Documenting My Marathon Training With Heart Rate Variability

Back in May 2013 I finally found a solution to a series of nagging knee injuries that had plagued me for 2 1/2 years restricting the volume of running I could manage to about 20 miles a week. Since that time I have gradually been able to increase that volume up to the 60 miles a week I am currently running (yippee  ).

I run primarily for the joy it brings me which is closely linked to the health benefits running provides, but I am now also looking forward to the challenge of attempting to qualify for the Boston marathon. I signed up for a May marathon thinking that the qualifying time for my age group was 3 hrs 35 minutes. Last week however, I discovered that I actually would need to run the marathon in less than 3 hrs and 30 minutes to potentially qualify. It will be a difficult challenge to reach the necessary fitness level in the time before the marathon while also losing about 10 pounds.

In order to attempt this challenge in a healthy way I will be monitoring my heart rate variability (HRV). HRV is a catch all term for a number of measures that capture the hearts subtle response patterns at rest (see this link for more description). A high HRV score is generally associated with good health and well-being. The primary measure of HRV recommended for use in studies of endurance is the root mean square of successive differences in heartbeats (RMSSD). I am also including  two other measures HRV which have shown to correlate with cognitive stress (Shannon Entropy  and Correlation Dimension (D2)) mostly for my own interest to see how they vary with training.

There are two primary uses of HRV data in endurance training.

1.  Day to day variations in HRV data in response to training are very useful to recognize when rest or an easy workout are a good idea to avoid over-training.
2. Weekly averages can be compared to identify gradual improvement in fitness over time.

It is my view that while an objective measure such as HRV can be an extremely useful tool we should also keep in mind it's limitations. If the subtle beating pattern of a heart at rest projects the state of harmony between the body, mind, and environment, we must remember that HRV data is still a partial map of the territory, an abstraction from the reality of the continuous ongoing relationships. Our subjective perceptions of our internal state of wellness are also limited and prone to any number of biases. Therefore the objective HRV data is most useful as a complement to our own internal perceptions, not as a replacement of those perceptions. Without further ado lets get to the results.

Week beginning Jan 6

I began collecting data on Thursday January9. To set the stage I had run 61 and 63 miles in the two previous weeks respectively. I count my weeks starting on Monday as my marathon will take place on a Sunday. I ran 3 miles on Monday, 12 on Tuesday, and 12 on Wednesday leading to the beginning of HRV data collection Thursday morning.



Thursday Jan 9

After waking 7:02 AM
Beats Per Minute = 47.9
RMSSD=61.2, Shannon Entropy =2.82, Correlation Dimension (D2)=3.57

The results indicate that I was responding well to the previous running this week as morning HRV indicators suggest full recovery from prior workouts. An RMSSD of 61.2 is a high HRV score and is consistent  with my typical score when I am well rested and feeling good.

I taught a 2 hr Tai Chi class in the AM, then ran an easy 9 miles on Thursday (4.5 to work, 4.5 back) carrying my backpack and clothes for the day. I felt especially good on the return run given the miles I had run on Tuesday and Wednesday.

Friday  Jan10

After waking 6:12 AM:
Beats Per Minute = 46.5
RMSSD=54.5, Shannon Entropy =2.94, Correlation Dimension (D2)=3.48

Again the HRV data suggest good recovery from the prior workout. My legs however felt a bit stale/tired during the easy 7 miles I ran to and from work on Friday.


Sat Jan 11

After waking 5:48 AM:
Beats Per Minute = 48.9
RMSSD=42.0, Shannon Entropy =2.70, Correlation Dimension (D2)=3.62

The RMMSSD result suggests that the accumulation may have resulted in less then full recovery this morning.

Nevertheless I had an excellent (if difficult) workout  with my running group the LA Leggers. We ran 11 miles. I ran the first 4.6 miles at an 8:20 pace uphill (over 400 foot gain), the next 1 1/2 miles at 7:40 pace, and the last 5 miles at 8 minute per mile pace. I would not have been able to maintain those paces on the same course a month ago.

To demonstrate the effect that a hard workout can have resting HRV data I collected my heart rate data twice more on Saturday (4 and 8 hrs after the run)

After Hard Run 12:36 PM:
Beats Per Minute = 59.7
RMSSD=24.4, Shannon Entropy =3.18, Correlation Dimension (D2)=0.85

Later 4:34 PM:
Beats Per Minute = 56.6
RMSSD=28.1, Shannon Entropy =2.83, Correlation Dimension (D2)=3.43

It is clear that HRV measures drop sharply after a hard workout and recovery takes time. This is the time when beneficial fitness adaptations are taking place.

Sunday Jan 12 (Happy Birthday #51 to me )

I also collected my heart rate data twice more on Sunday to further document the recovery/adaptation process. In the am I still had not fully recovered. This shows why it is generally not a great idea to string hard workout efforts on consecutive days if you are running on most days.

By 4 pm however my HRV data was at it's high point for the week.

After waking 7:13 AM:
Beats Per Minute = 42.0
RMSSD=45.5, Shannon Entropy =2.77, Correlation Dimension (D2)=3.95

Later after teaching Tai Chi 4PM:

Beats Per Minute = 49.0
RMSSD=61.7, Shannon Entropy =2.56, Correlation Dimension (D2)=3.35

 I then ran an easy 6 mile run in the evening.