Why SpO₂ and heart rate fail on the move, and how a wearable can tell
Optical sensors work beautifully on a still finger. On a chest or wrist during a walk, motion can be larger than the pulse itself. A good wearable knows when to say “not sure”.
Pulse oximetry shines red and infrared light into the skin and compares how much of each comes back as the blood volume changes with every heartbeat. The pulse part of the signal is tiny: often only 1 to 2 percent of the light that returns. Anything else that changes the light path, such as the sensor shifting on the skin, a tendon moving or ambient light leaking in, can be much larger.
What motion does to the signal
- Walking and running create a regular rhythm at the step rate, usually 1.5 to 3 steps per second. That is right in the range of a heart rate of 90 to 180 bpm, so an algorithm can lock on to the steps instead of the pulse.
- Strength training creates irregular bursts and pressure changes under the sensor that distort both heart rate and SpO₂.
- Pressure from a tight strap or lying on the sensor squeezes blood out of the tissue and flattens the pulse.
SpO₂ is more sensitive than heart rate, because it depends on the exact ratio between two small signals. A small error in either one can move the reading by several percent.
Use the accelerometer as a witness
A 3-axis accelerometer next to the optical sensor records the movement that causes the noise. Firmware can use it in three ways:
- Detect. If acceleration energy is above a threshold, mark the optical reading as unreliable.
- Compare. If the strongest frequency in the optical signal matches the step frequency from the accelerometer, the “heart rate” is probably the cadence.
- Correct. Adaptive filters can subtract the motion part from the optical signal. This works well for regular motion like walking and less well for irregular movement.
Cross-check with the ECG
A patch that records both ECG and optical pulse has a second, independent heart rate. The ECG is far less affected by motion at the heart-rate level, because the R wave is sharp and large. When the two disagree, the optical reading is usually the one to doubt. This is one reason Hridya combines both sensors with motion in one patch.
Report quality, not just numbers
The most useful thing a wearable can do during motion is to be honest. Our app marks each trace as good, fair or poor and names the reason, such as movement, noise or no skin contact. Try the activity picker in Vitals in motion on our home page to see how quality changes from rest to running.
Rule of thumb: trust resting SpO₂, treat SpO₂ during heavy motion as a hint, and always store the motion data next to the optical data so a reviewer can see why a reading looks odd.
Want an ECG and SpO₂ patch under your own brand?
Our prototype, Hridya, already records ECG, SpO₂ and motion and streams them to its own app. Develop it with us into your product, or let the same team build your own device.