test it.

If you have a health condition, please read Who should be careful with breathing practices before you start.

Resonance Heritage Qi is a personal, music-assisted practice inspired by Liu Zi Jue: connected to it, not a replacement for the traditional form or for study with a teacher of that tradition. No study has tested it, and we make no claims about its effects.

Enjoyed this article? Get my free practice guide to the Six Healing Sounds. Just sign up below, confirm your email, and it’s on its way.

Sources

1. de Witte M, Spruit A, van Hooren S, Moonen X, Stams GJ, 2020. Effects of music interventions on stress-related outcomes: a systematic review and two meta-analyses. Health Psychology Review 14(2):294–324. https://doi.org/10.1080/17437199.2019.1627897
2. Thoma MV, La Marca R, Brönnimann R, Finkel L, Ehlert U, Nater UM, 2013. The effect of music on the human stress response. PLoS One 8(8):e70156. https://doi.org/10.1371/journal.pone.0070156
3. Bradt J, Dileo C, Shim M, 2013. Music interventions for preoperative anxiety. Cochrane Database of Systematic Reviews 6:CD006908. https://doi.org/10.1002/14651858.CD006908.pub2
4. de Witte M, Pinho ADS, Stams GJ, Moonen X, Bos AER, van Hooren S, 2022. Music therapy for stress reduction: a systematic review and meta-analysis. Health Psychology Review 16(1):134–159. https://doi.org/10.1080/17437199.2020.1846580
5. Wood L, Egger M, Gluud LL, Schulz KF, Jüni P, Altman DG, Gluud C, Martin RM, Wood AJ, Sterne JA, 2008. Empirical evidence of bias in treatment effect estimates in controlled trials with different interventions and outcomes: meta-epidemiological study. BMJ 336(7644):601–605. https://doi.org/10.1136/bmj.39465.451748.AD
6. Adiasto K, Beckers DGJ, van Hooff MLM, Roelofs K, Geurts SAE, 2022. Music listening and stress recovery in healthy individuals: a systematic review with meta-analysis of experimental studies. PLoS One 17(6):e0270031. https://doi.org/10.1371/journal.pone.0270031

Source quality note:

• [1] and [4] are peer-reviewed meta-analyses by the same research group. We read both in full, in the versions on the University of Amsterdam’s repository (open access, CC BY-NC-ND). The search for [4] ran to 8 May 2019, and it rated study quality with the EPHPP quality assessment tool, not GRADE.
• [2] is a single randomised laboratory trial, read in full (open access).
• [3] is a Cochrane systematic review, read as its published abstract; its most recent version is from 2013.
• [5] is a large methodological study (146 meta-analyses), also cited in Study types explained.
• [6] is a peer-reviewed meta-analysis of 14 laboratory experiments in healthy adults, read in full (open access). Its search was run in April 2021; the authors rated all included studies as of moderate quality.
• All six were checked in their PubMed records, and all DOIs resolve. None studied qigong, the Six Healing Sounds or particular tunings, and this article makes no health claims.
• The quotation “an invitation to resonance, not a health claim” is from the main Resonance Heritage Qi site, as quoted in our Solfeggio article.

Disclaimer

This article is for general education only. It is not medical advice and does not replace diagnosis or treatment by a doctor or other qualified health professional. Qigong and the Six Healing Sounds are traditional practices; we make no claims that they cure, treat or prevent any illness. If you have a health condition, are pregnant, or are unsure whether a practice suits you, please talk to your doctor first.