HRV Support

[Darren Cope]

I know HRV support/tracking has been kicked around before (see issue 304/old 456) but it's come up a couple of times recently in posts (both by by Martin Wifling I believe.)

• Martin--care to share exactly how you're dealing with HRV data in GC?
• Mark and team, any comment on where this fits in terms of priorities, etc.?

Cheers,

Darren

[Mark Liversedge]

Depends :)

We *are* planning to fix the 'measures' code in 3.1, and will become more central to the way physiological indicators / markers are managed.

This means things like Weight, Sleep, Fatigue (SB) and possibly HRV will be in there, but also include CP, W' and possibly a couple of other 'estimated' capacity / performance measures as well e.g. LTS, Estimated vo2max.

So there is scope to include it .. still got 3-4 months to go !

Mark

[Martin Wifling]

Darren,

HRV issue is twofold:

1) HRV can be a powerful measure to identify your recovery status, ie to check whether you are ready to first sustain and secondly to process a hard workout. For example ithlete.com provides a similar service, but analysis can be done on your own as well (software tool: kubios hrv, see instruction video here: http://www.youtube.com/watch?v=_CdEoXGNt_c )
The markers are pretty well described and theres a lot of science behind and available.
I track some important markers (RMSSD, pnn50, LF/HF ratio, SD1, SD2) on a daily basis same as any other metric, by creating a manual workout (zero time, TSS etc.) which I name HRV for easier identification. A storage of beat by beat recordings is not necessary in GC, just storage of the metrics.
At this moment, this is what I do with GC.

2) In recent research, and this is probably the more promising fact for endurance athletes, there is a relationship between Maximum Lactate Steady State (MLSS) and some HRV metrics has been found.
This is described for example in this doctoral thesis (unfortunately it is in German): http://www.students.informatik.uni-luebeck.de/zhb/ediss1118.pdf

In this thesis, on 20 middle-trained riders their "individual anaerobic threshold, IANS" could be identified with similar precision than actual procedures by a single step test (start at 50W, onset 17W/min, all out). Probably our experts here (Dr. Skiba) can better explain the physiological backgrounds. At this point in time, there are still some difficulties in measurement, but it seems promising to have a tool, where an athlete can do something where normally lab testing is required at home with very small efforts. I also think, that this relationship probably could be described mathematically.
This gives the necessity to RECORD heart rate beat by beat, therefore it would be essential that GC has the ability to store that format. This would be nice anyway, because a lot of current HR-monitors support that already (I use RS800cx).

In the TOUR.de forums, there is a training group with actually approx. 160 riders (a lot of them are using GC as well) where HRV is tracked by a lot of them to monitor the recovery. The initiator of this group would like to test also this protocol to check if this protocol is useful.

I hope this givesyou a short insight.

cheers
Martin

[Mark Liversedge]

What resolution is useful for HRV ?

The ANT+ message supports up to 1/1024th of a second time since last heart beat, and sends 240 times a second, so faster than most human hearts will beat.

I don't think it would be too difficult to write a tool in GC that captures the beat-to-beat data, presents it on a chart as its received then stores it and calculates a metric or two (but the new measures code will need to have been written first!).

Is that what we need?

Mark

[Mark Liversedge]

240 times a minute. lol.

[Brett Dikeman]

On Tue, Jan 14, 2014 at 5:27 PM, Mark Liversedge <liver...@gmail.com> wrote:
> What resolution is useful for HRV ?

Not sure if this helps, but this link provides some background info
and notes the utility of BTLE straps. Apparently the mio alpha applies
some smoothing that interferes with the data.

http://www.marcoaltini.com/2/post/2013/12/heart-rate-variability.html

I read somewhere that Garmin HRM straps transmit interbeat timing info
and they used it in a number of units for some sort of calorie
calculation, but it sounds like only a few Garmin units record it -
the 610 and 910XT, I think?

> I don't think it would be too difficult to write a tool in GC that captures
> the beat-to-beat data, presents it on a chart as its received then stores it
> and calculates a metric or two (but the new measures code will need to have
> been written first!).

Maybe this would help?

https://code.google.com/p/hrvtracker/

[Nathan Townsend]

Thanks for this post Martin!

I'm well versed in this area.  I've measured HRV for years using a suunto T6 and also polar (in the past). More recently I have been sleeping with a bodyguard 2.  I also use kubios for analysis, however at present I'm trying to get started on learning matlab so I can fully explore the capability of the wide array of open source tools which can be found at physionet.org

One of the things I'm trying to investigate at present is the relationship between sleeping HRV and exercise (esp exercise recovery) HRV indicies. I want a device that can accurately record RR interval to be time synced (to the nearest second at least) with a power meter.  The problem I'm having at present though is that neither polar nor suunto ANT signals can be picked up standard ANT+ enabled devices such as a garmin edge 500.  Secondly, even if they could, few devices record with greater frequency than 1Hz, which of course simply isn't good enough for HRV analysis.  You'd think that the newer generation of bluetooth + ANT+ enabled HR monitors on the market eg: 4iiii would have an android app, but no... of course not.  You'd also think that they would record with enough resolution to determine RR interval.  But again this is a problem with the 4iii, it only samples at 120Hz, whereas there is research that shows a minimum of 200Hz is required.  

The problem with the ithlete is that you cannot export the raw data, so you're stuck with whatever index they're using to denote "HRV".  I don't know whether this measure is total spectral power, or some global time domain measure such as SDNN.  

So if anyone knows of a power meter head unit / HRM combo that can record RR interval with a sampling rate of at least 200Hz then please let me know!!!

[Nathan Townsend]

Thanks for this also.  Two words to tech developers.....

ANDROID PLEASE!!

Everything is iPhone only it seems right now in the BTLE HRM world.

[Brett Dikeman]

On Sun, Jan 19, 2014 at 6:02 PM, Nathan Townsend
<nathant...@gmail.com> wrote:

> Thanks for this also. Two words to tech developers.....
>
> ANDROID PLEASE!!
>
> Everything is iPhone only it seems right now in the BTLE HRM world.

There are several HRV apps on Android that support BTLE, both paid and
unpaid - although some are camera-based, which is a bit rubbish for
this.

When BTLE came out, Google lagged on implementation and an API.
Samsung released a proprietary API for their BTLE hardware, which is
why you saw a bunch of stuff that was only compatible with Samsung
phones at first. Despite the fact that Apple charges developers for
anything beyond very basic bluetooth access (yeah, you read that
right...), lot of companies have thrown up their hands and gone for
iPhone support because it's so much easier to target.

If you must have mobile support, my recommendation is to get a Sony
($$$$) or Samsung phone that has ANT+ support built-in, use a bridging
device like the 4iiii, or wait for the next generation of stuff that's
supposedly coming "in March." Or get an Android phone that supports
USB OTG mode (with power) and attach an ANT+ dongle such as the Suunto
or the OEM mini adapter. Note that battery life will take a hit,
though I have no idea by how much; I doubt the USB device is optimized
for power consumption.

-B

[Christian Del Rosso]

Hi,

the ANT+ heart rate profile specifies the max frequency as 4 messages/seconds.
And the documents provide data for R-R calculation based on the difference between the last heart beat event time and the previous heart beat event time.
The information can be found at thisisant.com website.

For Bluetooth Low energy there are R-R intervals in the standard, see the Heart Rate Profile and the Heart Rate Service documents:
https://www.bluetooth.org/en-us/specification/adopted-specifications

Overall both standards provide a way to find HRV.

[Luther Gulseth]

Brett, can you point us in the direction of an HRV Android app that is Ant+? I have the Galaxy S3 (USB OTG ready) and the only apps I can find require BT. I'm not ready to buy a BTLE strap as a close friend hasn't gotten good results out of the Viiiva sensor.

Luther

--
--
_______________________________________________
Golden-Cheetah-Users mailing list
golden-che...@googlegroups.com
http://groups.google.com/group/golden-cheetah-users?hl=en
---
You received this message because you are subscribed to the Google Groups "golden-cheetah-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to golden-cheetah-u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Nathan Townsend]

Thanks Christian.  I didn't know that about ANT+.  4Hz is clearly far too low for accurate HRV analysis which means BTLE is the only way other than the currently available proprietary protocols used by polar and suunto.  Therefore, for a product such as the 4iiii which acts as a ANT+ to wifi bridge, then you use the ANT+ to sync with your head unit, but you would use BTLE to communicate with the smartphone, NOT a samsung ANT+ enabled phone.  4iiii is the only HRM that can do that at present, however it hasn't got around to producing an android app yet.  Additionally, I've asked them to consider updating their firmware to sample at a higher frequency (at least 200Hz) but who knows how long that could take if ever.

Upon re-visiting where ithlete currently stands, it really is the best product on the market at present. It is designed specificlly for that purpose which means that the sampling frequency would be at least 200Hz.  Simon Wegerif, (the founder) has a background in electronic engineering and signal processing, so I trust he knows what he is doing there!  I discovered that their index (ie: the number they report) is the Ln RMSSD muliplied by 20 (which makes it look like a number out of 100 or so).  Actual values tend to be in the range 3 to 5.  This is a very good choice for a global HRV index and it reflects vagal control.  Once you know this, then suddenly the product becomes far more attractive because you know exactly what you're getting and you can examine how it responds to training in comparison to data from the scientific literature. 

[Nathan Townsend]

The new players to this market who are in development mode still and need to reduce their costs to eventually break into the consumer market are manufacturers of "smart garments"  ie: fabric and textiles that act as sensors such as ECG, skin temperature, respiratory rate.  It would be like wearing a thin, tight fitting (but elastic) base layer singlet top, and you get all that physiological information.

Who knows maybe even one day non-invasive blood lactate measures could be possible!!

Anyway, this is getting off topic.  Getting back to GC, in order for this support to be possible, therefore requires the development of a smartphone app which communicates with an appropriate BTLE HR strap.  The polar H7 does NOT do this so the link in Marco Altini's webpage above is incorrect.  You cannot conduct HRV analysis using this HRM.  I think it is a bit out of their capacity at present.

What could be beneficial is to simply implement user entered data.  I'll make another post on this later today in its own dedicated thread (or search for one already here)

[Mark Liversedge]

The 4hz is the rate at which you get data, the granularity of the actual measurement is 1/1024th of a second. 4hz would be ok for BPM up to 240 !

Mark

On Tuesday, 21 January 2014 21:49:37 UTC, Nathan Townsend wrote:

Thanks Christian.  I didn't know that about ANT+.  4Hz is clearly far too low for accurate HRV analysis [snip]

[Nathan Townsend]

Thanks!  This is why I love these forums. You can get all this excellent info much faster than by emailing manufacturers!

This is great because it means that ANT+ enabled devices should be able to record RR interval.   iThlete actually uses an analog strap and the little dongle thingy does the sampling.

It also means that at some later point in time (or a sooner point in time with the appropriate programming man power) Golden Cheetah could implement HRV analysis.  The requirements are thus as follows:

1. HRM signal (whether it be analog, BTLE or ANT+) needs to be sampled at a minimum of 200Hz (preferably > 500Hz).

2. This data needs to be available for export in an appropriate format so that it can be imported into GC.  

3. GC requires the coding for relevant HRV indices implemented.

For some very basic (but good) HRV indices #3 is a piece of cake.

[Stefano]

Yes, and present the data in a user friendly format!

This is what I do every morning (it takes me 7 minutes):
1) Acquire the data with my PC and a Garmin ANT+ key using HRVtracker (https://code.google.com/p/hrvtracker/)
2) Perform the analysis with Kubios (http://kubios.uef.fi/)
3) Elaborate a bit the results with Excel (automatically) following any scientific guidelines (e.g. http://www.ncbi.nlm.nih.gov/pubmed/17849143)
4) Present the outcome as an easy-to-use indication on the kind of training to be performed (see attached pic).

It would be nice to have everything implemented in GC.

Stefano

[Mark Liversedge]

Hi Darren,

Can you advise on an ANT+ HR strap that you /know/ supports the R-R info ?

I was looking to buy one so I could play with HRV ...

Mark 

On Tuesday, 14 January 2014 16:24:47 UTC, Darren Cope wrote:

[Brett Dikeman]

On Mon, Jan 27, 2014 at 2:11 PM, Mark Liversedge <liver...@gmail.com> wrote:

> Can you advise on an ANT+ HR strap that you /know/ supports the R-R info ?
> I was looking to buy one so I could play with HRV ...

All Garmin ANT+ straps do, reportedly. HRV has been integrated into
various Garmin products for a while for calorie-counting stuff
("Firstbeats" or something like that) but the HRV data itself isn't
actually recorded in the FIT files (I don't think) except for a very
small number of Garmin products, mostly the newer running watches and
one or two of the tri watches.

-B

[Mark Liversedge]

Thanks Brett,

I've ordered a Garmin HRM3.

Mark

[Jon Escombe]

Getting slightly off topic for GC, but just wondering..

Is optical measurement a viable option for HRV metrics (assuming the
waveform is tracked with a high enough sample rate)?

I would presume there is some small delay between the actual heartbeat
and what is picked up from say a fingertip sensor, but am wondering
whether there is also any/enough variability in this delay that it would
skew the sampled beat-to-beat timings?

Thanks,
Jon

[Nathan Townsend]

I've seen a publication which examines this.  I'll see if I can find it....

easy took me 10sec

http://www.ncbi.nlm.nih.gov/pubmed/23751411

[Nathan Townsend]

This one is even better and explains the potential source of error (ie: delayed propagation of the pulse pressure wave from heart to finger)....

Assessment of heart rate variability derived from finger-tip photoplethysmography as compared to electrocardiography

http://www.ncbi.nlm.nih.gov/pubmed/18663635

[Mark Liversedge]

Hi,

I just looked into the ANT code and saw that the data page is being read and decoded, and the infrastructure is there to just treat the r-r data as another data series. I've ordered an HRM3 strap so I can play with this, it all looks rather straight forward.

I would be interested in ideas about how a HRV dialog would work;

- connect device

- confirm data capture

- record against a protocol (which protocol is best?)

- summarise results

- record data and metrics for future analysis

I /really/ like the idea of exploiting the 'realtime' code to run tests that can be fed into the new measures and analysis functions.

Like having a home lab ;)

Mark

[Darren Cope]

Mark,

Yes, your dialog ideas sound about right. My thoughts (more or less the same as yours):

- Launch the dialog
- Have confirmation that the device is connected/transmitting/receiving properly (maybe a retry button in case it is not).
- Have dialog walk you through an appropriate collection protocol (not sure what this is, but, for example, "Sit quietly and don't move for XXX amount of time" and have a countdown timer showing how much longer to go (experts that have more experience in this please weigh in on what the proper data collection protocol actually is!)
- record the data
- provide pretty summary numbers/graphs/etc. of current readings and longer-term trends
- record data and metrics for future analysis within GC
- perhaps a way to automatically add the 'current' HRV metrics to a ride when downloading/importing a new activity

[Jon Escombe]

Perfect, thanks!

Jon.

[Brett Dikeman]

On Tue, Jan 28, 2014 at 6:48 PM, Jon Escombe <jesc...@googlemail.com> wrote:

> Is optical measurement a viable option for HRV metrics (assuming the
> waveform is tracked with a high enough sample rate)?

See my prior post on the 16th for mention the smoothing on the mio
alpha someone found interfered with HRV measurement.

A certain famous noted that the heartrate data from his mio alpha
seemed very reliable. This is why - they're smoothing the data quite a
bit to hide the unreliability of wrist optical HRM.

I don't think this is terribly off topic; for example, I'm curious as
to how smoothing by an HRM would affect various metrics GC calculates.
It'd also be nice to figure out how much smoothing takes place.

--B

[Mark Liversedge]

ZERO smoothing for all the regular metrics.

Its a philosophical thing -- we want to be absolutely transparent.

But then, when data is guff the metrics are too.

We do attempt to decrit to 1s recording for the mean max values on the CP chart and also for the W'bal curve.

Mark

[Brett Dikeman]

On Wed, Jan 29, 2014 at 2:32 PM, Mark Liversedge <liver...@gmail.com> wrote:
> ZERO smoothing for all the regular metrics.
> Its a philosophical thing -- we want to be absolutely transparent.
> But then, when data is guff the metrics are too.

Sure. But more optical HRMs are hitting the market in March, this time
with some handy features (ANT<->BT bridging) and people are going to
invariably ask, "what does that affect?", or it'd be nice to have
documentation so they know in advance. It's a fair question, and it's
also fair to use a device that smooths data, as long as you understand
what GC calculations will be impacted, and how significant that impact
would be. Maybe it turns out that aside from HRV, the smoothing
doesn't really impact anything else?

-B

[Jon Escombe]

My question was prompted by the actual measurement process.

While getting the metrics/analysis into GC would be fantastic for having
everything in one place, I was thinking that if I'm going to sample a
couple of minutes resting heartrate and HRV data each morning, then
using a laptop and HRM strap is possibly not the most elegant
arrangement. A smartphone app would be a big step up in ease of use, but
even then, it's often sat downstairs charging..

So I thought I could fairly easily prototype a dedicated device, using
an optical sensor and microcontroller. As long as the inbound sensor
data is sound enough, it should be fairly trivial to push a couple of
minutes worth of intervals to a web service and do any subsequent
analysis/processing/retrieval from there.

I've ordered an optical sensor to play with, so will see how the
intervals measured from this setup compare against ANT HRM strap data..

Regards,
Jon

[Mark Liversedge]

I received a mail off list saying pretty much the same thing.

iphone / android apps may be better than a tool in GC ...

in which case being able to import the data is of more value ...

[Darren Cope]

Hi all,

I agree that in principal, an app would make things easier. However, it may become an issue in that it requires a) an ANT+ 'dongle' for iPhones, or b) a SelfLoops ANT+ Accessory for Android Phones (which I think is the only way to get ANT+ data on an Android?), or c) an BT4.0 capable phone and HR Strap, all of which would likely be an additional expense for many people. However, for those that already have Golden Cheetah and a USB ANT+ dongle (for example to use Train mode,) it'd be nice to have support directly in GC.

Obviously there's no one solution that will work for everyone, but I feel the desktop environment is a bit more 'stable' for developing something like this, as opposed to iOS or Android, where things are constantly changing.

Open to thoughts...

Darren

[Christian Del Rosso]

Hi,

All current iPhones and iPads,  starting from the iPhone 4s have bluetooth LE.
Starting with Android 4.3 there is bluetooth LE support in Android. Going forward they will have BTLE.

If you want to use ANT+,  for iPhone there the Wahoo dongle,  for Android you can use the ANT accessory you use with Goldencheetah with an additional USB otg cable.

Overall the situation is getting more stable in the smartphone world.

Having a smartphone application makes things easier in my opinion.  Your phone is always with you and always on.  Of course,  the same functionality can be implemented in GC.

--

--
_______________________________________________
Golden-Cheetah-Users mailing list
golden-che...@googlegroups.com
http://groups.google.com/group/golden-cheetah-users?hl=en
---

You received this message because you are subscribed to a topic in the Google Groups "golden-cheetah-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/golden-cheetah-users/iyX3Qso8U44/unsubscribe.
To unsubscribe from this group and all its topics, send an email to golden-cheetah-u...@googlegroups.com.

[Mark Liversedge]

Lets have both.

Its not like GC is going to be on an iPhone any time soon anyway :)

I've been playing with this and certainly its easy to get the R-R data from a suitable ANT+ HR strap, I've done that already with a couple of extra lines of code.

The question, I guess, is what metrics do we want to calculate and how will we use the data captured...

Mark

[Serban Mestecaneanu]

Hi Mark,

I know this is an old topic but as you are currently supporting R-R data from Polar, which I guess is the same as HRV data, I would like to give my 2 cents here.

Re to your question what kind of data do we users expect, I can say for myself that after I have being diagnosed with a minor heart issue I have to monitor the frequency of my missing beats. And this can be done by monitoring HRV data.

I currently have a Fenix 3 which records HRV data, is there a way to bring this into GC?

Thanks,

Serban

[tfk]

appreciate old topic and i'm not sure what is being done with this development as I am new to this forum

1. As the HRV data is supported by the Garmin F3 (920 also supports it) then I would have thought best method of capture for GC is to effectively outsource the capture to a standard environment ie the ANT/FIT world and/or sml and/or hrm. I would assume that the data would then be imported as normal into GC...surely GC can store that level of RR/HRV granularity??? Providing the subsequent analysis then becomes the issue for the GC developers.

2. I have used an app (bioforceHRV) for well over a year and it is good and useful. the downsides are several in that I am tied to the analysis the app provides which is less than what some in this thread talk about and compared to what ithlete and EliteHRV offer. I have to use Bluetooth. I can't get the data out in a standard file format. For a wking HRV reading it would be no more or less easier to use a bedside mobile phone app or a bedside Garmin.

3. developers might want to check the patent status on 20LnRMSSD and other measures they are contemplating using, that specific calculation I understand is patented and others might be too.

4. Be wary of some dual-band HR straps. HRV will not necessarily work on both frequencies. Be wary of optical HRV as of Jan2016 it doesn't exist in a commercial product AFAIK. (watch Valencell for closest to market IMO)

5. I'm not sure of the support for HRV across Garmin Edge devices. From memory it's not great. so triathletes are prob OK but maybe less so cyclist users of GC