6/22/21 - I have updated this post with Core body temperature data and repeat ramps. As you will see, although it appears that lack of fan cooling dramatically alters the HRVT and DFA a1 behavior, until I had a Core unit, it was not clear if this was a core body temp vs a skin temp phenomena. With the help of the Core, we now know! Please start by reading the post below, and at the end, the new data will be presented. Keep in mind that the original post was written before our perspective review, validation study and the knowledge of a .75 DFA a1 threshold concept. In addition, I did not have a smart trainer or Zwift to perform ramps and was using a Hexoskin vest for HRV which is not as precise as the Polar H10 or Movesense ECG.
Sometimes when you have a spell of bad luck, unexpected positive outcomes can arise. As an example, a couple of years ago, I had a chance to monitor some interesting physiologic changes after getting a flat tire. Fast forward to a few weeks ago when I was all set to do some more DFA a1 ramps (as part of the last post). I was all prepped (Hexoskin on, well rested, calibrated power meter, timed and regulated breakfast) when a typical Florida thunder storm knocked out the electricity to my end of town. Fortunately, I have a backup generator but it does not power my cycling room and the house temperature/humidity rose a few degrees. I decided to still do the ramps (by manual reckoning since Zwift was offline) and see what happened. In the past I was always curious what DFA a1 behavior would look like with heat exposure but never got around to looking. However, let me emphasize, the following tests were done just a few degrees above the usual 73F and slightly higher humidity. The main difference from usual is the lack of a good fan. With the standard set up I have, there is never any sweat on the floor due to the excellent fan air flow rate. So, this is not a test of high thermal strain, simply one of thermal discomfort and higher skin temp (I think). In addition, I was well hydrated pre test and drank 500 cc over the 60 minutes or so of the ramps.
Before going over the data, I'd like to briefly review what happens to heart rate, stroke volume and cardiac output in the face of high skin temperature. In the past I reviewed a very clever study from the lab of Dr Coyle. This was their introduction:
The progressive decline in stroke volume (SV) is one of
the most salient cardiovascular responses to prolonged
exercise, and the degree of decline in SV is exaggerated
when exercising in a hot environment (1,2). There are
two theories of the mechanisms of how SV is lowered during
hyperthermic exercise. The traditional and possibly prevailing
theoretical mechanism is that the decline in SV is caused by an
increase in cutaneous blood flow (CBF), which leads to an increase
in cutaneous venous volume (1–3). Blood accumulated
in the cutaneous veins might reduce ventricular filling pressure,
end-diastolic volume, and subsequently SV, whereas
HR increases just to maintain cardiac output (CO) and blood
pressure (1–3). The alternative mechanism proposes that the
HR drift during prolonged exercise reduces the ventricular filling
time and possibly filling pressure and thus decreases SV
during exercise, especially with heat stress (4–7).
- the prevailing one consisting of a primary redistribution on blood flow and a rise in heart rate to compensate for the lower stroke volume vs
- the possibility that the heart rate rise is primary and the lower stroke volume is a consequence of decreased ventricular filling pressure.
If the heart rate rise was primary, normalizing the rate with a beta blocker (BB) should prevent the change in stroke volume. However if the heart rate rise was secondary, due to higher CBF, preventing heart rate rise would should not change the stroke volume.
They compared 3 groups doing moderate exercise control, hot and hot-BB:
Here is the result:
I put red arrows on each top (cardiac output), middle (heart rate) and bottom (stroke volume) panel at the end of exercise (the rest is the cooling phase).
The results clearly show that:
- Stroke volume is impaired by hot, but normalized by BB
- Heart rate is higher in hot but normalized by BB
- Cardiac outputs are all about the same at the end of exercise.
In separate graphs they did show that CBF was higher in hot conditions but:
The present study demonstrated that the decline in SV wasThe bottom line is that exercise in the heat leads to a higher heart rate that secondarily results in lower SV but maintained cardiac output.
unrelated to changes in CBF and CVC in several ways. First,
βB restored SV of HOT-βB to the same level as MOD despite
significantly higher CBF and CVC
This heart rate elevation is presumably due to changes in both hormonal and autonomic effects on the sino atrial node pacemaker.
Back to DFA a1 behavior in the heat:
My case report did look at beta blocker (BB) effects on DFA a1 in normal conditions and as expected, heart rate was much lower, but the DFA a1 to watt power relation was not altered. Both Ve, lactate and muscle O2 were looked at as well with no change between control and BB, implying VO2 was similar between conditions. My conclusion was that DFA a1 was mainly affected by cycling power/VO2 and not absolute heart rate.
In regards to the lack of a fan, I was honestly curious what would happen with skin temp elevation. My thought (given the above study by Dr Coyle's group) was that although heart rate would be higher, net cardiac output and VO2 would be the same so there would be no major shift where DFA a1 hits the .5 or .7 range. On the other hand, multiple metabolic, cardiac and neurologic inputs can lead to changes in the balance of the sympathetic and parasympathetic effects on the sino atrial node (the pacemaker cells) responsible for heart rate and HRV. Therefore the possibility that DFA a1 would drop prematurely at lower power should be entertained.
- The wrist temp with a fan blowing is about 72F.
- Room temp was about 73F.
- Temp is the gray line, power gray background
Here is the session with no fan (and slightly warmer house):
- Power in light gray.
- Gray line is wrist temp running between 85 and 90F over the last 2 sessions.
- The red line is a wrist temp of 85F.
- Room temp was about 76F
Now to the interesting part - DFA a1 behavior. Each point on the graphs below represent the average of the 3 separate segments of the interval power levels.
Watts vs DFA a1 decline - both "Normal conditions (Ctrl)" and Heat are graphed:
- At each power level the DFA a1 was lower in the heat. At each power value below VT1+5 watts (177 watts), DFA a1 is markedly down from control.
- Although DFA a1 drops with increasing power in both, it had already decreased below .7 at the first power level (with control at .95).
- Regression values are .95 or better.
- By the final power level tested (VT1+25 watts), both DFA a1 values were quite low, below .4 - if the ramp was continued at higher power levels, little further change would occur. This is good for repeat testing in that higher intensities are not going to needed.
- The slope is not as steep in the hot condition since the start value was down and the end was already at nadir. This may represent a curve shift to the right.
Heart rate vs DFA a1 decline:
- As expected, heart rate was higher in the hot condition at each measurement point, almost 10 bpm higher at the start and stayed that way throughout.
- The values of DFA a1 at a heart rate near 134 were very different (.63 Ctrl vs .39 Heat).
What can we conclude from this so far?
- Skin temp elevation causes heart rate drift - no surprise and well known.
- From previous studies, the higher heart rate will drive down the SV but cardiac output is preserved (and as long as dehydration does not occur, VO2 should be similar).
- DFA a1 decline related to intensity is accelerated in the heat and occurs at lower much watt levels.
- DFA a1 decline related to heart rate is altered and may lead to misjudging ones aerobic threshold. I did not plot out DFA a1 at even lower intensities which would have been more revealing (see new data below).
- This also raises the question whether the limits of zone 1 intensity change according to external factors such as heat, skin temp.
A popular theory of DFA a1 behavior is that of it being a final common pathway of multiple sensory inputs that alters autonomic/neural balance to the cardiac rhythm.
The next question is whether the change in DFA a1 decline with skin temp elevation is an effect from a particular branch of the autonomic system (parasympathetic withdrawal + sympathetic stimulation both cause heart rate elevation). If the HRV index decline with skin temp is related to the sympathetic activation side, then blocking this with Atenolol should prevent both the heart rate elevation, stroke volume drop as well as the change in DFA a1. On the other hand, if the DFA a1 index change is mostly caused by parasympathetic withdrawal, we would see no change in behavior between hot and hot-BB, although heart rate would be lower.
What about the balance between parasympathetic and sympathetic influences in regards to heart rate (and HRV)?
What about the balance between parasympathetic and sympathetic influences in regards to heart rate (and HRV)?
A good review on ANS balance during dynamic exercise showed the following figure:
- Their main point was that there is not a simple on and off occurring in the opposing systems.
- At each level of intensity there is some sort of balance. However, there does come a time when parasympathetic activity drops and plays a minor role at higher intensities.
With their summary -
In conclusion: (i) increases in exercise workload-related HR are not caused by a total withdrawal of the PSNS followed by an increase in sympathetic tone; (ii) reciprocal antagonism is key to the transition from vagal to sympathetic dominance, and (iii) resetting of the arterial baroreﬂex causes immediate exercise-onset reﬂexive increases in HR, which are parasympathetically mediated, followed by slower increases in sympathetic tone as workloads are increased.
The next step - Atenolol trial
As an aside, the RPE doing the ramps without a fan were much higher and it was not a pleasant experience. In the name of science I recreated the power outage scenario (without turning off the main breakers) by turning off the AC in the cycling room, not using a fan and intentionally making the room a bit stuffy by having my friendly Great Dane next to me (pictured below - hard at work snoring away).
The conditions were the same except for taking Atenolol 25 mg, 1 hour pre test.
Here is the wrist temp profile for the Atenolol trial:
- The red line is 85F, making the wrist temp very similar to the initial set up.
- Room temp was about 76F.
- Background light gray are the 3 series of intervals
DFA a1 vs Power:
Orange - Control
Blue - Heat
Red - Heat Atenolol
Orange - Control
Blue - Heat
Red - Heat Atenolol
- The Aten+Heat (red) vs Heat (blue) DFA a1 curves are virtually identical. Therefore, the same accelerated DFA a1 decline with heat is present, with no effect from blocking sympathetic input.
- Both Heat tracings still deviate substantially from the Control (orange).
- R values are high, above .95, with .99 for the Aten+Heat.
- Nadir values at VT1+25w (197 watts) are all comparable.
DFA a1 vs Heart rate:
- Heart rate is markedly reduce by Aten and is now about the same as in the control group (BB has done it's job).
- Despite a reduction in heart rate, the DFA a1 values are about the same as in the original Heat only group (blue arrows to compare each power level).
- R is .99 for both the Aten Heat ramp as well as the Heat only ramp.
- Even though the regression equations look very similar they yield diverse values - at a heart rate of 130 for Heat only, the DFA a1 is .41 but at the same heart rate in Heat Aten the DFA a1 is .71.
Before discussing what this may mean, a look at Ve may be of some value. The Hexoskin is capable of measuring minute ventilation with some accuracy (within 10%). Since Ve is generally related to VO2, similar Ve between all three ramps can put the issue of variable VO2 to rest.
Ve vs Power:
Control - Green
Heat - Orange
Heat Atenolol - Blue
Control - Green
Heat - Orange
Heat Atenolol - Blue
- The Ve at each power level is within 10% of each separate condition.
- The center green curve is control with the blue Heat Aten, above and the orange Heat only, below. Therefore, no trend in Ve is present for a shift in the heat.
- R values are all above .97.
- Although perhaps not rigorous enough for publication it seems that Ve is not different during the 3 conditions at each watt level.
- If the Ve to VO2 relation is valid, no major change in VO2 is present between Heat, Heat Aten and Ctrl.
- I did put in error bars that represent 1 standard deviation however paired t tests with unequal variance show a p=.1 for difference between the widest differential of the values at each power level.
Summary so far:
Some possible consequences of "mild skin temp elevation" (stuffy room with no fan):
- Does not seem to affect the Ve to Intensity relation. By inference, the VO2 to intensity (watts) relation is unchanged. However at higher thermal loads it is recognized that VO2 max is reduced.
- Causes heart rate elevation at each power level.
- Results in an accelerated decline in DFA a1 (or curve shift) with cycling power.
- Beta blockade will prevent the heart rate elevation/drift (and by inference the SV drop).
- Beta blockade does not alter the accelerated decline in DFA a1 with heat.
- Since DFA a1 decline is a result of sympathetic/parasympathetic balance, the lack of any effect of beta blockade on it's heat related decline (reducing the sympathetic side) seems to indicate that this HRV index is more dependent on parasympathetic withdrawal than sympathetic stimulation when responding to exercise load.
Multiple inputs from various body centers/systems (exercise power, emotions, skin temperature, recent food intake, altitude, hydration, glycogen availability, etc) are fed into the CNS/ANS resulting in a change in Sympathetic and Parasympathetic outflow to the SA node. In the above scenario, skin temp elevation causes a direct stimulation of the sympathetic side and a withdrawal of the parasympathetic.
This leads to a heart rate rise, a primary change in HRV with a secondary lower stroke volume (a la the Coyle data).With Atenolol blocking the sympathetic side, we still see the effects of parasympathetic withdrawal on HRV (since that is a primary effect). Since we have blocked the sympathetic side, heart rate and SV do not change.
Is there precedence for DFA a1 change with parasympathetic withdrawal in the literature?
Yes and No.
Like everything else it turns out to be complex and depends on what fractal scaling is used, how much atropine or beta blockade is done as well as activity level of the subjects.
From a recent study:
Sympathetic control seems to be capable of maintaining short-term fractal properties. When only 1-adrenoceptors are blocked, RR series become rougher and the dynamics of RR intervals tend to the random regime ( 0.5). In contrast, the vagal control is important to maintain the mid-term fractal properties of HRV, especially in the presence of sympathetic control, as midis not altered during double blockade. When only muscarinic receptors are blocked, RR series are smoother and their correlation properties cease to be a power law.
While short is not affected by parasympathetic blockade, cardiac sympathetic blockade with atenolol decreases short. On the other hand, mid is not affected by sympathetic blockade with atenolol but increases with parasympathetic blockade with methylatropine . For long window sizes( long), neither of the cardiac receptor blockades affected the scaling exponents for atenolol and for methylatropineThe above animal study looked at the effects of both atenolol and/or atropine on various fractal scaling exponents (short, medium, long) and found varying effects.
Depending on the window time scales, different effects on DFA were seen in both beta blocker and atropine use:
B is baseline
Ate is Atenolol
Atr is atropine
Short is close to a1
Here is a look at atropine only depending on window length:
It does appear that with a short window length (a1), atropine lowers the scaling exponent.
Also other studies have shown that atropine causes DFA a1 to increase at rest.
Finally - could age of the subject (me) play a role?
A recent article brings that point up. They found that older subjects exercising in the heat had lower DFA a1 values than younger ones:
The green arrows show the effect of age and moderate intensity on DFA a1.
However, if we look at their methods, the exercise intensity was higher in the older subjects - potentially skewing the DFA a1 results:
Therefore the results of this study may not be totally valid.
Quick look at a published study of thermal comfort and HRV:
Further insight into DFA a1 trends during thermal change come from this fascinating study. Although done at rest, the investigation looked at various HRV parameters at neutral, hot and cold ambient conditions.
Cutting to the results:
The study found that HRV evidently varies from oneBefore we get too confused here, this actually may make sense - at least with respect to heat. Although we found DFA a1 lower with heat, my data was under significant load, where the above study was at total rest.
environment to another (p < 1 × 10−4). The short-term DFA
coefficient is consistently highest in the hot environment and
lowest in the cold environment
If we hypothesize that a "shift" in the "DFA a1 vs intensity curve" takes place under higher skin temp, then that should apply at rest as well. Although I generally don't emphasize this, the DFA a1 trend rises from rest to low levels of exercise - below is from Gronwald's excellent cycling study:
- The red arrow indicates that from rest to low intensity there is a rise in DFA a1.
- Higher skin temps causing a higher DFA a1 may simply be from moving the values further right on the curve. At higher work rates (VO2) this would of course turn downward as noted in my testing.
A review by Wingo brings up the conundrum of how to incorporate heat stress into training limitation guidelines. Although beyond the scope of this post, several points are made including which metric to follow - power vs heart rate:
- As expected, maintaining a given power causes higher HR but keeping a stable heart rate is associated with a lower VO2/power. Which do we choose as our zone demarcating indicator?
- In any event, these are questions that will be sure to arise in the future. If one truly believes non linear HRV indexes such as DFA a1 reflect net internal stress, spending large amounts of time with suppressed values seems inappropriate.
- If you do read this review, the references to a lowering of VO2 max with heat exposure are related to much higher temps than seen with my data. In addition the author seems to be a proponent of the conventional HR drift theory - that of a lowering of blood volume/SV from flow to the cutaneous beds with a secondary rise in HR.
- Since many VO2 max tests are done without a fan (to prevent gas measurement error from blowing air), ramps done without a fan may not correspond to ones with good air cooling with respect to DFA a1 behavior.
- Interpretation of DFA a1 "curve" results between different published studies may be affected by skin temp and air flow.
- Exercise prescription recommendations based on DFA a1 suppression performed at one temperature condition may not correspond to another. Therefore if you derive numbers at home in a cool room based on power, they may not work for a road ride at high temp at that power (or visa versa).
- However, it seems heart rate may still function as a sufficient integrated metric combining intensity as well as thermal effects such as skin temperature.
- From the academic standpoint, the use of beta blockade seems to have the potential to help safely elucidate the basic physiology surrounding HRV with heat and dynamic exercise.
- Numerous factors appear to feed into (and maybe feed forward) the non linear HRV DFA a1, an index of self similarity/fractal complexity. Although this does make result consistency difficult across studies, it does provide what appears to be an excellent index indicating overall body demands.
New Data 6/2/21
With the aide of the Core body temp unit, we now can decide if this is a core vs skin temp issue.
Methods - 20 minute warm up then 20 minute ramp at 5w/min rise (Zwift), Polar H10, artifact below 1%. Room temp 73 degrees - only difference is a fan!
- There is a whopping 35 watt differential in HRVT (power at a1 = .75)
The big question - is this from a core or skin temp change?
- The difference between ramps is not core temp - it's skin temp only!
- In fact, the core temp was a bit lower without the fan (possibly just measurement variation).
Was heart rate markedly affected?
- The heart rate was a bit higher to start with but quite close over the last part of the ramp. Makes sense from the intensity/core temp standpoint.
HRVT heart rate comparisons:
- Very interesting results. One may have predicted that the HRVT (when DFA a1 crosses .75) in the heat/no fan would have been at much higher rates than control.
- Although the heart rate was higher in the heat/no fan condition (see other plots), the HRVT occurred so early in the heat ramp that the HRVT HR was lower than the control ramp.
- This makes reliance on HR to access effort in the heat somewhat questionable and unrelaible.
- Important take home point - do your ramps with a fan if you are indoors. This includes VO2 incremental ramp tests done as part of a research protocol.
What happens outdoors?
Does the skin temp differential effect invalidate our HRVT when exercising outdoors? Test conditions with same gear, bike and wearing a long sleeve jersey with full compression tights, cycling outside in Florida with an air temp of 80 to 85F, high humidity (85%) and direct sunshine:
- Core temp is no different than that of the indoor sessions - with or without a fan.
- Skin temps are below that of indoors with a fan (32 vs 34C).
- Short or lengthy HIT intervals did not markedly affect either skin or core temps.
- However, conditions in excess of the above certainly can lead to issues.
- DFA a1 behavior will change if skin temp rises above it's "normal" range. This is not a core temp or hydration issue. It is probably related to altered autonomic balance and premature parasympathetic withdrawal.
- It appears that air flow is the main requirement as even outdoor warm/humid conditions do not lead to either core or skin temp elevation. However, if either becomes elevated, DFA a1 will be unusually low at corresponding watt/power efforts.
- A fan is highly recommended for indoor testing to maintain skin temps. Even with a fan and comfortable indoor conditions, skin temps may be above what they are outdoors in the heat and humidity.
- Outdoor riding in moderate heat/humidity does not affect skin or core temps to a major extent. However, prolonged efforts at high power in conditions in excess of the above could shift the balance between heat generation and heat loss.
- Firstbeat VO2 estimation - valid or voodoo?
- Heart rate variability during exercise - threshold testing
- Exercise in the heat and VO2 max estimation
- DFA alpha1, HRV complexity and polarized training
- HRV artifact avoidance vs correction, getting it right the first time
- VT1 correlation to HRV indexes - revisited
- DFA a1 and Zone 1 limits - the effect of Kubios artifact correction
- HRV artifact effects on DFA a1 using alternate software
- A just published article on DFA a1 and Zone 1 demarcation
- DFA a1 vs intensity metrics via ramp vs constant power intervals
- DFA a1 decline with intensity, effect of elevated skin temperature
- Fractal Correlation Properties of Heart Rate Variability (DFA a1): A New Biomarker for Intensity Distribution in Endurance Exercise
- Movesense Medical ECG V2.0 Firmware brief review
- Movesense Medical ECG - improving the waveform and HRV accuracy
- DFA a1 and the aerobic threshold, video conference presentation
- DFA a1 - running ramp and sample rate observations with the Movesense ECG
- DFA a1 calculation - Kubios vs Python mini validation
- Frontiers in Physiology - Validation of DFA a1 as a marker of VT1
- Real time Aerobic thresholds and polarized training with HRV Logger
- Active Recovery with HRV Logger
- DFA a1 and exercise intensity FAQ
- DFA a1 agreement using Polar H10, ECG, HRV logger
- DFA a1 post HIT, and as marker of fatigue
- DFA a1 stability over longer exercise times
- DFA a1, Sample rates and Device quirks
- DFA a1 and the HRVT2 - VT2/LT2
- Low DFA a1 while running - a possible fix?
- Runalyze vs Kubios DFA a1 agreement
- DFA a1 - Runalyze vs Kubios vs Logger results in a cyclist