Category Archives: Heart Rate Variability

HRV Tutorial – Variable Interval Heart Beats

Your heart beats at varying intervals based on the state of your nervous system. If you are nervous, scared, or exercising your heart will beat regularly and fast. If you are relaxed your heart will beat at irregular intervals and more slowly.

This is how that works. Your Autonomous Nervous System is made up of two subsystems, the Sympathetic and Parasympathetic Nervous Systems. The Sympathetic Nervous System (SNS) is your accelerator. It raises your heart rate and pushes you into fight/flight mode. When you are in fight/flight mode your Heart Rate Variability (HRV) is low as your heart is pumping blood regularly and fast in order to get you through the immediate danger.

Your Parasympathetic Nervous System (PNS) acts as a brake on your SNS. When there is no danger, the PNS is braking your SNS, slowing the beats and the results is a lower heart rate and higher HRV. When this is happening you are resting and digesting.

The variability in your heart rate is the interplay of your body’s accelerator (SNS) and brake (PNS). As you start examining your HRV readings, it will be a bit awkward because low HRV numbers mean you are stressed and high HRV numbers mean you are relaxed. HRV does not read stress, it measures your heart’s variability and the measures are inversely correlated with stress. High variability is low stress, low variability is high stress.

Find out more by reading about RR Intervals.

Sign up for the QuantXLaFont Newsletter
Get our lifestyle tips and studies delivered to your inbox.
Thank you! We don't spam :)

HRV Tutorial – Introduction

Heart Rate Variability (HRV) is the most accessible measurement of stress and health for someone who is not in the medical profession, and there are a shortage of plain English explanations of what it is, how to capture it, and how to use it to improve your life.

It took me almost a year to playing with HRV measurements using numerous devices and different apps before I felt I understood it enough to explain it to others. Along that way I had some very generous help from colleagues and friends who shared key concepts with me. And I went down many unproductive paths that wasted a fair bit of time. My goal here is to help you past those unproductive paths and give you the summary benefit of the great advice I have gotten.

With the purchase of some inexpensive items and reading through this summary you should be able to have some fun taking readings that are specific to your physiology. The interesting element of measuring your own HRV is that the baselines and tolerances are uniquely yours. Aggregate data from others gives you tips on where to look for those readings that are yours and yours alone.

The best thing to do is get some simple kit and start playing with it. You’ll learn as you go and I am happy to share as many tips as possible here. Enjoy!

Find out more by reading about Variable Interval Heart Beats.

Five Meeting Heart Rate Variability Compared

I had an opportunity to compare five meetings that had similar content and the same attendees over a one month period. My colleagues and I were preparing a big launch and we were looking at the plan in a series of review meetings. In the first meeting I had created a first draft plan and had to present to executives I was meeting for the first time. I was not sure of how they liked to consume information and was on edge. So the first meeting looked like this:


What you see here is that I was in “overdrive” 33.8% of the meeting time.Overdrive means I my parasympathetic nervous system had stepped aside and my sympathetic nervous system had me in fight/flight mode. The blue lines in the chart are those heartbeats where the difference in time between beats was under 17 milliseconds for at least 10 consecutive beats. This meeting was almost two hours long, I was answering a lot of detail and we were finding our way together so I was in overdrive for one third the time.

We returned to review the progress from the first meeting a week later. In this I had my materials memorized and I knew how the executives consumed information. The meeting went very well, and we still had a lot of work to do. Using the same definition of Overdrive here is the chart:


This second meeting was almost two hours long and because I was so prepared I was in Overdrive only 10.8% of the time. As you can see from the chart there were only periodic physiological accelerations. Big difference. In the next meeting, the executive I was supporting and I did not have a lot of time to prepare for the meeting. We went in without synchronizing. You can see the chart here:


I was in Overdrive 15.2% of the meeting. You can see that my physiological fight/flight lines are concentrated early in the meeting as the executive that I was supporting and I were synching up. We found our way pretty quickly and you can see the blue lines even out.

In the fourth meeting we had taken another week to make progress on the launch. A lot of the details were worked out and we were in pretty good shape. When we got together the same executives were in the room and my supervising executive and I had a chance to coordinate. You can see the results in this chart:

Overdrive was only 4.4% of the time and the meeting was smooth. I felt good in the meeting and the readings show things went smoothly. We had one last meeting to get final check off an approval. I would through this data out because the environmentals of the meeting completely threw things off. You can see the data is very different:


Here you see a complete physiological meltdown as I was in Overdrive 87.2% of the time. Turns out the office I was taking the meeting from was extremely hot. I was perspiring and uncomfortable. It was a distracting situation. The meeting went well. We got approval and the communications afterward were universally positive. I believe the physical discomfort overrode the comfort with the materials.

So it appeared that reviews of familiar material with the same team of people did results in less time in Overdrive. Comfort with the material and people improved my performance. The last meeting is odd and I can’t definitively explain it with the hot temperature. But the first four seem to indicate improvements.

Parasympathetic Flatline Talking versus Listening

I wanted to look at how often I entered the Parasympathetic Flatline while in a 1:1 conversation with a colleague by phone. For the discussion I read by heart beat intervals using a Polar H7 heart rate belt and the Heart Rate Variability Logger app for iOS. I also recorded my side of the meeting on a smart phone. When the meeting was complete I downloaded my heart beat intervals via csv file and pulled them into excel. Once in excel I used a formula mechanism I created that graphs segments where more than 10 consecutive interbeat intervals are less than 17 milliseconds apart.

During the 60 minutes session I measured 4,451 heart beats and the intervals between them. Of those intervals, 14% were in groups of consecutive intervals that were close together, meaning during 14% of the meeting I was in what I call Parasympathetic Flatline. This measured the periods where I was in fight/flight mode during the discussion.

Here is a vizualization of the meeting:


In the session the forty-seven stress events triggered. Of these, 22 of 47 occurred when I was talking and presenting information to my colleague. 25, or 53%, occurred when I was listening to my colleague.  When listening to the recording, it is clear that the stress event, even when occurring when I am talking, begin when I was no agreeing with my colleagues response or trying to move him to a different position. The stress response was a result of not liking the direction the conversation was going.

Again physiology has shown that anticipating and trying to shape another person’t response is the source of stress in a 1:1 interaction. I once thought presenting my own opinion was a source of stress but that has turned out not to be the case. The stress, it appears, is not agreeing with someone else presenting their opinion.

Added clarification: From Twitter, fellow QS’er Gustavo (@GGlusman) asked the percentage of time I was talking versus not. Pushed by the question I went back beat by beat and looked at the session. As I reported above 736 beats were “in stress ” meaning that those beats were in a grouping with more than ten beats that occurred with a difference in beat interval less in 17 milliseconds to the adjacent beat. Of those beats in stress, I found that 238 were while I was talking and the remainder while listening. So that means 38.5% of the stress beats occurred while I was talking, 61.5% while I was listening. Impatience while listening was clearly more stress creating than flapping my gums. Thanks to Gustavo for asking the clarifying question!

Using the Parasympathetic Flatline

Using the Parasympathetic Flatline I analyzed a discussion with a colleague. I was following up on a topic that was not controversial. We had discussed this topic about a month prior. The colleague and I get along in a positive way.  So this should have been a relatively stress free and short meeting.

I used Marco Altini’s Heart Rate Variability Logger and a Polar H7 heart rate monitor to gather the base data while recording the meeting with my smart phone. Pulling that data into a spreadsheet I used my Parasympathetic Flatline model to determine at what point in the meeting I was experiencing physiological stress. I pulled the recording, the heart rate variability readings and the transcript into a timeline graph.


What we see here is when I talked (green), when my colleague talked (blue) and when I was experiencing a physiological event of what I call Parasympathetic Flatline (red), or stress. There are specific points in the discussion when I was amped up, but they were not what I expected.

I had a hypothesis that I was entering these states when I was putting myself “out there,” however I had one moment about 3/4 of the way through the meeting where I really pushed the boundary of a sensitive topic but I did not experience stress. I was synched with my colleague and the discussion did not trigger stress at that point.

There were four points in the discussion, however, that I entered a heightened state when I wanted the conversation to go in a different direction. My agenda was not being followed. In the first two, early in the conversation, I wanted to hear the answer to the question more directly. I was impatient. In the second two, I had the information I needed and wanted to wrap up.

It appears that difficult topics are not stress inducing when discussing them with a colleague when we are in synch, but my overall judgements about the progress of the discussion seem to trigger an aroused state. It is our judgements about the situation that may be the source of stress.

Testing the Parasympathetic Flatline

Last week I proposed measuring a Parasympathetic Flatline (PS Flatline) where at least ten successive heart beat intervals were close together. This differs from using 30 second averages for rMSSD, a time based measure of average intervals. I wanted to to a side by side comparison to see which of these methods more accurately could point me to moments when I triggered a shutdown of my Parasympathetic Nervous System which allowed the fight or flight reaction to run things.

I measured several meetings where I was able to audio record the proceedings while measuring my heart rate. I used my Android phone to record the audio, a Polar H7 heart rate belt to pick up my heart rate and Marco Altini’s Heart Rate Variability Logger to capture the data. Afterward, I downloaded the csv files for both the rr intervals and the 30 second rMSSD.

The meeting was 55 minutes long and included one person in the room with me and one person on the phone. We were discussing a topic I was comfortable with and an activity I had experience doing. I was briefing my colleagues on a time schedule and details. I was in a relaxed state going into the meeting and did a short two minute breathing reset using BreatheSync prior to the meeting.

In previous studies I had determined that a 30 second rMSSD under 48 was probably a stress state. I had used several stress events to create this baseline. When looking at this meeting, however, I was disappointed to see that using this method definitely overstated the number of stress states. The graph below shows blue bars where I had a 30 second rMSSD under 48 and you can see it shows I was in that state for much of the meeting.


This was not my experience of the meeting. I had the facts to hand, we had a good discussion and overall it was a friendly, informative discussion. So the 30 second rMSSD had not done the job for this meeting. I then looked at the PS Flatline that would show when my beat to beat intervals were less than 17 milliseconds for 10 consecutive beats. Here is what the meeting looked like using the PS Flatline.


Listening to the audio I could definitely see that this method caught moments that had me shifting into high gear. I chose this meeting because there was a point at the fifteen minute mark where my information was completely wrong and I felt the flush of embarrassment. This came through accurately in the chart. It also showed several moments where I was trying to calculate sums with an audience and I needed to shift into gear and focus. You can see on the chart different elements of the meeting I heard in the audio and my memory of what was happening.

So the PS Flatline approach seems to be much more accurate though I have to do further analysis on other meetings to ensure it is catching all of the stress trigger events. I have 10 additional meetings recorded with audio and HRV so i will start crunching those number next.

Finding the Parasympathetic Flat Line

From the start I have wanted to pinpoint the moment I was stressing out and identify the causes. I started with logging Upsets, then moved to using different devices to read Heart Rate Variability (HRV). I was always bothered by the lack of precision in how the devices give insight into HRV. They never said “you were stressed from the 15th to the 45th second.” Rather, they gave an average score over a longer and generalized period of time. I want to nail down the specific time my physiology starts and stops going berserk. To understand precisely when this is happening I have to look at  RR Interbeat intervals (pictured below) and find those periods in the readings where I have multiple consecutive intervals with very little variability.


When in stress mode the distance between successive beats for multiple beats remains very nearly the same. This occurs when the Parasympathetic Nervous System (rest & digest) flat lines and lets the Sympathetic Nervous System (fight or flight) run the show. Here are graphs of my RR Intervals for a similar time period using slow breathing to create a calm state described in last week’s blog post and a session on an elliptical where I was exercising and my heart rate was 145 beats per minute.


You can see that the RR Intervals vary while calm, and there is no variability at all while exercising. While running the Sympathetic Nervous System has the hammer down. In relaxation the Parasympathetic Nervous Systems is braking the machine and providing periodic slowdown. That means that even while resting and digesting our RR interbeat intervals are close to the same values for 32% of the time (red circled areas).


When reading HRV the fundamental output is the RR interval. All analysis is derived from that one string of numbers which are simply the number of milliseconds between beats. So it is straightforward to find periods in readings where those intervals are close together. Looking at the raw data I hypothesized that the Parasympathetic Nervous System is flat lined when the variance is small for 10 consecutive beat to beat intervals.


I then looked at how a rule of 10 consecutive intervals would work for my readings of the calm state and while exercising. The maximum number of low variance intervals in the calm session reading was 7 consecutive beats, and while exercising there was no variance in more that 5 beats. So if I gave each interval a value of “1” if it was in a group of of 10 ore more intervals with low variance and a value of “0” if it was not in such a group, the graphs of the calm and exercising sessions would be as seen below. No intervals are in a group of 10 low variance readings in the calme state, all intervals in are a group of 10 or more for exercising.


The second half of the calculation is the definition of “low variance.” I proposed in my post on HRV and Stress Free State that 25 milliseconds was low. So I took the rule set that the I was in berserk status when 10 consecutive intervals were under 25 ms and graphed if for a meeting I participated in last week. That graph shows more of the meeting in high vibration than I remember and didn’t quite look right to me. I lowered the number to 15 milliseconds and the amount of unrestrained Sympathetic activity seemed to get too small. Not a very rigorous sensitivity analysis I realize, but I have to pick a start point that seems to somewhat resemble what I remember happening.



I set the calculation for 17 seconds and the graph started to look like general stress cadence of the meeting as I remembered it. Fortunately, I had audio recorded the meeting!  So I went back and listened to those portions where it looked like my Parasympathetic Nervous System had stepped aside.



In those three portions of the recording I could hear in my voice that I was in a state of high vibration. In the first case I was presenting something and I sounded unsure of myself. In the second I sounded confident, but my cadence was noticeably slower and it sounded like I was searching for words at times. In the final case I actually said “I don’t understand your question” and there was a bit of confusion.

So I have a base that I am now going to start running data through to see if I can validate the 10 consecutive beat, 17 millisecond ruleset. If it starts identifying points of stress with precision I will have a framework that can help me start creating preparation regimes for 1:1 interactions based on precise knowledge of stressor that flatline my Parasympathetic Nervous System.


Stress Free State Through Lens of HRV

I have been collecting data on my Heart Rate Variability (HRV) to determine how to improve my personal performance and reduce stress. In doing so I have come to doubt the usefulness of “stress measurements” touted by some wearable devices and apps. They claim to be able to measure your physiology and indicate when you are stressed. My experience has been that the readings and the physiology is much more complex than a simple “red, amber, green” meter can indicate. It came to a head recently when I had a nicely engaged conversation with a colleague and my physiology readings indicated signs of stress. In fact, the experience had not been stressful at all. We were working hard on solving a problem and I felt engaged.

So I wanted to get a baseline by looking at data where I knew there was no stress at all. I had an HRV reading from an eight minute meditation session where I was completely relaxed and had nothing to distract me. The graph below shows how my heart rate stayed in the range from sixty-four to seventy-four beats per minute, a normal relaxed range for me.


The heart rate is negatively correlated to HRV, meaning that the body when in a state of arousal or stress will elevate the heart rate and naturally lower the amount of variability between beats. The variability in heart rate is examined by looking at the intervals between beats. Here is a graphic showing the “R-R Intervals” and how they vary. You can see in this example how the heart speeds up and slows down beat by beat.

interbeat intervals

For the eight minute session my RR interval steadily rose and fell. In this graphic I display the distance between beats for the entire period. You can see the distance increase and decrease regularly.


What you see here is my heart rate following my inhalations and exhalations. When we breathe in our heart accelerates and variability decreases. When we breathe out our heart decelerates and variability increases. During this exercise I used a breath pacer set for one breath every ten seconds. This graphic shows one ten second interval where I took one full breath in and out. You can see how the intervals drop on inhalation, rise on exhalation.


When all of the interval changes are put side by side you can see that in a stress free state the changes are regular. The following graph shows the changes both positive and negative in the differences in intervals during the entire eight minute session. Notice that on the outward breath, when the variability is positive, that the scale of the changes are larger than on the inward breath when the variability is negative.This means that the breathing outward has a powerful effect on the overall measured averages of HRV readings.


The averages of the readings, expressed best in the time based measure rMSSD, show for this session a settling at at rMSSD of 55 which is relaxed as any average reading for me indicates no stress when above 48.


But this reading does not give a good picture of what is actually happening physiologically. It looks like in a relaxed state the rMSSD stays above 48 and implies that if I drop below 48 I will be stressed. So a natural reaction would be to try and devise methods that would keep the rMSSD above 48. But this is an average and obscures how variable the underlying heart beats are.

Looking at those successive heart beats where the interval change is “close” or under 25 milliseconds. These beats would indicate a stress response for a lot of the session, but we know this period was stress free. In this graph you see the blue lines are those intervals under 25 milliseconds, and the white spaces are those intervals over 25.


My conclusion is this. In the drive to make wearable devices and apps more palatable to a mass audience the natural tendency is to bring the simplest interface to indicate stress or relaxation, be it a color or a score. This is useful for directing relaxation activity for beginners, and as the quantified self enthusiast digs into the underlying data much more nuanced picture emerges. What the last graph shows is the in the most relaxed and meditative state possible my beat to beat “score” was low for 41% of the beats. The fact is that these low score beats are part of healthy variability that goes both high and low. It isn’t right to see it as “59% good and 41% bad” –  it more correct to understand the mix of highs and lows, short term variability and the incredible speed with which the body becomes aroused and can calm itself. Sadly, the Apple Watch will not solve your stress problem. Data collection, repeated measurements and comparison with experience is the only way forward.

How Meetings Go – Physiologically

I have been measuring my rMSSD during work meetings to see what factors impact my performance when engaging face to face with others. I recently had two meetings with the same group of people on the same topic about a week apart. Before each meeting I also took my blood sugar to see if there was any information to be gleaned there. Here are the charts:

Slide1My blood sugar was abnormally high at 135 before the first meeting. I can’t account for it as I had salmon for lunch two hours before. My average rMSSD was much lower at 28.7 which is a reading of high stress. My stress point for rMSSD is 48, when I am below that reading it is an indicator that I am in stress. My experience during the meeting was of being overly excited and I breathed regularly during the meeting.

During the second meeting I was careful about my food choices for the day and had a good blood glucose level of 105. My average rMSSD for the meeting was higher at 38.8. I felt more in the zone in the second meeting physiologically. We were digging into more details in the second meeting and I felt challenged at a few points, and you can see where the reading drops to a very low rMSSD at a few points.

If asked I would have said the first meeting was more successful based on the discussion. And I would have said the second meeting was more challenging. However physiologically the second meeting was far less stressful than the first. When less stressed I imagine my actual performance was better. So my perception of the outcome was very different than the physiological reality.

Next step is to measure outcomes and see if the results correlate with the physiological state occurring during the meetings. I can’t rely on my perception of the situation so further readings will determine outcomes with physiology.