Heart rate monitors and HRV - accuracy?

[Stefan Litzenberger]

Dear community!

For a small project we need to monitor the Heart Rate Variability (HRV) of a number of subjects (approx. 25 concurrently) before, during and after prolonged exercise (cycling) in-situ (i.e. we are looking for a mobile solution). We are well aware that a correct assessment is best done with ECG and that - concerning mobility - several ECG belts are available on the market (e.g. https://www.pulse7.at/shop/index.php...rsatzgurt.html). However I have found out that some of the newer heart rate monitors are capable of R-R recording (polar RS800cx, Garmin forerunner 910xt). This would be highly convenient for our project because riders are going to be equipped with HR-monitors anyway, hence no additional equipment would be needed (and costs could be kept low(er)).
However I am interested in these monitors' accuracy. There are some papers on a Polar device (http://www.setantacollege.com/wp-con...%20Monitor.pdf, http://www.scielo.br/scielo.php?pid=...rttext&tlng=es) however they are not really up-to-date (2006 and 2008 respectively).

So: Does anyone have any information on the R-R measurement accuracy of these devices (and especially the Garmin 910xt)?

Glad for your help as always! Thanks!
Kind regards from Vienna
Stefan

[Dudley Tabakin]

Re: Heart rate monitors and HRV - accuracy?

Hi Stefan,

I cannot comment specifically on the Garmin 910xt for RR interval detection but:
When performing HRV assessments, it is always advisable to work from a signal that has the ECG waveform and with software where you can simultaneously visualize the ECG and the detected R-waves and subsequent R-R tachogram.
HRV metrics are extremely sensitive to artifact and even a very small amount of artifact can significantly bias the measurements. Fortunately there are several ways of addressing artifact but one must first confirm without doubt that there is indeed artifact.
Without having the ECG to visualize, there is no way of knowing if the beat detection is accurate or not and you are at risk of serious errors. No matter how good the sensor or QRS detection algorithm, there is always risk of artifact or detection errors. Especially so during prolonged exercise.
I would strongly advise you to choose a sensor solution that will provide the full ECG waveform. Several sensors that can provide this may be found on http://vivonoetics.com/products/sensors/ .VivoSense software (http://vivonoetics.com/products/vivosense-hrv/) also contains powerful visualization and artifact editing features that will ensure you high quality and accurate HRV results.

Please contact me if you would like to know more.

Dear community!

For a small project we need to monitor the Heart Rate Variability (HRV) of a number of subjects (approx. 25 concurrently) before, during and after prolonged exercise (cycling) in-situ (i.e. we are looking for a mobile solution). We are well aware that a correct assessment is best done with ECG and that - concerning mobility - several ECG belts are available on the market (e.g. https://www.pulse7.at/shop/index.php...rsatzgurt.html). However I have found out that some of the newer heart rate monitors are capable of R-R recording (polar RS800cx, Garmin forerunner 910xt). This would be highly convenient for our project because riders are going to be equipped with HR-monitors anyway, hence no additional equipment would be needed (and costs could be kept low(er)).
However I am interested in these monitors' accuracy. There are some papers on a Polar device (http://www.setantacollege.com/wp-con...%20Monitor.pdf, http://www.scielo.br/scielo.php?pid=...rttext&tlng=es) however they are not really up-to-date (2006 and 2008 respectively).

So: Does anyone have any information on the R-R measurement accuracy of these devices (and especially the Garmin 910xt)?

Glad for your help as always! Thanks!
Kind regards from Vienna
Stefan

[Matthew Tenan]

Re: Heart rate monitors and HRV - accuracy?

Stephan,
I've used HRV in a number of studies. Dudley is correct; without visualizing the waveform the analysis is EXTREMELY prone to errors. I've not seen a detection program that is robust to the point that I would trust it without seeing the ECG wave.

Vivonetics makes quality products. I might also suggest investigating Zephyr. They make a wireless ECG for continual monitoring (http://www.zephyranywhere.com/products/biopatch/) which also outputs the tachogram (R-R intervals).

Best of Luck,
Matt