Breathing and heart rate from Wi-Fi: what a trend can tell you

Stillsense now offers breathing-rate trends during sleep and heart-rate trends, sensed from Wi-Fi with nothing worn. Both are available, and neither has been independently validated yet. This article explains what that combination means and how far the published research goes.

Two signals of very different size

Wi-Fi sensing works by reading how movement changes the radio signal in a room. A person walking is a large change. A person lying still is a small one, but not zero.

Breathing moves the chest by several millimetres, slowly and regularly, a dozen or so times a minute. In a quiet bedroom that rhythm stands out. A heartbeat moves the chest surface by a fraction of a millimetre, about five times as often, and it sits underneath the breathing movement. Turning over in bed, a second person or a pet can swamp it.

So the two capabilities should not be read as equally mature just because they arrive together.

What published research shows

A 2025 study in the IEEE Journal of Translational Engineering in Health and Medicine tested Wi-Fi sensing of vital signs on older adults in their own homes and in a living lab. Against a respiration belt, breathing rate was off by about 2 breaths per minute on average. Against a pulse sensor, heart rate was off by about 9 beats per minute, and the authors report "high variability" for heart rate in uncontrolled settings. The validation rested on eight measurements from seven people. Read the study.

For comparison, dedicated bedside radar is further along. A validation of the Norwegian sleep monitor Somnofy in 37 healthy adults found breathing rate within 0.18 breaths per minute of the reference on average. The device reported a value 84 percent of the time. Read the Somnofy validation.

Two things follow. Contactless breathing rate during sleep is achievable. And even a well-validated device has periods with no reading, so the gaps matter as much as the accuracy.

These are other teams' results on other systems. They describe what the field has shown, and they are not measurements of Stillsense.

A trend is not a measurement

What Stillsense provides is a trend: how a person's breathing rate during sleep, or their heart rate, compares with their own previous nights.

That can be useful to a care team as a prompt. A resident whose nights look different this week than last may deserve a conversation or a closer look. It is information of the same kind as noticing that someone has been restless or is getting up more often.

It should not be used for anything a clinical instrument is used for:

• No diagnosis. A trend cannot establish sleep apnoea, an arrhythmia or an infection.
• No alarm on a single reading. One unusual value is weak evidence. So is one normal value.
• No replacement for a check. If staff are worried about a person, they measure with approved equipment.
• No reading is not a normal reading. Periods without a value need to be visible to whoever looks at the trend.

What validation has to show

Our independent validation study is planned and has not started. Until its results are published, we describe these capabilities as available but not yet independently validated, and we make no clinical claim.

For breathing rate and heart rate, a useful study has to answer four questions. How close are the values to an accepted reference across a full night? How much of the night produces a value at all? Does it hold in lived-in rooms with a partner in the bed, a pet or a fan running? And does it hold for older people, who are the ones it is meant for?

We will publish the limitations along with the results.

Where the medical-device line sits

Stillsense's care features are offered today as non-medical safety technology. Clinical vital-sign monitoring is a different product category with its own rules. If we offer it, we will take the medical-device route under EU MDR and do the clinical validation it requires.

Until then, a care service evaluating Stillsense should treat presence, movement and bed-exit as the capabilities to build routines on, and the vital-sign trends as supporting information. Our evidence and trust page shows the status of each capability.

Book a pilot conversation if you want to look at how the trends could fit into a night routine without being asked to carry more weight than they can.

The study figures come from published research on other systems, not from Stillsense. Fall detection, breathing rate and heart rate are available with independent validation planned; none carries a safety-grade or clinical guarantee.