DFA α1, explained

Journal · 2026-09-15 · R. Klug · All notes

Read in: Deutsch · Français · Español · Italiano

DFA α1 — detrended fluctuation analysis, short-term scaling exponent — is a heart rate variability index computed from the intervals between consecutive heartbeats. Unlike RMSSD or SDNN, it does not summarise how much the intervals vary. It describes how they are organised: whether successive beats correlate with the beats before them, or drift toward randomness.

What the number tracks

At rest and during easy exercise, beat intervals show fractal, correlated patterns — α1 sits around 1.0 or higher. As intensity climbs and both branches of the autonomic nervous system are recruited, that correlation breaks down and α1 falls in a fairly monotonic way, through roughly 0.75 toward 0.5 and below. That continuous fall is what makes the index interesting for endurance training: it behaves like an internal gauge of how hard the whole system is working, not just how fast the heart beats.

The 0.75 and 0.50 markers

Two working landmarks came out of the research group around Rogers, Gronwald and colleagues. In their 2021 papers, an α1 value of about 0.75 corresponded on average to the first (aerobic) threshold, and about 0.50 to the second threshold, when compared against gas-exchange testing in trained subjects. Later independent checks found the correspondence holds on average but with real individual spread — for some athletes the crossing sits meaningfully above or below the lab value.

That is why ampulse shows the 0.75 and 0.50 crossings as estimates. They are useful precisely because they are cheap to repeat — every ramp or steady session yields another data point — but they are not a substitute for an individually measured laboratory threshold, and the honest way to use them is as a repeated observation, not a verdict.

What the method demands from your sensor

DFA α1 is computed from RR intervals, so the strap must transmit them — and cleanly. Two practical consequences:

• A pulse-only sensor is not enough. Most optical wrist and arm sensors send a smoothed heart rate without beat-to-beat intervals. A chest strap that transmits RR over the standard Bluetooth heart rate service — Polar H10, Bluetooth-capable Garmin HRM models, CooSpo, Wahoo TICKR — is the entry ticket. The H10 is the model with the deepest validation record.
• Artifacts move α1 more than they move most HRV metrics. A missed or doubled beat injects exactly the kind of abrupt jump the scaling analysis is sensitive to. Published methodology recommends strict artifact correction and discarding windows that fail it. ampulse gates every α1 window on signal quality and would rather show a gap than a number computed from a noisy window.

How it is used in training

The common pattern is intensity discipline in long, easy work: keeping α1 above ~0.75 as an internal check that an “easy” session is actually easy, independent of pace on a hot day or a hilly route. The second pattern is threshold estimation from ramp tests, repeated over weeks — the trend of where your crossings land is more informative than any single test.

Both uses stay inside the same boundary: α1 describes today’s session. It does not diagnose anything, and a single unusual value is a reason to check the strap contact before it is a reason to change anything else.

Where to read the science

• Rogers B. et al. (2021), Frontiers in Physiology — α1 vs the first ventilatory threshold; origin of the 0.75 marker.
• Rogers B. et al. (2021), IJERPH — α1 ≈ 0.5 vs the second threshold.
• Gronwald T. & Hoos O. (2020), Journal of Sport and Health Science — review of correlation properties of heart rate variability during exercise.

ampulse computes DFA α1 live during sessions from a compatible RR strap, stores the underlying intervals as files on your iPhone, and marks the 0.75 and 0.50 crossings. The raw record stays available, so the math can be checked — in the app, or in any tool you export to.