By Don Moxley, Co-Founder and Chief Science Officer, Mode + Method
Sound is an autonomic input, not just an auditory one. In a 2026 analysis of Apple Hearing Study data, every 10-decibel rise in environmental noise tracked with a 6.8% drop in HRV over a 20-minute window and a 16.0% drop over 160 minutes — well below any volume that would damage hearing.
Key Takeaways
- Noise reaches your stress response before it reaches your judgment. The pathway runs through the amygdala and the HPA axis, releasing catecholamines and cortisol (Münzel et al., 2025, EXCLI Journal).
- The effect is dose-dependent and it lingers. Per 10 dB, modeled SDNN reductions ran 6.8% at 20 minutes and 16.0% at 160 minutes; at 60 dBA the long-term figure reached 27.4% (Zhang et al., 2026, JESEE).
- Headphones behave differently than traffic. Headphone sound showed 7.1% and 7.4% reductions — the effect barely grew with duration, unlike environmental noise.
- Your first reaction is the opposite of your lasting one. SDNN rose slightly for the first three to six minutes of louder sound, then declined for the rest of the window.
- Noise explains a small slice of the variance. In these models, noise alone accounted for under 2% of SDNN variation, while individual differences accounted for 33–40%. That is an argument for tracking yourself, not for ignoring sound.
- Birdsong helps and traffic quietly cancels part of it. HRV was significantly lower when traffic noise was mixed into natural soundscapes, at volumes below WHO daytime guidance (Uebel et al., 2026, People and Nature).
The Input You Never Scheduled
You schedule training. You defend sleep, or at least feel guilty when you don't. Both show up in your numbers and you know it.
Nobody schedules a noise session. It just arrives — the commute, the open office, the shop floor, the volume you chose for a podcast on a run. Because you never chose to do it, it doesn't feel like a thing that was done to you.
Your autonomic nervous system does not share that distinction. It registers inputs, not intentions.
Why Does Noise Affect the Nervous System?
Because your auditory system evaluates unexpected or rising sound as a possible threat before any deliberate reasoning gets involved. That evaluation is fast, and it runs through the same hardware as every other stress response you have.
HRV (heart rate variability): the beat-to-beat variation in the timing between heartbeats, reflecting the balance between your sympathetic branch (mobilization) and parasympathetic branch (digestion, repair, recovery).
SDNN: the standard deviation of intervals between normal heartbeats. A global measure of autonomic regulation, and the metric the Apple Watch reports.
Thomas Münzel and colleagues at University Medical Center Mainz laid out the pathway in a 2025 review in EXCLI Journal, drawing on decades of human and animal research on transportation noise. Noise activates central stress pathways via the amygdala, hippocampus and HPA axis, driving sympathetic activation and the release of catecholamines and cortisol. Those mediators promote systemic inflammation and oxidative stress, largely through NADPH oxidase 2 activation and nitric oxide synthase uncoupling, and elevated reactive oxygen species then drive endothelial dysfunction and blood pressure increases.
None of that chain requires the sound to be loud enough to damage hearing. It requires only that your nervous system keeps deciding the sound is worth responding to.
The review makes one point that should change how you think about your bedroom: the adverse effects are most pronounced at night, when noise fragments restorative sleep and amplifies the neurohormonal imbalance. Your baseline is already shifted toward recovery then. It has further to fall.
What Does Everyday Noise Do to HRV?
It lowers it, in proportion to the volume and the duration, and the effect outlasts the noise.
Xin Zhang, Sung Kyun Park, Lauren Smith and Richard Neitzel at the University of Michigan published the analysis in September 2026 in the Journal of Exposure Science & Environmental Epidemiology. They pulled from the Apple Hearing Study, a nationwide cohort running since November 2019, and built four analytical cohorts — environmental noise versus headphone sound, crossed with a 20-minute and a 160-minute exposure window. Each cohort used a demographically stratified sample of roughly 980 to 1,180 participants, drawn from tens of thousands of eligible adults.
Relative to a 40 dBA reference — roughly a quiet library — here is what a 10-decibel increase predicted:
| Sound source | 20-minute window | 160-minute window |
|---|---|---|
| Environmental noise | −6.8% SDNN | −16.0% SDNN |
| Headphone sound | −7.1% SDNN | −7.4% SDNN |
Two things in that table deserve more attention than the headline number.
Duration matters enormously for ambient noise and barely at all for headphones. Environmental noise more than doubled its effect as the window widened. Headphone sound did not. The researchers' proposed explanation is annoyance: ambient noise is largely involuntary and unwanted, while headphone sound is chosen. The body appears to treat sound you selected differently from sound imposed on you — though that remains a hypothesis their design cannot test.
The effect scales with volume. At 60 dBA — an ordinary conversation, a busy café — the long-term environmental reduction reached 27.4%, more than double the short-term figure at the same level.
The lag pattern is the detail I keep coming back to. When noise rose, SDNN went up slightly for the first three to six minutes, then turned and declined for the remainder of the window. Your immediate reaction to a noise is not your body's verdict on it. The verdict arrives later, and it arrives quietly.
Who Does Noise Affect Most?
The subgroup findings are the most practically useful part of the paper, and the pattern is not what most people would guess.
| Group | Pattern |
|---|---|
| Older adults | Larger SDNN reductions at the same sound level. At 60 dBA, roughly −31% in the 65+ group versus −21% in the 18–25 group |
| People with tinnitus | Larger reductions than people who never experience it, at any frequency of occurrence |
| Excellent self-rated hearing | Larger reductions from loud environmental noise |
| Poor self-rated hearing | Larger reductions from headphone sound |
| Higher perceived stress | Slightly larger reductions, most visibly with headphone sound |
That hearing split runs in opposite directions by source, which is the kind of finding that suggests a real mechanism rather than noise in the data. The authors propose that headphone listening imposes greater cognitive effort on people with hearing difficulty — effortful listening as a stressor in its own right — while people with excellent hearing may simply be more bothered by unwanted ambient sound.
If you track recovery for a parent or grandparent alongside your own, the age row is the one to sit with. The television left on all evening, the busy dining room, the window facing a main road. Same biology, different volume.
What This Study Cannot Tell You
More than the draft version of this story would suggest, and I would rather you hear it from me.
It is observational. No one randomized anybody to a quiet room. The association is real and dose-responsive; causation is not established.
Apple funded it. Salaries for three of the four authors were supported by Apple's contract with the University of Michigan. The funder reports no role in design, analysis or the decision to publish. That is the standard disclosure and I take it at face value — but you should know it.
The devices were not individually calibrated. Validation work found Apple Watch SDNN reading systematically lower than reference, by up to about 8 ms, and the Noise app measuring 2–3 dB below reference. Participants are all Apple product users, which limits generalizability.
Only one HRV metric was available. SDNN, from one-minute readings taken when the watch judged you still. Richer measures might tell a different story.
And the honest one: noise explains very little of the total variation. In these models, sound exposure alone accounted for between 0.7% and 1.8% of SDNN variance. Individual baseline differences accounted for 33% to 40%.
Read that last one correctly. It does not mean noise doesn't matter. It means noise is one modest input among many, and that you differ enormously from the average participant. Which is the argument for measuring yourself rather than importing a population number.
Can Natural Sound Undo It?
Partly. And the part it can't is instructive.
Konrad Uebel, Melissa Marselle and Eleanor Ratcliffe ran two connected studies published in People and Nature. More than 1,500 people rated nine simulated soundscapes varying in acoustic complexity and amplitude. A second group of 63 listened to the same recordings in a lab wearing heart rate monitors.
Birdsong produced a psychological wellbeing benefit regardless of the soundscape's other acoustic characteristics. But when traffic noise was present — even below WHO daytime guidance for road traffic — that benefit shrank, and heart rate variability was significantly lower, indicating a greater physiological stress response.
The worst combination was loud traffic paired with acoustically complex soundscapes. The researchers read that as overstimulation: too much total sound energy for the system to sort. Not simple loudness. Capacity.
So the birdsong outside your window is still doing something. The delivery truck two streets over is quietly taking a portion of it back, at a volume your city considers perfectly acceptable.
How Do You Manage Sound as a Recovery Input?
Not by chasing silence. Neither study measured silence, and perfect quiet is not available to most people most of the time. Treat sound the way you already treat training load: an input you can adjust, then verify against your own data.
| Adjustment | Why it follows from the evidence |
|---|---|
| Drop headphone volume a few notches | Headphone effects showed up at 20 minutes and didn't need hours to accumulate |
| Take the quieter route when the cost is equal | Environmental noise effects more than doubled as exposure lengthened |
| Audit your default workspace | Open offices, shop floors and kitchens run loud by default, all day, involuntarily |
| Protect the sleeping environment first | Effects are most pronounced at night, when your baseline is already low |
| Choose genuinely quiet or natural wind-down | Birdsong helped; birdsong with traffic underneath helped less |
| Give yourself one real quiet window | The lag data shows the decline building over 160 minutes, not resolving within it |
Then look at your HRV trend over weeks. Not tomorrow morning's number. One morning tells you almost nothing, and with noise explaining under 2% of variance in a population model, a single reading certainly won't isolate the effect of your commute.
Where The HRV Book Fits
Everything above is a population average dressed up in decimal places. The part that matters is what happens in your nervous system, which is not the average one.
The HRV Book goes further into separating a real shift in your recovery from ordinary day-to-day static — sound included — and into building the habit of reading your own trend rather than someone else's mean. That's the skill this research actually calls for. The study gives you a hypothesis. Your own data is what tests it.
What Thirty Years Taught Me About Background Variables
I spent thirty years in sports performance, including as the first sports scientist at Ohio State, and I watched a lot of very disciplined people optimize two variables and treat everything else as scenery.
Sleep and training load deserve the attention they get. They're large, they're well studied, and they move the numbers. But the reason they get that attention is partly that they're easy to see as variables at all. Sound isn't. It has no start time, no session, no entry in a log.
The thermostat-and-thermometer relationship holds here the way it holds everywhere. The habit is the thermostat. HRV is the thermometer. And you cannot adjust a setting you have not noticed is a setting.
Frequently Asked Questions
How does noise affect the nervous system? Unexpected or rising sound is evaluated as a potential threat before deliberate reasoning engages. That activates central stress pathways through the amygdala and the HPA axis, driving sympathetic activation and the release of catecholamines and cortisol, which in turn promote inflammation and oxidative stress in blood vessels.
Does everyday noise lower HRV? In a 2026 analysis of Apple Hearing Study data, each 10-decibel increase in environmental noise was associated with a 6.8% SDNN reduction over 20 minutes and 16.0% over 160 minutes. It is observational data, so it shows association with a dose-response pattern rather than proven cause.
Is headphone volume as bad as traffic noise for HRV? Different, not simply worse or better. Headphone sound showed roughly 7% reductions at both short and long windows, while environmental noise more than doubled its effect over the longer window. Researchers suggest involuntary, unwanted sound may provoke more annoyance and stress than sound you chose.
Does noise have to be loud enough to damage hearing to affect you? No. The autonomic and vascular pathways described in this research operate well below hearing-damage thresholds. They require only that your nervous system keeps responding to the sound.
Who is most affected by noise? Older adults showed larger HRV reductions at identical sound levels, as did people who experience tinnitus. People with excellent self-rated hearing showed larger reductions from loud ambient noise, while people with poorer hearing showed larger reductions from headphone sound.
Does birdsong actually help with stress? In a study of more than 1,500 listeners plus a 63-person lab group, birdsong produced psychological wellbeing benefits regardless of other acoustic qualities. Adding traffic noise — even below WHO daytime guidance — reduced that benefit and lowered heart rate variability.
Should I try to make my environment silent? Neither study tested silence, and it isn't available to most people. The practical approach is to treat sound as an adjustable input, make a few specific changes, and check your own HRV trend over weeks rather than chasing a single morning's number.
Why does noise matter if it only explains 2% of HRV variance? Because that figure is a population average across a highly varied group, and individual differences accounted for 33–40% of the variance in the same models. Small average effects can be much larger in specific individuals and environments, which is precisely why individual tracking beats population estimates.