Optical heart rate sensor vs bioimpedance sensor?

Question

I didn't know a better place to ask this, so here it goes. For activity trackers, Fitbit has the Charge HR with an optical heart rate monitor, while Jawbone has the UP3 with a bioimpedance sensor for heart rate.

Can you please explain the differences between the two. Also, diagrams would make it more understandable.

Info about optical sensor: electronicdesign.com/displays/build-your-own-optical-heart-rate-sensor Also, are any diagrams from here relevant?

On Jawbone's blog, there is some info:

Bioimpedance measures the resistance of body tissue to tiny electric current to enable the capture of a wide range of physiological signals including your heart rate. If you've ever measured your body composition such as fat content, this is very similar. Our choice to use bioimpedance technology sets us apart in three key ways:

Battery life. Because bioimpedance requires significantly less power compared to optical sensors for same level of accuracy, we can deliver a smaller form factor and longer battery life Physiological signals. A single platform utilizing bioimpedance sensors captures a wide range of signals: heart rate, respiration rate and galvanic skin response (commonly known as skin conductance) Updatable technology. Given the versatility of the sensor platform, we are able to (and we will) unlock exciting new features with a simple, free, over-the-air firmware updates in the coming months

Answer 1

The optical sensors (photoplethysmography) rely on the change of volume of blood in a digit or earlobe (based upon the pulse) changing the light absorption characteristics as detected by an LED/photodiode pair. They're pretty easy to use, and hard to get wrong.

Bioimpedance plethysmography is much the same, but relies on a very small electrical signal provided at one point on the body and received at another.

It's pretty well described in A Bio-Impedance Measurement System for Portable Monitoring of Heart Rate and Pulse Wave Velocity Using Small Body Area Min-Chang Cho, Jee-Yeon Kim, and SeongHwan Cho, in Proc. IEEE EMBC, 1997, pp. 2072–2073. (though this will be hard to get outside of a university)

My understanding is that the bioimpedance measurement is trickier, as electrode placement becomes pretty important.

Answer 2

A bioimpedance "sensor" is a circuit attached to ECG-like electrodes on the skin. For the heart rate, they don't use an impedance measure - that'd be kind of pointless when you have the wonderful, huge ECG (electrocardiographic) signal.

They simply use ECG! Impedance measurements are done through the same electrodes used to measure ECG, and provide breathing rate and galvanic skin response measures. It is possible to use an impedance measurement to acquire heart rate, but it's a pointless endeavor, since EKG is an overwhelming signal present all over your body - it's actually an artifact in other measurements.

Answer 3

The ppg measurement is very inaccurate if sensor is moved in relation to the skin. Fx if you walk or run the sensor will bounce around. The sensor takes a sample of the reflection from a flashing led. When blood flow in waves on heartbeats, blood vessels expand, and this is reflected in changes of measured samples from the sensor. If the sensor bounce, you are not mesureing on the same place all the time. The heart rate is calculated on the difference on the ebbs and flood of the blood stream - over time. This obviously is a problem for a fitness tracker.

Answer 4

Mark my words: While bioimpedance can (sometimes) be used to measure accurate biometrics under the exact right conditions, there is no evidence that bioimpedance can work for continuously accurate monitoring in wearables.

The reasons: 1) Electrode placement is challenging and not universal. 2) Electrode physical stability is not scalable to a mass population (again, for continuous monitoring that is). 3) Electrode conditioning is problematic (the electrodes are prone to corrosion). 4) The bioimpedance signal is affected by numerous other factors that have nothing to do with biological signals at all (motion artifacts, electrical interference, etc.) and other biological signals that are much stronger than the miniscule HR and BR signals. 5) Even if you could get a continuously accurate signal from bioimpedance it would almost certainly be on a customizable basis only, and the wearabililty would be characterized as "extremely uncomfortable" for continuous donning.

I have seen credible research papers where, for spot checking and analysis under certain conditions (sleep for example), bioimpedance can be very interesting. And if researchers focus on particular "rest" use cases, I think the technology can go quite far. But for labile living conditions, there is no real hope for bioimpedance in wearable devices.

Of course, only time will tell if I'm proven right or wrong... But as a scientist, the next best thing to being proven right is to be proven wrong.