Key Takeaways
- Higher environmental and headphone sound levels were associated with lower heart rate variability in adults participating in the Apple Hearing Study.
- In the environmental-noise models, an increase from 40 to 50 dBA was associated with an estimated 6.8% reduction in SDNN across the 20-minute exposure window and a 16.0% reduction across the 160-minute window.
- For a 10-dB increase in modeled headphone sound level, researchers estimated SDNN reductions of 7.1% in the short-term analysis and 7.4% in the longer analysis.
- These percentages are population-level model estimates. They do not mean every person’s HRV will fall by the same amount when exposed to louder sound.
- The study measured a short-term physiological marker, not heart attacks, strokes, or cardiovascular disease. It cannot establish that noise or headphone use causes heart disease.
Introduction
The sounds people encounter during everyday life may be accompanied by measurable changes in how the nervous system regulates the heart.
Researchers analyzing data from the Apple Hearing Study found that higher environmental and headphone sound levels were associated with lower heart rate variability, or HRV, in subsequent measurements (Zhang et al., 2026).
The study, published September 4, 2026, in the Journal of Exposure Science & Environmental Epidemiology, used sound and heart measurements collected through Apple devices while participants went about their normal lives.
The findings are intriguing because lower HRV has been associated in other research with cardiovascular risk. But that connection requires careful interpretation.
Heart rate and heart rate variability are not the same thing. Heart rate measures how fast the heart beats, while HRV measures how much the time between individual beats varies. In general, greater HRV reflects greater autonomic flexibility, while lower HRV can occur during physiological strain and has been associated with poorer cardiovascular health in some studies. However, a temporary drop in HRV does not by itself mean that the heart has been harmed.
The new study did not measure cardiovascular disease. It also cannot establish that noise caused the observed HRV changes or that temporary changes in HRV will translate into future heart problems.
What it does provide is a large-scale look at how real-world sound exposure during preceding minutes and hours was associated with subsequent measurements of a physiological marker related to autonomic nervous system activity.
What the Study Examined
Xin Zhang and colleagues at the University of Michigan analyzed observational data from consenting adults enrolled in the nationwide Apple Hearing Study between November 2019 and December 2024.
Participants had agreed to share environmental or headphone sound measurements and HRV data collected through compatible Apple devices. They also completed demographic and hearing questionnaires.
To be eligible for the analysis, participants had to contribute at least 20 consecutive SDNN measurements.
SDNN stands for the standard deviation of normal-to-normal heartbeat intervals. It is a time-domain measure of variation in the intervals between normal heartbeats.
In this study, SDNN was derived from one-minute wearable recordings. It should therefore be interpreted as a short-term physiological marker rather than a standalone diagnosis of autonomic dysfunction or cardiovascular disease.
The researchers created four analytical cohorts based on two factors: the source of sound and the length of the exposure window.
Environmental sound level was examined over 20-minute and 160-minute windows. Headphone sound level was analyzed over the same two time frames.
Because of computational demands and concerns about demographic representation, the researchers did not simply analyze every eligible participant. Instead, they divided participants into groups based on age, sex assigned at birth, and ethnicity, then randomly sampled 20 people from each eligible group.
The environmental-noise analyses each included 1,180 participants. They were drawn from 49,983 people eligible for the 20-minute analysis and 47,339 eligible for the 160-minute analysis.
The headphone analyses included 1,080 participants in the 20-minute model and 980 in the 160-minute model. Those samples came from eligible pools of 19,006 and 10,726 participants, respectively.
How Sound and Heart Rate Variability Were Measured
Environmental sound levels were collected through the Noise app on Apple Watch during participants’ normal daily activities.
The researchers treated these ambient measurements as environmental noise, although they acknowledged that real-world environmental sound can include both wanted and unwanted sounds.
Headphone sound levels came from Apple’s Health app while participants were using headphones to listen to material such as music, podcasts, or videos.
The distinction mattered because environmental noise may often be involuntary, while headphone listening is more commonly self-selected. Perception, annoyance, and control over sound could potentially influence physiological responses.
Heart rate variability was measured using one-minute recordings from Apple Watch. According to the paper, these SDNN measurements were generally collected every two to four hours while the participant was still.
The models estimated how sound exposure during preceding minutes and hours was associated with subsequent one-minute SDNN measurements.
The statistical analyses accounted for repeated measurements from the same person and adjusted for time of day and physical activity.
The researchers also investigated whether the association differed according to factors including age, sex, body mass index, perceived stress, self-rated hearing ability, and tinnitus.
What Researchers Found
Higher environmental sound levels were generally associated with lower SDNN.
In the short-term environmental-noise model, increasing the modeled sound level from the 40 dBA reference to 50 dBA was associated with an estimated overall 6.8% reduction in SDNN across the 20-minute exposure window. The 95% credible interval ranged from a 5.1% to an 8.4% reduction.
The timing of the modeled association was not completely uniform.
During approximately the first six minutes, the model predicted small increases in SDNN. Little meaningful change was observed around seven to eight minutes, followed by predicted reductions during the later portion of the 20-minute window.
The longer environmental-noise analysis produced a stronger overall association.
At 50 dBA compared with the 40 dBA reference, the model estimated that SDNN was 16.0% lower across the 160-minute exposure window. The 95% credible interval ranged from a 13.5% to an 18.3% reduction.
In the longer model, predicted reductions were more consistent throughout the exposure window.
Higher headphone sound levels were also associated with lower SDNN.
For a 10-dB increase in modeled headphone sound level, researchers estimated an overall 7.1% reduction in SDNN in the short-term analysis and a 7.4% reduction in the longer analysis.
Unlike the environmental-noise results, the magnitude of the estimated headphone association changed little between the two exposure windows.
These percentages are population-level model estimates. They should not be interpreted as meaning that every person’s HRV will fall by the same percentage whenever the sound around them or in their headphones becomes 10 dB louder.
Some Groups Showed Larger Changes
The modeled relationship between sound level and SDNN also varied between participants.
Older adults generally showed larger changes in response to environmental noise than younger adults.
For example, in the longer environmental-noise model at 60 dBA, the estimated mean SDNN change was a 20.9% reduction among participants aged 18 to 25 and a 31.3% reduction among those aged 65 or older.
Participants reporting tinnitus also tended to have somewhat larger estimated reductions at higher environmental sound levels than participants who reported never experiencing tinnitus.
Hearing ability produced a more complicated pattern.
At high environmental sound levels, participants who rated their hearing as excellent showed larger modeled SDNN reductions than those reporting poor hearing. With headphone sound, the direction was reversed, with the largest estimated reductions occurring among participants reporting poor hearing.
The authors proposed several possible explanations, including differences in sound perception and listening effort, but these mechanisms were not directly tested.
These subgroup results require particular caution.
The individual characteristics included in the secondary models explained less than 20% of the variation between participants. The authors also noted that uncertainty from the first-stage models was not fully carried into the subgroup analyses, meaning the reported intervals may be too narrow.
No significance testing was performed for these comparisons. They are therefore better viewed as exploratory patterns than firm evidence that particular groups are more susceptible.
How the Findings Fit With Earlier Research
The results are consistent with previous research linking noise exposure to short-term changes in autonomic cardiovascular activity.
In a much smaller 2016 study involving 10 healthy men, researchers examined cardiovascular responses during repeated 40-minute periods involving no noise, low-frequency noise, or high-frequency noise. Low-frequency noise exposure was associated with a 16% reduction in SDNN compared with no noise (Walker et al., 2016).
That experiment was extremely small and involved controlled exposures rather than the varied sounds people encounter during everyday life.
Other observational research has found that lower HRV can be associated with long-term cardiovascular outcomes.
In the Atherosclerosis Risk in Communities study, lower HRV measured from short rhythm strips was associated with higher subsequent risks of coronary heart disease and death from several causes among middle-aged adults (Dekker et al., 2000).
That does not mean the temporary SDNN reductions observed in the Apple Hearing Study carry the same risk.
A short-term physiological change associated with recent sound exposure and persistently low HRV used as a cardiovascular risk marker are not interchangeable findings.
Evidence concerning environmental noise and cardiovascular disease also extends beyond HRV. A 2025 systematic review and meta-analysis revisiting evidence used for World Health Organization environmental-noise guidance found evidence of associations between transportation noise and ischemic heart disease, although estimates varied according to factors such as noise source and study methods (Minkin et al., 2025).
The new Apple Hearing Study contributes a different type of evidence. Instead of relying primarily on estimates of long-term residential noise exposure, it paired sound measurements collected during everyday life with repeated HRV measurements from wearable devices.
What the Results May Mean
One plausible explanation is that higher sound exposure is associated with short-term changes in autonomic nervous system activity.
The autonomic nervous system helps regulate involuntary functions, including heart rate and cardiovascular responses to physical and psychological demands.
HRV reflects variation in the intervals between heartbeats and is influenced by this regulatory system.
Environmental noise could potentially produce physiological arousal through pathways involving stress, disturbance, or annoyance. The headphone findings are particularly interesting because headphone listening is often voluntary, meaning annoyance alone may not explain the observed association.
However, the study did not directly establish the biological mechanism responsible for the changes.
It also does not show that reducing headphone volume would improve cardiovascular health. Testing that question would require a different type of study, ideally an intervention that changes sound exposure and measures clinically meaningful health outcomes.
The study also does not establish a safe or unsafe headphone-volume threshold for an individual listener. Its findings should not be used to diagnose cardiovascular problems from a person’s wearable HRV readings.
Important Limitations
Several limitations substantially affect how these findings should be interpreted.
First, the study was observational.
Participants were not randomly assigned to different sound levels. Although the researchers examined the timing of exposure relative to subsequent SDNN measurements, the design cannot establish that sound itself caused the changes.
The hierarchical models helped account for stable differences between people, including some unmeasured characteristics, and the researchers adjusted for physical activity and time of day.
However, short-term factors such as caffeine, alcohol, sleep, emotional state, or other changing behaviors could still influence HRV and potentially confound the observed associations.
Second, both exposure and outcome measurements came from consumer wearable devices rather than individually calibrated research equipment.
Previous validation research cited by the authors found that Apple Watch SDNN measurements could be systematically lower than reference measurements by as much as 8.31 milliseconds.
The authors also reported that the Apple Watch Noise app underestimated sound levels by 2.1 and 3.2 dB under controlled conditions in a previous validation study.
Other artifacts could potentially push sound measurements upward. Clothing rubbing against the microphone or water on the watch, for example, could affect readings.
Third, SDNN was the only HRV measure available for the analysis. Other HRV measurements can provide additional information about autonomic activity.
The investigators also did not directly include heart rate as a separate variable in their principal model. Instead, they partially accounted for heart rate through active-energy expenditure, which itself incorporates heart-rate information.
Fourth, participants were Apple device users who voluntarily joined the Apple Hearing Study and agreed to share their data.
That population is not necessarily representative of the United States as a whole.
Some demographic groups were also excluded from the primary stratified sample because fewer than 20 eligible participants were available in those strata. This mainly affected some older groups.
Finally, the outcome was a physiological marker rather than disease.
The study provides no evidence that participants with larger noise-associated SDNN reductions subsequently experienced more heart attacks, strokes, arrhythmias, cardiovascular deaths, or other cardiovascular events.
Apple Funded the Study
Apple Inc. funded the study through a contract with the University of Michigan.
According to the published paper, Apple had no role in the study design or conduct, data collection, analysis or interpretation, the decision to submit the manuscript for publication, or manuscript preparation, review, or approval.
The authors also disclosed that salaries for Xin Zhang, Lauren M. Smith, and Richard L. Neitzel were supported by the study funding Apple provided to the University of Michigan.
The authors reported no other conflicts of interest for those researchers, while Sung Kyun Park declared no potential competing interests.
The financial relationship does not determine whether the findings are valid, but it is relevant context when evaluating research involving Apple devices and data.
Final Thoughts
The study provides an unusually detailed real-world look at how environmental and headphone sound levels are associated with short-term changes in heart rate variability.
Higher environmental sound levels were associated with lower SDNN, with a larger modeled reduction when exposure was examined across 160 minutes rather than 20 minutes. Higher headphone sound levels were also associated with lower SDNN, although the estimated association changed little between the two exposure windows.
Repeated measurements from wearable devices allowed researchers to investigate patterns that would be difficult to capture at this scale in a laboratory.
At the same time, the results should not be stretched beyond what was actually measured.
The study does not show that headphones damage the heart. It does not demonstrate that everyday noise causes cardiovascular disease. And it cannot tell us whether the short-term HRV changes observed in this analysis have lasting clinical consequences.
The central finding is narrower: among selected Apple Hearing Study participants, higher environmental and headphone sound levels during preceding exposure periods were associated with lower SDNN in subsequent measurements.
Whether repeated changes of this kind contribute meaningfully to long-term cardiovascular risk remains an important question for future research.
References
Dekker, J. M., Crow, R. S., Folsom, A. R., Hannan, P. J., Liao, D., Swenne, C. A., & Schouten, E. G. (2000). Low heart rate variability in a 2-minute rhythm strip predicts risk of coronary heart disease and mortality from several causes: The ARIC study. Circulation, 102(11), 1239–1244. https://doi.org/10.1161/01.CIR.102.11.1239
Minkin, M., Woodland, L., Williams, O. A., Hamilton, S., Hansell, A. L., Vienneau, D., Gong, X., & Fenech, B. (2025). Revisiting the association between transportation noise and heart disease reported in the World Health Organization Environmental Noise Guidelines for the European Region: A systematic review and meta-analysis. Environment International, 202, 109667. https://doi.org/10.1016/j.envint.2025.109667
Walker, E. D., Brammer, A., Cherniack, M. G., Laden, F., & Cavallari, J. M. (2016). Cardiovascular and stress responses to short-term noise exposures: A panel study in healthy males. Environmental Research, 150, 391–397. https://doi.org/10.1016/j.envres.2016.06.016
Zhang, X., Park, S. K., Smith, L. M., & Neitzel, R. L. (2026). Association between environmental and headphone noise and heart rate variability: Observations from Apple Hearing Study cohort. Journal of Exposure Science & Environmental Epidemiology. https://doi.org/10.1038/s41370-026-00970-8
