Part III: Heart rate variability
My dear, where to start. There’s so much to explain about heart rate variability (HRV) that a blogpost is almost too short. Anyway, there must be a way to bring some structure in the what’s written on the topic and today I’m going to give it a try. We will start of with a word on the physiology, this time more specific with respect to HRV. Then, we will cycle into the forest of calculation methodologies, and I will end with a guided tour through some applications. So hold on thight, because here we go!
Variation in the interbeat-intervals (IBI’s) of the heart, can be linked to the sinoatrial (SA) node. This node is influenced by many inputs like eating, physical activity, the baroreflex, thermoregulation, hormones and… mental stress. All these inputs together define HRV through the SA node. But according to what the theories claim, HRV can kind of turn the whole thing around and give a quantitative indication of the inputs by calculating it in the right way.
But before we start digging into the calculation methods, lets first take a look at the picture on the left. What you see is the so called QRS complex. Ever seen an ECG? This consist of subsequent QRS complexes. The Q, R and S waves in the complex exist because of electrical activation of the heart ventricles. When a heart beats at 60 beats-per-minute (bpm), then the distance between two consequetive R waves would be exactly 1 second (in theory). However, a theory is not always reality and there exists variation in the so called RR-intervals, even when the hearts beats a 60 bpm, ranging from 0.5 to 2.0 seconds. This can be caused by the inputs we talked about in the previous paragraph.
Calculation of HRV
It’s very important to understand that HRV can be calculated in many ways, since each way has it’s own properties. Some of them are claimed to relate to the activity of the autonomic nervous system, and therefore, to mental status of humans and animals. Others have not been proven to relate to mental status. However, very often they are claimed to be linked to other phenomena, ranging from metabolic to liver problems.
When we look at the calculation methodologies, we can distinguish two main branches. The first and oldest calculation techniques result in time domain HRV features. The second branch consists of calculation methods that result in frequency domain features. The time domain features are easiest to calculate, however, they often contain less information than the frequency domain features. When we take a look at literature, we find that mainly the low frequency (LF) and high frequency (HF) measures of HRV have been used to quantify mental status. For a long time in history, LF has been related to the activity sympathetic nervous system. Today however, the interpretation is a more challenging since it is known to consist of influences of both the sympathetic and parasympathetic nervous system. The HF is typically linked to the activity of the parasympathetic nervous system. Also the ratio LF/HF has been used to quantify at the balance between activity of the sympathetic and parasympathetic nervous system. But since a blogpost is supposed not to be too long, I will now focus on the frequency domain measures. But remember, there is more…like geometric and non-linear methods.
But how do we then calculate these features? Well, luckily there is help! Researchers often use the freeware Kubios. It’s developed for non-commercial research and personal use and is especially designed to analyse HRV. You just plug in heart rate data from a Polar, Suunto, Garmin, … and calculates the most important features for you, so that you can use it for assessment of stress and recovery. But of course you can also calculate LF and HF yourself. Typically, LF is the total spectral power of all normal-to-normal (NN) intervals between 0.04 and 0.15 Hz, HF is the total spectral power of all NN intervals between 0.15 and 0.4 Hz, and LF/HF is simply the ratio of the two measures. However, caution is advised, since a typical Fourier transform only allows evenly sampled data, and HRV isn’t because not every interval between to beats is equally long. This means that we don’t have a nice interval every second, but it is way more irregular. But moreover, there are abnormal intervals that have to be excluded, since HRV is only calculated with NN-intervals (normal to normal). Or there are gaps or extreme noise, all reasons not to use the data. Therefore, a fast fourier transform (FFT) is not ideal. But fortuately, the Lomb periodogram for unevenly sampled data can be used. This bypasses the disadvantages of the FFT.
There are so many applications of HRV that it’s hard to make a selection. Therefore, I will focus on one applications for humans and one for animals. So let’s take a short break from the reading with a video. The Inner Balance from HeartMath is a commercial product based on research. It’s designed to optimise your HRV patterns for a better mental life by means of bio-feedback. See it for yourself in the video below.
Rietmann et al. (2004) provide the second application of this blogpost. The aim of this academic study was to examine whether HRV can distinguish between different levels of excitement in warmblood horses. To induce the different levels of excitement in the horses, a challenging ground exercise task was used in which they for had to walk backward for 3 minutes. Then, the horses were exposed to two training sessions, after which the backward walking was re-evaluated. Th idea was that excitement is lower during the second backward walking period. Several time and frequency domain features were calculated from the collected data. And audio-visual scoring of video’s was used as the Gold Standard (or reference). According to the authors, the results were the following:
“Compared to rest and forward walk, the ﬁrst backward experiment induced a signiﬁcant rise in HR, LF and LF/HF and a signiﬁcant decrease of HF compared to baseline. After the training sessions, the parameters of the sympathetic branch of the autonomoic nervous system (LF, LF/HF) were decreased and the vagal tone (HF) increased compared to before; all changes were signiﬁcant.”
In conclusion, HF is an indicator of parasympathetic (vagal) tone. The higher HF, the more influence does the parasympathetic nervous system have on the body, and the body is calmed down. This is especially true when the LF is decreasing, which is here claimed to be a sign of the sympathetic influence(fight-or-flight) being reduced. But as we have seen, there is some controversion about whether this is true or not. Some researchers claim that there is also a vagal influence in the LF component of HRV.
Today, there’s a lot to do about HRV and mental status. Many interesting and upcoming applications, but also still some controversy about what everything might mean. So more than enough reason to keep challenging each other to keep looking for the truth. Next week’s episode: Decomposition of heart rate.
This blog was originally posted here.